Fiber biodurability and biopersistence: historical toxicological perspective of synthetic vitreous fibers (SVFs), the long fiber paradigm, and implications for advanced materials

Abstract

Extensive toxicology studies of synthetic vitreous fibers (SVFs) demonstrated that fiber dimension, durability/dissolution, and biopersistence are critical factors for risk of fibrogenesis and carcinogenesis. Lessons learned from the SVF experience provide useful context for predicting hazards and risk of nano-enabled advanced materials. This review provides (1) a historical toxicological overview of animal and in vitro toxicology studies of SVFs, (2) key findings that long durable fibers pose a risk of fibrogenic and tumorigenic responses and not short fibers or long soluble fibers, (3) in vitro and in vivo test methods for biodurability and biopersistence and associated predictive thresholds for fibrosis or tumors, and (4) recommendations for testing of advanced materials. Generally, SVFs (fiber lengths >20 µm) with in vitro fiber dissolution rates greater than 100 ng/cm²/hr (glass fibers in pH 7 and stone fibers in pH 4.5) and in vivo fiber clearance less than WT_1/2 40 or 50 days were not associated with fibrosis or tumors. Long biodurable and biopersistent fibers exceeding these fiber dissolution and clearance thresholds may pose a risk of fibrosis and cancer. Fiber length-, durability-, and biopersistent-dependent factors that influence pathogenicity of mineral fibers are also expected to affect the biological effects of high aspect ratio nanomaterials (HARN). Only with studies aimed to correlate in vitro durability, in vivo biopersistence, and biological outcomes will it be determined whether similar or different in vitro fiber dissolution and in vivo half-life thresholds, which exempt carcinogenicity classification of SVFs, can also apply to HARNs.

Keywords:

• Synthetic vitreous fibers
• glass
• stone
• rock
• inhalation
• dissolution
• biodurability
• biopersistence
• nanomaterials

Introduction

The 3Ds of fiber toxicology (dose, dimension, and durability) as the primary drivers for influencing hazard and disease risk have been well accepted for several decades. Recognition that fiber diameter, length, and durability can influence the extent to which fibers are deposited and retained in the lungs and present a risk of fibrogenic and carcinogenic disease emerged from the toxicological and epidemiologic experience with asbestos. The historical occupational and environmental experience with asbestos prompted caution in the research, regulatory, and industrial communities as new man-made fibers, including synthetic vitreous fibers (SVFs), were introduced into the market. It was through extensive animal and in vitro toxicology research of SVFs that the interplay between fiber dimension, durability, dissolution, and biopersistence (retention in the lungs) were identified as critical factors for SVFs not presenting a risk of fibrogenesis and carcinogenesis in occupational settings. Also, through this research, lung fiber clearance/biopersistence was correlated with fibrotic and tumorigenic responses in animals and fiber dissolution rates in vitro to identify critical thresholds for predicting hazards of novel SVFs in acellular in vitro systems and potentially limiting unnecessary animal testing. The use of acellular in vitro dissolution assays and animal testing, where needed, have been leveraged by industry as a safe-by-design tool to assure novel SVFs do not exceed dissolution/biopersistence thresholds and do not pose a health risk in occupational and end-use settings. Biopersistence is a determinant for hazard and risk that can generally be applied to all particles.

A reflection of the historical toxicological journey of SVFs is pertinent to modern day as well as industries outside SVF manufacturing. Advanced materials, such as engineered nanomaterials, are designed to have unique sizes, shapes and properties, which in some cases may include high-aspect ratio fibers (WHO respirable fibers – fiber length of >5 µm, diameter <3 µm, length:width ratio >3:1; generally long high-aspect ratio fibers have lengths >10–20 µm and aspect ratios >3–5). The Organization for Economic Co-operation and Development (OECD) recognized the need to design nanomaterials with safety in mind and formed a Working Party on Manufactured Nanomaterials (WPMN) in 2006 with the aim to ensure that the hazard, exposure, and human health risk assessment of nanomaterials is high quality and science-based through internationally harmonized standards (OECD 2018). Since the formation of the WPMN, the OECD has published nearly 100 reports outlining various methods to characterize the physicochemical properties, exposure and potential health risk of nanomaterials. The lessons learned from the historical perspective of the SVF experience provide useful context for evaluating and developing predictive tools to assess the hazards and risk of nano-enabled and advanced materials in a safety-by-design framework. As such, this review will (1) provide a historical toxicological overview of the animal and in vitro toxicology studies SVFs, (2) discuss key aspects of the historical research, which contributed to the understanding that long durable fibers pose a risk of fibrogenic and tumorigenic responses and not short fibers or long soluble fibers, (3) summarize available in vivo and in vitro test methods for biodurability and biopersistence and thresholds for solubility and lung clearance which are not associated with fibrosis or tumors, and (4) offer recommendations for hazard and toxicological testing of new advanced materials (e.g. high aspect ratio nanoparticles of sufficient length and durability).

Overview of fiber toxicology

Fiber dimension influences the location and extent to which fibers are inhaled, deposited in, and cleared from the lungs. While particles with an aerodynamic diameter of 10 µm and 4 µm will have 50% penetration to the thoracic and respiratory regions of the lung (Brown et al. 2013), the aerodynamic properties of fibers are driven by fiber diameter. The propensity of fibrous particles (with an aspect ratio [length-to-width ratio] of 3:1 or larger) to deposit in the respiratory tract is largely determined by fiber diameter, and, to a lesser extent, fiber length. Fibers with diameters less than 3 µm align with the airstream and aerodynamically behave like small isometric particles and deposit in the deep lung (Hesterberg and Hart 2000). While fibers up to 3.5 µm in diameter have been detected in the lungs of asbestos workers, fibers of this dimension may represent the very upper-bound limit of respirability (Gross et al. 1971). A number of studies have shown that nearly all fibers deposited in the pulmonary region of the lung are thinner than 0.7 µm (Harris and Timbrell 1975; Sussman et al. 1991a, 1991b; Strom and Yu 1994; Yu et al. 1994). It is noteworthy that most commercial fibrous glass products are comprised of fiber dimensions that do not penetrate the lungs to any great extent (Lippmann 1990). Fiber length has little impact on respirability up to a length of about 20 µm and the deposition of longer fibers is generally inversely related to fiber length. However, it is possible for fibers 100–300 µm in length (with sufficiently thin diameters) to reach the lower lung (Eastes et al. 1996).

Once fibers are deposited within the respiratory tract, the mechanism by which they are cleared from the lungs depends not only on the site of deposition, but also on the size and chemistry of the fiber itself (Figure 1). In general, solid particles are cleared from the lungs through a variety of mechanisms: (1) sneezing, coughing, and removing mucus from the nasopharyngeal region; (2) direct or macrophage-mediated transport along the mucociliary escalator and subsequent elimination by the gastrointestinal tract; (3) direct or macrophage-mediated transport across the bronchiolar or alveolar epithelium and subsequent clearance by the systemic circulation or interstitial lymphatics; and (4) physicochemical processes, including dissolution, leaching, and physical breakdown of particles (Madl et al. 2018). Figure 1 illustrates the various mechanisms by which deposited fibers can be removed from the lungs.

Figure 1. Mechanisms by which respirable short or long fibers clear from or persist in the lungs and pose a risk of non-carcinogenic and carcinogenic health effects. Respirable long biodurable fibers accumulate and persist in the lungs because of reduced dissolution and breakage and evasion of direct or macrophage-mediated transport out of the lungs. Whereas respirable short fibers or respirable biosoluble fibers undergo dissolution and cellular/acellular transport mechanisms to clear from the lungs (in non-overload conditions) and not pose a risk pulmonary health effects.

Macrophage-mediated clearance of deposited fibers is critical for clearance of “short” fibers, but also dissolution and breakage of “long” fibers. Human alveolar macrophages are approximately 14–21 µm in diameter and can readily engulf and clear fibers 15–20 µm in length, whereas rat pulmonary macrophages are 10.5–13 µm in diameter and are generally limited to clearance of fibers less than 7 µm in length (fibers >17 µm in length are generally too long to be completely engulfed) (Crapo et al. 1983; Lum et al. 1983; Sebring and Lehnert 1992; Stone et al. 1992; Krombach et al. 1997; Maxim et al. 2006). Fibers too long to be engulfed and cleared by macrophages must undergo dissolution and breakage in lung lining fluid (pH 7.4) and/or macrophage lysosomal fluid (pH 4.5) to be cleared from the lungs. Fibers that persist and are not completely engulfed by macrophages may cause “frustrated phagocytosis”, which triggers release of reactive oxidant species and acidic phagolysosomal contents into the surrounding tissue, stimulates inflammation, and initiates histopathological changes under chronic conditions (e.g. fibrosis, neoplasm). Thus, the biodurable and biopersistent nature of a fiber in the lungs will directly influence the extent to which a fiber will pose a hazard and risk of disease. While biodurability and biopersistence are terms that have been used interchangeably, they have distinct meanings. Biodurability refers to the fiber dissolution rate, which is typically measured in in vitro systems, and differs from biopersistence which encompass additional fiber removal mechanisms in vivo (Maxim et al. 2006). Biopersistence is defined as the ability of a fiber to remain in the lungs despite the various clearance mechanisms involved (e.g. alveolar macrophage clearance, dissolution, breakage, transport, mucociliary escalator) (Muhle et al. 1994; Muhle and Bellmann 1997; Figure 1). Biopersistence is measured in in vivo studies of laboratory animals or in human lung samples. Generally, the biopersistence and clearance half-life of fibers are measured by lung fiber burden analysis in which lung tissue samples are digested and the remaining fibers are counted. A detailed summary of methods for determining fiber dissolution rates in vitro and fiber biopersistence (and clearance) in vivo are provided in later sections.

When evaluating the effects of fibers in animal toxicology studies, there are a number of key anatomical and physiological differences between rodents and humans that need to be considered. First, as mentioned previously, the size of rat alveolar macrophages is smaller than that human alveolar macrophages. Thus, fibers which may elicit an inflammatory or fibrotic response in rats may not be expected to occur in humans because human alveolar macrophages are larger and can engulf longer fibers than alveolar macrophages in rodent species. Second, the rodent lifespan is 35-fold less than a human, and rats have a faster rate of aging and developing cancer than humans (Berry 1999). As such, the predicted tumor incidence in humans (e.g. asbestos and mesothelioma) is highly dependent on the clearance rate of durable fibers, whereas in rats, a similar dependence of fiber biopersistence-tumor incidence occurs at 17 times higher rate of elimination corresponding to less durable fibers (Berry 1999). Third, the anatomical features of the respiratory tract of a rat can limit, to some extent, the delivery and deposition of fibers in the lungs. Rats have highly convoluted nasal turbinates, are nose breathers, and have an asymmetrical airway branching pattern, which can enhance fiber interception deposition patterns in the nasal cavity and upper respiratory tract and limit fiber deposition in the lower respiratory tract. Generally, fibers with physical diameter of 1 µm or less are considered respirable in rats, whereas fibers with a diameter of 3 µm or less are respirable in humans (Hesterberg and Hart 2001). Fourth, macrophage-mediated clearance of insoluble particles is faster in rats compared humans, with reported half-times of 45–99 days in rats and 33–2500 days in humans (Muhle et al. 1990; Schlesinger 1995; Snipes 1995; ILSI 2005). Fifth, the method and extent of fiber exposures in animal toxicology studies may result in “overload” conditions, in which the lung’s normal mechanisms of particle clearance are overwhelmed by the sheer volume of material and result in responses that may be an artifact of dose rather than physicochemical properties of the inhaled material. All of these factors add complexity to understanding the physicochemical dynamics of fiber biodurability and biopersistence in evaluating effects in rodents and predicting responses in humans. A review of the historical animal toxicology studies of SVFs in the sections below provide a historical account as to how the state-of-the-science evolved from intrapleural and intraperitoneal studies to intratracheal instillation and inhalation studies that eventually led to the understanding that (1) long durable fibers (but not short fibers or long soluble fibers) pose a risk of fibrogenic and tumorigenic responses, and (2) correlations between fiber biopersistence and tumor incidence in vivo and fiber dissolution in vitro provide a basis for fiber biopersistence safety thresholds for newly developed SVFs.

The compilation of over three decades of fiber and SVF research has led to a regulatory framework (in the European Union) in which newly developed SVFs are exempted from classification as a carcinogen if a substance meets one of the following criteria (European Economic Community 1997):

1. A short-term biopersistence test by inhalation has shown that the fibers longer than 20 µm have a weighted halftime less than 10 days, or

2. A short-term biopersistence test by intratracheal instillation has shown that the fibers longer than 20 µm have a weighted halftime less than 40 days, or

3. An appropriate intraperitoneal test (single dose of 1 × 10⁹ WHO fibers per animal) has shown no evidence of excess carcinogenicity, or

4. Absence of relevant pathogenicity or neoplastic changes in a suitable long-term inhalation test.

In the modern era of the 3Rs (reduce, refine, replace) of animal toxicology research, the application of acellular in vitro test methods to measure fiber biodurability or dissolution and predict in vivo biopersistence is paramount to designing SVFs for safe use. The historical experience and lessons learned with SVFs have important implications for the testing and safety evaluation of advanced materials, such as high-aspect ratio nanoparticles or nanofibers.

Classification and physicochemical characteristics of SVFs

SVFs are a class of materials that have major uses for commercial and residential insulation against heat and sound. SVFs are produced by melting various types of rock (stone), sand, slag, clay, or ceramic fibers and then blowing or extruding these raw materials into fibers of specific properties. Specific chemical and physical characteristics are driven by the end-use needs of each product, such as needs for high strength, high electrical resistivity or resistance to chemical agents. All SVFs are composed of a silicate backbone but vary considerably with respect to other components between classes of fibers and within an individual class (WHO 2000). In contrast to the crystalline structure of asbestos fibers, SVFs have no discernable crystalline structure and are comprised of an amorphous elemental/molecular arrangement making them susceptible to dissolution and breakage in the lung (NEHC 1997; Bernstein 2007). As a result of their amorphous nature, SVFs lack longitudinal cleavage planes and thus do not split lengthwise. Rather, in the fiber dissolution process, SVFs exhibit horizontal fractures, resulting in shorter fibers of the same diameter (Assuncao and Corn 1975; IARC 2002). SVFs remain vitreous when used at temperatures below 500 °C; as temperatures increase, depending upon composition, SVFs can flow, melt or devitrify. High-silica and low-alkali metal oxide compositions such as refractory ceramic fibers, alkaline earth silicate (AES) wools, and some rock wools will start to devitrify at temperatures above 900 °C (NEHC 1997).

SVFs are also manufactured with a variety of coatings, binders, sizing, and dedusting agents (Hamilton et al. 1994). The quantity of binder ranges from 0.5 to 1.5% by mass and varies depending on the intended use of the product (IARC 2002). The exception is high-density insulation wool products which may contain up to 25% binder by mass (IARC 2002). The composition and proportion of binder differs between products. Different types of oil are used as a lubricant in rock and slag wools, while glass fibers are usually a complex formulation containing lubricant, resin for binding, and one or more cation active agents for adhesion (WHO 2000). SVFs are categorized by filaments and wools, where filaments (e.g. continuous glass filaments) are drawn continuously during manufacturing and wools (e.g. glass wool, rock/stone wool, slag wool, refractory ceramic fibers, AES) are spun (WHO 2000).

Glass fibers

Glass fibers utilize finely powdered sand as the major silica source; aluminum oxides from either kaolin clay or synthetic sources are commonly added as well as calcium and magnesium oxides from dolomite and boric acid derived from calcium borate (NIOSH 1980; IARC 1988; TIMA 1990; NEHC 1997). Water soluble glass fibers typically contain a lower percentage of alkali metal oxides (K₂O, Na₂O₃, Li₂O) and a higher percentage of alkali earth metals (Ca, Mg, Fe), which confer lower chemical resistance (TIMA 1990). Stabilizers including oxides from aluminum, titanium and zinc contribute to the fiber’s durability; levels are dependent upon the intended use of the product (NRC 2000).

In addition, there are glass subtypes specific to end use needs. Electrical glass (E glass, 99% of which is continuous filament glass) has a low alkali content but a higher aluminum content (14.8% as Al₂O₃) making it insoluble in hydrochloric acid (NEHC 1997). Chemical glass (C glass) is designed to be chemically resistant to acids; high tensile strength glass (S glass) has 33% more tensile strength than E glass. Another version is alkali resistant (AR glass), which is capable of reinforcing 20–30 times its weight due to its zirconium content and is often found in cement reinforcement (NEHC 1997).

Rock wool

Rock or stone wools are produced by melting various igneous rocks including basalt, olivine, and diabase which are typically composed of 40–60% calcium and magnesium carbonate and are dissolvable in hydrochloric acid (NIOSH 1980; TIMA 1990; NEHC 1997). Rock wool is primarily composed of silicon dioxide (45–52%), calcium oxide (10–12%), magnesium oxide (8–15%), and aluminum oxide (8–13.5%) (NEHC 1997). Rock wool products also have a substantial non-fibrous component making up between 20–50% of the total mass of the product. Rock wool fibers have tensile strength ranging from 70 to 100 × 10³ PSI and are optimal for high temperature insulation applications given their high melting point range (NEHC 1997; IARC 2002).

Slag wool

Slag wool recycles blast furnace waste (iron-ore slag is a common starting material) with the final composition dependent upon the metallic content of the slag starting material. Slag wools lack significant sodium and are typically slightly soluble in hydrochloric acid. Other materials are added to compensate for compositional deficiencies; for example, if acid oxides (silica) predominate, then limestone or a slag rich in calcium oxide is added (NEHC 1997). Similar to rock wools, slag wools have tensile strength ranging from 70 to 100 × 10³ PSI (NEHC 1997). Non-fibrous material is slightly higher in slag wool than rock wool, with a range of 30–50% and also ideal for high temperature insulation applications as a result of their higher melting points (NEHC 1997).

Refractive ceramic fibers

Refractive ceramic fibers (RCFs) are the most specialized of the SVFs, with varied composition including kaolin clay based, aluminum silicate and metallic oxide blends (chromous or zirconia), and high purity aluminum silicates. Trace amounts of metal oxides, commonly boron, titanium, magnesium, and/or chromium are also present (TIMA 1990; NEHC 1997). RCFs are specifically designed for uses requiring high-temperature tolerance (1000–1460 °C), with higher percentages of alumina and zirconia oxides to allow these fibers to retain their physical properties at high temperatures (Maxim et al. 1999; EIPPCB 2013; Mast et al. 2000). AES wools were developed as an alternative to the aluminosilicate composition of traditional RCF, although they lack the high heat tolerance of traditional RCF (IARC 2002). One limitation, however, to increased levels of aluminum is a decreased dissolution rate (Maxim et al. 2006).

While RCF are manufactured as amorphous fibers, they will devitrify at higher temperatures. As temperatures reach 980–1100° C, the alumina-silica matrix is chemically changed to mullite, an aluminosilicate crystalline compound. Further increases in temperature result in excess silica crystallizing to cristobalite, with maximum conversion at 1200 °C and a viscous liquid at 1400 °C (IARC 1988; TIMA 1990; NEHC 1997).

Other fibers

Other fibers do not fit neatly into traditional fiber categories, such as X607, which is a high-silica (57.9%) fiber with glass-like properties that is manufactured similar to rock and slag wools (Hesterberg, Hart, et al. 1998). Originally developed to withstand higher temperatures than glass fiber (1300°F, as compared to 900^°F for glass fibers), it is less expensive than RCF fibers and can be used to augment RCF insulation, which would still be in contact with higher heat surfaces (Hesterberg, Hart, et al. 1998). In contrast to RCF fibers, X607 (discussed in a later section) dissolves rapidly in vitro at pH 7.0, is cleared rapidly from the lungs, and does not demonstrate fibrogenic or tumorigenic activity (Hesterberg, Hart, et al. 1998).

Animal toxicology studies of SVFs

Early studies by Pott (Pott and Friedrichs 1972; Pott et al. 1987) and Stanton et al. (1972, 1981b), exposing rodents to fibrous dusts found a relationship between fiber dose, dimension, durability and resulting toxicity. From these experiments came the Stanton-Pott hypothesis, stating the severity of the biological effect of a given fiber is directly related to its dimensions, with longer and thinner fibers being more harmful (Hesterberg and Hart 2001). Early in vivo studies delivered fibers based on mass; although it was soon discovered that suspensions of fibers are rarely homogeneous with respect to size or mass. As study methodology evolved, the importance of size separating bulk material to obtain fibers suitably respirable for the target species was evident and became standard practice (Hesterberg et al. 1993). Fibers were also more clearly characterized in terms of fiber numbers, diameter and length distribution, and non-fibrous content. Later aerosol generation systems were also developed to minimize fiber breakage and non-fibrous dust and were monitored for temperature, relative humidity, and oxygen concentration throughout exposures (Hesterberg et al. 1993). The historical account of the toxicology studies of SVFs by method or route of exposure are provided in the sections below, whereas detailed summaries of individual studies are provided in Appendices Tables (Table A1 – Intraperitoneal Injection, Table A2 – Intrapleural Injection, Table A3 – Intratracheal Instillation, Table A4 – Whole-Body Inhalation, and Table A5 – Nose-Only Inhalation).

Table 1. Summary of in vitro dissolution, in vivo retention and biological responses of SVFs.

Fiber type	Category	Inhalation WT1/2 >20 µm (days)	Intratracheal WT1/2 >20 µm (days)	Kdis @ pH 7.4 (ng/cm^2/h)	Kdis @ pH 4.5 (ng/cm^2/h)	Inhalation Fibrosis	Inhalation Lung Tumors	References
Glass fibers
MMVF10	Glass fiber	37		122.4–300		–	–	Hesterberg et al. 1993, Searl et al. 1999, Maxim et al. 2002, Eastes et al. 1995
MMVF10	Glasswool	44		122.4 (pH 7.0)				Hesterberg et al. 1996, Searl et al. 1999
MMVF10.1	901 Glasswool	14.6				–	–	Maxim et al. 2002
MMVF11	Glass fiber	13		71	0.7	ND	ND	Bernstein et al. 1996
MMVF11	Glasswool	8.7–11	84	71–100		–	–	Hesterberg et al. 1993, Bernstein et al. 1996, Maxim et al. 1999, 2002, 2006, Eastes et al. 1995
CertainTeed glass	MMVF11 Glass fiber	9		100		–	–	Hesterberg, Chase, et al. 1998
JM E-glass	Glass fiber	79		9		+	+	Hesterberg, Chase, et al. 1998
MMVF32	E Glasswool	79	222	6.5		+	+	Davis et al. 1996, Maxim et al. 1999, 2002, 2006
MMVF33	JM 475 glass	49		12		+	+/–	Hesterberg, Chase, et al. 1998
MMVF33	475 Glasswool	48.5	206	53		+	+/–	Davis et al. 1996, McConnell et al. 1999, Maxim et al. 1999, 2002, 2006
JM 901 glass	Glass fiber	15		300		–	–	Hesterberg, Chase, et al. 1998
A	Glass fiber	3.5		129–250	2.4			Bernstein et al. 1996, Hesterberg, Hart, et al. 1998
A	Glasswool	3.5		129	2.4	ND	ND	Bernstein et al. 1996
B	Glass fiber	2.4		320	11.8	ND	ND	Bernstein et al. 1996
B0109	Glasswool	2 − 2.4	7	>500				Hesterberg, Chase, et al. 1998, Maxim et al. 2006
C	New glasswool	4.1 − 4.6	11	309–>500	6.2			Bernstein et al. 1996, Hesterberg, Hart, et al. 1998, Maxim et al. 2002, 2006
C	Glasswool	4.1		309	6.2	ND	ND	Bernstein et al. 1996
M	Glasswool	5	5.7	>500				Hesterberg, Hart, et al. 1998, Maxim et al. 2002, 2006
P	Glasswool	6	11	>500				Hesterberg, Hart, et al. 1998, Maxim et al. 2002, 2006
100/475				9.1 (pH 7.0)		ND	ND	Searl et al. 1999
Rock/SlagWools
MMVF21	Rockwool	78–88.6	190	17–28.9				Maxim et al. 1999, Searl et al. 1999, Maxim et al. 2006
MMVF21	Rockwool (96)	88.6				+	–	McConnell 1994, Maxim et al. 2002
MMVF21	Rockwool (98)	66.5				+	–	McConnell 1994, Maxim et al. 2002
MMVF21	Rockwool	53		28.9 (pH 7.0)				Hesterberg et al. 1996, Searl et al. 1999
MMVF21	Stonewool		196					Bellmann et al. 1995
MMVF21	Stonewool	66.5–88.6		20		+	–	Hesterberg et al. 1996; Hesterberg, Chase, et al. 1998
MMVF21	Rockwool			23	59	NS	NS	Kamstrup et al. 1998
MMVF22	Stonewool	28–140				+	–	Bellmann et al. 2003
MMVF22	Slagwool	5		52.8 (pH 7.0)		ND	ND	Hesterberg et al. 1996, Searl et al. 1999
MMVF22	Slagwool	8.5–9	55	52.8–400		–	–	Hesterberg, Chase, et al. 1998, McConnell 1994, Searl et al. 1999, Maxim et al. 1999, 2002, 2006, Eastes and Hadley 1996
MMVF30	Stonewool		133					Bellmann et al. 1995
MMVF31	Stonewool		123					Bellmann et al. 1995
MMVF34	Stonewool	5.4–6	70	55-59	620	–	–	Hesterberg, Chase, et al. 1998, Kamstrup et al. 2001, Maxim et al. 1999, 2002
F	Stonewool	8.5		96–160	5.9	ND	ND	Bernstein et al. 1996, Hesterberg, Hart, et al. 1998
G	New stonewool	5.0–5.4	12	129–250	4.1			Bernstein et al. 1996, Maxim et al. 1999, Hesterberg, Hart, et al. 1998, Maxim et al. 2002
H	Stonewool	13		169–270	5.5	ND	ND	Bernstein et al. 1996, Hesterberg, Hart, et al. 1998
Insulfrax	AES wool	7		150				Maxim et al. 1999, 2002
Isofrax	AES wool	7		255				Maxim et al. 1999, 2002
L	Stonewool	45		20–23	342	ND	ND	Bernstein et al. 1996; Maxim et al. 1999
O	Stonewool	6	12	>500				Hesterberg, Hart, et al. 1998, Maxim et al. 1999, 2006
RCF
RCF 1	Refractory			4.4 (pH 7.0)		ND	ND	Searl et al. 1999
RCF1	Refractory	53.3–55	266.5	3–7.6		+	+	Hesterberg, Hart, et al. 1998, Maxim et al. 1999, Maxim et al. 2006
RCF1a	Refractory	54.8				+	+	Mast et al. 1995a; Mast et al. 1995b, Maxim et al. 2002
RCF 2	Refractory			3.1 (pH 7.0)		ND	ND	Searl et al. 1999
RCF 4	Refractory			0.5 (pH 7.0)		ND	ND	Searl et al. 1999
Other
SiC1				0.2 (pH 7.0)		ND	ND	Searl et al. 1999
612	Calcia magnesia silica			62.7				Maxim et al. 2006
J (X-607)	Hybrid Fiber	9.8		104	4.3	ND	ND	Bernstein et al. 1996
X607	Hybrid fiber	9.6–9.8	26	104–990		–	–	Hesterberg, Chase, et al. 1998, Maxim et al. 2002, Maxim et al. 2006
Asbestos
Crocidolite	Asbestos	817		<1		+	+	Hesterberg, Chase, et al. 1998
Crocidolite	Asbestos	536				ND	ND	Bernstein et al. 1996
Amosite	Asbestos			0.2 (pH 7)		ND	ND	Searl et al. 1999
Amosite	Asbestos	418		<1		+	+	Hesterberg, Chase, et al. 1998

Note: –: negative; +, positive; +/–: equivocal; ND: no data; NS: not statistically significant.

Table 2. Glass SVF biological effects corresponding to in vitro dissolution at pH 7.4, lung retention following inhalation or intratracheal instillation in Figure 3.

Fiber	Fibrosis	Tumors	Mesothelioma	Comments	Reference
Inhalation
MMVF10	N/A	NS	None	Non-specific inflammatory response; lung burdens higher than occupational exposure (200 f/mg dry tissue)	Hesterberg et al. 1993
MMVF11	N/A	NS	None	Non-specific inflammatory response; lung burdens (6.0E + 05 f/mg/dry tissue) higher than occupational exposure (200 f/mg dry tissue)	Hesterberg et al. 1993
MMVF32 (E-glass)	+	+	N/A	Pathology from Davis et al. 1996 (whole body exposure)	Davis et al. 1996; Hesterberg, Chase, et al. 1998b
MMVF33 (JM475)	Mild up to 13 wks; moderate at 26 wks; marked at 78 and 84 wks	NS	+/– (1/63 hamsters [NS]; none in rats)	18 mo Exposure	McConnell et al. 1999
	+	N/A	1/50 Hamsters (NS)	18 mo Exposure	Hesterberg et al. 1997
901 Glass (MMVF10a)	Mild at all time points	N/A	–	18 mo Exposure	Hesterberg et al. 1997
Intratracheal instillation
MMVF11	Alveolitis noted at 6 and 40 d; no lesions noted up to 720 d	–	ND	Study in sheep	Dufresne et al. 1999

Note: +: positive effects; –: negative effects; +/–: equivocal effects; NS: not statistically significant; N/A: not applicable or evaluated; ND: no data

Table 3. Stone SVF biological effects corresponding to in vitro dissolution at pH 4.5 or 7.4 and lung retention following inhalation in Figure 4.

Fiber	Fibrosis (Wagner Score)	Tumors	Mesothelioma	Comments	Reference
Inhalation
MMVF21	Minimal at 28 mo (4.0)	NS	–	D/R fibrosis; NS tumor incidence; no meso	McConnell 1994
	Minimal at 18 mo (4.0)	N/A	N/A	D/R cellular changes/fibrosis	Kamstrup et al. 1998
	Minimal at 24 mo (4.0)	NS	NS	D/R fibrosis; incidence of non-neoplastic bronchoalveolar hyperplasia in 30 mg/m^3 group was significant	Kamstrup et al. 2001
MMVF22	–	NS	–	2 tumors in 3 mg/m^3 group, none in 16 mg/m^3 group, 3 in 30 mg/m^3 group; NS from control	McConnell 1994
MMVF34	Minimal cellular change at 18 mo (2.8)	NS	NS	Only one dose tested (30 mg/m^3)	Kamstrup et al. 1998
	Mild cellular change at 24 mo (3.2)	NS	NS	Only one dose tested (30 mg/m^3)	Kamstrup et al. 2001

Note: +: positive effects; –: negative effects; +/–: equivocal effects; D/R: dose-response; NS: not statistically significant; N/A: not applicable or evaluated; ND: no data.

