Aerosolization of Mycoplasma synoviae compared with Mycoplasma gallisepticum and Enterococcus faecalis

Abstract

In order to study the airborne transmission of an arthropathic strain of Mycoplasma synoviae, preliminary aerosol experiments were performed. They were conducted in duplicate in an empty isolator (1.3 m³) to assess the yield and viability of M. synoviae with time compared with Mycoplasma gallisepticum and Enterococcus faecalis. After aerosol generation air samples were taken with two different devices using gelatine or cellulose nitrate filters. There was no difference between the devices, but cellulose nitrate filters yielded very low bacterial counts. The aerosolized dose per isolator for M. synoviae was 3.4×10¹⁰ colony-forming units (cfu), for M. gallisepticum was 2.6×10¹⁰ cfu and for E. faecalis was 3×10¹⁰ cfu. Immediately after aerosolization, concentrations of about 10⁶ to 10⁷ cfu/m³, 10⁷ to 10⁸ cfu/m³ and 10⁸ to 10⁹ cfu/m³ air of M. synoviae, M. gallisepticum and E. faecalis were found, respectively. At 25 min M. synoviae concentrations dropped below the detection level (<4×10⁴ cfu), while 10⁵ to 10⁶ and 10⁸ to 10⁹ cfu were found for M. gallisepticum and E. faecalis, respectively. The average M. synoviae concentration during the experiment was estimated at 10² to 10³ cfu/l. The M. gallisepticum and E. faecalis aerosol generated an average of approximately 10³ to 10⁴ cfu/l air and 10⁵ to 10⁶ cfu/l air, respectively. Thus mycoplasma and E. faecalis aerosols were successfully generated despite considerable initial loss as measured by culture. The loss was greater in the mycoplasma aerosols, especially those of M. synoviae.

Résumé

Nébulisation de Mycoplasma synoviae comparé à M. gallisepticum et à Enterococcus faecalis

Dans le but d’étudier la transmission par voie aérienne d'une souche arthropathique de Mycoplasma synoviae, il a été réalisé préliminairement des expérimentations de nébulisation. Ces essais ont été effectués deux fois de suite dans un isolateur vide (1,3 m³) pour évaluer la quantité et la viabilité de M. synoviae récolté à différents moments, comparé à mycoplasma gallisepticum et Enterococcus faecalis. Après la génération de l'aérosol, des échantillons d'air ont été prélevés avec deux appareils différents utilisant de la gélatine ou des filtres de nitrate de cellulose. Il n'y a pas eu de différence entre les appareils, mais les filtres en nitrate de cellulose ont récolté des quantités très faibles de bactéries. La quantité de germes, nébulisée par isolateur, a été de 3,4 x 10¹⁰ colonies formant unité (cfu) pour M. synoviae, 2,6 x 10¹⁰ cfu pour M. gallisepticum et 3 x 10¹⁰ cfu pour E . faecalis. Immédiatement après la nébulisation, les concentrations ont été de 10^6-7, 10^7-8 et 10^8-9 cfu/m³ d'air pour respectivement M. synoviae, M. gallisepticum et E. faecalis. Après 25 minutes, les concentrations de M. synoviae ont chuté en dessous de la limite de détection (<4 x 10⁴ cfu) alors que des concentrations de 10^5-6 et 10^8-9 cfu ont été trouvées pour respectivement M. gallicepticum et E. faecalis. La concentration moyenne de M. synoviae durant l'expérience a été estimée à 10^2-3 cfu/l. Les aérosols de M. gallisepticum et E. faecalis ont généré approximativement 10^3-4 et 10^5-6 cfu/l d'air. Ainsi, les aérosols de mycoplasmes et de E. faecalis ont été produits avec succès malgré une perte initiale considérable mesurée par culture. La perte a été plus importante pour les aérosols de mycoplasmes, principalement pour ceux de M. synoviae.

