Sterile intra-amniotic inflammation in asymptomatic patients with a sonographic short cervix: prevalence and clinical significance

Abstract

Objective: To determine the frequency and clinical significance of sterile and microbial-associated intra-amniotic inflammation in asymptomatic patients with a sonographic short cervix.

Methods: Amniotic fluid (AF) samples obtained by transabdominal amniocentesis from 231 asymptomatic women with a sonographic short cervix [cervical length (CL) ≤25 mm] were analyzed using cultivation techniques (for aerobic and anaerobic as well as genital mycoplasmas) and broad-range polymerase chain reaction (PCR) coupled with electrospray ionization mass spectrometry (PCR/ESI-MS). The frequency and magnitude of intra-amniotic inflammation [defined as an AF interleukin (IL)-6 concentration ≥2.6 ng/mL], acute histologic placental inflammation, spontaneous preterm delivery (sPTD), and the amniocentesis-to-delivery interval were examined according to the results of AF cultures, PCR/ESI-MS and AF IL-6 concentrations.

Results: Ten percent (24/231) of patients with a sonographic short cervix had sterile intra-amniotic inflammation (an elevated AF IL-6 concentration without evidence of microorganisms using cultivation and molecular methods). Sterile intra-amniotic inflammation was significantly more frequent than microbial-associated intra-amniotic inflammation [10.4% (24/231) versus 2.2% (5/231); p < 0.001]. Patients with sterile intra-amniotic inflammation had a significantly higher rate of sPTD <34 weeks of gestation [70.8% (17/24) versus 31.6% (55/174); p < 0.001] and a significantly shorter amniocentesis-to-delivery interval than patients without intra-amniotic inflammation [median 35, (IQR: 10–70) versus median 71, (IQR: 47–98) days, (p < 0.0001)].

Conclusion: Sterile intra-amniotic inflammation is more common than microbial-associated intra-amniotic inflammation in asymptomatic women with a sonographic short cervix, and is associated with increased risk of sPTD (<34 weeks). Further investigation is required to determine the causes of sterile intra-amniotic inflammation and the mechanisms whereby this condition is associated with a short cervix and sPTD.

Keywords:

• Broad-range real-time polymerase chain reaction with electrospray ionization mass spectrometry
• cervical length
• infection
• pregnancy
• prematurity
• premature birth
• preterm delivery
• ureaplasma urealyticim

View correction statement:

Correction

Introduction

A sonographic short cervix in the midtrimester is a powerful predictor of spontaneous preterm delivery (sPTD) [1–11]. The shorter the sonographic cervical length (CL) in the mid-trimester, the higher the likelihood of spontaneous preterm labor/delivery [12–32]. Patients with a sonographic short cervix can be offered treatment with vaginal progesterone, which reduces the rate of sPTD and neonatal morbidity, regardless of whether they have a previous history of sPTD [33–35]. Cervical cerclage is one option for women with a short cervix and history of preterm delivery [36,37], even though vaginal progesterone is as efficacious in this subgroup of patients [38]. Placement of a cervical pessary has also been proposed [39–42], although the evidence for its efficacy has been contradictory [43].

A short cervix can be considered as a syndrome caused by multiple pathologic processes [1,44]. For example, patients can have a short cervix because of prior cervical surgery (conization, loop electrosurgical excision procedure) [45–50], congenital disorders (e.g. cervical hypoplasia [51–54], exposure to diethylstilbestrol (DES) in utero [55–61], etc.), genetic syndromes (such as Ehlers-Danlos syndrome) [62], subclinical intra-amniotic infection [63–65] or intra-amniotic inflammation [66,67], or the elusive condition referred to as cervical insufficiency [68–70]. A decrease in progesterone action has been implicated in untimely cervical ripening [71–78], as this hormone plays a key role in maintaining the cervix as long and closed during pregnancy [77–88].