Intraperitoneal injection studies

Intraperitoneal (IP) and intrapleural injection ([IPL], discussed in the next section) involve injection of bolus doses into the intraperitoneal or intrapleural cavity, respectively. While this route of exposure and bolus dose method are not representative of inhalation exposures, early studies utilizing this method provided fiber hazard ranking information and was the foundation upon which fiber dimension (especially long thin fibers) was recognized as a driving factor for lung pathogenesis (Vorwald et al. 1951). While a brief review of early SVF studies are described herein, it is important to keep in mind that IP and IPL studies have a number of inherent limitations. IP or IPL injection or implantation studies generally deliver exceedingly high doses (typically defined by fiber mass rather than fiber number) in body compartments well beyond that would be experienced through inhalation and that bypass mechanisms of dissolution, breakage, and removal of deposited fibers from the lungs. As a result, IP and IPL dose administration bypass the normal clearance mechanisms of the lung (mucociliary clearance, pulmonary macrophages), and deliver doses directly to compartments (peritoneum, intrapleural space) that may overwhelm clearance and repair mechanisms and lead to tumors (which would not otherwise occur through inhalation). These early IP and IPL studies often resulted in false-positive results, producing a significant increase in mesotheliomas for most fibers (Roller et al. 1996; Miller et al. 1999), for which later studies conducted by inhalation or intratracheal instillation showed that certain non-biodurable and non-biopersistent fibers were non-carcingenic (IARC 2002; NTP 2011; OEHHA 2011). Further, these early studies lacked specifications for fiber selection and characterization, which was subsequently codified in guidelines for standardization of fiber selection and IP studies in the European Commission (EC) Directive 97/69/EC, adopted in 1997 (European Economic Community 1997). With careful fiber characterization, IP injection studies can be comparable in terms of hazard ranking to studies of carcinogenicity following chronic inhalation of fibers, provided that the delivered fibers are of similar biopersistence and length (Bernstein et al. 2001a, 2001b). Overall, evaluation of biopersistence of long fibers indicated that fibers longer than 20 µm are a good predictor of the tumor response in chronic IP studies (Bernstein et al. 2001b).

Glass fibers

While early IP injection studies of SVFs (MMVF10, MMVF11, JM104 and JM112 glass wool fibers) reported tumors in rodent species, these studies generally delivered high bolus doses that likely overloaded natural defenses and did not characterize fiber size distribution of injected fibers (Pott et al. 1976; 1988; Hesterberg et al. 1993; Pott et al. 1994; Bernstein et al. 1996; Roller et al. 1996; Miller et al. 1999; Searl et al. 1999). However, despite these shortcomings, there were indications that fiber size and dissolution may be important determinants of tumorigenicity. For example, acid or base treatment of glass fibers prior to IP injection showed a different propensity of tumor formation compared to untreated glass fibers (Pott et al. 1988). Pott et al. (1988) showed that acid treatment of JM104 glass fibers (90% of fibers <8.4 µm long and <0.44 µm diameter) influenced the rodent tumor incidence following IP injection. While JM104 fibers treated with NaOH showed a similar tumor incidence (approximately 80%) as untreated fibers, JM104 fibers treated with HCl for either 2 or 24 h showed a lower tumor incidence of ∼60% and <10% respectively, suggesting that fiber durability (and perhaps biopersistence) was decreased for fibers in an acidic environment. Interestingly, SEM imaging did not show acid/base treatment altered fibers (Pott et al. 1988). JM475 glass fibers (median dimensions 3.2 μm × 0.18 μm) treated for 24 h with HCl and IP injected in rats resulted in a higher tumor incidence (50% versus 17–26%) compared to untreated JM475 glass fibers.

With other types of glass fibers, intraperitoneal injection studies comparing fibers of different sizes also showed that fiber dimension influences the incidence of tumors. Specialty glass fibers including either a thin, shorter B-09-0.6 (median dimensions 0.49 × 3.3 μm) or thicker, longer B-09-2.0 (median dimensions 1.2 × 10.5 μm) experimental glass wool had incidences of peritoneal mesotheliomas of 3% (100 mg B-09-0.6), 10% (300 mg B-09-0.6), 23% (150 mg B-09-2.0) and 53% (450 mg B-09-2.0); 0 to 1.4% (1/69 rats) mesotheliomas were detected in saline controls (Roller et al. 1996, 1997). Under the study conditions, exposure to longer and thicker fibers resulted in a greater number of tumors (Roller et al. 1996, 1997). In addition, IP studies showed that larger diameter glass fibers resulted in a lower tumor incidence compared to smaller diameter fibers. In these studies, JM475 glass fibers, which are typically used for filtration of air and liquids as opposed to insulation, were used. Although often grouped with E-glass, these fibers are both non-fibrogenic and non-carcinogenic and should not be equated to E-glass fibers (Bernstein 2007). Generally, these fibers are coded according to mean fiber diameter, with larger numbers indicating larger diameters (e.g. Johns Manville [JM] 110/475 fibers have a greater nominal diameter (1.9–3.0 μm) than JM100/475 fibers (0.28–0.38 μm). A single IP injection of 0.5 mg JM104/475 fibers resulted in a tumor rate of 17% (Muhle et al. 1987), whereas an IP injection of 8.3 mg 100/475 glass resulted in a mesothelioma incidence of 33% (Miller et al. 1999). Tumor incidences in these studies appear to be consistent with the slow dissolution rates of 9.1 ng/cm²/h for 100/475 fibers and 12 ng/cm²/h for 475 glass fibers (nominal diameter not provided) (Hesterberg, Chase, et al. 1998; Miller et al. 1999).

Other IP studies of glass fibers with higher dissolution rates show a lack of tumorigenicity. Intraperitoneal injection studies of biosoluble glass fibers (M, P, and V), and one soluble glass fiber (B) were administered by repeated intraperitoneal injection at doses of 0.5, 2.0, or 5 × 10⁹ WHO fibers to female rats (Grimm et al. 2002). The tumor response in this study was not statistically significant or treatment related for any of the glass fibers. Considering that the dissolution rates for M, P, V, and B glass fibers are 103.7, 610, 450, and 580 ng/cm²/hr, respectively (Bernstein et al. 1996), the findings from Grimm et al. (2002) are consistent with highly soluble fibers having low biopersistence and tumorigenic potential.

Rock and slag wools

Studies of rock and slag wools also appear to show similar findings as glass fibers in that fibers with higher dissolution rates show a lower propensity for tumor formation following IP injection. For example, mesothelioma incidences of MMVF21 (30 mg IP) showed a 97% mesothelioma incidence for two injections per week and 87% incidence for five injections per week, whereas R-stone E3 resulted in no tumor incidence for four weekly injections and 11% tumor incidence for nine weekly injections (Pott et al. 1993; Davis et al. 1996; Roller et al. 1996). While MMVF21 is known to be a more durable fiber (K_dis=21.8 ng/cm²/h in pH 7.4), results from equivalent dimension and higher dose suggest that R-stone E3 fibers (no K_dis data available) are less durable. These findings were further supported by a study by Miller et al. (1999), in which MMVF21 stone wool (K_dis=21.8 ng/cm²/h in pH 7.4) showed a higher mesothelioma incidence (95 vs. 59%) compared to the more biosoluble fiber, MMVF22 (K_dis range of 52.8–400 ng/cm²/h in pH 7.4).

Other historical IP studies of stone wool fibers showed mixed results and did not always have corresponding in vitro fiber dissolution rates or fiber size characterization to put the in vivo biological effects into appropriate context of biodurability and biopersistence. For example, M stone wool fibers (Pott et al. 1993; Davis et al. 1996; Roller et al. 1996) and experimental rock wool B-20 fibers (Pott et al. 1993; Davis et al. 1996; Roller et al. 1996) do not have reported corresponding dissolution rates and showed a range of peritoneal mesothelioma incidence rates (6–63% for M-stone fibers depending on dose, 6–87% for B-20 fibers depending on dose) following IP injection. When administered by IP injection to female rats, finer B-20 fibers (B-20-0.6, median diameter, 0.3 µm) showed generally a greater incidence of mesotheliomas (30, 43, or 75% for 3.5, 8.5, or 25 mg single doses, respectively) compared to B-20-2.0 fibers with wider fiber dimensions (6 or 42%, 22 or 35%, for 6 or 18 single doses to female or male rats, respectively; or 60% for two 30 mg injections in males only) (Pott et al. 1993; Davis et al. 1996; Roller et al. 1996). The explanation for the greater tumorigenic response of B-20-0.6 fibers compared to B-20-2.0 fibers may likely be attributed to the larger doses of WHO fibers of the B-20-0.6 fiber treatment group.

Other stone wool fibers with demonstrated high solubility and low biodurability show a lack of tumorigenic potential. The highly biosoluble stone wool fiber (O) administered by repeated intraperitoneal injection (0.5, 2.0, or 5 × 10⁹ WHO fibers per injection with either 2, 8, or 20 weekly injections) did not show statistically significant tumorigenic responses at any dose (Grimm et al. 2002). With an in vitro dissolution rate of K_dis = >500 ng/cm²/h, the Grimm et al. (2002) study illustrates that highly soluble, non-biodurable fibers do not have a propensity for tumorigenesis (Hesterberg, Hart, et al. 1998; Maxim et al. 2006).

Refractory ceramic fibers

The tumorigenic potential of RCFs in IP toxicology studies is similarly dependent on fiber dissolution and biodurability. Early IP injection studies of Fiberfrax RCF wool and Manville RCF wool show varying tumor incidence rates in IP injection studies (Pott et al. 1987; Smith et al. 1987), however dissolution rates have not been reported on these fiber types to put the biological effects into context. In a 1999 study by Miller, RCF1 and RCF2 showed mesothelioma incidence of 88% and 72%, respectively, whereas RCF4 fibers showed no incidence of tumors. The dissolution rate of RCF1 and RCF2 of 4.4 and 3.1 ng/cm²/h at pH 7.4, respectively, signifies a relatively low dissolution rate and high biodurability, which is consistent with the high tumor incidence observed in IP animal studies. Interestingly, the dissolution rate for RCF4 is slower than either RCF1 or RCF2 at a pH of 7.4 at 0.5 ng/cm²/h, which is inconsistent with the observation of no tumors in these animals (Miller et al. 1999). However, RCF4 is a heat-treated fiber that had fewer long thin fibers compared to RCF1. The heat treatment of RCF4 may have altered the chemical structure and explain the discrepancy between the low dissolution rate in pH 7.4 in vitro and lack of tumors in animals.

In summary, IP studies of glass, stone, and RCF SVFs provide some but limited insight on the relationship between fiber size, durability, and risk for tumors. IP studies of SVFs were limited by extreme administered doses and lack of fiber size characterization. However, retrospectively considering these IP studies with more modern knowledge of fiber dissolution rate, there were indications that some SVFs (e.g. M, P, V, and B glass fibers) with high dissolution rates (K_dis of 100–400+ ng/cm²/h) had a low propensity for tumorigenicity.

Overall, the utility of the intraperitoneal assay for screening potential human hazards is limited. The route of exposure does not mimic any possible route in humans and the suggested, but arbitrary, dose of 10⁹ WHO fibers is unrealistically high. In many cases, fiber burdens in earlier studies exceeded this level. In addition to the fiber clearance mechanisms being completely different between the peritoneal cavity and the lung, the IP methodology of toxicity testing has also been criticized as overly sensitive with a time scale of tumor production that is significantly shorter than that in humans. While IP toxicology studies may be generally insightful for hazard ranking of different fiber types, utilizing these assays for prediction or classification of the carcinogenicity is more tenuous. For intraperitoneal results to have significance, some relationship should exist between biopersistence, dose, and dose distribution of fibers in the lung. As discussed in more depth later, in the lung, the diameters of the long glass fibers (>20 µm) declined at a rate consistent with their exposure to a neutral pH, while the diameter of shorter fibers declined at a slower rate, consistent with exposure to a more acidic environment (phagolysosomes of alveolar macrophages). In the peritoneal cavity, regardless of length, glass fibers dissolved at the same rate (IARC 2002). At higher doses, excess material (greater than 1.5 mg) formed clumps of fibers that were either free in the peritoneal cavity or loosely bound to peritoneal organs. These nodules resulted in foreign body reactions with a granulomatous inflammatory response, which could be a nidus for subsequent neoplastic reactions. Thus, differences in both durability of a fiber in the peritoneal cavity and the presence of these nodules have implications for the use of intraperitoneal injections to assess potential carcinogenicity (Collier et al. 1994).

Intrapleural injection studies

Early IPL studies either injected or surgically implanted fibers directly into the pleural space to evaluate hazard of various fibers and model risk for pleural mesothelioma (Wagner 1963; Wagner and Berry 1969; Stanton and Wrench 1972). These studies had a number of limitations, such as the use of large, non-physiologically relevant quantities of fibers and bypassing normal defense mechanisms following inhalation. These early studies also typically lacked information on chemical composition of fibers and characterization of fibrous versus non-fibrous composition (IARC 2002). While IPL studies provide limited context for risk assessment, the information gained from these studies are useful from the perspective that biodurable fibers longer than 10–20 μm are most likely to induce a biological response (IARC 2002).

Glass fibers

While most studies directly placing fibers into the pleura used injection, Stanton implanted 40 mg of one of 17 glass fibers in 1.5 ml gelatin smeared onto coarse fibrous glass pledgets surgically into the left thoracic cavity. The incidence of pleural mesothelioma in animals that survived for more than 52 weeks varied from 0/28 to 20/29, and was dependent upon fiber size with the most carcinogenic fibers <1.5 μm in diameter and >8 μm in length. A key finding of this study was fibers less than or equal to 8 μm in length were inactivated by phagocytosis (Stanton et al. 1977). This research was basis of the Stanton hypothesis which indicates that the “carcinogenicity of inorganic particulates depends on dimension and durability rather than on physiochemical properties” (Stanton et al. 1981a). Increase in carcinogenicity was correlated with fibers longer than 8 μm, suggesting biodurability plays a key role for longer fibers. Interestingly, Wylie et al. (1987) conducted a follow-up evaluation of the fiber size distribution characteristics of samples administered in the Stanton et al. (1977) study because of the recognized variability of the dose-response relationship between long, thin fibers and carcinogenicity. After evaluating the fiber size distribution, sample reproducibility, and fiber-mass-number conversion of the Stanton samples, Wylie and colleagues concluded that additional factors besides size, shape, and abundance play a role in the development of tumors (Wylie et al. 1987).

Two studies investigated tumor incidence following a single intrapleural injection of various borosilicate fibers in mice or rats (Wagner et al. 1973; Davis et al. 1978). Wagner and colleagues administered to rats 20 mg by intrapleural injection of either borosilicate with 30% of fibers 1.5–2.5 μm in diameter (maximum diameter of 7 μm) and 60% >20 μm in length or borosilicate glass powder (with a diameter <8 μm). No tumor incidence was noted with glass fibers, one mesothelioma was identified in the glass powder exposed rats (Wagner et al. 1973). Davis and colleagues administered to mice 10 mg of two borosilicate glass fiber samples with average diameters of 0.05 and 3.5 μm, which had fiber lengths of several hundred microns or lengths of <20 μm. No pleural tumors were found in the glass fiber treated mice (Davis et al. 1978). For both of these studies, the large burden of fibers did not result in marked tumor incidence; variations in diameter also did not have an impact.

Single intrapleural injections of 20 mg of either a fine fraction (99% <0.5 μm diameter with a median diameter of 0.12 μm; 2% >20 μm in length with a median length of 1.7 μm) or course fraction (17% of fibers <1 μm in diameter with a median diameter of 1.8 μm; 10% >50 μm in length and a median length of 22 μm) JM100 glass fiber were delivered to rats (Wagner 1976). No pleural tumors were noted with the coarse fibers; 12.5% of rats injected with the fine fibers developed mesotheliomas (Wagner 1976). In a later study by the same group, 20 mg either English glass fiber dust with resin coating (70% fibers <5 μm in length; 85% <1 μm in diameter), English glass fiber without resin coating (57% <5 μm in length; 85% < 1 μm in diameter), or US JM 100 glass fiber (88% <5 μm in length; 98.5% ≤1 μm in diameter) were given via a single intrapleural injection. One mesothelioma occurred with English glass fiber dust; it was not clear if this occurred in the coated or uncoated group. Mesotheliomas were found in 4/48 (8.3%) animals in the glass fiber group (Wagner et al. 1984). Similar to the 1976 study, the finer JM100 glass fibers had a greater impact on tumor incidence. Another study delivered a single intrapleural injections of 20 mg JM104 glass fiber (mean length, 5.89 μm; mean diameter, 0.229 μm) to rats. Animals that received JM104 glass fiber had a mesothelioma incidence of 13.3% (Monchaux et al. 1981). With regard to glass fibers, a single exposure to finer fibers (JM100 or JM104) delivered by intrapleural injection induced a relatively low tumor incidence rates. Presence or absence of resin coating did not appear to impact fiber diameter although removal of coating appeared to result in an overall shorter fiber fraction.

Rock and slag wools

Rats received 20 mg either Swedish rock (stone) wool with resin coating (70% fibers <5 μm in length; 52% <0.6 μm in diameter), without resin coating (70% <5 μm in length; 58% <0.6 μm in diameter) or German slag wool with resin (67% <5 μm in length; 42% <0.6 μm in diameter) or without resin (80% <5 μm in length; 62% <0.6 μm in diameter), injected intrapleurally. Similar tumor incidences were seen in the rock wool groups with 3/48 with mesotheliomas in the resin group and 2/48 in the group without; no tumor induction was noted in the slag wool groups, either with or without resin (Wagner et al. 1984).

Refractory ceramic fibers

Rats received single intrapleural injections of 20 mg suspended solids of ceramic aluminum silicate RCF fiber, prepared by ball mill grinding (diameters between 0.5 and 1 μm, no length given). Of the 31 animals injected with RCF, three were found to have pleural mesotheliomas (9.7%). RCF fiber length was not described and it is unclear if ball milling resulted in lower levels of long fibers, which could have contributed to low tumor incidence. Electron microscopy images for RCF show large amounts of non-fibrous particulate, which may have confounded findings (Wagner et al. 1973).

Rats received a single intrapleural injection of 20 mg suspended solids of either fiber A made from kaolin (diameter ≤3 μm, 66%; length ≥10 μm, 80%) or fiber B made from alumina and silica (diameter ≤3 μm, 92%; length ≥10 μm, 46%), prepared by grinding, then sieving to remove large particles. No mesotheliomas were observed in animals treated with fiber A; fiber B rats had a single pleural (2.1%) and two peritoneal mesotheliomas in (4.1%), although the latter were deemed a result of a partial deposition of the dose in the peritoneum (Pigott and Ishmael 1992).

In summary, while there are numerous limitations to the use of IPL technique for SVF exposure, including the non-physiological loads injected, some important data has come from the collection of historical studies. The end goal for these early studies was to build an understanding of the comparative effects of various kinds of fibers and understand which fiber characteristics, especially diameter and length, were responsible for fiber-induced disease. Stanton found increase probability of tumorigenesis was correlated with fibers longer than 8 µm and less than 0.25 µm in diameter and also noted fibers less than or equal to 8 μm in length were inactivated by phagocytosis (Stanton et al. 1977). The large 40 mg dose used in the Stanton study would result in a greater concentration than would be inhaled by an animal in a two-year bioassay and would be equivalent to 11 g fibrous material into the pleural cavity of a 70 kg human (Hesterberg et al. 2012). These studies also indicate fibers longer than 10–20 μm are those most likely to induce a biological response (Stanton et al. 1977; IARC 2002).

Intratracheal instillation studies

Intratracheal instillation (IT) studies have been used for many decades as a means to evaluate the pulmonary effects of precisely delivered substances to the respiratory system. As described elsewhere (Madl and Pinkerton 2009), IT studies have both advantages and disadvantages to characterizing the toxicology of a substance, in particular particles and fibers. One primary advantage is that the exact dose of the delivered substance can be characterized, however the main shortcomings to this type of administration is the bolus nature of the dose delivery and liquid media required for delivery that does not replicate the deposition or clearance patterns of inhaled particles or fibers. Further, the bolus dose delivered by IT instillation is often non-uniform and may result in fibers being deposited in small clumps in the small airways with few fibers reaching the alveolar region (Hamilton et al. 1994; Davis et al. 1996). This bolus effect often results in the formation of granulomatous lesions in the upper airway (Bernstein et al. 1980) and may have significant adverse effects on lung clearance mechanisms (Davis et al. 1996), complicating data interpretation (Maxim et al. 1999). Despite these limitations and challenges, IT studies may provide useful mechanistic and hazard ranking information. In accordance with the 1997 EU Directive, IT studies can be used to exempt SVFs from classification as a carcinogen providing that short-term biopersistence test by IT has shown that fibers longer than 20 µm have a weighted halftime less than 40 days (European Economic Community 1997). The following sections review the historical IT toxicology studies on SVF biodurability, biopersistence, and respiratory effects; a compilation of the in vitro dissolution, in vivo half-life, and biological effects from IT administration of SVFs are summarized in Table 1.

Glass fibers

Early IT studies provided limited or no information about fiber size distribution, dissolution kinetics, or lung clearance half-life, which in turn offered limited understanding for biodurability and biopersistence of different SVF types and dimensions. IT studies of JM100 glass fibers (1 instillation/wk for 5 wks, >94% ≥20 µm in length) showed no pulmonary tumors in rats or hamsters (Smith et al. 1987), whereas JM 104/475 glass fibers (one 0.5 mg instillation/wk for 20 wks) showed a 14.7% incidence of pulmonary tumors in rats (Pott et al. 1987). However, Pott et al. (1987) did not report fiber size distribution (particularly the proportion of fibers >20 µm in length) and both studies did not provide information on dissolution rates or half-time data. While Pott et al. (1988) also did not report dissolution or half-time data, acid treatment of JM104 glass fibers were observed to influence the lung fiber burdens following IT instillation of HCl or NaOH treated fibers. For example, fiber burdens of untreated, NaOH-treated, and HCl-treated JM104 glass fibers were 295 × 10⁶/lung, 76 × 10⁶/lung, and 31 × 10⁶/lung, respectively, 9 months following IT instillation. Fiber burden data suggest greater biopersistence of untreated JM104 than of NaOH- or HCl treated JM104, however tumor incidences of the three test fibers did not follow a dose-response in terms of extent of retained fibers. While the Pott et al. (1988) study illustrated that acid or base treatment can influence the extent that SVFs are retained in the lungs, the lack of fiber distribution, dissolution, and half-time data provided limited information about the physicochemical characteristics of SVFs that preclude or produce inflammatory, fibrotic or tumorigenic responses.

Additional studies during this same time period showed that fibers which had slower fiber dissolution rates in vitro translated to greater biopersistence and longer retention times in the lung. 100/475 and MMVF10 glass fibers have dissolution rates (K_dis) of 9.1 ng/cm²/hour and 122.4–300 ng/cm²/hour, respectively (Searl et al. 1999). Following a 4 day instillation in male rats using size-selected 100/475 or MMVF10 glass fibers, it was shown that the more durable 100/475 fibers were cleared from the lung more slowly than the more soluble MMVF10 fibers (no weighted half-time (WT_1/2) values reported). As of note, clearance half-times are usually represented by a faster clearance phase (generally associated with tracheobronchial clearance) followed by a slower clearance phase (generally associated with pulmonary clearance). WT_1/2 is an index of the complete clearance that includes both fast and slow clearance half-times (Bernstein et al. 1996). Interestingly, longer fibers cleared more rapidly than shorter fibers, supporting the hypothesis that long fibers undergo dissolution and disintegration into shorter fibers, which in turn are cleared by cellular, physical, and dissolution processes.

In a study of sheep, animals received a single IT instillation of MMVF11 and were evaluated 6, 40, 60, 180, 360, or 730 days post-treatment (Dufresne et al. 1999). Lung clearance decreased with time according to both a fast and a slow kinetic component. The diameter of MMVFs decreased over the course of the study, which was not observed with crocidolite fibers (which showed no change). No typical ferruginous bodies (particle or fiber with a coating of protein and iron regarded as support for metal-catalyzed oxidative stress in cell and tissue injury (Ghio et al. 2004)) were found in the group exposed to MMVF11 fibers, but were observed with the crocidolite fibers. Clearance of MMVF11 was thought to proceed through dissolution and macrophage translocation (Dufresne et al. 1999). The IT half-time of MMVF11 was calculated to be 84 days, and the dissolution rate at pH 7.4 was reported to range 71–100 ng/cm²/hour (Hesterberg et al. 1993; Bernstein et al. 1996).

Rock and slag wools

Similar to glass SVFs, historical IT studies of rock or slag SVFs demonstrated that long fibers with greater dissolution rates were more readily cleared from the lungs compared to durable SVFs with slower dissolution rates. Two studies evaluated biopersistence of HT stone wool (K_dis of 59 ng/cm²/hr at pH 7.4), MMVF30 or MMVF31 experimental stone wools (the latter two with chemical composition modified to increase biosolubility) in comparison to MMVF21 (K_dis range of 18–28.9 g/cm²/hr at pH 7.4) (Bellmann et al. 1994, 1995; Kamstrup et al. 1998; Searl et al. 1999; Maxim et al. 2002). Fiber burdens categorized by fiber count by length and diameter fractions were used to characterize whether dissolution, breakage or mechanical clearance were driving mechanisms for clearance from the lungs. While at 2 days post instillation MMVF21 lung burdens were higher than HT rock wool, fiber burdens were 41% for MMVF21 and 10% for HT stone wool of the initial dose at 12 months, and 31% of MMVF21, 39% of MMVF30 and 29% of MMVF31 of the initial dose at 18 months (Bellmann et al. 1994, 1995). Based on the fiber kinetics over the duration of these studies, the WT_1/2 values for fibers >20 µm in length were 196, 133, 123 and 92 days for MMVF21, MMVF30, MMVF31, and HT wool, respectively (Bellmann et al. 1994, 1995). When considering available K_dis values, (range of 18–28.9 ng/cm²/hr for MMVF21 and 59 ng/cm²/hr for HT wool at pH 7.4), the relative dissolution rates did not rank (in relative terms) with the respective WT_1/2 values, likely due to dissolution rates being measured in pH 7.4 instead of pH 4.5. Further, pathology data for IT instillation is limited; MMVF21 delivered to sheep resulted in alveolitis (beginning at 6 days and continued through 720 days), but no lesions throughout the entire 720 post instillation follow-up period (Dufresne et al. 1999). Although there were no lesions, the inflammatory response (alveolitis) may have impacted the in vivo clearance, thus potentially explaining the disconnect between the in vitro dissolution rates and in vivo clearance rates.

In another study of MMVF21 rock wool and MMVF22 slag wool fibers, faster in vitro dissolution rates followed a trend of faster in vivo fiber half-life (Searl et al. 1999). In a biopersistence study, male rats received 0.5 mg fibers per day for 4 days of size-selected MMVF21 rock wool or MMVF22 slag wool. More durable MMVF21 long fibers (>20 µm) cleared from the lung more slowly while longer fibers (>20 µm) of the less durable MMVF22 cleared more rapidly (consistent with dissolution rates of 28.9 and 52.8 ng/cm²/hr at pH 7.4 for MMVF21 and MMVF22, respectively). At 12 months, lung burdens of all fibers longer than >0.4 µm was 37% for MMVF21 and 21% for MMVF21 of the initial dose. Results from the Searl et al. (1999) study supported that SVFs with a faster dissolution rate translated to faster clearance from the lungs, which occurred through a process of fiber dissolution, disintegration, and cellular removal.

Refractory ceramic fibers

Similar to the other SVF types previously reviewed, pulmonary clearance of RCF fibers also generally tracked with in vitro dissolution rates. In a biopersistence study, male Fischer 344 rats were exposed by IT instillation (0.5 mg fibers per day for 4 days) to size-selected RCF1, RCF2, and RCF4 fibers. The biopersistence of the different fibers was influenced by their dimensions and solubility; long fibers (>20 µm) tended to clear from the lung more slowly than shorter fibers (those >0.4 µm diameter and <20 µm length) for the more durable RCF1, RCF2 and RCF4 fibers, consistent with the hypothesis that short fibers are predominantly cleared by direct transport and cellular processes, whereas long fibers are cleared by a combination of dissolution, disintegration, and cellular mechanisms (Searl et al. 1999). While there is limited data on half-times for IT instillation of RCF fibers, one study reported a WT_1/2 of 266.5 days for IT of RCF1 fibers which is consistent with slow dissolution of 3–7.6 ng/cm²/hour at pH 7.4 (Hesterberg, Hart, et al. 1998; Maxim et al. 1999). For RCF2 and RCF4, K_dis values of 3.1 and 0.5 ng/cm²/hr at pH 7.4, respectively, were identified.

Other fibers

In the 1990s, researchers became more sophisticated in their approaches to prepare and characterize fiber size selective IT doses, quantify clearance half-times in the lungs, and benchmark in vivo fiber clearance with in vitro fiber dissolution. The biopersistence of the hybrid fiber X607 was evaluated following a single intratracheal instillation of 2 mg sized fraction in 0.3 ml to Wistar rats. Fiber elimination correlated with fiber diameters, which supported dissolution, fiber breakage, and physical clearance as the mechanism of the fiber elimination. The half-time of fiber elimination from the lung following IT exposure was 47 days (all fiber lengths) (Muhle et al. 1994), however in another study of the same fibers delivered by inhalation, half-time of 9.8 days (fibers >20 µm in length) and a dissolution rate of 104 ng/cm²/hour at pH 7.4 and 4.3 ng/cm²/hour at pH 4.5 was observed (Bernstein et al. 1996).