Zusammenfassung

Aerosolierung von Mycoplasma synoviae im Vergleich zu Mycoplasma gallisepticum und Enterococcus faecalis

Um die aerogene Übertragung eines arthropathischen Stammes von Mycoplasma (M.) synoviae zu untersuchen, wurden Vorversuche mit dem Aerosol durchgeführt. Sie wurden zweimal in einem leeren Isolator (1,3m³) zur Bestimmung der Ernte und Lebensfähigkeit von M. synoviae im zeitlichen Verlauf im Vergleich mit M. gallisepticum und Enterococcus (E.) faecalis wiederholt. Nach der Erzeugung des Aerosols wurden mit zwei verschiedenen Geräten unter Verwendung von Gelantine- oder Zellulosenitratfiltern Luftproben entnommen. Es gab keine Unterschiede zwischen den Geräten, aber die Zellulosenitratfilter erbrachten nur eine sehr geringe Anzahl von Bakterien. Die Sprühdosis pro Isolator betrug bei M. synoviae 3,4 x 10¹⁰ Kolonie bildende Einheiten (cfu), bei M. gallisepticum 2,6 x 10¹⁰ cfu und bei E. faecalis 3 x 10¹⁰ cfu. Unmittelbar nach der Aerosolierung wurden etwa 10^6-7, 10^7-8 und 10^8-9 cfu/m³ Luft von M. synoviae, M. gallisepticum bzw. E. faecalis gefunden. Nach 25 min fielen die Konzentrationen von M. synoviae unter die Nachweisgrenze (<4 x 10⁴), während 10^5-6 und 10^8-9 cfu/m³ Luft von M. gallisepticum bzw. E. faecalis festgestellt wurden. Die durchschnittliche M. synoviae-Konzentration während des Experiments wurde auf 10^2-3 cfu/l Luft geschätzt. Die M. gallisepticum- und E. faecalis-Aerosole erreichten einen Durchschnitt von 10^3-4 bzw. 10^5-6 cfu/l Luft. Auf diese Weise konnten, wie kulturell nachgewiesen, trotz erheblicher Anfangsverluste erfolgreich Mycoplasmen- und E. faecalis-Aerosole erzeugt werden. Der Bakterienverlust war bei den Mykoplasmen-Aerosolen, insbesondere bei denen mit M. synoviae größer.

Resumen

Aerosolización de Mycoplasma synoviae comparada con M. gallisepticum y Enterococcus faecalis

Se realizaron una serie de experimentos preliminares con aerosol para estudiar la transmisión aérea de una cepa artropática de Mycoplasma synoviae. Se realizaron por duplicado en un aislador vacío (1.3 m³) para evaluar la producción y la viabilidad de M. synoviae a lo largo del tiempo en comparación con Mycoplasma gallisepticum y Enterococcus faecalis. Tras la generación del aerosol, se tomaron muestras con dos aparatos diferentes utilizando filtros de gelatina o de nitrato de celulosa. No se encontraron diferencias entre los aparatos, pero los filtros de nitrato de celulosa dieron contajes bacterianos muy bajos. La dosis de aerosol por aislador de M. synoviae fue de 3.4 x 10¹⁰ unidades formadoras de colonia (ufc), de M. gallisepticum 2.6 x 10¹⁰ ufc y de E. faecalis 3 x 10¹⁰ ufc. Inmediatamente tras la aerosolización, se encontraron concentraciones de 10^6-7, 10^7-8 y 10^8-9 ufc/m³ aire de M. synoviae, M. gallisepticum y E. faecalis respectivamente. A los 25 minutos, las concentraciones de M. synoviae cayeron por debajo del nivel de detección (<4 x 10⁴ ufc), mientras que se encontraron niveles de10^5-6 y 10^8-9 ufc para M. gallisepticum y E. faecalis, respectivamente. La media de la concentración de M. synoviae durante el experimento fue estimada en 10^2-3 ufc/l. Los aerosoles de M. gallisepticum y E. faecalis generados presentaron una media de aproximadamente 10^3-4 y 10^5-6 ufc/l de aire respectivamente. En definitiva los aerosoles de micoplasma y E. faecalis se generaron de forma satisfactoria a pesar de las considerables pérdidas iniciales que fueron medidas mediante cultivos. La pérdida fue mayor en los aerosoles de micoplasma, especialmente en aquellos de M. synoviae.