Since intra-amniotic inflammation (with or without detectable microorganisms) is associated with adverse pregnancy outcomes in preterm labor with intact membranes [89–102], preterm prelabor rupture of membranes (PROM) [95,103], and cervical insufficiency [69], we and others have explored the frequency of intra-amniotic inflammation in patients with a sonographic short cervix [66,67,104,105]. Several investigators have reported that a sonographic short cervix (CL ≤25 mm) in the midtrimester is associated with intra-amniotic inflammation, and that patients with this condition are at risk for adverse pregnancy outcome [66,67,104,105]. However, it is not clear whether this inflammatory process is due to the presence of microorganisms or represents a sterile inflammatory process. Consequently, we have used a combination of cultivation and molecular techniques for the identification of bacteria and a group of viruses frequently involved in human infections [106–135]. The purpose of this study was to determine the frequency and clinical significance of sterile and microbial-induced intra-amniotic inflammation in asymptomatic patients with a sonographic short cervix.

Materials and methods

Study population

A retrospective cohort of asymptomatic women with a sonographic short cervix was selected from the clinical database and Bank of Biological Samples of Wayne State University, the Detroit Medical Center, and the Perinatology Research Branch of the Eunice Kennedy Shriver National Institutes of Child Health and Human Development (NICHD) if they met the following criteria: (1) singleton gestation; (2) sonographic CL ≤25 mm; (3) amniocentesis (trans-abdominal) performed for microbiological studies; and (4) known pregnancy outcome. Patients were excluded from the study if they had: (1) rupture of the chorioamniotic membranes before AF collection; (2) a chromosomal or structural fetal anomaly; or (3) placenta previa.

In our institution, patients diagnosed with a sonographic CL ≤ 25 mm in the midtrimester are offered an amniocentesis to determine the microbial status of the amniotic cavity, as previous observations indicate an association between a sonographic short cervix, intra-amniotic infection, and adverse pregnancy outcome [65,136]. Moreover, treatment with antibiotics in patients with a sonographic short cervix and a positive AF culture was shown to result in eradication of intra-amniotic infection and frequently associated with delivery at term [65,137]. Patients were counseled by their treating physicians, and those who agreed to undergo an amniocentesis were asked to donate AF and allow collection of clinical information for research purposes. Further management of these patients was at the discretion of the attending physician. All patients provided written informed consent and the use of biological specimens as well as clinical and ultrasound data for research purposes were approved by the Institutional Review Boards of NICHD and Wayne State University.

Clinical definitions

Gestational age was determined by the last menstrual period and confirmed by ultrasound examination, or by ultrasound examination alone if the sonographic determination of gestational age was not consistent with menstrual dating by more than 1 week. Gestational age at diagnosis was defined as the earliest gestational age at which a sonographic CL ≤ 25 mm was documented. Patients with an a priori risk for sPTD (“high risk”) were defined as those having at least one of the following: one or more previous spontaneous preterm deliveries (≤35 weeks); one or more late mid-trimester spontaneous miscarriages (≥16 weeks); and prior cervical surgery (loop electrosurgical excision procedure or cone biopsy) [49]. Spontaneous preterm labor was defined by the presence of regular uterine contractions occurring at least twice every 10 min, associated with cervical change before 37 completed weeks of gestation. Early sPTD was defined as delivery before 34 completed weeks of gestation. Patients defined as having sPTD included those with spontaneous preterm labor, prelabor PROM and those in whom labor was induced or augmented as a result of clinical chorioamnionitis (defined according to the diagnostic criteria proposed by Gibbs et al.) [138]. Intra-amniotic inflammation was diagnosed when the interleukin-6 (IL-6) AF concentration was ≥2.6 ng/mL [93,139]. Patients who were lost to follow-up and those for whom no delivery data were available were censored from the last documented follow-up visit. Microbial invasion of the amniotic cavity (MIAC) was defined according to the results of AF culture and PCR/ESI-MS analysis [140–143]. Based on the results of AF culture, PCR/ESI-MS and AF concentration of IL-6, patients were classified as: (1) no intra-amniotic inflammation or microorganisms in the amniotic cavity by either AF culture or PCR/ESI-MS; (2) MIAC (identification of microorganisms without the presence of intra-amniotic inflammation); (3) microbial-associated intra-amniotic inflammation (combination of MIAC and intra-amniotic inflammation); or (4) sterile intra-amniotic inflammation (intra-amniotic inflammation without evidence of microorganisms using cultivation and molecular methods). Acute histologic chorioamnionitis was diagnosed based on the presence of inflammatory cells in the chorionic plate and/or chorioamniotic membranes [144], and acute funisitis was diagnosed by the presence of neutrophils in the wall of the umbilical vessels and/or Wharton’s jelly, using the criteria previously described [144–146].