In summary, IT toxicology studies in the 1990s began to illustrate a clear association between fiber dissolution and biopersistence of SVFs; higher fiber dissolution rates in vitro translated to a faster clearance from the lungs. IT studies of SVFs generally did not correlate in vitro dissolution kinetics to in vivo clearance kinetics or to biological effects. It was not until additional studies using inhalation delivery of SVFs (as described next) that researchers began to appreciate that fiber biodurability and dissolution not only influenced biopersistence, but also impacted the propensity of fibers to elicit lung inflammatory, fibrotic and tumorigenic responses.

Inhalation studies

Before inhalation studies of SVFs were initiated by the Research and Consulting Co (RCC, Itingen, Switzerland), many earlier studies were confounded by study design limitations. For example, fiber aerosols were delivered by whole body exposure, which could result in oral exposure through preening and potential confounding of digestive tract tumors. Aerosol concentrations were characterized by mass concentrations rather than fiber number concentration by fiber dimension. Significant non-fiber particulate concentrations were delivered with SVFs, which increased the potential for overload conditions and biological effects due to the extent of dose rather than the characteristics of the SVFs. Limitations also included lack of size selective fiber aerosol preparations that allow for inhalation and delivery of fibers to the deep rodent lung; fibers with diameters greater than 1 µm are unlikely to have appreciable deposition in the rodent lung. In the Smith et al. (1987) study, RCF, slag wool, Certainteed glass, High Temp, and Manville insulation fibers of mean fiber diameters 1.8–6.1 µm were delivered by nose-only inhalation to rats and hamsters. While no significant tumor induction was noted with any of the SVFs, the relatively large fiber diameters suggest a low possibility of fibers reaching alveoli (Smith et al. 1987). Further, a high proportion of non-fiber particulate to fiber in the aerosol (4:1 for JM100, 33:1 for RCF, 38:1 for high temp glass wool) also suggests that any biological response would be highly confounded by exposure to non-particulate matter (Smith et al. 1987). Thus, for meaningful toxicological comparisons of different fiber types, comparable aerosol fiber dimensions and doses of respirable fibers (with low interference of non-fiber particulates) need to be delivered by inhalation. Based on experience up to the late 1980s, this was not a small technological and methodological undertaking.

By 1989, rodent inhalation studies began to address these limitations. The criteria, as described in Bernstein et al. (1995), included: (1) highly rat respirable fibers (geometric mean diameter <1 µm), (2) large proportion of long fibers (arithmetic mean length >20 µm), (3) representative of fiber dimensions in occupational exposure, (4) aerosol generating system that does not destroy fibers, (5) nose-only flow past inhalation, and (6) characterization of fiber aerosol and lung burden by fiber dimension. During this time, key studies were initiated at the RCC in Switzerland to evaluate biological effects and biopersistence of fibers representative of each class of SVF, with a focus on three main protocols, including chronic inhalation to evaluate thoracic fibrogenesis and tumorigenesis during and following a 5 day exposure to characterize the maximum tolerated dose (MTD) and biopersistence (18 months for hamsters; 24 months for rats). While many studies included multiple species when evaluating the effects of SVFs, some animal models, such as hamsters, have less sensitivity for evaluating the pulmonary carcinogenicity of inhaled fibers (Maxim and McConnell 2001). For example, hamsters (but not rats) appear to be resistant to development of lung tumors from exposure to RCF and sensitive to the induction of mesothelioma from exposure to amphibole asbestos (McConnell 1994; McConnell et al. 1999). In the context of evaluating the potential for fibrosis, lung cancer, or mesothelioma, Warheit and Hartsky proposed that the rat is a preferred model, because the fibrogenic responses (at least in response to asbestos) in rats are qualitatively similar to humans, rats have low spontaneous tumor rates, and incidence of mesotheliomas in asbestos- and erionite-exposed rats mirrors that observed in humans (Warheit et al. 1994).

Early studies illustrated the importance of fiber length and also illustrated inhalation of more soluble fibers was less likely to result in a pathogenic response. To standardize the measurement of the clearance rate of fibers from the lung and better understand the relationship between biopersistence and toxicity, standardized biopersistence protocols (“Ispra Protocols”) were developed by the European Commission at the European Chemicals Bureau (ECB) (Bernstein and Riego Sintes 1999). These protocols specify male rats should be exposed to a well-defined rat respirable aerosol SVF for 6 h a day for a total of 5 days. The aerosol should have at least 100 fibers at lengths >20 μm/cm³ with diameters less than ∼0.8 μm. The clearance half-times are determined by fitting the data to either a single or double exponential curve based on criteria in the protocol. The clearance half-time of the fibers longer than 20 μm in length is reported either directly from the single exponential curve or as a weighted half-time combining the double exponential curves. Within the context of this protocol, the clearance halftime of SVFs longer than 20 μm ranged from a few days to less than 100 days. For SVFs, the European Commission has established a directive that states that if the inhalation biopersistence clearance half-time of a fiber is less than 10 days, it does not need to be classified as a carcinogen (Bernstein and Riego Sintes 1999). A compilation of the in vitro dissolution, in vivo half-life, and biological effects from inhalation exposure to SVFs are summarized in Table 1.

Glass fibers

A wide range of glass fibers have been tested in inhalation toxicology studies to understand the correlation between fiber lung clearance, biopersistence, fiber dissolution, and biological effects (including inflammation, fibrogenesis, and tumorigenesis). Among the historical studies, tested glass fibers have included MMVF10, MMVF11, MMVF10a (a thinner version of MMVF10), MMVF32, JM475 glass fiber (MMVF33), 104E, 100/475, JM902, JM901F, E-glass microfiber, and other glass fibers coded “A”, “B”, and “C” (Hesterberg et al. 1993; Bernstein et al. 1996; Hesterberg, Chase, et al. 1998; Hesterberg, Hart, et al. 1998; McConnell et al. 1999; Cullen et al. 2000; Hesterberg and Hart 2001; Hesterberg et al. 2002; Bellmann et al. 2003; Bernstein 2007). Many of these studies exposed the animals using flow-past nose only inhalation, whereby airborne fibers were generated by a rotating brush feed system which aerosolized fibers with minimal breakage (Bernstein et al. 1996; Hesterberg, Chase, et al. 1998; Hesterberg, Hart, et al. 1998; Hesterberg et al. 2002). The following collection of studies demonstrate that glass SVFs with low biopersistence, relatively fast clearance rates from the lungs, have correspondingly high dissolution rates in in vitro systems, and a negligible risk for lung fibrosis or tumors.

In 1993, Hesterberg et al. tested two standard glass fibers, MMVF10 and MMVF11, by aerosolizing and delivering size selected fibers to rats by nose-only inhalation for 6 h/day, 5 days/week with post-treatment follow-up up to 24 months. Macrophage infiltration was observed beginning at 3 months exposure, which was transient and resolved by 6 to 12 months after exposure cessation, and neither SVF significantly increased the incidence of mesotheliomas or lung tumors (Hesterberg et al. 1993). Lung clearance (WT_1/2) of fibers longer than 20 µm has been reported ranging from 39 to 44 days for MMVF10 and 9 to 38 days for MMVF11; in vitro dissolution rates were considerably lower for MMVF11 either 100 and 25 ng/cm²/hr at pH 7 and 4.5 compared to MMVF10 at 300 and 329 ng/cm²/hr at pH 7 and 4.5, respectively (Bernstein et al. 1996; Hesterberg et al. 1996; Hesterberg, Hart, et al. 1998). While MMVF10 in vitro dissolution rates were higher and in vivo half-life was lower than MMVF11, both SVFs demonstrate a sufficient extent of lung clearance to not present a risk of tumors.

McConnell et al. (1999) exposed male hamsters by nose-only inhalation 6 h a day, 5 days a week for 18 months to MMVF10a, (a thinner version of MMVF10) or JM475 glass fiber (MMVF33). MMVF10a resulted in lung inflammation with recovery at six weeks and no neoplasms, whereas MMVF33 showed inflammation and mild fibrosis by 26 weeks, progressing in severity by 52 weeks with a single incidence of mesothelioma. Fiber lung burdens (>20 µm length) for MMVF10a were 37% lower than MMVF33 at 6 h post-exposure. After 78 weeks of exposure and six weeks of recovery, MMVF10a fiber (>20 µm length) burden reduced by 95% of the initial dose, while the MMVF33 fiber burdens were only reduced by 40%. Cumulative fiber burdens were shown to be inversely related to in vitro dissolution rates for MMVF33 (K_dis 12 ng/cm²/hr at pH 7.4 and K_dis 12 ng/cm²/hr at pH 4.5) with higher cumulative lung burdens and more severe pulmonary effects; no dissolution data was available for the finer MMVF10a (McConnell et al. 1999).

MMVF11 glass wool biopersistence was compared to one of several experimental glass wools labeled A, C, or B-01/09 (Bernstein et al. 1996). Aerosolized fibers were delivered to Fischer 344 rats by nose-only inhalation for 6 h a day, 5 days a week and followed at post-exposure time points of 1 h, 1 or 5 day(s), or 4, 13, or 26 weeks. Generally, the WT_1/2 were correlated with in vitro dissolution:

• MMVF11 – WT_1/2 13 days (fibers >20 µm long) – K_dis 71 ng/cm²/hr at pH 7.4 and 0.7 ng/cm²/hr at pH 4.5

• C Glass – WT_1/2 4.1 days (fibers >20 µm long) – K_dis 309 ng/cm²/hr at pH 7.4 and 6.2 ng/cm²/hr at pH 4.5

• A Glass – WT_1/2 3.5 days (fibers >20 µm long) – K_dis 129 ng/cm²/hr at pH 7.4 and 2.4 ng/cm²/hr at pH 4.5

• B-01/09 – WT_1/2 3.5 days (fibers >20 µm long) – K_dis 320 ng/cm²/hr at pH 7.4 and 11.8 ng/cm²/hr at pH 4.5

Cullen et al. (2000) evaluated two special purpose glass microfibers, 104E and 100/475, in an inhalation study in which fibers were delivered to rats by nose-only inhalation for 12 months with recovery follow-up for an additional 12 months. At the end of the 12-month recovery period, fibers of all fiber lengths were 30% of the initial fiber burden for 104E fibers, resulting in lung tumors (7 carcinoma, 3 adenoma) and mesotheliomas. Persistence of 100/475 fibers was 28% of the initial burden but resulted in minimal fibrosis, lung adenomas and no mesotheliomas. While the chemical composition of 104E fibers were not significantly altered by up to 24 months of residence in lung tissue, the chemical composition of 100/475 was substantially altered (leaching of some components resulted in modified surface properties for 100/475 fibers) over the same time period. Overall, despite similar dissolution rates (9.1 ng/cm²/hr for 100/475 and 9 ng/cm²/hr for JM E-glass fibers [similar to 104E]), pathogenicity outcomes were markedly different. This is likely partly due to differences in numbers of long fibers (at 12 months of exposure, 83 × 10⁶ 104E fibers and 11 10⁶ 100/475 fibers >20 µm were measured) but also due to proportionately greater leaching of 100/475 fibers (Cullen et al. 2000).

Two types of glass wools developed for optimal biosolubility in the lung, JM902 (used for insulation and filtration) and JM901F (used for standard thermal and acoustical insulation) were compared to previous evaluations of JM901 glass wool and JM475 (MMVF33) (Hesterberg and Hart 2001). Rats received JM902, JM901F or JM475 by nose-only inhalation for 6 h a day over a total of 5 days, with evaluation at day 1, 14, and days 29–30 post treatment. After 5 days total exposure to JM902, WT_1/2 was 6.8 days (fibers >20 μm length) and 90% clearance time (T90) was 33 days; for JM901F fibers, WT_1/2 (fibers >20 μm in length) was 8.1 days and T90 was 38 days and, for JM475 fibers, WT_1/2 (fibers >20 μm in length) was 49 days and T90 was 240 days. Inhalation of either JM902 and JM901F fibers induced pulmonary inflammation (primarily macrophage infiltration), which returned to normal at 3 weeks post-exposure. Lung clearance half-times for 902 and 901 F did not exceed the European Union (EU) criteria for noncarcinogenic fibers (WT_1/2 F > 20 μm, <10 days), while the JM475 half-times of 49 days did exceed the EU criteria. Dissolution rates for JM902 and JM901F (K_dis- 500 and 150 ng/cm²/hr at pH 7.4, respectively) compared to JM475 at a substantially slower rate of 12 ng/cm²/hr at pH 7.4, further demonstrated that biosoluble (non-biodurable) fibers, such as 902 and 901 F, with a fast rate of lung clearance (WT_1/2 <10 days for fibers >20 μm) should be less persistent than JM475 in the rat by inhalation (Hesterberg et al. 2002).

In a study of E-glass microfibers (Bellmann et al. 2003), lung retention was evaluated following 3 months of nose-only inhalation exposure in rats. A WT_1/2 of 67 days (range of 56–78 days) was reported and after three months of exposure, increases in lung weight, polymorphonuclear leukocytes (PMN) in the bronchoalveolar lavage fluid (BALF), cell proliferation (BrdU-response) in terminal bronchiolar epithelium, and interstitial fibrosis were evaluated. The authors concluded that the cell proliferation assay may offer important predictive value for evaluating the potential for carcinogenicity (Bellmann et al. 2003).

In general, glass fibers have lower biopersistence than other forms of SVFs. The newer Fibers A, B, and C have the lowest reported WT_1/2 values and are less biopersistent than either MMVF10 or MMVF11, while special purpose fibers MMVF33 and MMVF32 have the highest WT_1/2 values (Bernstein 2007). Of the less biopersistent fibers, Fibers C and B have no alumina and have high levels of alkaline and alkaline earth metals, contributing to an increase in solubility. The more biopersistent fibers MMVF32 and MMVF33 have higher concentrations of alumina and silica (54.3% SiO₂ and 13.9% Al₂O₃ and 58.6% SiO₂ and 5.87% Al₂O₃, respectively), which contributes to slower dissolution rates (Bernstein 2007).

Based on these key inhalation studies, the risk of lung fibrosis or lung cancer following inhalation of glass fibers is negligible for biosoluble fibers; some tumor induction was found with continuous exposure at concentrations high enough to maintain a high lung burden, especially with high concentrations of fibers longer than 20 µm. Long fibers with higher concentrations of alumina and silica were associated with lower dissolution rates and slower clearance from the lungs. However, the risk of tumors was typically associated fibers with lung clearance rates (WT_1/2) greater than 40–50 days and in vitro dissolution rates less than 100 ng/cm²/hr at pH 7.

Rock and slag wools

Historical animal inhalation studies of rock and slag wools show similar findings as glass fibers in that faster fiber lung clearance and dissolution rates (although at pH 4.5 instead of pH 7.4) are associated with a decreased risk of lung fibrosis and tumors. In these studies, a wide range of different stone and slag wools were tested, including MMVF21 (rock wool), MMVF22 (slag wool), MMVF34 (HT, high alumina, low-silica rock wool fiber), and experimental rock wool fibers (labeled F, G, H, and L). MMVF21 stone wool, with longer biopersistence (WT_1/2 64.5 days), resulted in persistent inflammation and fibrosis, while other newer fibers like MMVF34 HT rock wool with a substantially lower WT_1/2 of 6 days and inflammatory or fibrotic responses were not observed (Bernstein et al. 1996; Kamstrup et al. 1998). Both fibers have comparable in vitro dissolution rates at pH 7.5, but MMVF34 has a substantially higher K_dis at pH 4.5 of 620 ng/cm²/hr compared to that of MMVF21 at pH 4.5 of 59 ng/cm²/hr indicating MMVF34 fibers should be more easily cleared as long as doses do not exceed clearance mechanisms managed by alveolar macrophages (e.g. do not reach overload conditions) (Bernstein et al. 2005).

In a study of MMVF21 (rock wool) or MMVF22 (slag wool), male Fischer 344 rats were exposed by nose-only inhalation for 6 h/day, 5 days per week for up to 104 weeks at fiber concentrations of 3, 16 and 30 mg/m³ (McConnell 1994). Groups of six randomly selected animals were either evaluated for lung fiber burdens or removed from exposure (recovery groups) at 3, 6, 12, 18, and 24 months. For the 24-month exposure group, animals were held for lifetime observation until 20% survival (28 months). While there were some lung tumors documented with the MMVF21 and MMVF22 treatment groups, there was a difference in the tumor incidence rate at the different dose levels. For example, four lung adenomas (3.5%) and one lung carcinoma (0.8%) each were observed in the 3, 16 and 30 mg/m³ MMVF21 groups. In the MMVF22 group, one adenoma (0.9%) and one carcinoma (0.9%) were noted in the 3 mg/m³ dose; no tumors were observed in the 16 mg/m³ dose, and two pulmonary adenomas (2.6%) and one lung carcinoma (0.9%) was observed in the 30 mg/m³ treatment group. While size-selected stock fiber numbers were roughly equivalent (slightly more in MMVF21 samples), the retained lung burdens/mg of dry lung tissue after 24 months were greater for MMVF21 compared to MMVF22-treated animals. Across doses, lung burdens for MMVF22 were 72, 82, and 73% lower than MMVF21 animals for fibers >20 µm in length at the 30, 16, or 3 mg/m³ doses, respectively. In later studies, WT_1/2 was determined to be 64.5 days for MMVF21 and 8.7 days for MMVF22, consistent with the lower lung burden seen at 24 months with MMVF22 (McConnell 1994; Bernstein et al. 1996). In general, the faster in vitro dissolution rate (K_dis 52.8 ng/cm²/hr at pH 7.4) and in vivo clearance (WT_1/2 8.7 days) for MMVF22 translated to a lower tumor incidence rate compared to MMVF21, which showed a slower in vitro dissolution rate (K_dis- 23 and 59 ng/cm²/hr at pH 7.5 and 4.5, respectively), slower in vivo clearance rate (WT_1/2 64.5 days) and modestly higher tumor incidence of 4.4–4.5% for 3, 9, and 30 mg/m³ (MMVF21) compared to 1.7% (3 mg/m³) and 2.6% (30 mg/m³) for MMVF22 (McConnell 1994; Bernstein et al. 1996; Kamstrup et al. 1998).

Two separate studies evaluated the pulmonary and fiber kinetic effects of either MMVF34 (HT, high alumina, low-silica rock wool fiber) or MMVF21 (rock wool) following nose-only inhalation exposure to male rats for 6 h/day, 5 days per week for 104 weeks with sacrifices at 13, 26, 52, and 104 weeks (Kamstrup et al. 2001). Minimal cellular changes were seen with MMVF21 at three months, minimal fibrosis at 18 months, and a 0.5% incidence of fibrosis in the second study at 24 months. Normal to minimal fibrosis was noted with MMVF34 in the first study, whereas in the second study, fibrosis incidence was 5.04% at 18 months but appeared to be resolving by 24 months. Tumor incidence was only evaluated in the later study and was found not to be statistically significant from controls (MMVF21 − 3.5% adenomas, 0.9% carcinoma; MMVF34 − 4.7% adenomas, 0 carcinomas) (Kamstrup et al. 2001). Lung burdens steadily increased for both fibers at each time point with MMVF34 fibers peaking at 18 months then declining by 24 months, while MMVF21 lung burdens continued to increase through 24 months. MMVF34 demonstrated WT_1/2 values for WHO fibers and long fibers (L > 20 µm) of 25 and 6 days, respectively, while MMVF21 stone wool demonstrated a much slower clearance rate, with elimination halftimes of 65 days (WHO fibers) and 92 days (L > 20 µm) (Kamstrup et al. 1998). Lung fiber clearance of long fibers (L > 20 µm) was faster for MMVF34 compared to MMVF21, mild fibrosis occurred with MMVF21, whereas MMVF34 inhalation resulted in minimal to mild cellular changes (macrophage infiltration). The tumor incidence was similar for MMVF21 (3.5% adenoma, 0.9% carcinoma) and MMVF34 (4.7% adenomas, no carcinomas) with neither significantly different from controls (Kamstrup et al. 1998; 2001). In another study, MMVF34 showed significant leaching from fibers, further demonstrating characteristics of a fiber with low biopersistence (Hesterberg, Chase, et al. 1998). MMVF21, with a high silica and low alumina content, was significantly less soluble in vitro at pH 4.5 and therefore less susceptible to dissolution and breakage when exposed to an acidic environment. HT stone wool (MMVF34) was observed to have low biopersistence in the rat lung (weighted retention half-time of fibers >20 μm, 6 days), but has a slow dissolution rate in vitro at pH 7.4 (59 ng/cm²/h) and a substantially faster dissolution rate at pH 4.5 (620 ng/cm²/h) (Etherington et al. 1981; Kamstrup et al. 2001; IARC 2002).

Biopersistence of four experimental rock wools (labeled F, G, H, and L) was evaluated in rats (5 days of exposure, for 6 h per day) with post-exposure monitoring up to 1 h, 1 or 5 days, or 4, 13, or 26 weeks (Bernstein et al. 1996). The highest lung burdens were found at both 1-h and 26-weeks post exposure for F stone wool at 15.6 × 10⁶ and 0.62 × 10⁶ total fibers/lung, respectively; the lowest in L stone wool at 5.7 × 10⁶ total fibers/lung at 1 h and 0.30 × 10⁶ fibers/lung at 26 weeks. It is important to note, however, that both F and G stone wools had higher aerosol fiber numbers compared to L fibers. The WT _1/2 ranged from 5.4–45 days for fibers >20 μm in length, with L stone wool having the greatest lung persistence and G stone wool the shortest. Values for WT_1/2 and K_dis at pH 7.4 and pH 4.5:

• F rock wool – WT_1/2 8.5 days (fibers >20 µm long) – K_dis 96 ng/cm²/hr at pH 7.4 and 5.9 ng/cm²/hr at pH 4.5

• G rock wool – WT_1/2 5.4 days (fibers >20 µm long) – K_dis 129 ng/cm²/hr at pH 7.4 and 4.1 ng/cm²/hr at pH 4.5

• H rock wool – WT_1/2 13 days (fibers >20 µm long) – K_dis 169 ng/cm²/hr at pH 7.4 and 5.5 ng/cm²/hr at pH 4.5

• L rock wool – WT_1/2 45 days (fibers >20 µm long) – K_dis 23 ng/cm²/hr at pH 7.4 and 342 ng/cm²/hr at pH 4.5

The clearance rate of WHO fibers was found to closely reflect the clearance rate of fibers in the 5–20 μm range and disappearance of fibers longer than 20 µm mediated by their dissolution rates at pH 7.4. Lack of correlation for clearance of shorter fibers at pH 4.5 is likely due to clearance mediated by macrophages and lymphatic transport to the lymph nodes and BALT as well as dissolution (Bernstein et al. 1996). Dissolution rates are also impacted by composition of these fibers, with L rock wool having a higher percentage of aluminum oxide (13.5%) as compared to G, F, and H rock wools (0.45, 3.15, or 3.9%, respectively), which is a factor known to decrease in vitro dissolution. In contrast, presence of higher concentrations of alkali and alkali Earth metals including magnesium and calcium oxides will increase dissolution. For example, F, G, and H stone wools have 2-2.5x calcium oxide content compared to L stone wool (Bernstein et al. 1996; Maxim et al. 2006).

In a study by Bellmann et al. (2003), Wistar rats were exposed by nose-only to fiber aerosol concentrations of approximately 15, 50, and 150 fibers/cm³ (fiber length >20 μm) MMVF21 stone wool or 150 fibers/cm³ of a new high-temperature application fiber (calcium–magnesium–silicate fiber for CMS) and followed for 3 months post-exposure. After 3 months of recovery, long fiber fraction concentration was decreased to 63.9 and 3.0% compared to original lung burden for MMVF21 and CMS, respectively. Dose-dependent effects of MMVF21 exposure include increase in lung weight, in measured biochemical parameters and polymorphonuclear leukocytes (PMN) in the bronchoalveolar lavage fluid (BALF), in cell proliferation (BrdU-response) of terminal bronchiolar epithelium, and interstitial fibrosis. As CMS fibers are biosoluble (WT_1/2 9.8 days, compared to 67 days for E-glass and 79 days for MMVF21), fibrogenic potential was anticipated to be low but was nonetheless detected. A possible explanation for fibrosis observed with the CMS fibers (despite its rapid in vivo clearance) is the presence of a nonfibrous particle fraction (mass of lung fibers 1 wk after exposure was 1.02 mg as compared to 1.879 mg for particles) of CMS (with the same chemical composition as the fibrous fraction) (Bellmann et al. 2003).

Overall, the historical inhalation toxicology studies of rock and slag wools show that some fibers can elicit fibrosis and tumors, however this generally associated with fibers with lung clearance half-times (WT_1/2) greater than 40 or 50 days (MMVF21). With respect to in vitro dissolution data, most studies tested stone and slag wool fibers in pH 7.4 and far fewer studies conducted in vitro dissolution in pH 4.5. Given that in vivo fiber half-life of stone wool fibers is not well correlated with in vitro fiber dissolution at pH 7.4, further research is needed to characterize stone wool fibers and the correlations between in vivo fiber clearance and in vitro dissolution at pH 4.5. Despite this, fibers such as MMVF34 HT rock wool showed fast in vivo clearance (WT_1/2 6 days), rapid in vitro dissolution (K_dis 620 ng/cm²/hr at pH 4.5), which translated to a lack of inflammatory and fibrotic responses. Thus, while more research is needed to bridge the gap between in vitro dissolution of stone and slag wool fibers at pH 4.5 and in vivo biopersistence, available studies suggest that stone and slag wool fibers with a fiber clearance half-life less than 10 days did not show fibrotic or tumorigenic responses.

Refractory ceramic fibers

A number of different RCF were evaluated in historical animal inhalation studies, including RCF1, RCF1a, kaolin-based RCF, zirconia based RCF, high-purity RCF, after-service RCF, and high purity aluminum zirconia silica (AZS) RCF (Hesterberg et al. 1993; Mast et al. 1994; Mast et al. 1995a, 1995b; Everitt et al. 1997; Hesterberg, Chase, et al. 1998; Hesterberg, Hart, et al. 1998; Brown et al. 2000; Bellmann et al. 2001; Muhle and Mangelsdorf 2003; Bernstein et al. 2005; Brown et al. 2005). While there are several inhalation animal studies which evaluated the pulmonary effects of RCF, very few studies assessed the relationship between in vivo fiber kinetics, in vitro fiber dissolution, and fibrogenic and tumorigenic potential of RCF.

Early studies of RCF1 showed pulmonary fibrosis, lung tumors (both adenomas and carcinomas), and mesotheliomas (Hesterberg et al. 1993; Everitt et al. 1997; Hesterberg, Chase, et al. 1998). RCF1 has negligible amounts of alkaline earth metals (slightly over 1%) that would normally confer increased solubility and relatively high levels of SiO₂ and Al₂O₃ (47.7 and 48%, respectively) which increases biopersistence (Bernstein et al. 2005). While RCF1 and RCF1a have comparable compositions, RCF1 has the presence of a relatively large percent of non-fibrous particles (25%) as compared to other RCF1a formulations (2% non-fibrous particles) (Brown et al. 2000; Bellmann et al. 2001). The presence of the non-fibrous particles can not only have a significant effect on the clearance rates from the lungs, but also influences the risk of inflammation, fibrosis and tumors. For example, in a study by Bellmann et al. (2001), female Wistar rats were exposed 6 h/day for 3 weeks to either RCF1a or RCF1 fiber aerosol of approximately 125 fibers (>20 µm long)/cm³. The results from this study indicated that particle fraction of RCF1 significantly enhanced adverse effects (Bellmann et al. 2001). In another Bellmann et al. study, a 3 week nose-only inhalation exposure to RCF1 and RCF1a (730 fibers/cm³ >20 µm length) resulted in more persistent inflammation with RCF1 compared to RCF1a, which showed transient increases in lymphocytes and decreases in macrophages that were resolved by 90 days post-exposure (Brown et al. 2000; Bellmann et al. 2001). The differential biological effects of RCF1 versus RCF1a were related to the increased biopersistence of RCF1 (and confounding of non-fibrous particulate). Alveolar clearance of tracer particles was slowed after RCF1 exposure (1,200 days clearance half time) compared to the exposure to RCF1a (80 days clearance half time) (Muhle and Mangelsdorf 2003). Thus, the significant composition of non-fibrous particles (25%) in the RCF1 formulation likely contributed to a lung overload condition, which slowed tracer particle clearance and further enhanced the fibrogenic and tumorigenic response. Additionally, in vivo half-life of RCF1a was 55 days (WT_1/2, L > 20 µm; 90% clearance (T90) was 227 days for L > 20 µm) and in vitro dissolution (K_dis) at pH 7.4 was 3 ng/cm²/hr (which is still a relatively slow dissolution rate) (Hesterberg, Chase, et al. 1998).

Exposure to other RCF resulted in fibrosis (kaolin-based RCF, kaolin RCF, After Service RCF) and tumors (kaolin RCF, zirconia RCF, High-Purity RCF, After Service RCF), however there were no apparent correlations between lung burdens (retention) and tumor incidence (Mast et al. 1994). In another study by Mast et al. (1995a, 1995b), kaolin-based RCF, zirconia-based RCF, high purity AZS size-selected RCF fibers, or an after-service heat-treated (2400 °F for 24 h) kaolin-based fibers resulted in inflammation, interstitial fibrosis (all fibers except after-service fiber), pulmonary neoplasms (bronchoalveolar adenomas and carcinomas combined), and pleural mesotheliomas (kaolin, AZS, high-purity, and after-service fibers). No biopersistence or dissolution data were available and no chemical composition data were listed. It is noteworthy that mathematical modeling data indicate impaired clearance occurred at concentrations at or exceeding 30 mg/m³ with an approximate half-life of 200 days; lower doses had estimated half-lives of 140 days (Mast et al. 1994).

Other fibers

A rapidly dissolving fiber with high silica and glass-like characteristics, X607, was evaluated for biopersistence (Hesterberg, Hart, et al. 1998). Following two years of nose-only inhalation exposure to rats, mean levels were 275 × 10³ WHO fibers/dry lung (48 × 10³ fibers longer than 20 µm/dry lung). X607 was neither fibrogenic nor tumorigenic and induced only minimal lung cellularity that reversed after exposure was terminated (78 weeks of exposure, 26 weeks recovery). Evidence of rapid leaching (loss of calcium), a WT_1/2 of 9.6–9.8 days, and a rapid rate of in vitro dissolution (K_dis = 990 ng/cm² per hour) as well as evidence of transverse fragmentation are indicative of the low biopersistence of X607 fibers (Hesterberg, Hart, et al. 1998).