Introduction

Spontaneous cases of Mycoplasma synoviae-associated amyloid arthropathy in brown layers have been described previously (Landman & Feberwee, 2001), while in the same study induction of the condition with one of the arthropathic M. synoviae field isolates was reported.

In order to investigate the pathogenesis of M. synoviae-induced amyloid arthropathy, an animal model based on a natural infection route (aerosol) was developed (W. J. M. Landman, unpublished results, 2002). Preliminary aerosol studies were conducted prior to the animal experiments and are reported here. The studies were performed in empty isolators to assess the yield and viability of M. synoviae aerosol particles with time compared with Mycoplasma gallisepticum and Enterococcus faecalis. M. gallisepticum was included in order to compare with another mycoplasma species, while the arthropathic and amyloidogenic E. faecalis strain 6085.94 was used as a reference because it had been used in previous aerosol studies (Landman & Van Eck, 2001; Landman et al., 2001).

Aerosol transmission of M. synoviae has been reported previously (Kleven et al., 1972, 1975; Wyeth, 1974; Vardaman & Drott, 1977). However, in these earlier studies the aerosol efficiency was not controlled, the half-life, initial loss and theoretical bacterial uptake were not calculated, and the aerosol techniques were not fully described. In the present study, a more precise characterization of M. synoviae aerosols compared with M. gallisepticum and E. faecalis was attempted.

Materials and Methods

Experimental design

Six aerosol experiments, performed in duplicate for each bacterial species (M. synoviae, M. gallisepticum and E. faecalis), were carried out in the same empty isolator (Beyer & Eggelaar, Utrecht, the Netherlands). The volume of the isolator was 1.3125 m³ (1.40 m long, 0.75 m wide and 1.25 m high).

In all experiments the concentration of bacteria (per m³ air) with time and the half-life after aerosol generation was studied. Air samples were captured on filters immediately after aerosol production to assess initial loss and then again after 25 min to calculate the half-life.

The ventilation of the isolator was switched off during the whole experimental period. After each aerosol experiment the ventilation was switched on for 20 min. The ventilation capacity was approximately 40 m³/h.

Aerosol fluids, application and assessment

Aerosol fluids

M. synoviae (chicken/NL/Dev/801979Rob/00) originating from a spontaneous case of M. synoviae-associated amyloid arthropathy in brown layers (Landman & Feberwee, 2001) and M. gallisepticum (chicken/NL/Dev/SP-1608Vin/99), a field isolate from broilers with respiratory symptoms (A. Feberwee, unpublished results, 2000), were used. Both mycoplasmas had been isolated on ME (Mycoplasma Experience) agar (Mycoplasma Experience, Reigate, UK), identified by immunofluorescence (Kiernan, 1990) and stored at −70°C in ME broth.

For each mycoplasma species, one drop of the stored broth (−70°C) was added to 20 ml ME broth. The broths were incubated for 2 days at 37°C until a change of colour was observed. Bacterial concentrations in the inocula were counted following an international standard (ISO 7402, 1985). Briefly, decimal dilutions (10⁻¹ to 10⁻⁷) of 2-day cultures were made in ME broth and 20 μl each dilution were plated on ME agar. Incubation was at 37°C until mycoplasma colonies were visible, and then the initial bacterial concentration was calculated.