Sonographic assessment of the cervix

Transvaginal ultrasound examinations were performed using commercially available ultrasound systems (Acuson Sequoia, Siemens Medical Systems, Mountain View, CA; Voluson 730 Expert™ or Voluson E8, GE Healthcare, Milwaukee, WI) equipped with endovaginal transducers with frequency ranges of 5–7.5 MHz and 5–9 MHz, respectively. All sonographic examinations of the cervix were performed using a previously described technique [2,4].

Sample collection

AF was transported in a capped sterile syringe to the clinical laboratory where it was cultured for aerobic and anaerobic bacteria, including genital mycoplasmas. Evaluation of white blood cell (WBC) count, AF glucose concentration and Gram stain of AF were also performed shortly after collection. AF not required for clinical assessment was centrifuged for 10 min at 4 °C shortly after the amniocentesis, and the supernatant was aliquoted and stored at −70 °C until analysis. Following delivery, the placenta, umbilical cord, and chorioamniotic membranes were collected and the presence or absence of acute histologic chorioamnionitis and/or funisitis was determined.

Detection of microorganisms with cultivation and molecular methods

AF was analyzed using cultivation techniques (for aerobic, anaerobic and genital mycoplasmas) and with PCR/ESI-MS (Ibis Technology – Athogen, Carlsbad, CA). Briefly, DNA was extracted from 300 uL of AF using a method that combined bead-beating cell lysis with a magnetic-bead based extraction method [147,148]. The extracted DNA was amplified on the bacterial artificial chromosome (BAC) spectrum assay according to the manufacturer’s instructions. PCR/ESI-MS can identify 3400 bacteria and 40 Candida spp, which are represented in the platform’s signature database [120,131,149]. A total of 200 µL of extract was used per sample. For viral detection, nucleic acids were extracted from 300 µL of AF using a method that combined chemical lysis with a magnetic-bead based extraction method. The extracted RNA/DNA was amplified on the broad viral assay according to the manufacturer’s instructions. In the eight wells, 14 primer pairs were used to detect the following viruses: Herpes simplex virus 1 (HHV-1), Herpes simplex virus 2 (HHV-2), Varicella-zoster virus (HHV-3), Epstein-Barr virus (HHV-4), Cytomegalovirus (HHV-5), Kaposi’s sarcoma-associated herpes virus (HHV-8), Human adenoviruses, Human enteroviruses, BK polyomavirus, JC polyomavirus and Parvovirus B19 [149].

After PCR amplification, 30 -μL aliquots of each PCR product were desalted and analyzed via ESI-MS as previously described [110,120]. The presence of microorganisms was determined by signal processing and triangulation analysis of all base composition signatures obtained from each sample and compared to a database. Along with organism identification, the ESI-MS analysis includes a Q-score and level of detection (LOD). The Q-score, a rating between 0 (low) and 1 (high), represents a relative measure of the strength of the data supporting identification; only Q-scores ≥ 0.90 were reported for the BAC Spectrum assay [132]. The LOD describes the amount of amplified DNA present in the sample: this is calculated with reference to an internal calibrant, as previously described [109], and is reported herein as genome equivalents per PCR reaction well (GE/well). The sensitivity (LOD) of PCR/ESI-MS for the detection of bacteria in blood is on average 100 CFU/mL (95% CI, 6–600 CFU/mL) [131]. A comparison of detection limits between blood and amniotic fluid (AF) showed that the assays have comparable detection limits (100 CFU/mL). The sensitivity (LOD) for the broad viral in plasma ranges from 400 copies/mL to 6600 copies/mL [150]. A comparison of detection limits between AF and plasma was performed by spiking known amounts of a DNA virus (HHV-5) and an RNA virus (Human enterovirus) into AF and plasma. Detection limits in AF were similar to plasma, ranging from ∼800 to 1600 copies/mL (depending upon the specific microorganism).