In summary, the inhalation route of exposure is the most relevant for comparison to human exposure. Despite this, there has been controversy as to which route of exposure is most appropriate. Muhle and Pott (2000) opined that the inhalation model in rats is not sufficiently sensitive to predict cancer risk (other than for asbestos) and proposed intraperitoneal injections would be more appropriate. In contrast, Maxim et al. (2002) argued well-conducted inhalation studies of carcinogenicity are sufficiently sensitive and that rats may be more sensitive than humans to the carcinogenic potential of SVFs. Very-high-dose experiments in toxicology must always be suspected of causing effects through mechanisms irrelevant in more "normal" exposures (Hext 1994). In any case, available information suggest that in vitro fiber dissolution rates greater than 100 ng/cm²/hr of glass fibers in pH 7 and in vivo fiber clearance less than WT_1/2 40 or 50 days are not associated with fibrosis or tumors. While there is some suggestion that these same thresholds apply to stone wool fibers, more research is needed to correlate in vitro dissolution rates of stone SVFs at pH 4.5 with in vivo clearance half-life and biological effects.

Correlations between biodurability, biopersistence and health effects

Dissolution, physical breakage, and clearance of SVFs from the lungs is a complex interaction between fiber length and chemical composition. As described above, through a series of inhalation toxicology studies of SVFs, generally referred to as the RCC studies, the relationship between exposure to long, durable fibers, and the development pulmonary disease became apparent.

The dissolution of SVFs is dictated by its chemistry in aqueous and lung fluids. The amorphous nature of SVF in a randomly oriented silicon oxygen tetrahedron aid in the dissolution of silicon (H₄SiO₄) and associated fiber elements into aqueous solution (Bernstein 2007). Once a fiber is deposited in the lungs, it will encounter the neutral pH of the lung lining fluid and the acidic environment of the macrophage phagolysosome. Lung lining fluid is critical to the function of the lung; it provides an important barrier to the exogenous environment, is comprised of phospholipids and proteins to reduce surface tension and maintain alveoli in an open formation for gas exchange, and modulates innate immune function (opsonization, enhancement of phagocytosis, and regulation of immune signaling) (Martin and Poynter 2016). The pH extracellular lung fluid is maintained at pH 7.4–7.5, whereas the pH of the macrophage phagolysosome is pH 4.5–5 (Oberdorster 2000; Effros et al. 2005). Glass fibers generally dissolve in the neutral pH of the lung lining fluid, whereas rock wool fibers readily dissolve in the acidic environment of the macrophage phagolysosome (Figure 2).

Figure 2. Mechanisms of dissolution SVFs depend on the length of the fiber and chemical composition. Irrespective of the composition, short fibers are cleared by direct or macrophage-mediated transport out of the lungs. For long glass SVFs, dissolution and subsequent breakage into shorter fibers occurs in extracellular space and neutral pH 7.4 of the lung lining fluid. In contrast, stone SVFs generally undergo minimal dissolution at neutral pH, but are readily dissolved and broken into shorter fiber segments in the acidic environment (pH 4.5) of the macrophages phagolysosome.

The differential processes of fiber dissolution of glass versus rock wool SVFs is important in considering the correlation between dissolution rates measured in vitro and biopersistence and health effects observed in vivo. With glass SVFs, long fibers (>20 µm in length) undergo dissolution in the lung lining fluid, which causes breakage into shorter fibers which are in turn cleared from the lungs through physical transport and/or macrophage phagocytosis (Figures 1 and 2). Because rock wool fibers undergo dissolution in acidic rather than neutral conditions, lung clearance of long rock wool fibers (>20 µm in length) is driven by dissolution and breakage at or near macrophage engulfment sites along the fiber, which in turn will result in removal from the lungs through similar physical transport/phagocytosis processes (Figures 1 and 2).

Dissolution rates (K_dis) are a function of fiber composition, and at pH 7.4 represent dissolution of long fibers not subject to alveolar macrophage removal, and at pH 4.5 generally signify dissolution of fibers in associated with macrophage phagocytosis (Maxim et al. 2006). For glass SVFs, dissolution occurs faster at neutral pH outside of the macrophage than inside the acidic environment of the macrophage phagolysosome, and thus long fibers are dissolved into shorter fibers outside of the macrophage, resulting in short parts of fibers inside the macrophage. For stone wool SVFs, dissolution occurs faster inside the acidic environment of the macrophage rather than the neutral pH of the lung lining fluid. As a result, stone wool fibers are dissolved inside the macrophage and fiber parts outside of the macrophage are mostly intact, which in turn are phagocytized by and dissolved in macrophages (Figure 2) (Bellmann et al. 2010). Although, long fibers that are partially phagocytosed by macrophages may also experience breakdown and weakening due to a release of an acidic milieu in the macrophage extracellular microenvironment.

SVF composition can vary depending on the desired commercial properties (tensile strength, elasticity, thermal conductivity, heat capacity, and end use temperature), which in turn can influence the propensity for dissolution. For example, CaO, MgO, Na₂O, B₂O₃, and BaO can increase the dissolution rate of borosilicate glass fibers, whereas Al₂O₃ decreases the dissolution rate (Eastes et al. 2000). The ratio of alumina and silicon (Al³⁺/Al³⁺ + Si⁴⁺) can influence the dissolution behavior of SVFs. SVFs with an alumina content of less than 5% (ratio below about 0.07) corresponds to a decreasing dissolution rate at neutral pH, whereas for medium- and high-alumina fibers (ratio above 0.2), the dissolution rate increases with increasing Al³⁺/Al³⁺ + Si⁴⁺ ratio. In acidic pH 4.5, the dissolution rate is low until the Al³⁺/Al³⁺ + Si⁴⁺ ratio reaches about 0.25 (when Al³⁺ is about one-third of Si⁴⁺) (Guldberg et al. 2000). As a result of these physicochemical factors which influence biodurability and biopersistence, the EC adopted a criteria for composition for classification, labeling and exemption of SVFs as carcinogens referred to as “Note Q” (European Economic Community 1997). The basis for this regulation is that the chemical composition of vitreous fibers influences its biopersistence. SVFs are grouped according to the sum of weight of oxides of sodium, potassium, calcium, magnesium and barium. If this sum of oxides exceeds 18%, then carcinogen category 3 (suspected animal carcinogen) applies: if the sum is below 18%, then carcinogen category 2 (animal carcinogen) applies (European Economic Community 1997). Fibers classified as carcinogen category 3 can be exempt if it is demonstrated that biopersistence is low (T_1/2 F > 20 µm = 10 days by inhalation, T_1/2 F > 20 µm = 40 days by intratracheal instillation) (European Economic Community 1997).

Fiber length

Since the early 1970s, it has been acknowledged that long, thin fibers present a greater risk of disease than shorter fibers (Stanton 1973; Stanton et al. 1977; Wagner et al. 1980; Stanton et al. 1981b; Lippmann 1988; ATSDR 2002; ERG 2003). Stanton and colleagues conducted a series of pleural implantation experiments with asbestos and non-asbestos fibers and ultimately concluded that fibers less than or equal to 8 μm in length (with diameters of less than 1.5 μm) are inactivated by phagocytosis and have little or no mesotheliogenic potential (Stanton 1973; Stanton et al. 1977, 1981b). Following an analysis of fiber length distribution data from rat inhalation studies using amosite, brucite, chrysotile, crocidolite, erionite, and tremolite, Lippmann concluded that the concentration of fibers longer than 10 or 20 μm in length is a better predictor of lung tumor incidence than is the concentration of fibers longer than 5 μm (Lippmann 1994). Additionally, comparison of long (>6 µm length) versus short (3–5 µm length) erionite fibers (generally thought to be one of the most potent fiber types for mesothelioma risk) demonstrated that no cases of mesothelioma developed among 24 rats exposed by inhalation to shortened erionite and almost a 100 percent incidence of mesothelioma occurred among 27 rats exposed to longer-fiber erionite (Wagner 1990). From the Wagner research, it was suggested that short fibers will not cause mesothelioma regardless of the potency of longer fibers of the same fiber type. Further, the compilation of epidemiology, toxicology, and in vitro research of asbestos over the last 40 years has demonstrated that chrysotile and amphibole fibers with lengths <5–10 µm do not contribute to the health effects of asbestos (Bernstein 2022).

Studies by Donaldson et al. (2010) have provided additional insights as to how biodurable, biopersistent long fibers induce effects on the pleura, whereas degradable and easily cleared fibers do not. As evidenced by “black spots” on the parietal pleural wall at autopsy of urban dwellers, a fraction of all deposited particles reach the pleura and through normal pathways of clearance exit the pleura through stomata in the parietal pleura. The stomatal openings (<10 µm in diameter) on the parietal pleura drain fluid from the pleural space into lung lymph nodes. It was proposed by Donaldson and others that if particles are too large or too elongated to navigate through the stomatal opening, accumulation can occur leading to inflammation and pleural pathology (Donaldson et al. 2010). While the impetus for this original hypothesis and research was to explore how high aspect ratio nanoparticles (HARN), such as carbon nanotubes, may fit into the classic fiber toxicology paradigm, it also has implications for understanding the biological effects of biodurable long fibers retained in the pleura space (Donaldson et al. 2010).

Using intrapleural or intraperitoneal injection in mice as a model to evaluate effects on mesothelial cells of the lining of the lung or peritoneum, respectively, has consistently shown across different particle types that long fibers (greater than 15–20 µm length) induce greater inflammatory and granulomatous reactions than short fibers (less than 5 µm in length) (Murphy et al. 2011; Osmond-McLeod et al. 2011). In particular, long amosite fibers injected directly in the pleural cavity showed inflammation and granuloma reactions leading to progressive fibrosis, whereas short amosite fibers did not (Murphy et al. 2011). Interestingly, polystyrene beads (generally considered an innocuous and inert particle) injected as 10 µm diameter sized particles into the pleura cavity, resulted in an inflammatory response, whereas quartz (a highly surface reactive particle), coal mine dust, or polystyrene beads of 3 µm in diameter or less showed no inflammatory response (Murphy et al. 2011). These studies have supported and confirmed the hypothesis that although a portion of all deposited particles clear through the pleura, the pathogenicity of high aspect ratio particles arises as a result of length-dependent retention of durable fibers at the stomata on the parietal pleura (Murphy et al. 2011).

Test methods for biodurability and biopersistence

Animal toxicology research of mineral fibers, such as asbestos, and SVFs underscore the correlation between fiber dissolution (biodurability), lung clearance (biopersistence), and health effects. Long durable fibers result in increased persistence in the lungs, which in turn leads to an increased risk of fibrosis and tumors. The historical research on SVFs has served as a basis for developing and utilizing in vivo test methods to characterize fiber biopersistence and in vitro test methods to assess biodurability with the overall intent to evaluate safety and restriction of SVFs that may present a risk for disease. As the historical research has illustrated, long durable fibers (>20 µm length) produce an increased incidence of fibrosis and tumors in animals exposed by intratracheal instillation or inhalation, whereas short fibers (<20 µm) of the same fiber type do not produce these same effects. However, soluble (non-durable) SVFs >20 µm length are cleared from the lungs and do not result in fibrosis or tumors following chronic exposures. Figure 3 illustrates that there is an inverse relationship between in vivo fiber half-life and dissolution of glass fibers at pH 7.4 in that slow dissolution rates of fibers >20 µm length translate to a longer residence time in the lungs, longer fiber half-lives, and increased propensity for disease (Table 2). It should be noted, however, that dissolution may also be influenced by fiber width (thicker fibers take longer to dissolve than thinner fibers). SVF fibers evaluated in the literature do not all have the same diameter, and thus, variations in the correlation between in vitro dissolution and in vivo half-life may be related the fiber dimensions as well as experimental setup of the dissolution assay (as well as other factors as discussed later). While fewer studies are available for stone wool fibers, available studies of in vitro dissolution at pH 4.5, in vivo biopersistence, and biological effects of stone wool fibers show similar trends – increasing dissolution at pH 4.5 was associated with shorter fiber half-lives, and decreased biological responses (Figure 4, Table 3). No in vitro dissolution-in vivo biopersistence patterns were observed at pH 7.4 with stone wool fibers. From these studies (Table 1), in vitro dissolution rate of 100 ng/cm²/hr or greater and an in vivo half-life less than 40 or 50 days following inhalation exposure correlated with a low probability of producing fibrosis and tumors in animal inhalation studies (Eastes and Hadley 1996). Considering the low likelihood that fibrosis and tumors will develop from inhaled fibers of >20 µm length with a half-life of 40 or 50 days or less, the EU Nota Q Directive 97/69/EEC criteria of a weighted half-life of less than 10 days from a short-term inhalation biopersistence test (fibers >20 µm length) to exempt carcinogenic classification of fibers is conservative and provides a margin of safety.

Figure 3. Correlation between in vitro dissolution at pH 7.4 (K_dis, ng/cm²/hr), in vivo clearance (WT_1/2 fibers >20 µm in length) of glass SVFs, and corresponding biological effects (when available, Table 2). The vertical dotted line represent the EU Note Q WT_1/2 threshold (10 days for inhalation, 40 days for intratracheal instillation studies) requirement of exemption of classification for carcinogenicity. Glass SVFs (MMVF32, MMVF33) that demonstrated fibrotic and tumorigenic responses showed K_dis less than 100 ng/cm²/hr and WT_1/2 (inhalation and/or intratracheal instillation) greater than 40–50 days. The confluence of the K_dis (>100 ng/cm²/hr, as illustrated by horizontal dotted line) and WT_1/2 (>40–50 days) thresholds that are not associated with biological effects is represented by the blue shadowed area.

Figure 4. Correlation between in vitro dissolution at pH 4.5 (K_dis, ng/cm²/hr), in vivo clearance (WT_1/2 fibers >20 µm in length) of stone SVFs, and corresponding biological effects (when available, Table 3). The vertical dotted line represent the EU Note Q WT_1/2 threshold (10 days for inhalation, 40 days for intratracheal instillation studies) requirement of exemption of classification for carcinogenicity. Stone SVF (MMVF21) that demonstrated the most significant fibrotic responses (tumorigenic responses were not significant) showed K_dis less than 100 ng/cm²/hr and WT_1/2 (inhalation) greater than 40–50 days. The confluence of the K_dis (>100 ng/cm²/hr, as illustrated by horizontal dotted line) and WT_1/2 (>40–50 days) thresholds that are not associated with biological effects is represented by the blue shadowed area. It is noteworthy that correlations were observed between in vitro K_dis at pH 4.5 (but not pH 7.4) and in vivo WT_1/2 for stone SVFs, however limited data are available for in vitro dissolution at pH 4.5, in vivo clearance, and in vivo health effects (Table 3) for a given stone SVF.

Fiber dissolution test methods

In vitro dissolution methods have been used for at least a couple decades to characterize the biodurabilty and dissolution behavior of SVFs, but also used as a tool to predict the biopersistence and biological effects in vivo. Static and dynamic flow systems have been used to measure the dissolution constant of SVFs. The general principal of these test methods is that dissolved amounts of silicon and other elements are measured in the effluent at regular time intervals in a system in which a static or continuous constant flow (dynamic) of buffer is moved across the sample of fibers. While a static system (a single finite volume of fluid is mixed with the SVF sample) is generally considered a more simplistic approach than a dynamic system (fluid actively flows across the SVF sample), there are complexities with both systems which can significantly influence consistency of the fiber dissolution rate and potential correlation with in vivo measures of biopersistence and biological effects. For example, fiber sample preparation (fiber mass, dimension, and surface area), buffer solution composition (ionic strength, complexing agents, buffer capacity), buffer volume, pH, ratio of flow rate to fiber surface area, temperature, system design features (fiber cassette characteristics, agitation, solution sampling and analytical methods), and analytical method can greatly influence the assay results. While a governmental standardized method has yet to be developed for evaluating SVF dissolution in vitro, the industry has a test method guideline “In-vitro acellular dissolution of man-made vitreous silicate fibres” (Sebastian et al. 2002). In the early 2000s, the European Insulation Manufacturers Association (EURIMA) undertook a round robin testing program to evaluate the inter- and intra-laboratory variability of two round robin testing regimens among nine and six different laboratories with two glass wool samples (traditional and soluble SVFs at pH 7.4) and two stone wool samples (traditional and soluble [high alumina/low silica] at pH 4.5) (Guldberg et al. 2003). A specified mass was placed in filter holds (“flow cells”), artificial lung fluid (Gamble’s liquid) continuously flowed through the flow cells, and silicon in the effluent was measured in intervals up to 14 days. The results from this round robin testing demonstrated that there was relatively good reproducibility within a given laboratory, but the inter-laboratory variability was relatively high, ranging from 24 to 61 percent. It was concluded that despite the difference between laboratories (which may be relatively unimportant in the broader solubility assessment), the method still may be adequately valid for ranking different fibers with respect to dissolution coefficients. It was also noted that it is important to test fibers in both neutral and acidic pH, because the neutral pH of the lung lining fluid was important for the dissolution of glass wool fibers (low alumina fibers), whereas the acidic environment was important for dissolution of stone wool fibers (high alumina/low silica fibers) (Guldberg et al. 2003).

Regulatory acceptability of an acellular in vitro dissolution assay will require that the test method has robust reliability and predictability for in vivo biopersistence and biological effects. In vitro dissolution (K_dis) values can vary significantly across SVFs, particularly between stone and glass fibers at a given pH. In a publication by Bellmann et al. (2010) comparing calculated dissolution (K_dis) with in vivo biopersistence (WT_1/2) for multiple SVF types showed that a K_dis threshold of 100 ng/cm²/hr or greater for glass fibers was consistently associated with an in vivo half-life less than 40 days for fibers >20 µm in length (Figure 5). It is noteworthy that the K_dis values in Bellmann et al. (2010) were modeled and not measured. It is also noteworthy that few SVF types would meet the K_dis 100 ng/cm²/hr and WT_1/2 thresholds if WHO sized fibers were considered. This finding is consistent with the recommendation from Eastes and Hadley (1996) that an in vitro dissolution rate of 100 ng/cm²/hr correlates with an acceptably low probability of producing fibrosis and tumors in animal inhalation studies. However, Bellmann et al. (2010) showed that some stone wool fibers had a fast predicted K_dis (>100 ng/cm²/hr), but a relatively long half-life in in vivo biopersistence studies. While this disparity between calculated K_dis and in vivo clearance of some stone wool fibers needs to be further understood, it may be that actual measured in vitro dissolution of stone wool in buffered solution at pH 4.5 could better align with in vivo biopersistence. In fact, it was noted by Eastes et al. (2000) that K_dis equations and coefficients for weight of oxides may not adequately fit for all rock and slag wool compositions, because of the lack of boron, lower silica concentrations, and higher and more variable alumina concentrations (0–50%) compared to glass wool fibers. Irrespective of the alignment of measured versus calculated K_dis or the variability that is observed with measured in vitro dissolution assays across different laboratories, continuing to correlate in vitro dissolution and in vivo biopersistence with the presence or lack of biological effects in inhalation and intratracheal instillation animal studies will increase confidence in acellular in vitro systems and reinforce critical in vitro K_dis thresholds to adequately predict health effects and product safety. Further, it may be that there is a reasonable range of variance of in vitro dissolution results across laboratories that can still be within acceptable thresholds which are correlated with in vivo biopersistence and responses.

Figure 5. Correlation of fiber dissolution rate (K_dis, calculated) and fiber clearance from the lungs of animals (WT_1/2, weighted half-life) for multiple SVF types show that a K_dis threshold of 100 ng/cm²/hr or greater for glass fibers was consistently associated with an in vivo half-life less than 40 days for fibers >20 µm in length. This is consistent with the Note Q threshold requirements for exemption of fibers for carcinogenicity classification in the EU. Some stone wool fibers were predicted to have a relatively fast K_dis (>100 ng/cm²/hr), but showed a relatively long half-life. While this disparity between calculated K_dis and in vivo clearance of some stone wool fibers needs to be further understood, it may be that actual measured in vitro dissolution of stone wool in buffered solution at pH 4.5 could better align with in vivo biopersistence. It is noteworthy that few SVF types would meet the K_dis 100 ng/cm²/hr and WT_1/2 thresholds if WHO sized fibers were considered.

Biopersistence test methods

In vivo inhalation and intratracheal instillation studies to evaluate biopersistence and biological effects of SVFs are still the gold standard upon which to exempt classification and labeling of SVFs for carcinogenicity. However, with the mission to reduce the use of animal testing, international regulatory agencies, including in Europe and the U.S., have committed to supporting the development of alternative testing strategies in the reduction, refinement, and replacement of use of animals in research. In fact, in 2013, EU executed a complete ban on animal testing for finished cosmetics products and cosmetic ingredients (European Commission 2009). As a result of this directive, the European center for the validation of alternative methods (ECVAM) serves an important role in the development, validation, and international recognition of alternative methods to reduce, refine, and replace the use of animals in testing. While there do not appear to be any initiatives to replace the animal testing requirement of SVFs for exemption under the EU Nota Q Directive, standardization of in vitro methods to predict the effects in animals and humans would be beneficial to reduce resources and safeguard product safety and innovation in the event that regulators seek to reduce animal testing of SVFs.

In 1999 as a part of the Directive 97/69/EC, the European Commission issued animal testing protocols for four tests, the results of which exonerate synthetic mineral fibers for classification as carcinogens (Bernstein and Riego Sintes 1999). These four animal testing protocols include (1) Biopersistence of Fibers – Short term exposure by inhalation [ECB/TM/26 rev.7], (2) Biopersistence of Fibers – Intratracheal instillation [ECB/TM/27 rev. 7], (3) Carcinogenicity of synthetic mineral fibers after intraperitoneal injection in rats [ECB/TM/18(97) rev. 1], and (4) Chronic inhalation toxicity of synthetic mineral fibers in rats [ECB/TM/17(97) rev. 2]. For the inhalation study protocol, rats (Fischer 344 or Wistar) are exposed by nose-only inhalation 6 h/day for 5 days to fiber test atmospheres optimized for rat respirability (aspect ratio 3:1, 100 fibers/cm³ >20 µm length, gravimetric mean concentration of fibers 0.8 µm should not exceed 40 mg/m³, all particulates should not exceed 60 mg/m³). Subgroups of animals (at least 7 per group/time point) are evaluated for fiber and particulate burden (which involve counting of fibers/particles of a digested tissue sample) at various time intervals (1, 2, 3, 14 days, 4 weeks, 3, 6, and 12 months) for determination of the time for removing 50% of fibers (T_1/2) which are longer than 20 µm. T_1/2 of other fiber length fractions <5 µm and 5–20 µm (WHO fibers) should also be determined. MMVF21 and MMVF10a are recommended as a validation material (Bernstein and Riego Sintes 1999).

For the IT study protocol, rats (Fischer 344 or Wistar) are intratracheally administered fibers (0.5 mg and 2 mg) for 4 consecutive days and examined for lung fiber burden at post-exposure time points comparable to the inhalation study protocol (Bernstein and Riego Sintes 1999). Instilled fibers should be suspended in saline and have a mean aspect ratio of 3:1, an upper limit of 3 µm (95% less than 3 µm), and at least a 20% of WHO (L > 5 µm, D < 3 µm) fibers should have lengths >20 µm and geometric mean diameter as close to 0.8 µm as possible. Similar to the inhalation study protocol, IT studies should report the T_1/2 of fibers longer than 20 µm, as well as of lengths <5 µm and 5-20 µm (WHO fibers). The IP injection study protocol involves evaluation of the tumorigenic response in the peritoneal cavity in rats following one or more IP injections of a well characterized SVF. Animals (Wistar rats, 50 per treatment group) are followed throughout their lifetime until a total of 20% survival occurs in the exposure group. Criteria for the fiber sample preparation are the same as that outlined for the IT study protocol with delivery of a single dose of 1 x 10⁹ WHO fibers per animal. Histopathology is assessed by reporting the number, type, and percentage of each type of lesion.

Finally, the chronic inhalation toxicity study involves inhalation of SVF for 2 years for evaluation of fiber lung burden by fiber size fraction and prevalence of fibrosis, adenocarcinoma, and mesothelioma. Wistar rats (100 per group) are exposed by inhalation to at least 100 fibers/cm³ (L > 20 µm with GMD close to 0.8 µm) for 6 h/day, 5 day/week for 104 consecutive weeks to a well characterized fiber test atmosphere optimized for rat respirability and followed over their lifetime. Fibers such as E-glass, MMVF21, or MMVF10a are recommended as a validation material. Similar to the inhalation biopersistence protocol, fiber burden of fibers <5, 5–20 and >20 µm in length should be reported, along with the histopathology (Bernstein and Riego Sintes 1999).

The more recent inhalation studies of SVFs (as described in the previous sections) generally followed these EC protocols and support that a half-life of 10 days or faster is a conservative threshold for which the tumorigenic risk is negligible. Interestingly, when WT_1/2 of SVFs are compared for fibers administered by inhalation versus IT administration, there is a comparable relative ranking between the exposure methods, but the actual WT_1/2 values vary significantly. Figure 3 and Table 1 clearly show that WT_1/2 of MMVF32, MMVF33, MMVF11, and C Glass fibers by inhalation exposure have a similar ranking as exposures by IT instillation. However, the actual WT_1/2 values are dramatically different. For example, MMVF11 WT_1/2 >20 µm is 8.7–11 days by inhalation and 84 days by IT, whereas MMVF32 WT_1/2 >20 µm is 79 days by inhalation and 222 days by IT. The method of exposure for biopersistence studies may have implications for carcinogenicity exemption of SVFs. For example, MMVF11 by inhalation may be exempt according to Nota Q (based on criteria WT_1/2 >20 µm 10 days), but not by intratracheal instillation (based on criteria WT_1/2 >20 µm 40 days). While the EC protocols provide an opportunity to evaluate and exempt novel SVFs should they meet established thresholds, these types of studies require significant time and resources. Thus, standardizing and gaining regulatory acceptance of in vitro dissolution assays to predict in vivo effects would not only conserve animals for testing, but also provide more efficient research and development and safety assessment for novel SVFs for the marketplace. The lessons learned from the SVF research can also provide insights to other industries, such as engineered high aspect ratio nanomaterials, as they seek to design advanced but safe products for the marketplace. The following section will review the existing state of the science for biodurability and biopersistence methods of advanced nanomaterials and potential considerations and lessons learned that can be gained from the SVF industry experience on these topics. Additionally, considerations from the nanomaterials industry will be reviewed for the potential standardization of in vitro biodurability methods of SVFs for predicting biopersistence and health effects.

Applicability of SVF biodurability and biopersistence methods for advanced materials

The OECD WPMN has recommended testing methods to address biodurability and biopersistence of nanomaterials including: (1) dissolution and dispersion stability for environmental testing (OECD 2020b), (2) biodurability to characterize dissolution and biodegradation and predict toxicity and health effects (OECD 2018), and (3) biopersistence/biodurability to induce lysosomal membrane permeabilization (LMP) as a predictor of long-term toxic effects (OECD 2020a). LMP is the loss of membrane integrity which allows the release of luminal contents into the cytosol and can activate multiple stress response pathways including regulated cell death.

While biodurability and biopersistence methods developed for SVFs have some applicability for advanced nanomaterials, nanoparticles or nanofibers have exceptional physicochemical characteristics that require careful consideration. Nanomaterials span a seemingly endless array of sizes, shapes, surface chemistries, aggregation and agglomeration states, which can change from in situ conditions in their production to aqueous or airborne conditions in testing assays (e.g. in vitro or in vivo) or environmental exposure scenarios (e.g. occupational and consumer product settings). Thus, standardizing testing regimens to correlate dissolution behaviors of nanomaterials in simulated lung fluid to lung deposition-clearance-retention patterns (and ultimately correlate to biological effects) is no small challenge.

In the Guidance Document for the Testing of Dissolution and Dispersion Stability of Nanomaterials and the Use of the Data for Further Environmental Testing and Assessment Strategies, two methods are described for the determination of solubility of substances, a static batch test and a dynamic test (OECD 2020b). These test methods are intended to evaluate the dissolution rate and dispersion of nanomaterials in environmental aqueous media, which has limited applicability to the complex physicochemical milieu of the lung lining fluid. The primary limitation with static assays is with the restricted volume, of which the supernatant may become supersaturated inhibiting further dissolution (Campopiano et al. 2014). In the dynamic system, test media composition (pH, ionic or organic compounds), flow rates, test duration, sample concentration (mass, surface area), and test conditions (agitation) are study design features which can influence dissolution rates.

For both nanomaterials and SVFs, solubility (or biodurability) and dissolution rate can inform on further hazard testing strategies, but characterization (chemical composition, size distribution, morphology, surface modifications) of the test material is critical for appropriate interpretation of test results. Many of the dissolution studies of nanoparticles or nanofibers unfortunately do not report dissolution rates or constants so correlating physicochemical characteristics with dissolution kinetics is not possible (Utembe et al. 2015). Dissolution can be greatly influenced by media characteristics (pH, ionic strength, presence of dissolve organic matter, concentration of polyvalent ions, temperature, equilibrium time, analytical method), and intrinsic properties of the material (particle size, morphology, composition, crystallinity, surface modifications (Utembe et al. 2015). Additionally, ions and molecules (such as sulphides, chlorides, proteins, enzymes, polysaccharides) in the test media can influence dissolution (Utembe et al. 2015). As described in an extensive review of the chemistry and composition of simulated biological fluids, Innes and colleagues noted that particle and fiber dissolution is not always driven by pH and is not only driven by fluid composition (Innes et al. 2021). Thus, selecting an appropriate simulated biological fluid for in vitro dissolution studies is complex and needs validation, alignment and correlation with the in vivo experience (clearance, biopersistence). Additional challenges with nanomaterials are generating a representative sample for in vitro dissolution assays and inhalation studies particularly given that these materials have a high propensity for agglomeration. While the same laws of surface-area-normalized-dissolution-rate as is observed with SVFs and micro-sized particles may also apply to nanomaterials, resolution on the drivers for nanomaterial dissolution cannot be resolved until systematic material characterization is coupled with measured dissolution kinetics. Ultimately, correlations of in vitro nanomaterial dissolution kinetics with in vivo biopersistence kinetics will provide support of biodurability and biopersistence outcomes as predictive tools for health effects (similar to what have been established with SVFs). However, testing procedures and material characterization (as noted above) will likely need to be extensively investigated and harmonized before interpolations can be made for nanomaterial dissolution and health effects.