For reference purposes E. faecalis isolate 6085.94 was used. It had been isolated from a spontaneous case of amyloid arthropathy (Landman et al., 1994) and was known to induce this condition (Landman et al., 1997). Frozen beads containing the isolate (−70°C) were rolled on a sheep blood agar plate. After overnight incubation at 37°C, colonies were scraped off and suspensions of approximately 10⁹ colony-forming units (cfu)/ml were prepared in peptone saline (8.5 g sodium chloride and 1 g bacteriological peptone per litre of water) (ISO 6887, 1983). Bacterial concentrations were assessed as already described, except that a volume of 0.1 ml was inoculated onto sheep blood agar plates, which were incubated at 37°C for 48 h.

Aliquots of 10 ml M. synoviae, M. gallisepticum or E. faecalis culture containing approximately 10⁹ cfu/ml (Table 1) plus 10 ml buffered peptone water (BPW) were used as aerosol fluids. The peptone water contained 10 g peptone, 5 g sodium chloride, 3.7 g disodium phosphate and 1.5 g potassium dihydrogen phosphate per litre of water.

Validation of the bacteriology of filters

Gelatine filters

Filters were placed in Petri dishes and were inoculated with 0.5 ml culture containing M. synoviae, M. gallisepticum or E. faecalis. The cultures were prepared as described for aerosol fluids. Two filters were analysed for each micro-organism. After 5 min the filters were dissolved by adding 20 ml BPW kept at 36°C to the dishes. After gentle mixing this fluid was pipetted into small flasks that contained the remaining 30 ml BPW.

Cellulose nitrate filters

Two cellulose nitrate filters were tested for each micro-organism. The filters were placed on the filter-rinsing device and inoculated with 0.5 ml of the same culture as used with the gelatine filters. After 5 min the rinsing devices were closed and rinsed with 10 ml BPW.

Bacterial counts were performed as described previously. The counts of the stock cultures were compared with those obtained after inoculation and bacteriological analysis of both filter types as described earlier. Counts were expressed as cfu/ml.

M. synoviae, M. gallisepticum and E. faecalis aerosol application

For the aerosol application, an air compressor (Mecha Concorde, type 7SAX, 100l, 10 bar/max; SACIM, Verona, Italy) coupled to a spray-head (Walther Pilot I spray-head with 0.5 mm diameter; Walther Spritz- and Lackiersysteme, Wuppertal, Germany) was used. A volume of 20 ml was aerosolised in 3 min at a pressure of 2 bar in the isolator with an air yield of 30 l/min. The isolator temperature was 18 to 20°C.

Assessment of the M. synoviae, M. gallisepticum and E. faecalis aerosol production and calculation of the half-life

The aerosol efficiency for M. synoviae, M. gallisepticum and E. faecalis was measured by means of air sampling. The samples were taken using the Sartorius MD2 and MD8 airscan air sampling devices (Sartorius, Nieuwegein, the Netherlands). For bacteriology, three different filter types were used. On the MD2 device, sterile gelatine filters of 3 μm pore size and 50 mm diameter (type 12602-050-ALK; Sartorius) and sterile cellulose nitrate filters of 0.2 μm pore size and 50 mm diameter (type 11307-050-CAN; Sartorius) were applied. On the MD8 airscan, sterile gelatine filters of 3 μm pore size and 80 mm diameter (type 17528-80-ACD; Sartorius) were used. The sampling time was 2 min at 1500 l/h (50 l air) for the MD2 and at 2000 l/h (67 l air) for the MD8 airscan.

Bacteriological culture of gelatine filters was performed by dissolving them in 50 ml BPW kept at 37°C and preparing six decimal dilutions (10⁻¹ to 10⁻⁶) from this solution (ISO 6887, 1983). Thereafter, for the mycoplasmas 20 μl undiluted suspension and of each dilution was pipetted onto ME agar, while for E. faecalis a volume of 0.1 ml was pipetted onto sheep blood agar plates. Mycoplasma plates were cultured for 1 week at 37°C, and then colonies were counted. E. faecalis was cultured aerobically for 48 h at 37°C. Identification of mycoplasmas and E. faecalis was performed as described previously (Landman et al., 1994; Landman & Feberwee, 2001). The detection limit for mycoplasmas was 10^4.7 cfu/m³ air and 10^4.6 cfu/m³ air for the MD2 and MD8 airscan gelatine filters, respectively. For E. faecalis the detection limit was 10⁴ cfu/m³ air and 10^3.9 cfu/m³ air for these filters, respectively.

Bacteriology of MD2 cellulose nitrate filters was performed after rinsing the filters with 10 ml peptone water in a sterile filter rinsing device (type SM165 98; Sartorius GmbH, Göttingen, Germany). Ten decimal dilutions (10⁻¹ to 10⁻¹⁰) of the rinse fluids were made. Further procedures were exactly as already described. The detection limit was 10⁴ cfu/m³ air for the mycoplasmas and 10^3.3 cfu/m³ air for E. faecalis.

The detection limits were calculated as demonstrated in the following example for E. faecalis. If sampled on gelatine filters with the MD2 a sampling time of 2 min at 1500 l/h (50 l air) was used. Since the filters were dissolved in 50 ml peptone water (50 l air in 50 ml=1 l/ml peptone water), 0.1 ml (0.1 l air) is plated out on agar; if 0.1 ml contains 1 cfu, the detection limit will be 10 cfu/l of air (per m³ air it will be 10⁴ cfu). When using the MD8 a different volume of air is sampled, this slightly changes the detection limit (10^3.9 cfu/m³). If cellulose nitrate filters are used the micro-organisms within the sample of air are collected in 10 ml rinsing peptone water instead of 50 ml, and will give a different detection limit. In this example it is 10^3.3 cfu/m³ air.

The half-life of aerosols was calculated using the formula: t _1/2=0.3.X/log₁₀ C/E, where X is the time interval, C the bacterial concentration at the start and E the bacterial concentration at 25 min.

Characterization of the aerosol spectra

The droplet size distribution of the aerosol produced by three different Walther Pilot I spray-heads was measured with a laser diffraction particle size analyser (Mastersizer-S long bed; Malvern Instruments Ltd., Malvern, UK). Brain heart infusion (BHI) (Oxoid CM 225; 37 g/l water) was used as aerosol fluid. It was nebulized with the same flow rate and pressure as described earlier for the mycoplasma and E. faecalis application. Each aerosol was measured holding the spray-head at a distance of 4 cm from the laser beam and approximately 2 cm from the lens. A 300 mm lens with open bench configuration was used. The measurements were repeated three times per spray-head. A combination of ME broth (5 ml) and BPW (5 ml) was also tested. All experiments were performed at a temperature of 26±2°C and a relative humidity of 36%.

Results

Validation of the bacteriology of filters

The results of the validation of the bacteriology of filters are presented in Table 1. The bacterial counts of the stock cultures were similar to those of the gelatine and cellulose nitrate filters subjected to the bacteriological analysis described in Materials and Methods.

Table 1. Bacterial counts (log₁₀ cfu/ml) of the validation of bacteriology of filters

Sample	M. synoviae	M. gallisepticum	E. faecalis
Stock	8	7.6	8.7
Gelatine filter 1	9	8.1	8.9
Gelatine filter 2	9.4	7.4	8.9
Cellulose nitrate filter 1	9.1	7	8.8
Cellulose nitratefilter 2	9	7.2	8.8

Aerosol assessment

M. synoviae, M. gallisepticum and E. faecalis aerosol initial loss, concentration with time and half-life

The initial dose and loss are presented in Table 2. The initial loss was calculated using only the average of both gelatine filters, because the readings of the cellulose nitrate filters were much lower and those for the mycoplasmas were below the detection level. The dose per m³ isolator air for M. synoviae was 3.4×10¹⁰, for M. gallisepticum was 2.6×10¹⁰ and for E. faecalis was 3×10¹⁰ cfu. The initial loss in the aerosol, based on bacterial counts of air samples taken immediately after aerosol generation, was 4 log₁₀ cfu M. synoviae, 3 log₁₀ cfu M. gallisepticum and 2 log₁₀ cfu E. faecalis.