The OECD guidance document, Assessment of Biodurability of Nanomaterials and Their Surface Ligands, provides a compilation of biodurability information of pristine and functionalized nanomaterials in biological and environmental media in vitro and in vivo with a brief summary of methods for stability and halftimes (OECD 2018). Nanomaterial surface modifications (e.g. surfactants, ligands, coatings) were proposed as the primary determinants of nanomaterial biodurability. Nanomaterial surface modifications, such as use of surfactants, capping agents or attached ligands (small molecules or polymers that carry different functional groups) have been shown to influence dissolution or biodegradation (OECD 2018). Similar to larger particles and fibers, however, dissolution of nanomaterials is also expected to be affected by physicochemical properties, including size, shape, surface area, surface ligands, morphology, crystallinity, aggregation and agglomeration. Although there is not extensive research on correlations between nanomaterial dissolution rate, biopersistence, and health effects, there are examples (CeO₂ nanoparticles) in which low dissolution rate nanomaterials persist in the lungs and cause lung inflammation and granulomas (Keller et al. 2014).

As of present, in vivo biopersistence studies of high dissolution nanomaterials (e.g. ZnO, CuO) are generally lacking to determine whether nanomaterials follow a similar pattern as SVFs: high dissolution translates to low biopersistence and low risk of health effects, whereas low dissolution is associated with high biopersistence and increased risk of health effects. Nanomaterial dissolution and release of free or complexed metal ions (which induce cytotoxicity, oxidative stress, lysosome destabilization, inflammation) may suggest a more complex interplay between nanomaterial biodurability, biopersistence, and adverse health effects than what is observed with SVFs. Further research is needed on whether nanomaterial in vitro dissolution and toxicity assays may correlate and be predictive of in vivo persistence and long-term biological responses.

The OECD guidance document (as titled) further explores the Ability of Biopersistent/Biodurable Manufactured Nanomaterials (MNs) to Induce Lysosomal Membrane Permeabilization (LMP) as a Prediction of their Long-Term Toxic Effect as an alternative testing strategy for safety assessment (OECD 2020a). Lysosomal and autophagy dysfunction has been hypothesized as a mechanism of nanomaterial induced toxicity and pathology (Gulumian and Andraos 2018). Nanomaterials enter the cell through endocytic mechanisms into vesicular structures (endosomes, phagosomes) that merge into lysosomes. The nanomaterials are in turn degraded by different hydrolytic enzymes in an acidic environment. Biodurable nanomaterials, however, are resistant to biodegradation and may induce lysosomal and autophagy dysfunction and subsequent long-term toxicity. A common form of phagolysosomal dysfunction is LMP which has been associated with assembly and activation of inflammasome NLRP3 and chronic inflammation leading to pathological effects such as fibrosis and cancer (Sayan and Mossman 2016). This phenomenon is not specific to nanosized particles as crystalline micro-sized particles have been shown to induce LMP with subsequent NLRP3 inflammasome activation and has been suggested as the underlying mechanism for silica-induced fibrosis (Cassel et al. 2008; Dostert et al. 2008; Hornung et al. 2008; Franchi et al. 2009; Biswas et al. 2014). Because of the lack of understanding and potential complexities at play between nanomaterial durability, biopersistence, and health effects, research will need to first focus on the development of acellular in vitro dissolution assays that can predict nanoparticle biopersistence kinetics in vivo before cellular in vitro assays are able to predict long-term effects in humans.

As described by Donaldson et al. (2013), the biologically effective dose (BED) (internal dose that drives the critical pathophysiological outcomes) is critical for safety evaluation of any substance. For fibers, BED is driven by fiber dimension and durability, which influences the deposition, retention, persistence and fiber-cellular interaction that may or may not lead to toxicological effects. Whereas for nanomaterials, other physicochemical factors may come into play when considering the appropriate BED. For example, toxicity potential of metal ions released during nanoparticle dissolution in the acidic environment of the phagolysosome has been shown to influence inflammation. Soluble zinc ions (but not zinc nanoparticles or magnesium ions) elicited inflammation in rats following intratracheal instillation (Flink et al. 1992; Cho et al. 2011, 2012). However, it should be noted that the effects of metal nanoparticles versus metal ions have not been evaluated or validated in inhalation studies. These findings suggest that the relative toxicity of compositional ions of high solubility nanoparticles is important for evaluating the toxicity and BED for nanomaterials and their dissolved ions. Allocating the relative toxicity of released ions versus nanoparticles may be challenging due to in vitro assay differences (media, concentration, duration, ion-particle interactions) (Peijnenburg et al. 2020). Further, the electric and oxidative potential of a particle can impact cellular membrane integrity, oxidative balance, and proinflammatory responses (Donaldson et al. 2013).

Understanding the impact of fiber dimension of high aspect ratio nanoparticles or HARN is important for defining biopersistence and the BED. The rigidity of fibers to maintain a high aspect ratio or their ability to coil up into a bundle or to agglomerate (which will modify the length) may influence macrophage phagocytosis (frustrated phagocytosis and lysosome instability) and in turn in vivo biopersistence and pathogenicity (Donaldson et al. 2011; Kane et al. 2018). In fact, a study by Donaldson et al. (2011) demonstrated that long rigid carbon nanotubes (>15 µm length) produced inflammation and fibrosis in mice exposed by intraperitoneal injection, whereas short or tangled carbon nanotubes or carbon nanotubes incubated in Gambles solution for 10 weeks (which reduced the fiber mass and fiber lengths [<15 µm] and formed tightly agglomerated bundles) produced minimal inflammogenic responses (Osmond-McLeod et al. 2011).

While corollaries between asbestos and manufactured nanofibers, nanotubes, nanorods, and nanowires have been made, the fiber length-, durability-, and biopersistent-dependent factors that influence pathogenicity of mineral fibers are also expected to affect the biological effects of HARN. Additionally, unique nanomaterial shapes, such as graphene nanoplatelets, which can have a relatively low aerodynamic diameter (3 µm) but large actual size (30 µm), may present a risk of respiratory effects due to respirability and deposition of relatively large particles in the lungs that cannot be easily engulfed and cleared by macrophages (Donaldson et al. 2013). While the fiber toxicology paradigm (dose, dimension, and durability) which have been garnered from research of asbestos and SVFs are anticipated to apply to HARN, other factors such as toxicity of metal ions, surface reactivity (electric or oxidative potential), and protein corona formation in biological fluids (which appears to mitigate cellular toxicity) may also need to be considered particularly for HARN which have moderate or greater biopersistence (Arts et al. 2015). Only with toxicological studies aimed to correlate in vitro durability (and other measures of HARN reactivity), in vivo biopersistence, and biological outcomes will it be determined whether similar or different in vitro fiber dissolution and in vivo half-life thresholds, which exempt carcinogenicity classification of SVFs, can also be applied to HARNs.

Conclusions

In vitro dissolution assays in both neutral and acidic pH have been used for several decades to characterize the biodurability of SVFs and predict the biopersistence and biological effects in animals and humans. While in vitro dissolution data of SVFs were not always reported for correlation with in vivo clearance and biological effects, available information suggest that generally in vitro fiber dissolution rates greater than 100 ng/cm²/hr of glass fibers in pH 7 and stone fibers in pH 4.5 and in vivo fiber clearance less than WT_1/2 40 or 50 days do not present a risk of fibrotic and tumorigenic responses. The epidemiologic studies of SVF manufacturing worker cohorts have supported that SVFs do not present a risk of nonmalignant or malignant respiratory disease (Stone et al. 2004; Egnot et al. 2020), thus further reinforcing the utility of fiber physicochemical, dissolution, and biopersistence characterization methods that have been applied for understanding the safety and risks of SVFs.

While the fiber length-, durability-, and biopersistent-dependent factors that influence pathogenicity of mineral fibers are also expected to affect the biological effects of HARN, other factors such as toxicity of metal ions, surface reactivity, protein corona formation, particle shape (e.g. straight versus curly), and surface modifications may also need to be considered for HARN. As in vitro biodurability and biodissolution assays are being developed for HARN, the following are recommended for consideration:

• Material characterization – Complete physicochemical characterization, including concentration, mass, dimension, surface area, morphology, crystallinity, surface ligand modification, aggregation, agglomeration, fiber rigidity, electric and oxidative potential, protein corona formation;

• Assay design – System selection which aligns with the intended assay purpose (e.g. static system as a screening tool, dynamic system as representative biological systems), as well as characterization of design features (fiber cassette characteristics, flow rate, equilibrium time, agitation method/frequency, temperature, test duration, solution sampling and analytical methods) that may influence dissolution rates;

• Media composition – Characterization of buffer solution compositions representing both the neutral pH of the lung lining fluid and acidic pH of the macrophage phagolysosome, in addition to the understanding the impact of ionic strength, complexing agents, organic matter, buffer capacity, buffer volume, and flow rate on material dissolution rates;

• Particle versus ion interactions – Toxicity characterization of particles in pristine and degraded states, as well as released metal ions free or complexed with organic binders.

Understanding the primary factors which influence HARN dissolution and biodurability in in vitro systems will provide a foundation upon which to bridge in vivo biopersistence and biological responses, which in turn can be used to identify critical thresholds to adequately predict health effects and product safety. Only with toxicological studies aimed to correlate in vitro durability (and other measures of HARN reactivity), in vivo biopersistence, and biological outcomes will it be determined whether similar or different in vitro fiber dissolution and in vivo half-life thresholds, which exempt carcinogenicity classification of SVFs, can also be applied to HARNs.

Abbreviations
3Ds	=	dose, dimension, and durability
3Rs	=	reduce, refine, replace
AES	=	alkaline earth silicate
AR	=	alkali resistant
AZS	=	aluminum zirconia silica
BALF	=	bronchoalveolar lavage fluid
BALT	=	bronchus-associated lymphoid tissue
BED	=	biologically effective dose
BrdU	=	bromodeoxyuridine
CMS	=	calcium–magnesium–silicate
EC	=	European Commission
ECB	=	European Chemicals Bureau
EURIMA	=	European Insulation Manufacturers Association
ECVAM	=	European Center for the Validation of Alternative Methods
GI	=	gastrointestinal
GMD	=	geometric mean diameter
HARN	=	high aspect ratio nanomaterials
HT	=	high temperature
IP	=	intraperitoneal
IPL	=	intrapleural
IT	=	intratracheal
Kdis	=	dissolution rate
L	=	length
LMP	=	lysosomal membrane permeabilization
MΦ	=	macrophage
MMVF	=	man-made vitreous fibers
MTD	=	maximum tolerated dose
NLRP3	=	nucleotide-binding doman (NOD)-like receptor protein 3
SVF	=	synthetic vitreous fiber
OECD	=	Organization for Economic Co-operation and Development
PMN	=	polymorphonuclear leukocytes
PSI	=	pounds per square inch
RCC	=	Research and Consulting Co
RCF	=	refractive ceramic fibers
WHO	=	World Health Organization
WPMN	=	Working Party on Manufactured Nanomaterials
WT1/2	=	weighted half-life

Acknowledgements

We appreciate and acknowledge the four anonymous reviewers the Editor selected. Besides these anonymous reviewers, the Editor, and the authors, no individual has reviewed any drafts of the paper.

Declaration of interest

AKM and HCO are employed by ChemRisk, a consulting firm providing scientific advice to the government, corporations, law firms, and various scientific/professional organizations. Funding for this manuscript was provided, in part, by the North American Insulation Manufacturers Association (NAIMA). AKM has also been engaged by NAIMA to provide general scientific consulting and expert advice on scientific matters. This article was prepared and written exclusively by the authors without review or comment by NAIMA employees, counsel, or associated members. The preparation of the paper, including conduct of the literature review, review of the individual papers, integration and synthesis of the findings, the conclusions drawn, and recommendations made are the exclusive professional work product of the authors and may not necessarily be those of their employer or the financial sponsor.

Appendix A

Table A1. Summary of intraperitoneal injection toxicology studies of SVFs.

Fiber	Exposure	Fiber number and dimensions	Controls	Species	Outcomes	Reference
German glass wool	1x 2 or 10 mg in 2 ml	24, 120 × 10^6 fibers > 5 μm in length; 59% of fibers < 3 μm in length	Chrysotile (6/37 for 2 mg and 25/31 for 25 mg repeat) or 2 mg UICC crocidolite (15/39 tumors) or 50 mg granular corundum (3/37 abdominal tumors)	Rat/hamster	In rats, mesotheliomas and spindle-cell sarcomas were observed at incidences of 1/34, 4/36 for the exposures of 24, 120 × 10^6 German glass fibers. No abdominal tumors in hamsters	Pott et al. 1976; IARC 2002
German glass wool	4x (weekly) 25 mg in 2 ml	1,200 × 10^6 fibers > 5 μm in length; 59% of fibers < 3 μm in length		Rat/hamster	Rats 23/32 for 1200 × 10^6 German glass fibers. No abdominal tumors in hamsters
JM 104 glass fiber	1x for 2 or 10 mg; 2x 25 mg injections (2, 10, or 50 mg total) in 2 ml	Mean fiber dimensions, 0.2 μm x 10 μm		Rat	20/73 (2 mg), 41/77 (10 mg) and 55/77 (50 mg) with mesotheliomas, sarcomas and, occasionally carcinomas observed in rats
JM 112 glass fiber	1x 20 mg in 2 ml	Mean dimensions 1 μm x 30 μm		Rat	14/37 tumors
JM 100 glass fiber (expt 12)	1x 2mg or 10 mg in 2 ml	Median fiber dimensions, 2.4 μm x 0.33 μm	SA crocidolite 0.5 or 2 mg (18/32 low exposure, 28/32 high exposure). Three tumors were noted in rats that received two injections each of either 20 mg Mount St Helen’s volcanic ash or 20 ml saline alone	Wistar or Sprague Dawley rats	Tumor incidence: 2 mg: 21/54; 10 mg: 24/53 animals	Pott et al. 1987; IARC 2002
JM 104E glass fiber	1x 5 mg in 1 ml	Cut and ground for 1 h in an agate mill; median fiber dimensions, 4.8 μm x 0.29 μm			Adult exposure: 44/54; adolescent exposure: 20/45
JM 475 [code 104] glass fiber	1x 0.5 or 2 mg in 1 ml	Median fiber dimensions, 3.2 μm x 0.18 μm			Tumor incidence 0.5 mg: 5/30; 2.0 mg: 8/31
JM 475 glass fiber	2 mg treated w/ HCl in 1 ml	*not stated if differed from above			16/32 with acid-treated glass fiber
ES3 filament	50 and 250 mg/rat in 2 ml	16.5 μm x 3.7 μm			Tumor incidence 6% (50 mg) or 9% (250 mg)
ES5 filament	10, 40 and 250 mg/rat in 2 ml	39 μm × 5.5 μm			Tumor incidence 4% (10mg),11% (40 mg), or 7% (250 mg)
ES7 filament	40 mg/rat in 2 ml	6 μm x 7.4 μm			Tumor incidence 2%
Swedish rock wool	3x 25 mg (75 mg) in 2 ml	Median fiber length, 23 μm; median diameter,1.9 μm			Tumor incidence 45/63 (71.4%)
Swedish rock wool fine fraction	1x 10 mg in 2 ml	Median fiber length, 4.1 μm; diameter, 0.64 μm			Tumor incidence 6/45 (13.3%)
German Slag [Rheinstahl] (expt 3)	2x 20 mg (1x/wk) in 2 ml	Median fiber length of 26 μm and a median fiber diameter of 2.6 μm			Tumor incidence 6/99 (6.1%)
German slag [Zimmermann] (expt 3)	2x 20 mg (1x/wk) in 2 ml	Median fiber length of 14 μm and a median fiber diameter of 1.5 μm			Tumor incidence 2/96 (2%)
Fiberfrax RCF	9 mg x 5 (weekly) 45 mg total in 2 ml	Fiber length, 8.3 μm; fiber diameter, 0.91 μm			Tumor incidence 32/47 (68%)
Manville RCF	15 mg x 5 (weekly) 75 mg total in 2 ml	Fiber length, 6.9 μm; fiber diameter, 1.1 μm			Tumor incidence 12/54 (22%)
Untreated JM104 glass fiber	5 mg	Chopped and ground, 50% were < 4.8 μm in length.	No positive control; Included saline control	Wistar rats	Tumor incidences (estimated from the graph) approximately 80%.
HCl treated JM104 glass fiber		Soaked in 1.4 mol/L HCl at room temperature for 2 or 24 h. Fiber weight loss after 2h: 25%, 24h 33%			Tumor incidences (estimated from the graph) approximately ∼60% (2h) and <10% (24h).
NaOH treated JM104 glass fiber		Soaked in 1.4 mol/L NaOH at room temperature for 24 h			Tumor incidences (estimated from the graph) approximately 80%.
JM 100 glass fibers	25 mg in 0.5 ml	Geometric mean fiber dimensions, 4.7 μm x 0.4 μm; 19% of fibers > 10 μm in length and 0.2–0.6 μm in diameter	20/25 rats injected with 25 mg UICC crocidolite (5% ≥5 μm in length; mean, 3.1 μm), but in neither of the control groups. 40% of hamsters and 80% of rats injected with crocidolite had abdominal mesotheliomas. Saline controls did not have any tumors.	Osborne-Mendel rats	8/25 rats had abdominal mesotheliomas	Smith et al. 1987
Fibrefrax RCF	25 mg in 0.5 ml	GMD of 0.9 μm and a GML of 25 μm but length and diameter distributions and particulate counts for the elutriated material injected were unspecified			19/23 rats had abdominal tumors
Fibrefrax RCF	25 mg in 0.5 ml	GMD of 0.9 μm and a GML of 25 μm but length and diameter distributions and particulate counts for the elutriated material injected were unspecified		Syrian hamsters	Mesotheliomas were found in 2/15 hamsters in the first group and in 5/21 hamsters in the second group (**21/36 and 15/36 from groups 1 and 2 died within 30 d)
JM 104/475	0.5 mg in 1 ml	90% of fibers <8.4 µm long and <0.40 µm diameter; 28% longer than 5 µm	Crocidolite (90% <7.7 µm long and 0.36 µm diameter) or Calidria chrysotile (90% <5.9 µm long and 0.10 µm diameter) or UICC chrysotile (90% <3.6 µm long and 0.18 µm diameter). Tumor incidence: Crocidolite 55%, Calidria chrysotile 6%, UICC chrysotile 84%	Female Wistar rats	17% tumor incidence	Muhle et al. 1987
X7753 Glass Fiber	Single injection 2.5 mg/ml in 2 ml	mean diameter 2.05 µm and mean length 19.4 µm	No positive control	Female Fisher Rats	Retained mass 2 days post injection: 4.08 mg (range 1.83-5.41 mg)	Collier et al. 1994; IARC 2002
X7484 Glass Fiber		Not listed			Retained mass 2 days post injection: 5.16 mg (range 0.21-6.21 mg)
MMVF11 Glass Fiber (CertainTeed)	70 mg (2x 35 mg weekly) or 180 mg (6x 30 mg weekly) in 2 ml	0.4 × 10^9 or 1.0 × 10^9 fibers ; median length 14.6 μm × median diameter 0.77 μm	Crocidolite 0.5 mg (5 injections of 0.1 mg at two week intervals; males and females) or 0.5 mg (females only, 5 injections at weekly intervals. 0.042 x 10^9 fibers; median length of 1.8 μm and median diameter of 0.19 μm. Mesothelioma incidence: 25/32 (78%), 32/48 (67%), 20/39 (51%); Included vehicle controls	Female Rats	Mesothelioma incidence 12/40 (30%) and 16/23 (70%)	Roller et al. 1996
MMVF21 rock wool	60 mg (2x 30 mg weekly) or 180 mg (6x 30 mg weekly) in 2 ml	0.4 × 10^9 or 1.0 × 10^9 fibers; median length of 16.9 μm and median diameter of 1.02 μm		Female Rats	Mesothelioma incidence: 37/38 (97%) and 33/38 (87%)
MMVF22 slag wool	20 mg (1x 20 mg weekly), 50 mg (1 x 50 mg), or 150 mg (3x 50 mg weekly) in 2 ml	0.4 × 10^9,1.0 × 10^9 or 2.9 x 10^9 fibers; median length of 8.7 μm and median diameter of 0.77 μm		Female Rats	Mesothelioma incidence: 4/40 (10%), 8/40 (20%), 18/38 (47%)
R-stone E3	114 mg (4x 28.5 mg) or 256.5 mg (9 x 28.5 mg) in 2 ml	0.4 × 10^9 or 0.9 x 10^9 fibers; median length of 16.9 μm and median diameter of 1.03 μm		Female Rats	Mesothelioma incidence: 0/30 and 4/35 (11%)
M stone	8.5 mg (1x), 25.5 mg (1x), 85 mg (2x 42.5 mg, 2 week interval between) in 2 ml	0.1 × 10^9, 0.3 x 10^9, or 1.0 x 10^9 fibers; median length 10.1 μm and median diameter 0.84 μm		Male and Female Rats	Mesothelioma incidence: 2/32 (6%) and 2/36 (6%) for females and males, 9/32 (28%) and 8/36 (22%) for females and males, and 22/35 (63%) for males
B-20-2.0	6 mg, 18 mg, or 60 mg (2x 30 mg with 2 week interval between) in 2 ml	0.08 × 10^9, 0.24 x 10^9, or 0.8 x 10^9 fibers; median length 7.8 μm and median diameter 0.77 μm		Male and Female Rats	Mesothelioma incidence: 2/32 (6%) females and 15/36 (42%) males, 7/32 (22%) females and 12/34 (35%) males and in 21/35 (60%) males
B-20-0.6	Single injection of 3.5, 8.5 or 25 mg or 75 mg (3 weekly 25 mg) in 2 ml	0.4 × 10^9, 1.0 x 10^9, 3.0 x 10^9, or 9.0 x 10^9 fibers; median length 3.6 μm and median diameter 0.30 μm		Female Rats	Mesothelioma incidence: 12/40 (30%), 17/40 (43%), or 30/40 (75%) and 27/31 (87%)
B-09-0.6 glass wool	100 (2 weekly 50 mg) or 300 mg (6 weekly 50 mg) in 2 ml	2.0 × 10^9 or 6.1 x 10^9 fibers; median length 3.3 μm and median diameter 0.49 μm		Female Rats	Mesothelioma incidence: 1/40 (3%), 4/40 (10%)
B-09-2.0 glass wool	150 (3 weekly 50 mg) or 450 mg (9 weekly 50 mg) in 2 ml	1.1 × 10^9 or 3.2 x 10^9 fibers; median length 10.5 μm and median diameter 1.2 μm		Female Rats	Mesothelioma incidence: 9/40 (23%) and 21/40 (53%)
B-01-0.9 glass wool	125 mg (5 weekly 25 mg), 250 mg (10 weekly 25 mg), 500 mg (20 weekly 25 mg), 1000 mg (40 weekly 25 mg) in 2 ml	2.5 × 10^9, 5.0 × 10^9, 10.0 × 10^9 or 20.0 x 10^9 fibers; median length ∼9.0 μm and median diameter 0.7 μm		Male and Female Rats	Mesothelioma incidence: 3/39 (8%) 125 mg females, 4/37 (11%), 250 mg females and 3/36 (8%) in 500 mg females; 10/48 (21%) 500 mg males and 33/50 (66%) 1000 mg males
753-104 glass fiber	17 or 50 mg in 2 ml	1 x 10^9 3 x 10^9 fibers; median length ∼3.3 μm and median diameter 0.22 μm		Female Rats	Mesothelioma incidence: 30/40 (75%) and 36/40 (90%)
100/475 glass fiber	8.3 mg in 2 ml	9 × 10^6 fibers >20 μm in length (diameter <0.95 μm); 23.7% persistence >20 µm fibers at 12 mo	6.1 mg of amosite asbestos (410 × 10^6 WHO fibers/rat; ∼ 99% of the fibers had diameters below 0.95 μm 21/24 (88%) for amosite treated rats	SPF Wistar rats; 12 month observation	Mesothelioma incidence 8/24 (33%); Kdis pH 7.0: 9.1 ng/cm^2/hr	Miller et al. 1999
MMVF10 glass wool	144 mg in 2 ml	119 × 10^6 fibers >20 μm in length (diameter <0.95 μm); 0.6% persistence >20 µm fibers at 12 mo			Mesothelioma incidence 13/22 (59%); Kdis pH 7.0: 122.4 ng/cm^2/hr
MMVF21 stone wool	183.1 mg in 2 ml	344 × 10^6 fibers >20 μm in length (diameter <0.95 μm); 43% persistence >20 µm fibers at 12 mo			Mesothelioma incidence 19/20 (95%); Kdis pH 7.0: 28.9 ng/cm^2/hr
MMVF22 slag wool	129.6 mg in 2 ml	142 × 10^6 fibers >20 μm in length (diameter <0.95 μm); 1.7% persistence >20 µm fibers at 12 mo			Mesothelioma incidence 13/24 (54.2%); Kdis pH 7.0: 52.8 ng/cm^2/hr
RCF1	110.9 mg in 2 ml	85 × 10^6 fibers >20 μm in length (diameter <0.95 μm); 50.5% persistence >20 µm fibers at 12 mo			Mesothelioma incidence 21/24 (88%); Kdis pH 7.0: 4.4 ng/cm^2/hr
RCF2	188.8 mg in 2 ml	111 × 10^6 fibers >20 μm in length (diameter <0.95 μm); 157.3% persistence >20 µm fibers at 12 mo			Mesothelioma incidence 13/18 (72%); Kdis pH 7.0: 3.1 ng/cm^2/hr
RCF4	90.4 mg in 2 ml	6 × 10^6 fibers >20 μm in length (diameter <0.95 μm); 142.7% persistence >20 µm fibers at 12 mo			Mesothelioma incidence 0/22; Kdis pH 7.0: 0.5 ng/cm^2/hr
100/475 glass fiber		Respirable samples collected from aerosol testing (diameter <3 μm, length >5 μm, aspect ratio >3:1)	Amosite (reported previously in Davis et al. 1996 and Miller et al. 1999)	Male Wister Rats	No carcinomas, 4/38 adenomas (10.5%); no mesotheliomas	Cullen 2000
104E Glass Microfiber	10^9 fibers, (diameter <3 µm, length >5 μm; aspect ratio >3:1) in 2 ml				7/43 carcinomas (16.3%), 3/43 adenomas (7%); 2/43 mesotheliomas (4.7%)
Biosoluble M glass wool	0.5, 2.0, or 5 x 10^9 total fibers (equivalent to 41, 164 and 410 x 10^6 WHO fiber) in 2.5 ml	Stock fibers: GMD 0.42 μm × GML 8.26 μm	Crocidolite Tumor incidence: 0.1 x 10^9 1/51, 1 x 10^9 0/51	Female Wistar Rat	4/50, 0/51 and 2/52 with mesothelioma; T1/2: 5.5 days (>20 µm long), 8.5 days for WHO fibers; Kdis/SiO2: 103.7 ng/cm^2/hr	Grimm et al. 2002
Biosoluble P glass wool	0.5, 2.0, or 5 x 10^9 total fibers (equivalent to 51, 205 and 512 x 10^6 WHO fiber) in 2.5 ml	Stock fibers: GMD 0.43 μm × GML 9.9 μm			1/51, 1/51 and 1/52 with mesothelioma; T1/2: 12 days (>20 µm long), 21 days for WHO fibers; Kdis/SiO2: 610 ng/cm^2/hr
Biosoluble V glass wool	0.5, 2.0, or 5 x 10^9 total fibers (equivalent to 72, 290 and 724 x 10^6 WHO fiber) in 2.5 ml	Stock fibers: GMD 0.8 μm × GML 10.07 μm			2/51, 1/51 and 2/51 with mesothelioma; T1/2: not measured; Kdis/SiO2: 450 ng/cm^2/hr
Biosoluble O stone wool	0.5, 2.0, or 5 x 10^9 total fibers (equivalent to 54, 215 and 536 x 10^6 WHO fiber); 2, 8, or 20 weekly injections in 2.5 ml	Stock fibers: GMD 0.42 μm × GML 11.45 μm			0/51, 1/51 and 1/51 with mesothelioma; T1/2: 12.4 days (>20 µm long), 8.5 days for WHO fibers; Kdis/SiO2: 523 ng/cm^2/hr
B Glass Wool	2.0 or 5 x 10^9 total fibers (equivalent to 216 and 541 x 10^6 WHO fiber); 2,8, or 20 weekly injections in 2.5 ml	Stock fibers: GMD 0.55 μm × GML 9.27 μm			1/51 and 2/53 with mesothelioma; T1/2: 4 days (>20 µm long), 17 days for WHO fibers; Kdis/SiO2: 580 ng/cm^2/hr
HT insulation wool	9 mg: 2.6 x 10^9 total fibers, with 2.1 x 10^9 WHO fibers in 2 ml; 39.6% SiO2; 21.5% Al2O3; 15.4% CaO; 10.7% MgO; 6.4% FeO; 2.0% TiO2	GMD 0.65 μm, GML 10.7 μm; 28% of WHO fraction >20 μm	No positive control; included vehicle controls	Female Wistar rats	No mesotheliomas; 6% of animals had fibrous nodules and 58% had adhesions	Kamstrup et al. 2002
D6 (MMVF21-like)	36 mg: 1.2 x 10^9 total fibers, with 0.9 x 10^9 WHO fibers in 2 ml; 45.9% SiO2; 13.8% Al2O3; 17% CaO; 9.5% MgO; 6.2% FeO; 3.0% TiO3; 2.5% Na2O	GMD 0.8 μm, GML 8.5 μm; 18% of WHO fraction >20 μm			Mesothelioma incidence 32/57 (56%); 88% fibrous nodules and 93% adhesions

Table A2. Summary of intrapleural injection toxicology studies of SVFs.