Table 2. Initial loss of M. synoviae, M. gallisepticum and E. faecalis in aerosols as measured by culture

Aerosol experiment	Dose (log10 cfu)	Time 0 min^a (log10 cfu/m^3 air)	Initial loss (log10 cfu/m^3 air)
M. synoviae 1	10.5	6.4	4.1
M. synoviae 2	10.5	6.7	3.8
M. gallisepticum 1	10.4	7.3	3.1
M. gallisepticum 2	10.4	7.2	3.2
E. faecalis 1	10.5	8.6	1.9
E. faecalis 2	10.5	8.6	1.9
^a Average of both gelatine filters.

In Table 3 the number of cfu of M. synoviae, M. gallisepticum and E. faecalis/m³ isolator air with time and half-life after aerosol generation are given. Taken together the results of both experiments and results of bacteriology of gelatine filters, immediately after the aerosol production, showed concentrations of about 10⁶ to 10⁷ cfu/m³ air, 10⁷ to 10⁸ cfu/m³ air and 10⁸ to 10⁹ cfu/m³ air for M. synoviae, M. gallisepticum and E. faecalis, respectively. At 25 min, concentrations dropped below the detection limit for M. synoviae, while for M. gallisepticum and E. faecalis 10⁵ to 10⁶ cfu/m³ air and 10⁸ to 10⁹ cfu/m³ air were found, respectively.

Table 3. M. synoviae, M. gallisepticum and E. faecalis log₁₀ aerosol concentration with time and half-life

Time after aerosol generation (min)	Device	Filter system	M. synoviae				M. gallisepticum				E. faecalis
			Experiment 1		Experiment 2		Experiment 1		Experiment 2		Experiment 1		Experiment 2
			Log10 cfu/m^3 air	Half-life (min)^a	Log10 cfu/m^3 air	Half-life (min)	Log10 cfu/m^3 air	Half-life (min)	Log10 cfu/m^3 air	Half-life (min)	Log10 cfu/m^3 air	Half-life (min)	Log10 cfu/m^3 air	Half-life (min)
0	MD2	Gelatine	6.5		6.8		7		6.6		8.3		8.5
0	MD8	Gelatine	6.3		6.7		7.7		7.8		9		8.8
0	MD2	Cellulose nitrate	<4		4		4.9		4		7.8		7.5
25	MD2	Gelatine	<4.7		<4.7		5.9		5.6		8.3		8.3
25	MD8	Gelatine	<4.6	<4	<4.6	<4	6	5	6.5	6	8.5	19	8.8	75
25	MD2	Cellulose nitrate	<4		<4		<4		<4		6.5		7
^a Half-life time was based on the average results obtained with both gelatine filters.

The results of M. synoviae and M. gallisepticum bacteriology of cellulose nitrate filters were always ≥ 2 log₁₀ lower than those of gelatine filters, and results of E. faecalis bacteriology of cellulose nitrate filters were 0.5–2 log₁₀ below those of gelatine filters.

The half-lives of M. synoviae and M. gallisepticum, based on average results obtained with both gelatine filters, were<4 min and 5 to 6 min, respectively. For E. faecalis the half-life ranged from 19 min (Experiment 1) to 75 min (Experiment 2). Cellulose nitrate filters were not used for half-life calculations.

Characterization of the aerosol spectra

Table 4 presents the particle size distribution (D) expressed as a volume diameter. The values D(v, 0.1), D(v, 0.5) and D(v, 0.9) refer to particle diameters below which 10%, 50% and 90% of the particle volume is contained, respectively. The values obtained for D(v, 0.1) ranged between 3.93 and 4.45 μm, while those for D(v,0.5) varied between 10.78 and 12.58 μm. A diameter of between 416.42 and 623.84 μm was found for D(v, 0.9). The mean spectrum obtained for each spray-head is shown in Figure 1.