Fiber	Exposure	Fiber number and dimensions	Controls	Species	Outcomes	Reference
Glass borosilicate	1x 20 mg in 0.4 ml	30% 1.5–2.5 µm diameter, max 7 µm; 60% >20 µm length	Two samples of chrysotile: tumors 23/36 and 21/32 (64–65%)	SPF rat; survival ranged from 568 (chrysotile) to 774 (glass fibers)	No tumor induction	Wagner et al. 1973
Glass borosilicate powder		Borosilicate max 8 µm diameter			1 mesothelioma (2.8%)
Suspended solids of RCF/ball milled		No dimensions given			9.7% pleural mesothelioma
2 types of borosilicate glass fibers each prepared by hand grinding or mechanical mortar and pestle	1x 10 mg in 0.5 ml	Separated samples with avg diameters between 0.05 and 3.5 µm into samples with lengths of 100+ µm and <20 µm long	Chrysotile and crocidolite, no exposure listed; 2/150 chrysotile and crocidolite animals had pleural tumors	Mouse; 18 month survival (37/52)	No pleural tumors in glass treated animals	Davis 1976
Fine JM100 glass fiber	20 mg in 0.4 ml	99% <0.5 μm diameter with a median diameter of 0.12 μm; 2% >20 μm in length with a median length of 1.7 μm	No positive control; Included vehicle control	Wistar rats; survival until natural death	4/32 (12%) mesotheliomas	Wagner et al. 1976
Coarse JM110 glass fiber	20 mg in 0.4 ml	17% of fibers <1 μm in diameter with a median diameter of 1.8 μm; 10% >50 μm in length and a median length of 22 μm			No pleural tumors were noted with the coarser US JM 110 glass fiber
Resin coated or uncoated glass fibers (17 types)	40 mg in 1.5 ml 10% gelatin (vehicle) on coarse fiber glass pledgets	Samples with avg diameters between 0.06 and 25 µm	No positive control	Osborne-Mendel rats; thoracic implantation; 2 year observation	Pleural mesothelioma in animals that survived for more than 52 weeks varied from 0/28 to 20/29, dependent upon fiber size. When two of the glass fiber preparations (diameter, >0.25 μm) were leached to remove all elements except silicon dioxide, the incidences of pleural mesotheliomas were 2/28 and 4/25	Stanton 1977; IARC 2002
Fine fraction JM104 glass fiber	20 mg in 2 ml	mean length, 5.89 μm; mean diameter, 0.229 μm	20 mg chrysotile A (mean dimensions, 3.21 μm x 0.063 μm) or crocidolite (mean dimensions, 3.14 μm x 0.148 μm), or varied concentrations of acid to leach chrysotile/crocidolite. Tumor incidence with untreated chrysotile was 15/33 and 21/39 with untreated crocidolite. Acid-leaching reduced tumor incidence (loss of 89.1% MgO resulted in no tumors/mesothliomas)	Sprague Dawley rats	JM104 glass fiber had a mesothelioma incidence of 6/45 (13%)	Monchaux et al. 1981
English glass wool dust	20 mg	Resin coated (70% fibers ≤5 μm in length; 85% ≤1 μm in diameter), or without resin coating (57% ≤5 μm in length; 85% ≤1 μm in diameter)	20 mg chrysotile (fibers longer than 5 µm: 196 x 10^8; particles > 1 µm: 8.8 x 10^8). 6/48 mesotheliomas (12.5%); Included vehicle control	Sprague Dawley rats	One mesothelioma (unclear if it this occurred in the coated or uncoated group)	Wagner et al. 1984
Swedish rock wool		Resin coated (70% fibres ≤5 μm in length; 77% ≤1 μm in diameter), Swedish rock (stone) wool after removal of resin (70% ≤5 μm in length; 81.5% ≤1 μm in diameter)			Mesotheliomas- 3/48 with resin; 2/48 without resin
German slag wool		Resin coated (67% ≤5 μm in length; 70% ≤1 μm in diameter), without resin (80% ≤5 μm in length; 2% ≤1 μm in diameter			No mesotheliomas
JM 100 glass fiber		88% ≤5 μm in length; 98.5% ≤1 μm in diameter			Mesotheliomas 4/48 (8.3%)
Saffil RCF	20 mg in 2 ml	Diameter ≤ 3 μm, 86%; length ³ 10 μm, 92%	20 mg chrysotile: 19/48 (40% tumor incidence, excluding mammary and pituitary; 7/48 (14.5%) pleural mesotheliomas. Included vehicle control – High background tumor incidence: saline control animals 10/48 (21%) (excluding pituitary and mammary tumors); no mesothelioma	Albino Wistar rats	No mesothelioma ; 10/48 (21%) tumors (excluding pituitary and mammary)	Pigott & Ishmael 1992
Aged (1200C) Saffil RCF		Diameter ≤ 3 μm, 82%; length ≥10 μm, 99.7%			No mesothelioma; 7/48 (15%) tumors (excluding pituitary and mammary)
RCF suspended solids of fiber A(kaolin)		Diameter ≤ 3 μm, 66%; length ≥10 μm, 80%			No mesothelioma; 4/46 (9%) tumors (excluding pituitary and mammary)
RCF suspended solids of fiber B (alumina/silica)		Diameter ≤ 3 μm, 92%; length ≥10 μm, 46%			3/48 (6%) mesothelioma; 15/48 (31%) tumors (excluding pituitary and mammary)
RCF- aluminosilicate fibers (High Duty grade), vitrified	20 mg in 0.4 ml	52–56 wt- % alumina and 44-48 wt-% silica; arithmetic mean fiber diameter is 2-3 µm; length not listed. 19-24 animals/group. Ground to a fine powder using a mortar and pestle	No positive control; Included vehicle control	Wistar-Porton rats	No mesothelioma tumor induction; four pleural hyperplasias	Carthew et al. 1995
RCF- aluminosilicate fibers (High Duty grade), devitrified 2 1200C					No mesothelioma tumor induction; six pleural hyperplasias
RCF- aluminosilicate fibers (High Duty grade), devitrified at 1400C					No mesothelioma tumor induction; two pleural hyperplasias

Table A3. Summary of intratracheal instillation toxicology studies of SVFs.

Fiber	Fiber prep	Binder/coating	Experimental design	Fiber number and dimensions	Controls	Outcomes	Reference
Glass Fibers	Ball Milled	Uncoated	Rats/Hamsters 3 x 3.5 mg in 1 ml	Average diameter at 0.5 µm and the average length of 10 µm with a range of 5-20 µm	Saline controls	Polyploid masses of inflammatory tissue in smaller bronchial divisions (4 days post exposure); no evidence of these masses at 12 mo. Alveoli around respiratory bronchioles & alveolar ducts contained dust filled macrophages	Gross et al. 1970
			Rats 10 x 3.5 mg in 1 ml
		Phenol-formaldehyde coated	Hamsters: 1 x 3.5 mg in 1 ml
			Hamsters: 2 x 1.75 mg in 1 ml
			Hamsters: 3 x 3.5 mg in 1 ml
			Rats, 3 x 3.5 mg in 1 ml
			Rats, 10 x 3.5 mg in 1 ml
		Starch Coated	Hamsters: 3 x 3.5 mg in 1 ml
			Rats, 3 x 3.5 mg in 1 ml
			Rats, 10 x 3.5 mg in 1 ml
JM100 Glass Fibers (no binder)	Wetted with ethanol then placed in a plug-packing mold assembly (dried overnight); aerosolized through a Timbrell generator	Not listed	Osborne-Mendel rats & Syrian Golden hamsters 1x/week for 5 weeks; 2 mg/instillation (10 mg) in 0.2 ml	Mean diameter 0.45 µm, GMD 0.4 µm and geometric mean length [GML] 4.7 µm; 94% of fibers were ≤20 µm	UICC crocidolite 74% of hamsters (13/20 benign, 7/20 malignant) and 8% (2 malignant) of rats had primary tumors; both hamsters and rats developed significant fibrosis	No tumors noted; 14% of rats had bronchoalveolar metaplasia; 32% of rats had fibrosis	Smith et al. 1987
RCF Carborundum Fibrefrax (no binder)				Mean diameter 1.8 µm with a GMD of 0.9 µm and GML 25 µm; approximately 86% of fibers had diameters <2.0 µm with 83% of fibers >10 µm long		No tumors noted; 27% of rats had bronchoalveolar metaplasia; 16% of hamsters and 9% of rats had fibrosis
JM104/475 Glass Fiber	Cut with scissors, ground in knife mill for 30 min	Not listed	Wistar rats, 20 weekly instillations of 0.5 mg in 0.3 ml (10 mg total dose)	50% length <3.2 µm and width <0.18 µm	South African crocidolite; median dimensions of fibers, (20 x 0.5 mg weekly) 2.1 μm x 0.2 μm; 15/35 rats developed lung tumors of which 11 were carcinomas and four had mixed tumors	1/34 adenoma, 2/34 adenocarcinoma, 2/34 squamous cell carcinoma total incidence 14.7%	Pott et al. 1987
Rock Wool	Not listed	Not listed	Syrian hamsters; 2 mg/dose 1 x/wk in 0.2 ml for 5 weeks (10 mg total dose); animals sacrificed 2 yrs after first instillation	6.1 µm x 296 µm	Included neg vehicle control; Positive control included amosite (1.05 µm x 15.5 µm) and crocidolite 0.88 µm x 17.4 µm), which resulted in 2/20 tumors (lung adenoma, mesothelioma) and 0/20 tumors, respectively.	No tumor induction; inflammation, fibrosis, and pleural thickening noted	Adachi et al. 1991; 1992
Glass Fiber				0.65 µm x 16.8 µm		Inflammation and pleural thickening; 1 squamous cell carcinoma (lung) and 1 adrenal gland neuroblastoma (2/20)
Potassium Titanate				0.36 µm x 7.17 µm		Inflammation and pleural thickening; no tumor induction
Calcium Sulfate				1.0 µm x 17.8 µm		Inflammation; 3/20 tumors: found in rib, heart or kidney
Basic Magnesium Sulfate				0.45 µm x 22.4 µm		Pleural thickening; 9/20 tumors of which 1 was unspecified lung and 2 were mesotheliomas
Metaphos-phate				2.38 µm x 64.1 µm		Inflammation; 5/20 tumors with 1 mesothelioma
X607 GlassFiber	Sized	Not listed	Wistar rats; single instillation 2 mg in 0.3 ml; serial sacrifices up to and including 24 mo	4 or 8 x 10^6 fibers; length >5 μm and a diameter <2 μm (aspect ratio >5:1; size distribution not given	No positive or negative controls	T1/2= all fibers 47 days (range of 41–54 days); WHO fibers 46 days (41-54 days)	Muhle et al. 1994
MMVF21 Rock Wool	Not listed	Not listed	Female Wistar rats, 2 mg single exposure in 0.3 ml. 5 rats/group sacrificed at 2 d, 2 wks, 1, 3, 6, or 12 mo	400 fibers of stock samples; critical fibers/ng (L > 5 µm, D <2 µm, L/D > 5/1): 29. 90% of fibers <15.3 µm long, <1.21 µm diameter	UICC Crocidolite Long (increased fraction of long fibers); 0.1 mg/0.3 ml saline. Groups sacrificed at 2d, or 6, 12 mo. Lung dose: Total fibers: 2d: 197.5 x10^6/lung, 6 mo: 99.9 x 10^6/lung, 12 mo: 135.2 x 10^6/lung. Mean T1/2: 695 d (L > 5 µm, D < 3 µm)	Total fibers: 2d: 124.1 x 10^6/lung, 1 mo: 111.9 x 10^6/lung, 3 mo: 92 x 10^6/lung, 6 mo: 45.2 x 10^6/lung, 12 mo: 50.8 x 10^6/lung; Mean T1/2 WHO fibers (L > 5 µm, D < 3 µm): 291d	Bellmann et al. 1994
HT Stone Wool				400 fibers of stock samples; critical fibers/ng (L > 5 µm, D <2 µm, L/D > 5/1): 17. 90% of fibers <27.9 µm long, <1.62 µm diameter		Total fibers: 2d: 40.2 x 10^6/lung, 1 mo: 30.8 x 10^6/lung, 3 mo: 22.7 x 10^6/lung, 6 mo: 9.4 x 10^6/lung, 12 mo: 4.5 x 10^6/lung; Mean T1/2 WHO fibers (L > 5 µm, D < 3 µm): 98d
M-104/475 Glass	Not listed	Not listed	0.8 x 10^9 fibers (10 instillations of 0.5 mg each) or 1.6 x 10^9 fibers (20 instillations of 0.5 mg each); 5 or 10 mg dose	length >5 μm and a diameter <2 μm (aspect ratio > 5:1; size distribution not given	Tremolite: 2.5 mg (5 doses of 0.5 mg each) or 7.5 mg (15 doses of 0.5 mg each); total number of tremolite fibers in the low dose was 70 × 10^6 and 300 × 10^6 fibers were in the high dose; (length, >5 μm; diameter, <2 μm; aspect ratio, >5:1). Animals had 6/38 non-lung tumors in the low dose group and 4/37 in the high dose group	5 mg: 8/55 and 10 mg: 4/38 with "other" tumors" (includes fibrosarcomas, lymphosarcomas, mesotheliomas and lung metastases	Pott et al. 1994
Rock wool			4 x 10^6 fibers (10 instillations of 0.5 mg each) or 20 weekly doses of 0.5 mg each; 4 or 8 x 10^6 fibers; 5 or 10 mg dose			5 mg: 5/59 and 10 mg: 4/40 with "other" tumors" (includes fibrosarcomas, lymphosarcomas, mesotheliomas and lung metastases)
MMVF21 Rock Wool	Size fractions	Not listed	Female wistar rats, single dose (2 mg in 0.3 ml). 5 rats/group sacrificed at 2d, 2wks, 1,3,6, or 12 mo	Stock fibers: 90% of fibers 15.3 µm long, 1.21 µm diameter	Studied in Bellmann et al. 1994	Total fibers: 2d: 103.9 x 10^6/lung, 1 mo: 93.7 x 10^6/lung, 3 mo: 80.8 x 10^6/lung, 6 mo: 40.9 x 10^6/lung, 12 mo: 46.7 x 10^6/lung, 18 mo: 32.1 x 10^6/lung; Mean T1/2: Fibers >20 µm long: 196 d; WHO fibers 326 d	Bellman et al. 1995
MMVF30 Rock Wool				Stock fibers: 90% of fibers 18.3 µm long, 1.46 µm diameter		Total fibers: 2d: 49.8 x 10^6/lung, 1 mo: 44 x 10^6/lung, 3 mo: 38.9 x 10^6/lung, 6 mo: 31.8 x 10^6/lung, 12 mo: 24.2 x 10^6/lung, 18 mo: 19.4 x 10^6/lung; Mean T1/2 Fibers >20 µm long: 133 d; WHO fibers 357 d
MMVF31 Rock Wool				Stock fibers: 90% of fibers 23.1 µm long, 1.59 µm diameter		Total fibers: 2d: 55 x 10^6/lung, 1 mo: 43.3 x 10^6/lung, 3 mo: 36.9 x 10^6/lung, 6 mo: 21.7 x 10^6/lung, 12 mo: 16.1 x 10^6/lung, 18 mo: 16 x 10^6/lung; Mean T1/2 Fibers >20 µm long: 123 d; WHO fibers 257 d
100/475 Glass Fiber	Size Selected	Not listed	Male Wistar rats; single instillation (0.5 ml of 2 mg/ml) into the deep lung using a blunt cannula. Lung burden evaluated at 3d and 12 mo	Recovered fibers: All fibers: mean diameter 0.2 µm (3 d) and 0.27 µm (12 mo). Fibers longer than 20 µm: 0.79 µm (3 d) and 0.51 µm (12 mo). Majority of fibers recovered were 0.4-5 µm long	Recovered fibers: All fibers: mean diameter 0.31 µm (3 d) and 0.35 µm (12 mo). Fibers longer than 20 µm: 0.5 µm (3 d) and 0.52 µm (12 mo). Majority of fibers recovered were 0.4-5 µm long. Fibers longer than 0.4 µm: 3d: 563 x 10^6 fibers/lung 12 mo: 39.2 x 10^6 fibers/lung; 7% of 3d remaining	Fibers longer than 0.4 µm: 3 d: 2,466 x 10^6 fibers/lung 12 mo: 255 x 10^6 fibers/lung; 10% of 3 d remaining. Kdis: 9.1 ng/cm^2/hr; Silica leached in NaOxalate pH4.6: 6.3% pH 7: 5.9%	Searl et al. 1999
MMVF10 Glass Wool				Recovered fibers: All fibers: mean diameter 1.39 µm (3 d) and 0.63 µm (12 mo). Fibers longer than 20 µm: 1.62 µm (3 d) and 0.95 µm (12 mo). Lengths were distributed across ranges with more fibers >20 µm long		Fibers longer than 0.4 µm: 3 d: 4.15 x 10^6 fibers/lung 12 mo: 0.64 x 10^6 fibers/lung; 15% of 3 d remaining. Kdis: 122.4 ng/cm^2/hr; Silica leached in NaOxalate pH 4.6: 30.7% pH 7: 15.3%
MMVF21 Rock Wool				Recovered fibers: All fibers: mean diameter 0.97 µm (3 d) and 1.14 µm (12 mo). Fibers longer than 20 µm: 1.13 µm (3 d) and 1.41 µm (12 mo). Lengths were distributed across ranges with more fibers >20 µm long		Fibers longer than 0.4 µm: 3 d: 6.1 x 10^6 fibers/lung 12 mo: 2.27 x 10^6 fibers/lung; 37% of 3 d remaining. Kdis: 28.9 ng/cm^2/hr; Silica leached in NaOxalate pH 4.6: 1.6% pH 7: 2.8%
MMVF22 Slag Wool				Recovered fibers: All fibers: mean diameter 0.89 µm (3 d) and 0.47 µm (12 mo). Fibers longer than 20 µm: 1.11 µm (3 d) and 1.08 µm (12 mo). Lengths were distributed across ranges.		Fibers longer than 0.4 µm: 3 d: 7.09 x 10^6 fibers/lung 12 mo: 1.5x 10^6 fibers/lung 21% of 3 d remaining. Kdis: 52.8 ng/cm^2/hr; Silica leached in NaOxalate pH 4.6: 51.3% pH 7: 52.7% *solution saturated at 14 d, data extrapolated
RCF1				Recovered fibers: All fibers: mean diameter 0.95 µm (3 d) and 0.93 µm (12 mo). Fibers longer than 20 µm: 1.18 µm (3 d) and 1.1 µm (12 mo). Lengths were distributed across ranges.		Fibers longer than 0.4 µm: 3 d: 7.39 x 10^6 fibers/lung 12 mo: 4.53 x 10^6 fibers/lung; 61% of 3 d remaining. Kdis: 4.4 ng/cm^2/hr; Silica leached in NaOxalate pH 4.6: 0.3% pH 7: 0.4%
RCF2				Recovered fibers: All fibers: mean diameter 1.11 µm (3 d) and 1.3 µm (12 mo). Fibers longer than 20 µm: 1.64 µm (3 d) and 1.46 µm (12 mo). Lengths were distributed across ranges.		Fibers longer than 0.4 µm: 3 d: 3.81 x 10^6 fibers/lung 12 mo: 2.62 x 10^6 fibers/lung; 69% of 3 d remaining. Kdis: 3.1 ng/cm^2/hr; Silica leached in NaOxalate pH 4.6: 0.2% pH 7: 0.4%
RCF4 (heat-treated RCF1)				Recovered fibers: All fibers: mean diameter 1.21 µm (3 d) and 1.29 µm (12 mo). Fibers longer than 20 µm: 2.1 µm (3 d) and 1.96 µm (12 mo). Majority of fibers recovered were <10 µm long.		Fibers longer than 0.4 µm: 3 d: 8.9 x 10^6 fibers/lung 12 mo: 7.08 x 10^6 fibers/lung; 80% of 3 d remaining. Kdis: 0.5 ng/cm^2/hr; Silica leached in NaOxalate pH 4.6: 0.1% pH 7: 0.2%
MMVF11 Glass Fiber	Not listed	Not listed	Sheep, single instillation (100 mg/animal). 10/group removed for analysis at days 6, 40, 60, 180, 360, and 730	Not listed	Crocidolite (100 mg/100 ml); All lengths: T1/2 (fast) 69-90 d; T1/2 (slow) 501 d; Fibers ≤5 µm: T1/2 (fast) 93-101 d; T1/2 (slow) 340 d. Pathology: Day 6, 40, 60: peribronchiolar inflammatory lesions, few fibers. Day 180: peribronchiolar inflammatory lesions or without lesions, few fibers. Day 360: peribronchiolar inflammatory lesions, beginning of fibrosis, notable amt of fibers. Day 720: no significant lesions, interstitial pneumonitis, few fibers	Days 6, 40: alveolitis, no lesions, none/few fibers. Days 60, 180, 360, 720: no lesions, none/few fibers. All lengths: T1/2 (fast) 33-42 d; T1/2 (slow) 462 d; Fibers >5 µm: T1/2 (fast) 31-35 d; T1/2 (slow) 572 d	Dufresne et al. 1999
MMVF21 Rock Wool						Day 6: alveolitis, no lesions, few fibers. Day 40: alveolitis, no lesions, notable amount of fiber. Days 60, 180: alveolitis, no lesions, few fibers. Day 360: alveolitis, no lesions, notable amount of fibers. Day 720: alveolitis, no lesions, few fibers. All lengths: T1/2 (fast) 45 d; T1/2 (slow) 433 d; Fibers >5 µm: T1/2 (fast) 45 d; T1/2 (slow) 517 d
RCF-1						Day 6: light alveolitis, few fibers. Day 40: alveolitis, notable amt of fibers. Days 60, 180: alveolitis, few fibers. Day 360: no lesions very few fibers. Day 720: no lesions, very few fibers. All lengths: T1/2 (fast) 67-156 d; T1/2 (slow) 537 d; Fibers >5 µm: T1/2 (fast) 61-157 d; T1/2 (slow) 254 d
MMVF10a Glass Fiber	RCC sourced, details not provided	Not listed	Male Syrian Golden hamsters, aerosol delivered (2 mg/ml) through a tracheal cannula for 5-12 minutes until a sufficient number of fibers for microscopic analysis were delivered. The number of inhaled and exhaled fibers were measured by pneumotachography and laserlight scattering photometry.	Avg length of 200 fibers measured by light microscopy: 16 µm (range 1-66 µm); avg diameter 1.1 µm (range 0.6-1.7); ratio 15.6 (range 1.1-64)	No positive or negative controls	Fibers on the surfaces of conducting airways/alveoli. 89% of the fibers were totally and 11% partially covered by lining-layer material in airways. 32% of the fibers were totally submersed; others touched the alveolar wall, stuck at one end, bridging the airspace in alveoli. Studies in a surface balance showed that fibers were immersed into the aqueous subphase by approximately 50% at film surface tensions of 20–25 mJ/m^2 and were immersed at approximately 10 mJ/m^2. Fibers were also phagocytosed by macrophages. Substantial number of particle profiles within alveolar blood capillaries.	Geiser et al. 2003
Rock Wool	Not listed	Not listed	Male Fischer rats, 2 mg (1x in 0.2 ml) or 2 mg (2 mg once/wk for 4 wks); 2 or 8 mg for 4 or 16 weeks	Not listed	No positive control; Included vehicle control	SOD activity significantly decreased in BALF (8 mg 4w). Ascorbic Acid significantly increased in lung tissue (8 mg 4 wks)	Kovacikova et al. 2005
MMVF10 Glass Wool
RCF	Not listed	Not listed	Male Wistar rats- 2 x 2 mg in 0.4 ml (once weekly) 6 mo recovery; 4 mg total dose	All fibers listed as > 20 µm long; only 18% ≤1 µm in diameter	No positive control; Included vehicle control	Statistically significant increase in proportion of lymphocytes and neutrophils, % immature alveolar macrophages (AM) and phagocytic activity of AM; Statistically significant decrease in viability of AM and proportion of AM as a % of total in comparison with the control group.	Hurbankov et al. 2005
Rock wool RW1	Not listed	Not listed	Male wild-type or homozygous -lacI transgenic F344 Rats instilled either 1x (1 or 2 mg) or 2 mg 1x weekly for 4 weeks. 10 mg/ml stock suspension. 1, 2, or 8 mg total dose	1 mg contained 1,700 fibers; mean 1.8 µm x 16.5 µm; median 3 µm x 23 µm; 85% of fibers <5 µm or diameter <60 µm long	No positive control; Included vehicle control	Significant increase in mutation frequency after 16 wk exposure; DNA strand breaks were increased in alveolar macrophages and lung epithelial cells; Neutrophils in BALF at greater level than MMVF10 at 16 weeks post-exposure; no increase in oxidative markers. T1/2=∼3-4 mo	Topinka et al. 2005
MMVF10 Glass Wool				1 mg contained 8,300 fibers longer than 5 µm and 3,300 fibers >20 µm long; mean 1.31 µm x 22.6 µm; median 2 µm x 26 µm; 85% of fibers <3 µm diameter or <66 µm long		No change in mutation frequency; DNA strand breaks were increased in alveolar macrophages and lung epithelial cells; lower levels of neutrophils in BALF compared to RW1 at 16 weeks post-exposure; No increase in oxidative markers. T1/2= ∼10-11 mo; W-T1/2= ∼37 days (from Hesterberg et al. 1998)

Table A4. Summary of whole-body inhalation toxicology studies of SVFs.

Fiber	Fiber prep & binder/coating	Exposure	Fiber dimensions	Exposure/respirable fiber conc	Controls	Outcomes	Reference
Glass Fibers	Ball milled; Uncoated	135 mg/m^3	Averaging 1 µm in diameter of fiber lengths of predominately 50 µm or less; 70-76% fibrous and 24-30% non-fibrous particles	Rats, hamsters; 5 hr/d, 5 d/wk for 24 months with interim sacrifices at 6 and 12 mo. Not determined (average concentration in dust chamber estimated at 100 mg/m^3)	Fresh air controls	No lesions noted in either rats or hamsters; alveoli around respiratory bronchioles & alveolar ducts contained dust filled macrophages	Gross et al. 1970
	Ball milled; Phenol-Formaldehyde	106 mg/m^3
	Ball milled; Starch Binder	113 mg/m^3
Glass Fiber	Ball Milled	0.7 × 10^6/L	Ball-milled diameters ranged from 0.2-6.5 µm; ∼80% of lengths ≤3 µm	Sprague Dawley rats, hamsters, guinea pigs 5 hr/d, 5 d/wk for 3 months. 0.424 mg/L (6.54 x 10^6 total fibers/L air; 0.73 x 10^6 fibers/L >5 µm long)	Amosite 0.30 mg/L with 3.1 x 10^6/L >5 µm; 3100 fibres/cm^3 > 5 μm in length; 38% with length < 3 μm or ∼1178 WHO fibers/cm^3 3/12 animals with tumors; Fresh air controls	Hamsters 0/61; Rats 2/61; guinea pigs 2/61 (bronchoalveolar adenomas)	Lee et al. 1981; IARC 2002
Fybex (Potassium Octanitate)	–	2.9 × 10^6/L	6.7 x 0.2 µm	Sprague Dawley rats, hamsters, guinea pigs 5 hr/d, 5 d/wk for 3 months. Expt 1: 0.079 mg/L; (Expt 2: 0.039, 0.082, or 0.37 mg/L; Expt 1: 8.53 x 10^6 total fibers/L; 41.76 x 10^6 fibers/L >5 µm long; Expt 2: 6.72, 36.10, or 101.5 x 10^6 total fibers/L air; 2.94, 13.49, or 41.76 x 10^6 fibers/L >5 µm long)		Expt 1: Hamsters 1/64 (pleural mesothelioma); Rats 1/64 (bronchoalveolar adenoma); guinea pigs 0/64. Expt 2: rats 3/63 and 1/58 in mid and high exposure, guinea pigs no tumors, hamsters 1/63 mesothelioma in the mid exposure group and 1/63 pleural mesothelioma and 2/58 [one with mesothelioma, one bronchoalveolar adenoma)
PKT (pigment potassium titanate)	–	2.0 × 10^6/L	4.2 x 0.2 µm	Sprague Dawley rats, hamsters, guinea pigs 5 hr/d, 5 d/wk for 3 months. 0.073 mg/L (6.53 x 10^6 total fibers/L air; 2.00 x 10^6 fibers/L >5 µm long)		Hamsters 0/64; Rats 0/64; guinea pigs 0/64
SG Glass Wool	Ring vibro-grinding mill; Timbrell Dust Generator; Resin Free	5 mg/m^3	68.7% of fibers with diameters < 1 µm, 41.5% of fibers ≤ 10 µm in length (65.5% respirable)	Wistar rats, 5 hr/d, 5 d/wk for 12 or 24 months. 48 WHO fibers/cm^3 (avg respirable dust concentrations at 12 months 5.20 mg/m^3; 4.96 mg/m^3 at 24 mo)	Canadian chrysotile (714-7D; 89.9% <5 µm in length; 93.9% respirable fibers <5 μm in length; 5901 WHO fibers/cm^3). Measured in respirable dust chambers at 12 mo: 4.99 mg/m^3 or 5.20 mg/m^3 at 24 mo; 5/24 males and 4/23 females with pulmonary tumors. Lowest lung retention by mass of 0.70 (12 mo) and 1.66 (24 mo) mg/g dry lung weight (83% clearance 16 mo post 12 mo exposure)	Lung: 12 and 24 mo: 2.12 and 3.19 mg/g dry tissue (67% clearance 16 mo post 12 mo exposure). 1/23 males and 0/22 females had pulmonary tumors; 3/22 females with digestive tumors	Le Bouffant et al. 1987
SG Rock Wool	Carded then ring vibro-grinding mill; Timbrell Dust Generator; Resin Free		22.7% <1 μm in diameter, 39.9% of fibers ≤10 μm in length (56.6% respirable)	Wistar rats, 5 hr/d, 5 d/wk for 12 or 24 months. 11 WHO fibers/cm^3 (avg respirable dust concentrations at 12 months 4.14 mg/m^3; 5.37 mg/m^3 at 24 mo)		Lung: 12 and 24 mo: 1.03 and 2.57 mg/g dry tissue (84% clearance 16 mo post 12 mo exposure). 0/23 males and 0/22 females with pulmonary tumors; 1/24 digestive tumor (male) **Rock wool had a high percentage of non-fibrous contamination
JM 100 glass fiber	Timbrell Dust Generator (binder not noted)		100% ≤ 1 μm in diameter (EM); length 94.1% <5 μm (electron microscopy-EM) 48.4% 5-10 μm (light microscopy-LM) (96.6% [EM] and 41% [LM] respirable)	Wistar rats, 5 hr/d, 5 d/wk for 12 or 24 months. 332 WHO fibers/cm^3 (avg respirable dust concentrations at 12 months 6.34 mg/m^3; 5.39 mg/m^3 at 24 mo)		Lung: 12 and 24 mo: 9.18 and 15.32 mg/g dry tissue (62% clearance 16 mo post 12 mo exposure). No pulmonary or digestive tumors
Rock Wool	Resin free; brief milling	10 mg/m^3	42.6% of the fibers ≤ 1 μm in diameter and 5-20 μm in length	Sprague Dawley rats, 7 hr/d, 5 d/wk for 12 months. 227 WHO fibers/cm^3	UICC Canadian chrysotile (10 mg/m^3 equivalent to 3832 WHO fibers/cm^3) 82.3% of fibers with a diameter ≤1 μm between 5-20 μm in length. 25% tumor incidence; BAH noted in 5/48 animals. Retained dust 12 mo: 0.42 mg f/lung; 24 mo: 0.4 mg f/lung	12 mo: 3.11 mg f/lung; 24 mo 1.5 mg f/lung. 2/48; one animal had bronchoalveolar hyperplasia (BAH)	Wagner et al. 1984
Glass Wool	Resin free; brief milling		32.0% of fibers with a diameter of ≤1 μm and length between 5-20 μm	Sprague Dawley rats, 7 hr/d, 5 d/wk for 12 months. 323 WHO fibers/cm^3		12 mo: 0.94 mg f/lung; 24 mo 0.2 mg f/lung. 1/47 (benign adenoma); one incidence of BAH.
Glass Wool	Resin coating; brief milling		42.2% of fibers with a diameter of ≤1 μm and length between 5-20 μm	Sprague Dawley rats, 7 hr/d, 5 d/wk for 12 months. 240 WHO fibers/cm^3		12 mo: 1.88 mg f/lung; 24 mo 0.5 mg f/lung. 1/48 (adenocarcinoma); 3 BAH
JM 100 glass fiber	Not noted; brief milling		89.7% ≤ 1 μm in diameter and length between 5-20 μm; 71.2% were shorter than 10 µm	Sprague Dawley rats, 7 hr/d, 5 d/wk for 12 months. 1436 WHO fibers/cm^3		12 mo: 4.45 mg f/lung; 24 mo: 2.1 mg f/lung. 1/48 adenocarcinoma; three incidences of BAH

Table A5. Summary of nose-only inhalation toxicology studies of SVFs.

Fiber	Fiber prep	Exposure	Fiber dimensions	Exposure/aerosol concentrations	Controls	Outcomes	Reference
JM100 Glass Fibers (no binder)	Wetted with ethanol then placed in a plug-packing mold assembly (dried overnight); aerosolized through a Timbrell generator	3.0 or 0.3 mg/m^3	Mean diameter 0.45 µm, GMD 0.4 µm and geometric mean length [GML] 4.7 µm; 94% fibers ≤20 µm long	Osborne-Mendel rats and Syrian Golden hamsters 6 hr/d, 5 d/wk for 24 months. 3000 or 300 fibers/cm^3; mean 2.4 mg/m^3; particle:fiber ratio 4:1	UICC crocidolite (7 mg/m^3, 3000 fibers/cm^3) Rats 1/59, 1/61, and 3/57 (1 day, 30 day or 24 month exposure, respectively) 10, 36, or 53% fibrosis (1 day, 30 day and 24 month exposure, respectively); Hamsters no tumors at any time point, 2, 22, or 53% fibrosis (1 day, 30 day and 24 month exposure, respectively)	3 mg/m^3: Rats: 1.87 × 10^6 fibers/mg dry lung weight; Hamsters: 0.96 × 10^6 fibers/mg dry lung; 0.3 mg/m^3: Rats: 1.23 × 10^5 fibers/mg dry lung weight; Hamsters: 0.52 × 10^5 fibers/mg dry lung. 3 mg/m^3: Rats 0/57, fibrosis 7% Hamsters 0/69, fibrosis 6%; 0.3 mg/m^3: Rats 0/57, fibrosis 5% Hamsters 0/70, fibrosis 4%	Smith et al. 1987
Certainteed Glass Wool (silicone lubricant)		10 mg/m^3	Mean diameter 3.1 µm, GMD 1.2 µm and GML 24 µm; 99% of fibers were >5 µm, 77% >20 µm, and 68% fibers 10–80 µm long	Rats and hamsters 6 hr/d, 5 d/wk for 24 months. 100 fibers/cm^3; mean 4.4 mg/m^3; particle:fiber ratio 6:1		Rats: 2.85 × 10^4 fibers/mg dry lung weight; Hamsters: 1.15 × 10^4 fibers/mg dry lung. Rats 0/52, fibrosis 8%; Hamsters 0/60, fibrosis 3%
Manville building insulation (phenol-formaldehyde binder)		9 mg/m^3	Mean diameter 6.1 µm, GMD 3.0 µm, and GML 83 µm; approximately 53% fibers <2.5 µm diameter and 36% 10–80 µm long	Rats and hamsters 6 hr/d, 5 d/wk for 24 months. 25 fibers/cm^3; mean 7.0 mg/m^3; particle:fiber ratio 31:1		Rats: 5.72 × 10^2 fibers/mg dry lung weight; Hamsters: 5.05 × 10^2 fibers/mg dry *Retained dose low. Rats 0/58, fibrosis 5%; Hamsters 0/61 and 0/38, no fibrosis
Owens-Corning High Temperature Range Glass Wool (phenol-formaldehyde binder)		12 or 1.2 mg/m^3	Mean diameter 5.4 µm, GMD 1.1 µm, and GML 20 µm; 74% fibers <2.0 µm diameter and 67% 10-80 µm long	Rats and hamsters 6 hr/d, 5 d/wk for 24 months. 100 or 10 fibers/cm^3; mean 9.9 mg/m^3; particle:fiber ratio 38:1		12 mg/m^3: Rats: 1.0 × 10^3 fibers/mg dry lung weight; Hamsters: 2.35 × 10^3 fibers/mg dry lung; 1.2 mg/m^3: Rats: 1.11 × 10^3 fibers/mg dry lung weight; Hamsters: 0.73 × 10^2 fibers/mg dry lung *Retained dose low. 12.0 mg/m^3: Rats 0/57, fibrosis 7% Hamsters 1/66, no fibrosis; 1.2 mg/m^3: Rats 0/61, 8% fibrosis Hamsters 0/65, fibrosis 2%
RCF Carborundum Fibrefrax (no binder)		12 mg/m^3	Mean diameter 1.8 µm, GMD 0.9 µm, and GML 25 µm; 83% >10 µm long and 86% <2 µm in diameter	Rats and hamsters 6 hr/d, 5 d/wk for 24 months. 200 fibers/cm^3; 88 fibers/cm^3 >10 µm long and ≤1.0 µm diameter; mean 10.8 mg/m^3; particle:fiber ratio 33:1		Rats: 2.18 × 10^4 fibers/mg dry lung weight; Hamsters: 0.86 × 10^4 fibers/mg dry lung. Rats 0/55, fibrosis 22% Hamsters 1/70 (mesothelioma), 1% fibrosis
JV Spinner Slag Wool (no binder)		10 mg/m^3	Mean diameter 2.7 µm with a GMD of 0.9 µm and GML 22 µm; 93% of fibers had diameters <2.5 µm with 75% of fibers >10 µm long	Rats and hamsters 6 hr/d, 5 d/wk for 24 months. 200 fibers/cm^3; 76 fibers/cm^3 >10 μm in length and ≤1.0 μm in diameter; mean 7.8 mg/m^3; particle:fiber ratio 28:1		Rats: 3.08 × 10^3 fibers/mg dry lung weight; Hamsters: 3.07 × 10^3 fibers/mg dry lung. Rats 0/55, fibrosis 16% Hamsters 0/69, 1% fibrosis
JM 104/475 Glass Fibers (-/+ SO2)	Knife milled; Vibrating Bed Aerosol Generator	3.0 mg/m^3	Dimensions of 104/475 starting material not listed	Female Wistar rats; 5 hr/d, 4 d/wk for 6 wk and followed for 24 mo. 576 fibers/mL, 252 were >5 µm long, 90% fibers <0.80 µm and <12.4 µm length	Calidria chrysotile (6.0 mg/m^3; 241 f/ml total fibers, 131 f/ml >5 µm in length; 90% of aerosolized chrysotile was <1.6 µm diameter and <14.0 µm length) or UICC crocidolite (2.2 mg/m^3; 2011 total fibers/ml, 162 f/ml >5 µm in length); 74% animals exposed to crocidolite developed bronchoalveolar hyperplasia also one squamous metaplasia and one adenocarcinoma. No tumors in chrysotile group	104/475 (-SO2): 183, 310, and 187 x 10^6 fibers/lung; 104/475 (+SO2): 208, 340, and 233 x 10^6 fibers/lung; crocidolite: 403, 557, 229 x 10^6 fibers/lung; chrysotile: 398, 347, 223 x 10^6 fibers/lung all at 6, 12, and 24 mo, respectively. 104/475 (-SO2): Bronchoalveolar hyperplasia 11% (NS), one squamous metaplasia, one squamous cell carcinoma (107 animals); 104/475 (+SO2): Bronchoalveolar hyperplasia 16%, one adenoma (108 animals)	Muhle et al. 1987
MMVF 10 glass wool	Size-selected fine respirable fractions from commercial bulk products	3, 16, or 30 mg/m^3	Size selected stock: GMD 1.13 µm and GML 14.4 µm	Fischer F344 rats; 6 hr/d, 5 d/wk for 24 mo (3 or 6 animals sacrificed at 3, 6, 12, and 18 mo). 29.1 mg/m^3 (3 and 16 mg/m^3 aerosol not measured) Aerosolized: GMD 1.26 µm and GML 13.1 µm	10 mg/m^3 chrysotile (GMD 0.08 µm and GML of 1.2 µm) pulmonary fibrosis, mesothelioma, and significant increases in lung tumors in chrysotile exposed animals of 13/69 (18.9%)	30 mg/m^3: Highest levels at 24 mo: 4.16, 2.69, and 0.37 x 10^5 total fibers/mg dry tissue with 2.88, 1.85, and 0.24 x 10^5 WHO fibers for 30, 16, and 3 mg/m^3, respectively. No tumors observed in 3 mg/m^3 exposure group; one adenoma and no carcinomas in 16 mg/m^3 group; six adenomas and one carcinoma in 30 mg/m^3 exposure group.	Hesterberg et al. 1993
MMVF 11 glass wool	Size-selected fine respirable fractions from commercial bulk products	3, 16, or 30 mg/m^3	Size selected stock: GMD 0.76 µm and GML 13.0 µm	Fischer F344 rats; 6 hr/d, 5 d/wk for 24 months (3 or 6 animals sacrificed at 3, 6, 12, and 18 mo). 28.9 mg/m^3 (3 and 16 mg/m^3 aerosol not listed) Aerosolized: GMD of 0.69 µm and a GML of 13.7 µm		30 mg/m^3: Highest levels at 24 mo: 6.42, 3.46, and 0.62 x 10^5 fibers/mg dry tissue with 5.03, 2.35, and 0.48 x 10^5 WHO fibers for 30, 16, and 3 mg/m^3, respectively. Three adenomas and 1 carcinoma observed in 3 mg/m^3 exposure group; six adenomas and three carcinomas in 16 mg/m^3 group; three adenomas and no carcinomas in 30 mg/m^3 exposure group.
RCF1	Size-selected fine respirable fractions from commercial bulk products	30 mg/m^3	Size selected stock: GMD 0.86 µm and GMD 16.5 µm	Fischer F344 rats; 6 hr/d, 5 d/wk for 24 mo (3 or 6 animals sacrificed at 3, 6, 12, and 18 mo). 29.1 mg/m^3; Aerosolized: GMD 0.82 µm and GML 15.9 µm		30 mg/m^3: Highest levels at 18 mo: 4.67 x 10^5 total fibers/mg dry tissue with 3.57 x 10^5 WHO fibers; 24 mo: 3.70 x 10^5 total fibers/mg dry tissue with 2.75 x 10^5 WHO fibers. Pulmonary fibrosis, mesothelioma, tumor incidence 16/123 (13.0%)
MMVF21 Rock Wool			GMD about 0.95 µm and GML about 14 µm; WHO fibers/cm^3 of 34, 150 and 243; fibers >20 µm long 13, 74 and 114, respectively	6 hr/d, 5 d/wk for 18 months (Syrian Golden hamsters); or 24 months (Osborne-Mendel or Fisher 344 rats). 3.1 mg/m^3: 44 total fibers/cm^3 (34 WHO fibers/cm^3; GMD 0.94 x GML 13.0 µm); 16.1 mg/m^3: 185 total fibers/cm^3 (150 WHO fibers/cm^3; GMD 0.90 x GML 15.4 µm), or 30.4 mg/m^3: 264 total fibers/cm^3 (243 WHO fibers/cm^3; dimensions not given)	10 mg/m^3 crocidolite (GMD of 0.28 µm, determined by SEM, and a GML of 4.1 µm, determined by light microscopy; of 106 crocidolite-treated rats, ten had adenomas, six had pulmonary carcinomas and two had mesothelioma; crocidolite lung burdens were 759 x 10^3 WHO fibers (41.1 x 10^3 fibers >20 µm)	24 mo: 242 x 10^3 WHO fibers (39.5 x 10^3 fibers >20 µm) for the 30 mg/m^3 exposure, 217 x 10^3 WHO fibers (25.6 x 10^3 >20 µm) for the 16 mg/m^3 exposure, and 55.7 x 10^3 WHO fibers (9.0 x 10^3 >20 µm) for the 3 mg/m^3 exposure. Four lung adenomas (3.5%) and one lung carcinoma (0.8%) each were observed in the 3, 16 and 30 mg/m^3 rock wool exposure groups, respectively. Minimal focal fibrosis noted.	McConnell et al. 1994; IARC 2002
MMVF22 Slag Wool			GMD 0.85 µm and GML 13 µm, WHO fibers/cm^3 of 30, 131 and 213	6 hr/d, 5 d/wk for 18 months (Syrian Golden hamsters); or 24 months (Osborne-Mendel or Fisher 344 rats). 3.1 mg/m^3: 33 total fibers/cm^3 (30 WHO fibers/cm^3; GMD 0.84 x GML 12.3 µm); 16.1 mg/m^3: 158 total fibers/cm^3 (131 WHO fibers/cm^3; GMD 0.84 x GML 13.2 µm), or 29.9 mg/m^3: 245 total fibers/cm^3 (213 WHO fibers/cm^3; GMD 0.87 x GML 15.2 µm)		24 months was 177 × 10^3 (11.0 × 10^3 fibers >20 μm), for the 30 mg/m^3 exposure, 96.7 × 10^3 (4.5 × 10^3 fibers >20 μm) for the 16 mg/m^3 exposure and 44.4 × 10^3 WHO fibers (1.8 × 10^3 fibers >20 μm) for the 3 mg/m^3 exposure. One adenoma and one carcinoma were noted in the 3 mg/m^3 exposure; no tumors were observed in the group that received the 16 mg/m^3 exposure. In the group treated with 30 mg/m^3, two pulmonary adenomas and one lung carcinoma
Kaolin-Based RCF	Pre-sized fiber (1 µm diameter 22-26 µm long)	3, 9, 16, or 30 mg/m^3	215 WHO fibers/cm^3, mean diameter 0.9 µm and mean length 22.1 µm; also tested 16, 9, and 3 mg/m^3	Fischer F344 rats: 6 hr/d, 5 d/wk for 24 mo (terminal sacrifice at 29 mo); Hamsters: 6 hr/d, 5 d/wk for 18 mo (terminal sacrifice at 20 mo). 30 mg/m^3, 215 WHO fibers/cm^3, mean diameter 0.9 µm and mean length 22.1 µm; also tested 16, 9, and 3 mg/m^3	11 mg/m^3 chrysotile; 10,600 WHO fibers/cm^3, mean diameter 0.09 µm and mean length 2.17 µm. Incidence of bronchoalveolar neoplasms following chrysotile exposure was 18.8%; a single mesothelioma was noted. Lung burden 24 mo: 1.08 x 10^7	24 mo: Hamsters 2.39 x 10^5 f/mg dry lung Rats: 24 mo: 3.70 x 10^5, 2.70 x 10^5, 1.80 x 10^5, and 5.5 x 10^4 f/mg dry lung for 30, 16, 9, and 3 mg/m^3, respectively. Fibrosis Wagner Grade 4 for 30 and 16 mg/m^3 (rats and hamsters), Grade 3.2 for 3 mg/m^3; 14.6% tumor incidence in rats and no tumors in hamsters. Mesothelioma 1.6% (rats) and 42% (hamsters)	Mast et al. 1994
Zirconia Based RCF	Pre-sized fiber (1 µm diameter 22-26 µm long)	30 mg/m^3	224 WHO fibers/cm^3, mean diameter 1.0 µm and mean length 18.6 µm	Fischer F344 rats: 6 hr/d, 5 d/wk for 24 mo (terminal sacrifice at 29 mo); Syrian Golden hamsters: 6 hr/d, 5 d/wk for 18 mo (terminal sacrifice at 20 mo). 29 mg/m^3, 224 WHO fibers/cm^3, mean diameter 1.0 µm and mean length 18.6 µm		24 mo: 9.60 x 10^5 f/mg dry lung. Fibrosis Wagner Grade 4; 8.8% tumor incidence and 2.5% mesotheliomas
High-Purity RCF	pre-sized fiber (1 µm diameter 22-26 µm long)	30 mg/m^3	172 WHO fibers/cm^3, mean diameter 1.0 µm and mean length 24.2 µm	Fischer F344 rats: 6 hr/d, 5 d/wk for 24 mo (terminal sacrifice at 29 mo); Syrian Golden hamsters: 6 hr/d, 5 d/wk for 18 mo (terminal sacrifice at 20 mo). 29 mg/m^3, 172 WHO fibers/cm^3, mean diameter 1.0 µm and mean length 24.2 µm		24 mo: 2.70 x 10^5 f/mg dry lung. Fibrosis Wagner Grade 4.3; 16% tumor incidence and 1.7% mesothelioma incidence
After-Service RCF	Pre-sized fiber (1 µm diameter 22-26 µm long)	30 mg/m^3	166 WHO fibers/cm^3, mean diameter 1.3 µm and mean length 12.7 µm	Fischer F344 rats: 6 hr/d, 5 d/wk for 24 mo (terminal sacrifice at 29 mo); Syrian Golden hamsters: 6 hr/d, 5 d/wk for 18 mo (terminal sacrifice at 20 mo). 30 mg/m^3, 166 WHO fibers/cm^3, mean diameter 1.3 µm and mean length 12.7 µm		24 mo: 5.90 x 10^5 f/mg dry lung. Fibrosis Wagner Grade 4; 4% tumor incidence and 0.8% mesothelioma
Kaolin-Based RCF (RCF1)	Size-selected (approximately 1 µm diameter and approximately 20 µm length)	30 mg/m^3	Stock fiber dimensions: GMD 0.86 µm and GML 16.5 µm	Fischer F344 rat: 6 hr/d, 5 d/wk for 3, 6, 9, 12, 18, or 24 mo (terminal sacrifice at 20% survival 30 mo). 29.1 mg/m^3, 234 total fibers/cm^3 (187 WHO fibers/cm^3, GMD 0.82 µm and GML 15.9 µm; 43.2% of fibers >20 µm long	10 mg/m^3 chrysotile; (gravimetric 10.1 mg/m^3) 1.05 x 10^5 total fibers/cm^3 (1.06 x 10^5 WHO fibers/cm^3, GMD 0.85 µm and GML 15.9 µm; 43.2% of fibers >20 µm long. Highest level 9 mo: 367 x 10^5 fibers/mg of dry lung tissue (32 x 10^5 WHO fibers) 30 mo: 85 x 10^5 fibers/mg of dry lung tissue (4.25 x 10^5 WHO fibers). The incidence of bronchoalveolar neoplasms following chrysotile exposure was 18.8%; a single mesothelioma was noted. Air controls	Highest level 18 mo: 4.67 x 10^5 fibers/mg of dry lung tissue (3.57 x 10^5 WHO fibers) 30 mo: 1.7 x 10^5 fibers/mg of dry lung tissue (1.33 x 10^5 WHO fibers). 13% incidence of bronchoalveolar adenomas and carcinomas; two mesotheliomas; cellular and fibrotic histopathology grade 2-4.3	Mast et al. 1995a
Zirconia Based (RCF 2)	Size-selected (approximately 1 µm diameter and approximately 20 µm length)	30 mg/m^3	Stock fiber dimensions: GMD 0.91 µm and GML 14.6 µm	F344 Rats: 6 hr/d, 5 d/wk for 3, 6, 9, 12, 18, or 24 mo (terminal sacrifice at 20% survival 30 mo). 28.9 mg/m^3 268 total fibers/cm^3 (220 WHO fibers/cm^3), GMD 0.88 µm and GML 12.8 µm; 32.4% fibers >20 µm long		Highest level 18 mo: 9.82 x 10^5 fibers/mg dry lung tissue (7.97 x 10^5 WHO fibers) 30 mo: 1.68 x 10^5 fibers/mg dry lung tissue (1.35 x 10^5 WHO fibers). 7.4% incidence of bronchoalveolar adenomas and carcinomas; three mesotheliomas; cellular and fibrotic histopathology grade 2-4.2
High-Purity RCF (RCF 3)	Size-selected (approximately 1 µm diameter and approximately 20 µm length)	30 mg/m^3	Stock fiber dimensions: GMD 0.94 µm and GML 17.7 µm	F344 Rats: 6 hr/d, 5 d/wk for 3, 6, 9, 12, 18, or 24 mo (terminal sacrifice at 20% survival 30 mo). 29.2 mg/m^3 213 total fibers/cm^3 (182 WHO fibers/cm^3), GMD 0.85 µm and GML 17.4 µm; 47.9% of fibers >20 µm long		Highest level 18 mo: 3.23 x 10^5 fibers/mg dry lung tissue (2.60 x 10^5 WHO fibers) 30 mo: 1.92 x 10^5 fibers/mg dry lung tissue (1.45 x 10^5 WHO fibers). 15.7% incidence for bronchoalveolar adenomas and carcinomas; two mesotheliomas; cellular and fibrotic histopathology grade 2.3-4.7
After-Service RCF (RCF 4)	Size-selected (approximately 1 µm diameter and approximately 20 µm length)	30 mg/m^3	Stock fiber dimensions: GMD 1.1 µm and GML 7.0 µm	F344 Rats: 6 hours a day, 5 days per week for 3, 6, 9, 12, 18, or 24 mo (terminal sacrifice at 20% survival 30 mo). 30.1 mg/m^3 206 total fibers/cm^3 (153 WHO fibers/cm^3), GMD 1.2 µm and GML 9.0 µm; 16.4% fibers >20 µm long		Highest level 24 mo: 5.95 x 10^5 fibers/mg of dry lung tissue (3.48 x 10^5 WHO fibers) 30 mo: 4.56 x 10^5 fibers/mg of dry lung tissue (2.85 x 10^5 WHO fibers). 3.4% tumor incidence for bronchoalveolar adenomas and carcinomas; one mesothelioma; cellular and fibrotic histopathology grade 1-4.3
Kaolin-Based RCF	Size selected (approximately 1 µm in diameter and approximately 20 µm in length)	3, 9, or 16 mg/m^3 (compared to 30 mg/m^3 from Mast et al. 1995a	Stock fiber dimensions: GMD 0.86 µm and GML 16.5 µm	Fischer 344 rats: 6 hr/d, 5 d/wk for 24 mo (terminal sacrifice at 20% survival 30 mo). 26, 75, 120, or 187 WHO fibers/cm^3, GMD 0.82 µm and GML 1.5-16.8 µm	Compared to Mast et al. 1995a; air controls	30 mg/m^3: Highest level at 18 mo exposure/18 mo sacrifice: 46.70 x 10^4 total fibers (WHO fibers: 35.70 x 10^4); 24 mo: 37.0 x 10^4 (27.50 x 10^4 WHO fibers) 16 mg/m^3: Highest level 24 mo exposure/24 mo sacrifice: 27.8 x 10^4 total fibers (22.1 x 10^4 WHO fibers). Minimal to mild inflammation; Minimal interstitial fibrosis by 12 mo (9 and 16 mg/m^3) progressed to mild at 24 mo (16 mg/m^3). Tumor incidence (adenoma/carcinoma) 1.6%, 3.9%, 1.6% for 3, 9, or 16 mg/m^3. One mesothelioma at 9 mg/m^3	Mast et al. 1995b
MMVF 11 glass fiber	Counting and sizing per Bernstein et al. 1995	30 mg/m^3	Size selected stock: GMD of 0.69 µm and GML of 12 µm	Fischer F344 rats, 6 hr/d for 5 d, post-exposure 1 hr, 1 d, 5 d, 4 wk, 13 wk, and 26 wk. Gravimetric concentration 30 mg/m^3; 451 total fibers/cm^3 (377 WHO fibers/m^3) GMD 0.75 µm and GML 13 µm	Not evaluated	Highest level at 5 days: 6.3 x 10^6 total fibers/lung with 5.1 x 10^6 WHO fibers; 26 weeks: 0.35 x 10^6 total fibers/lung with 0.23 x 10^6 WHO fibers W-T1/2=32 days (fibers 5-20 µm long) W-T1/2=28 days (WHO fibers) Kdis @ pH 7.4 = 71 ng/cm^2/h; @ pH 4.5 = 0.7 ng/cm^2/h	Bernstein et al. 1996
A Glass Wool (soluble)			Size selected stock: GMD 0.56 µm and GMD 9.5 µm	Fischer F344 rats, 6 hr/d for 5 d, post-exposure 1 hr, 1 d, 5 d, 4 wk, 13 wk, and 26 wk. Gravimetric concentration 35 mg/m^3; 1127 total fibers/cm^3 (944 WHO fibers/m^3) GMD of 0.59 µm and a GML of 10 µm		Highest level at 5 days: 13.8 x 10^6 total fibers/lung with 9.3 x 10^6 WHO fibers; 26 weeks: 0.89 x 10^6 total fibers/lung with 0.52 x 10^6 WHO fibers W-T1/2=16 days (fibers 5-20 µm long) W-T1/2=15 days (WHO fibers) Kdis @ pH 7.4 = 129 ng/cm^2/h; @ pH 4.5 = 2.4 ng/cm^2/h
B-01/09			Size selected stock: GMD 0.52 µm and GMD 12 µm	Fischer F344 rats, 6 hr/d for 5 d, post-exposure 1 hr, 1 d, 5 d, 4 wk, 13 wk, and 26 wk. Gravimetric concentration 29 mg/m^3; 619 total fibers/cm^3 (541 WHO fibers/m^3) GMD of 0.48 µm and a GML of 11 µm		Highest level at 1 hour: 7.7 x 10^6 total fibers/lung with 5.1 x 10^6 WHO fibers; 13 weeks: 0.35 x 10^6 total fibers/lung with 0.18 x 10^6 WHO fibers W-T1/2=11 days (fibers 5-20 µm long) W-T1/2=11 days (WHO fibers) Kdis @ pH 7.4 = 320 ng/cm^2/h; @ pH 4.5 = 11.8 ng/cm^2/h
C Glass Wool (soluble)			Size selected stock: GMD 0.56 µm and GMD 13 µm	Fischer F344 rats, 6 hr/d for 5 d, post-exposure 1 hr, 1 d, 5 d, 4 wk, 13 wk, and 26 wk. Gravimetric concentration 30 mg/m^3; 709 total fibers/cm^3 (589 WHO fibers/m^3) GMD of 0.43 µm and a GML of 11 µm		Highest level at 1 hour: 12.1 x 10^6 total fibers/lung with 7.9 x 10^6 WHO fibers; 26 weeks: 0.15 x 10^6 total fibers/lung with 0.07 x 10^6 WHO fibers W-T1/2=15 days (fibers 5-20 µm long) W-T1/2=14 days (WHO fibers) Kdis @ pH 7.4 = 309 ng/cm^2/h; @ pH 4.5 = 6.2 ng/cm^2/h
J (X-607, soluble)			Size selected stock: GMD 0.51 µm and GMD 11 µm	Fischer F344 rats, 6 hr/d for 5 d, post-exposure 1 hr, 1 d, 5 d, 4 wk, 13 wk, and 26 wk. Gravimetric concentration 31 mg/m^3; 804 total fibers/cm^3 (643 WHO fibers/m^3) GMD of 0.54 µm and a GML of 10 µm		Highest level at 1 hour: 7.4 x 10^6 total fibers/lung with 4.8 x 10^6 WHO fibers; 26 weeks: 0.48 x 10^6 total fibers/lung with 0.25 x 10^6 WHO fibers W-T1/2=24 days (fibers 5-20 µm long) W-T1/2=23 days (WHO fibers) Kdis @ pH 7.4 = 104 ng/cm^2/h; @ pH 4.5 = 4.3 ng/cm^2/h
F Stone Wool			Size selected stock: GMD 0.5 µm and GMD 9.4 µm	Fischer F344 rats, 6 hr/d for 5 d, post-exposure 1 hr, 1 d, 5 d, 4 wk, 13 wk, and 26 wk. Gravimetric concentration 31 mg/m^3; 899 total fibers/cm^3 (708 WHO fibers/m^3) GMD of 0.66 µm and a GML of 9.9 µm		Highest level at 1 hour: 15.6 x 10^6 total fibers/lung with 9.0 x 10^6 WHO fibers; 26 weeks: 0.62 x 10^6 total fibers/lung with 0.36 x 10^6 WHO fibers W-T1/2=28 days (fibers 5-20 µm long) W-T1/2=28 days (WHO fibers) Kdis @ pH 7.4 = 96 ng/cm^2/h; @ pH 4.5 = 5.9 ng/cm^2/h
G Stone Wool			Size selected stock: GMD 0.67 µm and GMD 11 µm	Fischer F344 rats, 6 hr/d for 5 d, post-exposure 1 hr, 1 d, 5 d, 4 wk, 13 wk, and 26 wk. Gravimetric concentration 30 mg/m^3; 981 total fibers/cm^3 (832 WHO fibers/m^3) GMD of 0.46 µm and a GML of 11 µm		Highest level at 1 hour: 10.0 x 10^6 total fibers/lung with 6.4 x 10^6 WHO fibers; 26 weeks: 0.33 x 10^6 total fibers/lung with 0.17 x 10^6 WHO fibers W-T1/2=23 days (fibers 5-20 µm long) W-T1/2=22 days (WHO fibers) Kdis @ pH 7.4 = 129 ng/cm^2/h; A pH 4.5 = 4.1 ng/cm^2/h
H Stone Wool			Size selected stock: GMD 0.66 µm and GMD 13 µm	Fischer F344 rats, 6 hr/d for 5 d, post-exposure 1 hr, 1 d, 5 d, 4 wk, 13 wk, and 26 wk. Gravimetric concentration 30 mg/m^3; 351 total fibers/cm^3 (299 WHO fibers/m^3) GMD of 0.68 µm and a GML of 13 µm		Highest level at 1 hour: 9.5 x 10^6 total fibers/lung with 6.7 x 10^6 WHO fibers; 26 weeks: 0.47 x 10^6 total fibers/lung with 0.33 x 10^6 WHO fibers W-T1/2=27 days (fibers 5-20 µm long) W-T1/2=26 days (WHO fibers) Kdis @ pH 7.4 = 169 ng/cm^2/h; @ pH 4.5 = 5.5 ng/cm^2/h
L Commercial Stone Wool			Size selected stock: GMD 0.63 µm and GMD 16.9 µm	Fischer F344 rats, 6 hr/d for 5 d, post-exposure 1 hr, 1 d, 5 d, 4 wk, 13 wk, and 26 wk. Gravimetric concentration 31 mg/m^3; 478 total fibers/cm^3 (420 WHO fibers/m^3) GMD of 0.51 µm and a GML of 15 µm		Highest level at 1 hour: 5.7 x 10^6 total fibers/lung with 4.0 x 10^6 WHO fibers; 26 weeks: 0.30 x 10^6 total fibers/lung with 0.23 x 10^6 WHO fibers W-T1/2=57 days (fibers 5-20 µm long) W-T1/2=54 days (WHO fibers) Kdis @ pH 7.4 = 23 ng/cm^2/h; @ pH 4.5 = 342 ng/cm^2/h