Figure 1. Average droplet size distribution measured with a laser diffraction particle size analyser of BHI aerosols generated by three different Walther Pilot I spray-heads. The average of three aerosols analysed per spray-head is illustrated.

Table 4. Particle size distribution (D) expressed as a volume diameter (μm) of BHI aerosols produced by three different spray-heads

Percentage particle volume	D(v, 0.1)	D(v, 0.5)	D(v, 0.9)
Spray-head 1	4.45	12.58	434.18
Spray-head 2	3.93	10.78	623.84
Spray-head 3	4.12	11.54	416.42
Average particle diameter	4.17	11.63	491.48
The spectrum per spray-head represents the average of three aerosol measurements.

The results obtained for the combination ME broth and BPW were similar to those achieved with BHI (data not shown).

Discussion

All aerosol experiments were performed in duplicate and showed good reproducibility. Both air-sampling machines (MD2 and MD8 airscan) performed equally well with gelatine filters but sampling with cellulose nitrate filters was less efficient. Results of air samples taken with cellulose nitrate filters were ≥2 log₁₀ and ≥0.5 log₁₀ lower than with gelatine filters for mycoplasma and E. faecalis aerosols, respectively. Although air sampling with cellulose nitrate filters and the use of a filter rising device has proven very useful for the recovery of aerosolized Newcastle disease vaccine virus (Van Eck, 1990), this system seems less suitable for bacteria, especially mycoplasmas.

The isolator was not decontaminated between experiments. This was not considered to be necessary since a ventilation rate of 40 m³/h was applied for 20 min, which would have reduced any remaining micro-organisms to insignificant numbers. This is because a ventilation rate of 40 m³/h is equivalent to 1 m³/1.5 min and since the volume of the isolator was approximately 1 m³, it was ventilated once every 1.5 min and therefore 13.3 times in 20 min. After 1.5 min the bacterial concentration would be reduced by one-half and the dilution factor in 20 min would be 0.5^13.3=9.9×10⁻⁵=10⁻⁴ cfu/m³. In a case where an aerosol concentration of 10⁹ cfu/m³ air was found in the isolator at 25 min (which only applied to E. faecalis), the remaining number of cfu after ventilating would be 10⁻⁴×10⁹=10⁵ cfu/m³. If the remaining 10⁵ cfu were to be added to the concentrations found immediately after producing the second aerosol, the following concentrations are obtained: 10⁹+10⁵=10 001×10⁵ as opposed to 10 000×10⁵ (=10⁹), making a difference of 1/10⁴ cfu, which is insignificant in this total.

The reduction of micro-organisms by means of ventilation as mentioned earlier is true only if adequate mixing of the entering air with the air already present is ensured. Fresh air entering the isolator was conducted through a 32 cm long protruding stainless-steel tube of 7 cm diameter. This tube, which was perforated with eight rows of 22 holes of 1 cm diameter, was placed in a corner of the isolator facing inwards. The exhaust was placed diagonally opposite 1.6 m from the inlet. At the ventilation rate used (1 m³/1.5 min=11 100 cm³/sec) the airspeed at the small holes of the inlet was 0.8 m/sec and air would take only 2 sec to travel from the inlet to the exhaust. Considering a ventilation period of 20 min (1200 sec), adequate air mixing should have been ensured.

Differences in inactivation of aerosolized micro-organisms can be explained by several factors including differences in their susceptibility to shear forces of the spraying device, differences in sedimentation, differences in susceptibility to dehydration due to droplet evaporation and differences in susceptibility to the sampling technique.