MMVF10 Glass Wool	Size-selected	30 mg/m^3	Details for each fiber type not given; approximate arithmetic mean 1 x 20 µm to simulate workplace exposure	Fischer F344 rats, 6 hr/d, 5 d/wk for 24 mo (3 or 6 animals sacrificed at 3, 6, 12, and 18 mo). 32 mg/m^3; Aerosolized: GMD of 1.1 µm and a GML of 11 µm	10 mg/m^3 Crocidolite (NIEHS) GMD 0.2 µm x GML 3 µm; Aerosolized: 11 mg/m^3; GMD of 0.3 µm and a GML of 4.2 µm. W-T1/2=986 days (fibers >20 µm long) W-T1/2=234 days (WHO fibers); 61% of WHO fibers retained (day 1 vs day 365)	1 d: 0.14 x 10^6 total fibers/lung >20 µm long and 2.8 x 10^6 WHO fibers; 365 d: 2.0 x 10^3 total fibers/lung >20 µm long with 3.0 x 10^5 WHO fibers W-T1/2=44 d (fibers >20 µm long) W-T1/2=89 d (WHO fibers); 11% of WHO fibers retained (d 1 vs d 365)	Hesterberg et al. 1996
CertainTeed B MMVF11 Glass Wool				Fischer F344 rats, 6 hr/d, 5 d/wk for 24 mo (3 or 6 animals sacrificed at 3, 6, 12, and 18 mo). 29 mg/m^3; Aerosolized: GMD of 0.7 µm and a GML of 12 µm		1 d: 0.97 x 10^6 total fibers/lung >20 µm long and 5.6 x 10^6 WHO fibers; 365 d: 2.0 x 10^3 total fibers/lung >20 µm long with 2.0 x 10^5 WHO fibers W-T1/2=6 d (fibers >20 µm long) W-T1/2=35 d (WHO fibers); 4% of WHO fibers retained (d 1 vs d 365)
Rockwool Intl MMVF21 Rock Wool				Fischer F344 rats, 6 hr/d, 5 d/wk for 24 mo (3 or 6 animals sacrificed at 3, 6, 12, and 18 mo). 31 mg/m^3; Aerosolized: GMD of 0.9 µm and a GML of 16 µm		1 d: 0.55 x 10^6 total fibers/lung >20 µm long and 1.8 x 10^6 WHO fibers; 365 d: 2.0 x 10^4 total fibers/lung >20 µm long with 1.0 x 10^5 WHO fibers W-T1/2=53 d (fibers >20 µm long) W-T1/2=111 d (WHO fibers); 8% of WHO fibers retained (d 1 vs d 365)
USG Interiors MMVF22 Slag Wool (oil treated)				Fischer F344 rats, 6 hr/d, 5 d/wk for 24 mo (3 or 6 animals sacrificed at 3, 6, 12, and 18 mo). 30 mg/m^3; Aerosolized: GMD of 0.9 µm and a GML of 15 µm		1 d: 0.40 x 10^6 total fibers/lung >20 µm long and 3.4 x 10^6 WHO fibers; 365 d: 1.0 x 10^3 total fibers/lung >20 µm long with 1.0 x 10^5 WHO fibers W-T1/2=5 d (fibers >20 µm long) W-T1/2=75 d (WHO fibers); no WHO fibers retained (d 1 vs d 365)
RCF-1	Not listed	Average of 45.6 mg/m^3 650 fibers/cc; 300 WHO fibers/cc	Fibers with an aspect ratio >3:1 GMD 1.8 µm and GML 2.5 µm (65% of aerosol) and particles with aspect ratio <3:1 (35% of aerosol), GMD 1.62 µm and GML 1.58 µm	Fischer F344 rats and Syrian Golden hamsters, 4 hr/d, 5 d/wk for 12 wk. GMD 0.63 µm and GML 5.3 µm; 35% non-fiber particle	No positive control	Pleural fiber burden - Hamster: 4 wks 17.5 x 10^3, 12 wks: 14.8 x 10^3, 24 mo: 15.6 x 10^3 total fibers; Rat: 4 wks 42.1 x 10^3, 12 wks: 41.4 x 10^3, 24 mo: 40.3 x 10^3 total fibers. Rats and hamsters - pleural inflammation. Mesothelial cell proliferation more pronounced in hamsters than in rats. Mesothelial cell labeling index was highest in parietal pleural mesothelial cells lining the surface of the diaphragm (rats and hamsters). Hamsters but not rats had significantly elevated collagen in the visceral pleura at 12 wk.	Everitt et al. 1997
RCF1a	Size-selected for rat respirability	Target 150 fibers/cc > 20 µm long	Stock fiber dimensions not listed	Fischer rats, 62 mg/m^3 351 WHO fibers/cc and 145 fibers >20 µm, GMD 0.79 µm x GML 13.6 µm	Amosite (17 mg/m^3, GMD 0.48 µm, GML 7.7 µm); WT1/2 >400 days	Highest level d 2: 20 x 10^5 fibers/lung >20 µm length, d 365: 0.6 x 10^5 fibers/lung >20 µm length (4% of d 1); Recovered fibers: GMD 0.54 µm x GML 7.0 µm W-T1/2=55 d 90% Clearance = 227 d (fibers >20 µm long) Kdis @ pH 7.4 = 3 ng/cm^2/h	Hesterberg et al. 1998b
MMVF21 Rock Wool				Fischer rats, 58 mg/m^3 466 WHO fibers/cc and 138 fibers >20 µm, GMD 0.86 µm x GML 11.9 µm		Highest level d 7: 12 x 10^5 fibers/lung >20 µm length, d 365: 1 x 10^5 fibers/lung >20 µm length (5% of d 1); GMD 0.61 µm x GML 7.1 µm W-T1/2=67 d 90% Clearance = 264 d (fibers > 20 µm long) Kdis @ pH 7.4 = 20 ng/cm^2/h
MMVF22 Slag Wool						Kdis @ pH 7.4 = 400 ng/cm^2/h
MMVF32 E Glass Fiber				Fischer rats, 51 mg/m^3 316 WHO fibers/cc and 146 fibers >20 µm, GMD 0.81 µm x GML 16.1 µm		Highest level d 2: 14 x 10^5 fibers/lung >20 µm length, d 365: 1 x 10^5 fibers/lung >20 µm length (10% of d 1); GMD 0.47 µm x GML 8.9 µm W-T1/2=79 d 90% Clearance = 371 d (fibers > 20 µm long) Kdis @ pH 7.4 = 9 ng/cm^2/h
MMVF33 475 Glass Fiber				Fischer rats, 51 mg/m^3 371 WHO fibers/cc and 163 fibers >20 µm, 51 mg/m^3 GMD 0.74 µm x GML 16.2 µm		Highest level d 1: 13 x 10^5 fibers/lung >20 µm length, d 365: 0.8 x 10^5 fibers/lung >20 µm length (6% of d 1); GMD 0.39 µm x GML 8.2 µm W-T1/2=49 d 90% Clearance = 240 d (fibers >20 µm long) Kdis @ pH 7.4 = 12 ng/cm^2/h
MMVF34 (HT Stone Wool)				Fischer rats, 60 mg/m^3 367 WHO fibers/cc and 160 fibers >20 µm, GMD 0.73 µm x GML 13.2 µm		Highest level d 1: 15 x 10^5 fibers/lung >20 µm length, d 365: <0.01 fibers/lung (0.04% of d 1); GMD 0.48 µm x GML 7.3 µm W-T1/2=6 d 90% Clearance = 19 d (fibers > 20 µm long) Kdis @ pH 7.4 = 59 ng/cm^2/h
MMVF21 Stone Wool	Size-selected for rat respirability	60 mg/m^3; Target 150 fibers/cm^3 > 20 µm long	Not listed	Male Fischer F344 rats, 6 hr/d for 5 d, post exposure 12 mo. 58 mg/m^3, 467 WHO fibers/cm^3, 147 fibers/cm^3 >20 µm long	No positive control	Highest level d 2: 8.0 x 10^6 WHO fibers/lung, 1.1 x 10^6 fibers >20 µm; d 365: 0.5 x 10^6 WHO fibers/lung (6% of day 1), 0.1 x 10^6 fibers >20 µm T 1/2=65 d for WHO fibers; 92 d for fibers > 20 µm Kdis @ pH 7.5 = 23 (16-30) ng/cm^2/h; @pH 4.5 = 59 (41-77) ng/cm^2/h (based on Si)	Kamstrup et al. 1998
MMVF34 HT Stone Wool				Male Fischer F344 rats, 6 hr/d for 5 d, post exposure 12 mo. 60 mg/m^3, 370 WHO fibers/cm^3, 161 fibers/cm^3 > 20 µm long		Highest level d 2: 10.6 x 10^5 WHO fibers/lung, 1.3 x 10^6 fibers >20 µm; d 365: 0 WHO fibers/lung or fibers >20 µm T 1/2=22 d for WHO fibers; 5 d for fibers >20 µm Kdis @ pH 7.5 = 59 (41-77) ng/cm^2/h; @ pH 4.5 = 620 (434-806) ng/cm^2/h (based on Si)
X-607	Size-selected fine respirable fractions from commercial bulk products	30 mg/m^3 (target conc 180 WHO fibers/cc)	Mean dimensions across all fiber types 1 x 20 µm	Fischer 344 rats, 6 hr/d, 5 d/wk for 2 years. 30 mg/m^3 average of 174 WHO fibers/m^3; 47 fibers >20 µm long; GMD 0.9 μm x GML 11 μm	10 mg/m^3 Chrysotile 10,600 fibers/cc Mild macrophage aggregation, alveolar bronchiolization, microgranulation and bronchoalvelar collagen deposition. Tumor induction 17.4%, 1 pleural mesothelioma.	2 yrs no recovery: 187 x 10^3 WHO fibers with 4 x 10^3 >20 µm; Recovered fiber dimensions 0.5 diameter x 6 length µm. Highest level: 78 wks/26 wks recovery 13.3 x 10^6 WHO fibers (16% of exposure retained) with no fibers >20 µm. Mild macrophage aggregation and cellularity; some microgranulation noted by 78 weeks. Tumor induction 1.6% and no pleural mesotheliomas	Hesterberg et al. 1998a
RCF1				Fischer 344 rats, 6 hr/d, 5 d/wk for 2 yr. 29 mg/m^3 average of 187 WHO fibers/m^3; 101 WHO fibers >20 µm long		2 yrs no recovery: 275 x 10^3 WHO fibers with 48 x 10^3 >20 µm; Recovered fiber dimensions 0.5 diameter x 8 length µm. Highest level: 78 wks/26 wks recovery 87.7 x 10^6 WHO fibers (51% of exposure retained) with 6.89 x 10^6 fibers >20 µm (33% of exposure counts retained). Mild macrophage aggregation, cellularity microgranulation, bronchoalveloar collagen deposition, minimal pleura collagen deposition. Moderate pulmonary change Wagner Score (4). Tumor induction 12.4%, 1.7% mesothelioma
MMVF21 Stone Wool	Size-selected for rat respirability	16 or 30 mg/m^3	Not listed	Male Fischer 344 rats, 6 hr/d for 5 d/wk for 2 yr, interim sacrifices at 3, 6, 12, 18, 24 mo. 16 mg/m^3 150 WHO fibers/cm^3 with 74 fibers/cm^3 >20 µm long ; 30 mg/m^3: 243 WHO fibers/cm^3 with 114 fibers/cm^3 >20 µm long	No positive control	16 mg/m^3 3 mo: 38.0 x 10^3 WHO fibers/lung, 8x 10^3 fibers >20 µm 6 mo: 85 x 10^3 WHO fibers/lung and 16 x 10^3 fibers >20 µm 12 mo: 210 x 10^3 WHO fibers/lung and 37 x 10^3 fibers >20 µm 18 mo: 233 x 10^3 WHO fibers/lung and 58 x 10^3 fibers >20 µm 30.4 mg/m^3 3 mo: 243x 10^3 WHO fibers/lung, 114x 10^3 fibers >20 µm 6 mo: 143 x 10^3 WHO fibers/lung and 23 x 10^3 fibers >20 µm 12 mo: 319 x 10^3 WHO fibers/lung and 55 x 10^3 fibers >20 µm 18 mo: 283 x 10^3 WHO fibers/lung and 62 x 10^3 fibers longer than 20 µm. 16 mg/m^3: Wagner scores ranged from 2.2 (minimal cellular change) at 3 mo to 4 at 18 mo (minimal fibrosis), 30.4 mg/m^3: Wagner scores ranged from 3.2 (mild cellular change) at 3 mo to 4 (minimal fibrosis) at 18 mo	Kamstrup et al. 1998
MMVF34 HT Stone Wool	Size-selected for rat respirability	30 mg/m^3 (3, 12, or 18 mo exposure)	Not listed	Male Fischer 344 rat, 6 hr/d, 5 d/wk for 2 yr, interim sacrifices at 3, 6, 12, 18, 24 mo. 31 mg/m^3 3 mo: 282 WHO fibers/cm^3 with 84 fibers/cm^3 >20 µm long; 12 mo: 264 WHO fibers/cm^3 with 82 fibers/cm^3 >20 µm long; 18 mo: 288 WHO fibers/cm^3 with 86 fibers/cm^3 >20 µm long		30.5 mg/m^3 3 mo: 288 x 10^3 WHO fibers/lung, 86 x 10^3 fibers >20 µm 6 mo: 147 x 10^3 WHO fibers/lung and 11 x 10^3 fibers >20 µm 12 mo: 152 x 10^3 WHO fibers/lung and 10 x 10^3 fibers >20 µm 18 mo: 222 x 10^3 WHO fibers/lung and 11 x 10^3 fibers >20 µm. Interstitial fibrosis Wagner scores ranged from 1.6 (normal approaching minimal) at 3 mo to 2.8 (approaching mild)
MMVF10a Glass Fiber	Size-selected to be rat respirable	30 mg/m^3	0.84 μm × 12.4 μm	Syrian Golden hamsters, 6 hr/d, 5 d/wk for 78 wk (18 mo); Select animals evaluated at 13, 26, 52, and 78 wk exposure. 29.6 mg/m^3 (mean diameter 0.95 x mean length 19.4 µm) 339 WHO fibers/cm^3, including 134 fibers >20 μm/cm^3	Amosite asbestos (36, 165, or 263 WHO fibers/cm^3 with a GMD of 0.48 µm and a GML of 7.7 µm); severe pulmonary fibrosis and multiple mesotheliomas; 6 hr lung burden: 20.8 × 10^5 WHO fibres/lung (2 fibres >20 μm) at 78 wk lung burden 144 × 10^6. No pulmonary neoplasms, mesotheliomas 3.6%, 25.9%, and 19.5%	6 hr deposition: 8.4 × 10^5 WHO fibers/lung (1.6 x 10^5 fibers >20 μm long); 78 wks: 76.7 × 10^6 WHO fibers (4.6 × 10^6 fibers >20 μm long) 78 wks+ 6 wks of recovery: 27.8 × 10^6 WHO fibers (0.23 × 10^6 fibers >20 μm long). Lung inflammation with recovery by 6 wks; no fibrosis or neoplasms	McConnell et al. 1999
JM475 (MMVF33) Glass Fiber	Size-selected to be rat respirable	37 mg/m^3	0.7 μm × 11.8 μm	Syrian Golden hamsters, 6 hr/d, 5 d/wk for 78 wk (18 mo); Select animals evaluated at 13, 26, 52, and 78 wk exposure. 37 mg/m^3 (mean diameter 0.9 x mean length 16.9 µm) 310 WHO fibers/cm^3, including 109 fibers >20 μm/cm^3		6 hr deposition: 11.5 × 10^5 WHO fibers, (2.2 × 10^5 fibers >20 μm) 78 wks: 234 × 10^6 WHO fibers (30 × 10^6 fibers >20 μm long) 78 wks+ 6 wks of recovery: 141 × 10^6 WHO fibers (11.3 × 10^6 fibers >20 μm long). Inflammation and mild fibrosis by 26 wk, more severe by 52 wk. Pleural hyperplasia, 1 mesothelioma
MMVF34 HT Stone Wool Fiber	Size-selected to be rat respirable	30 mg/m^3	Stock fiber dimensions: GMD 0.98 µm and GML 11.1 µm)	Male Fischer F344 rats 6 hr/d, 5 d/wk for 104 wk (interim sacrifices at 13, 26, 52, and 104 wk). 30.1 mg/m^3: mean concentration 400 fibers/cm^3 (291 WHO fibers/cm^3) with 85 fibers/cm^3 that were >20 µm in length; GMD 0.87 µm and GML 10.8 µm	No positive control group	Highest level at 12 mo 192.3 x 10^6 total fibers (WHO fibers highest 12 months: 80.3 x 10^6); 24 mo: 108.5 x 10^6 total fibers (60.2 WHO fibers). Interstitial collagen/fibrosis 0.5% at 24 mo; adenomas 4.7% (NS from control), no carcinomas	Kamstrup et al. 2001
MMVF21 Stone Wool	Size-selected to be rat respirable	3, 16, or 30 mg/m^3	Stock fiber dimensions: GMD 1.09 µm and GML 16.1 µm	Male Fischer F344 rats 6 hr/d, 5 d/wk for 104 wk (interim sacrifices at 13, 26, 52, and 104 wk). 30.4 mg/m^3: 264 total fibers/cm^3 (243 WHO) with GMD 0.98 µm and GML 16.5 µm, 43% of fibers were >20 µm in length; 16.2 mg/m^3: 185 total fibers/cm^3 (150 WHO), GMD 0.90 µm and GML 15.4 µm, 40% fibers > 20 µm length	No positive control group	30 mg/m^3: Highest level at 24 mo 110.1 x 10^6 total fibers (WHO fibers highest 12 mo: 94.1 x 10^6); 24 mo: 87.6 x 10^6 WHO fibers 16 mg/m^3: Highest level at 24 mo: 95.2 x 10^6 total fibers (24 mo 67.8 x 10^6 WHO fibers) ** 3 mg data not listed. 30 mg/m^3: interstitial collagen/fibrosis 5.04% at 18 mo (decreased to 3.84% at 24 mo); 3.5% adenomas, 0.9% carcinoma (NS from control) 16 mg/m^3: interstitial collagen/fibrosis highest level 1.68% at 24 mo; 3.5% adenomas, 0.9% carcinoma (NS from control)
RCF1	Standard for RCC studies	50 mg/m^3	Target concentration: 130 fibers/ml > 20 µm	Female Wistar rats; 6 hr/d, 5 d/wk for 3 wk. 51.2 mg/m^3; 10% fibers were >20 µm and 90% were <20 µm; GMD 0.79 µm x GML 6.8 µm	No positive control group	3d post inhalation: 3.3 x 10^6 /lung >20 µm and 46.5 x 10^6 /lung <20 µm; 102.4 x 10^6 non-fibrous particles. Biochemical and cytological analyses in BALF at days 3, 17, 31, 94, and 365 d post exposure showed inflammation in RCF1 animals was more persistent compared to RCF1a animals.	Bellmann et al. 2001; Brown et al. 2000 from IARC 2002
RCF1a	less "aggressive" preparation resulting in longer fibers longer and lower levels of nonfibrous particles	25 mg/m^3	Target concentration: 125 fibers ml > 20 µm	Female Wistar rats; 6 hr/d, 5 d/wk for 3 wk. 25.8 mg/m^3; 19% fibers were >20 µm and 81% were <20 µm; GMD 0.49 µm x GML 7.8 µm	No positive control group	3d post inhalation: 5.0 x 10^6/lung >20 µm and 32.7 x 10^6/lung <20 µm; 6.9 x 10^6 non-fibrous particles. Half time based on alveolar tracer clearance= 80 days (range 71-91 days)
JM901	Size-selected rat respirable (high concentration of long, thin fibers w/ diameter <1 μm and length >20 μm)	13 mgm^3	Stock fiber dimensions not given	Fischer F344 rats, 6 hr/d for 5 d, 13 mg/m^3, 403 WHO fibers/cm^3 (104 fibers >20 µm/cm^3), GMD 0.63 µm and GML 10.3 µm	Amosite 5 mg/m^3 Health outcomes: Day 1: Significantly increased LDH, total protein, eosinophils, leukocytes, and neutrophils; decreased macrophages in BALF. Significantly increased eosinophils in pleural lavage. Day 14: Significant increase in neutrophils in BALF and total protein, macrophages, eosinophils, and mast cells in pleural lavage. Day 29-30: Significantly increased leukocytes and neutrophils in BALF and total protein, macrophages, and eosinophils in pleural lavage.	90d: 2.9 x 10^6 WHO fibers/lung, peak burden 1d: 8.6 x 10^6 WHO fibers/lung (1.6 x 10^6 WHO fibers/lung at 181 d). W-T1/2 = 9.8 d (fibers >20 µm long) 90% Clearance = 69 d Kdis @ pH 7.4= NT. Day 1: Significantly increased LDH, total protein and neutrophils and significantly decreased macrophages in BALF; significantly increased eosinophils in pleural lavage. Day 14: Significant increase in neutrophils in BALF and macrophages and eosinophils in pleural lavage. Day 29-30: Significant increase in pleural macrophages.	Hesterberg et al. 2002
JM901F		12 mg/m^3		Fischer F344 rats, 6 hr/d for 5 d, 14 mg/m^3, 401 WHO fibers/cm^3 (173 fibers >20 µm/cm^3), GMD 0.71 µm and GML 12.7 µm		91d: 1.8 x 10^6 WHO fibers/lung, peak burden 1 d: 9.5 x 10^6 WHO fibers/lung (0.6 x 10^6 WHO fibers/lung at 182 d). W-T1/2 = 8.1 d (fibers >20 µm long) 90% Clearance = 38 d Kdis @ pH 7.4 = 500 ng/cm^2/h. Day 1: Significantly increased LDH, total protein and neutrophils in BALF and total protein in pleural lavage. Day 14: Significantly increased pleural total protein. Day 29-30: NS
JM902		32 mg/m^3		Fischer F344 rats, 6 hr/d for 5 d, 32 mg/m^3, 443 WHO fibers/cm^3 (119 fibers >20 µm/cm^3), GMD 0.51 µm and GML 10.9 µm		90d: 2.1 x 10^6 WHO fibers/lung, peak burden 1 d: 11.7 x 10^6 WHO fibers/lung. W-T1/2 = 6.8 d (fibers >20 µm long) 90% Clearance = 33 d Kdis @ pH 7.4 = 150 ng/cm^2/h. Day 1: Significantly increased LDH, total protein and neutrophils in BALF. Day 14: Significantly increased pleural macrophages and mast cells. Day 29-30: Significant increase in eosinophils in pleural lavage.
JM475 (MMVF33)		21 mg/m^3		Fischer F344 rats, 6 hr/d for 5 d, 21 mg/m^3, 405 WHO fibers/cm^3 (166 fibers >20 µm/cm^3), GMD 0.51 µm and GML 10.9 µm		90d: 1.6x 10^6 WHO fibers/lung, peak burden 1 d: 7.1 x 10^6 WHO fibers/lung (0.5 x 10^6 WHO fibers/lung at 365 d). W-T1/2 = 49 d (fibers >20 µm long) 90% Clearance = 240 d Kdis @ pH 7.4 = 12 ng/cm^2/h. Day 1: Significantly increased LDH, total protein, eosinophils, leukocytes, and neutrophils in BALF and total protein and eosinophils in pleural lavage. Significant decrease in BALF macrophages. Day 14: Significant increase in total protein, macrophages, mast cells and eosinophils in pleural lavage. Day 29-30: Significantly increased LDH, beta-glucuronidase, total protein, leukocytes in BALF and total protein, macrophages, and eosinophils in pleural lavage.
E-Glass Fiber	Counting and sizing per WHO guidance (1985)	15, 50, or 150 f/ml (fiber length >20 µm)	Size selected stock: GMD 0.34 µm and GMD 5.22 µm	Male Wistar rats, 6 hr/d, 5 d/wk for 3 mo. 17.2 mg/m^3	No positive control; Included air control	1 wk exposure: 17 × 10^6 fibers > 20 μm per lung. After 3 mo recovery, decreased to 38.4% of original burden. W-T1/2 =57-63 d (fibers >20 µm long) Increased lung weight, biochemical parameters and polymorphonuclear leukocytes (PMN) in BAL, in cell proliferation (BrdU-response) and interstitial fibrosis. The values observed in the proliferation assay on the carcinogenic E-glass microfiber indicate that this assay has an important predictive value with regards to potential carcinogenicity.	Bellmann et al. 2003
MMVF21			Size selected stock: GMD 0.92 µm and GMD 10.85 µm	Male Wistar rats, 6 hr/d, 5 d/wk for 3 mo. 37 mg/m^3		1 wk: 5.7 × 10^6 fibers >20 μm per lung. After 3 mo recovery, decreased to 63.9% of original burden. W-T1/2=28-140 d (fibers >20 µm long)
High Temp CMS fiber (calcium-magnesium-silicate)		150 f/ml (fiber length >20 µm)	Size selected stock: GMD 0.94 µm and GMD 10.1 µm	Male Wistar rats, 6 hr/d, 5 d/wk for 3 mo. 49.5 mg/m^3		1 wk: 0.88 × 10^6 fibers >20 μm per lung. After 3 mo recovery, decreased to 3% of original burden. W-T1/2 fibers >20 µm long = 8 d
MMVF10a Glass Fiber	As described in Gelzleicher et al. 1996a	Target mass 45 mg/m^3	Not listed	Fischer F344 rats, Syrian Golden hamsters: 4 hr/d for 5 d/wk for either 4 or 12 wk. 45 mg/m^3 (610 WHO fibers/cc)	No positive control	Hamster: 160 fibers/cm^2; Rats: 60 fibers/cm^2. Pleural inflammatory responses minimal. Rats: pleural inflammation w/ increased #s of macrophages, increases in mesothelial cell replication. Hamsters: increased #s of macrophages and lymphocytes, and mesothelial-cell replication elevated on parietal pleura; some responses persisting through 12 wk of postexposure recovery.	Bermudez et al. 2003
E-Glass	Not described	15, 50, or 150 f/ml	Stock fiber dimensions (fibers >20 µm): GMD 0.52 µm and GML 34.31 µm	Male Wistar rats; 6 hr/d, 5 d/wk for 3 mo. Gravimetric exposure: 150 f/ml, 17.2 mg/m^3; aerosol fibers >20 µm in length: 15 f/ml: GMD 0.56 µm x GML 31.4 µm; 50 f/ml: GMD 0.57 µm x GML 30.6 µm; 150 f/ml: GMD 0.57 µm x GML 31.1 µm	No positive control	For fibers > 20 µm in length: 17 x 10^6; after 3 mo recovery 38.4% remaining. Increase in lung weights; BAL LDH, b-glucuronidase, and protein; interstitial fibrosis-EPA score 2 (50/150 f/ml)	Bellmann et al. 2003
MMVF21	Not described	15, 50, or 150 f/ml	Stock fiber dimensions (fibers >20 µm): GMD 1.13 µm and GML 35.25 µm	Male Wistar rats; 6 hr/d, 5 d/wk for 3 mo. Gravimetric exposure: 150 f/ml, 38 mg/m^3; aerosol fibers >20 µm in length: 15 f/ml: GMD 0.88 µm x GML 33.0 µm; 50 f/ml: GMD 0.91 µm x GML 33.7 µm; 150 f/ml: GMD 0.93 µm x GML 33.1 µm	No positive control	For fibers >20 µm in length: 5.7 x 10^6; after 3 mo recovery 63.9% remaining. Increase in lung weights; BAL LDH, b-glucuronidase, and protein; interstitial fibrosis- EPS score 2 (50/150 f/ml),
CMS (Calcium-magnesium- silicate)	Not described	150 f/ml	Stock fiber dimensions (fibers > f 20 µm): GMD 0.92 µm and GML 30.2 µm	Male Wistar rats; 6 hr/d, 5 d/wk for 3 mo. Gravimetric exposure: 150 f/ml, 38 mg/m^3; aerosol fibers >20 µm in length: 150 f/ml: GMD 0.88 µm x GML 28.9 µm	No positive control	For fibers >20 µm in length: 0.88 x 10^6; after 3 mo recovery 3.0% remaining. Interstitial fibrosis (EPS score 1)