Initial losses of approximately 4 log₁₀ cfu M. synoviae and 3 log₁₀ cfu M. gallisepticum, as measured by culture, were higher than the losses for E. faecalis of approximately 2 log₁₀ cfu. These losses were probably due to the deposition of aerosol droplets (physical loss) in the isolator and to destruction of bacteria. Evaporation is known to affect the viability of bacteria, and may have played an important role in the loss of the aerosolized mycoplasmas. Virus inactivation (Newcastle disease and infectious bronchitis virus) due to evaporation has been estimated at 1 log₁₀ EID₅₀ at 15°C, increasing to ≥3 log₁₀ at 30°C. Inactivation of Escherichia coli at 15°C was even larger (2 to 3 log₁₀ cfu) (J. H. H. van Eck, unpublished data, 1994). The physical loss in our experiments would have been similar in all aerosols because the same aerosol generator, isolator and climatic conditions were used. Mycoplasmas may be especially susceptible to environmental factors and could have been inactivated as a result of the air sampling technique, which might have caused dehydration of bacteria when exposed to air speeds of 0.2 m/sec for the MD2 and 0.1 m/sec for the MD8 airscan. The short half-life of <4 min for M. synoviae and of 5 to 6 min for M. gallisepticum suggests that high losses of viable mycoplasmas were not caused by the sampling technique, but this needs further investigation. For E. faecalis the half-life ranged from 19 to 75 min. This variation was explained by small fluctuations in bacterial concentrations, which are of great influence where the half-life is long. The initial concentrations found immediately after aerosolization were similar to that found in another aerosol study with E. faecalis (Landman & Van Eck, 2001). However, when comparing concentrations at 25 min in the present work with those at 30 min in the previous study, a 2 to 3 log₁₀ higher concentration was found here, resulting in a longer half-life. Differences in climatic conditions (relative humidity) may have influenced in part the outcome of the experiments.

The higher loss of mycoplasmas versus E. faecalis as measured by culture might be a consequence of the lack of a cell wall in the former. However, a satisfactory explanation for the higher loss of cultivable M. synoviae particles versus M. gallisepticum was not found. Investigations into the survival of M. gallisepticum and M. synoviae (Christensen et al., 1994) showed a tendency towards a lower survival rate of M. synoviae on several materials despite the fact that the inoculum was greater than M. gallisepticum (10⁷ to 10⁸ cfu M. synoviae compared with 10⁵ to 10⁶ cfu M. gallisepticum).

If 1-day-old chicks were to be exposed to these aerosols for 1 h, the M. synoviae, M. gallisepticum and E. faecalis uptake would be approximately ≥10² cfu/chick, ≥10² to 10³ cfu/chick and ≥10⁴ to 10⁵ cfu/chick, respectively, based on the assumption that the tidal volume of a 1-day-old chick is 1.2 l air/h (A. J. H. Visschedijk, personal communication, 1991), and the amount of aerosol inhaled is ≥10%. If the high losses of mycoplasmas were partly attributable to the sampling technique, then the calculated uptake would be higher. A particle size of at least <7 μm is required for inhalation. Particles of 3.7 to 7 μm were shown to be deposited primarily in the anterior portion of the respiratory system of adult cockerels, while particles of 1.1 μm were deposited mainly in the posterior parts (Hayter & Besch, 1974). Therefore, according to the aerosol spectra established immediately after generation, approximately 10% of the mass of the aerosol was able to be inhaled. This percentage will increase with time after aerosol generation due to particle size reduction because of evaporation. On the other hand, evaporation, as mentioned previously, will affect the viability of aerosolized micro-organisms. The large particle size found expressed by D(v, 0.9) was attributed to the incidental occurrence of high volume particles; a single large particle contributes to the total volume to a much larger extent than numerous small particles.

Acknowledgments

The authors thank Brenda Vermeulen for assisting in the aerosol spectra analyses and critically reading the manuscript, and thank Dirk Mekkes for performing the validation of the bacteriology of air filters.