Determination of IL-6 in amniotic fluid

AF concentrations of IL-6 were determined to assess the magnitude of the intra-amniotic inflammatory response. We used a sensitive and specific enzyme immunoassay obtained from R&D Systems (Minneapolis, MN). Briefly, the immunoassay utilized was the quantitative sandwich enzyme-linked immunoassay technique, and the concentrations were determined by interpolation from the standard curves. The inter- and intra-assay coefficients of variation for IL-6 were 8.7% and 4.6%, respectively. The detection limit of the IL-6 assay was 0.09 pg/mL. AF IL-6 concentrations were determined for research purposes, and such results were not used in patient management. We have previously reported the use of IL-6 for the assessment of intra-amniotic inflammation [89,90,93,94,135,140,142,151–164].

Statistical analysis

The Kolmogorov–Smirnov test and visual plot inspection were used to assess the normality of continuous data distributions. Patients were stratified by gestational age upon diagnosis of a sonographic short cervix (24 weeks) and according to the presence of intra-amniotic inflammation or MIAC. Between-group comparisons were performed using the Kruskal–Wallis and Mann–Whitney U tests to examine the differences in arithmetic variable distributions. The χ² or Fischer’s exact test was used to determine differences in proportions, as appropriate. Kaplan–Meier survival analyses were performed to assess the amniocentesis-to-delivery interval, according to the gestational age (censoring observations for patients delivered for maternal or fetal indications), upon diagnosis of a sonographic short cervix, as well as by the presence of intra-amniotic inflammation. Logistic and Cox proportional hazard regression models were used to examine magnitudes of association. A two-tailed p value of <0.05 was considered statistically significant. The statistical package used was SPSS v.15.0 (SPSS, Chicago, IL).

Results

Characteristics of the study population

During the study period, 231 patients had a sonographic CL ≤25 mm between 16 and 32 weeks of gestation. All patients included in the study were asymptomatic at the time of amniocentesis. Demographic and clinical characteristics of the study population are shown in Table 1. Thirty-nine percent of the patients (91/231) were nulliparous, while 32% (74/231) had a history of at least one sPTD. The median [interquartile range (IQR)] gestational age at diagnosis of a sonographic short cervix was 24.3 (21.1–27.3) weeks. According to the gestational age at diagnosis, 46.8% (108/231) of the patients had a sonographic short cervix <24 weeks and 53.2% (123/231) at ≥24 weeks (Table 1). The CL at diagnosis was significantly shorter in patients with a short cervix diagnosed <24 weeks than in those diagnosed ≥24 weeks of gestation [12 (6–15.7) versus 14 (10–19) mm; (p = 0.01), Table 1]. In addition, pre-pregnancy body mass index (BMI) was significantly higher in patients with a CL ≤25 mm diagnosed <24 weeks than in those diagnosed ≥24 weeks of gestation [28.1 (23.5–34.6) versus 24.7 (21.3–31.5) kg/m²; (p = 0.02), Table 1]. There were no significant differences in race, cervical dilatation at admission, tobacco use, and previous history of sPTD between women diagnosed with a sonographic short cervix <24 weeks and those diagnosed ≥24 weeks of gestation (Table 1).

Table 1. Demographic and clinical characteristics of the study population.

		GA at diagnosis of short cervix
	Overall cohort (n = 231)	≥24 weeks of gestation (n = 108)	>24 weeks of gestation (n = 123)	p* value
Maternal age (years)	23.5 (20–28)	25.7 (22–31.7)	22 (19–26)	<0.001
Pre-pregnancy body mass index (kg/m^2)	26.2 (21.9–31.9)	28.1 (23.5–34.6)	24.7 (21.3–31.5)	0.02
Race				0.46
African-American	91.3% (211)	91.7% (99)	91.1% (112)
White	3% (7)	2.8% (3)	3.3% (4)
Hispanic	3% (7)	4.6% (5)	1.6% (2)
Asian	0.9% (2)	0	1.6% (2)
Other	1.7% (4)	0.9% (1)	2.4% (3)
Smoker	15.6% (36)	17.3% (18/104)	14.8% (18/122)	0.6
Nulliparity	39.4% (91)	38% (41)	40.7% (50)	0.67
History of spontaneous preterm delivery	32% (74)	35.2% (38)	29.3% (36)	0.34
Sonographic cervical length at diagnosis (mm)	13 (7–18)	12 (6–15.7)	14 (10–19)	0.01
GA at amniocentesis (weeks)	24.3 (21.1–27.3)	20.9 (19.2–22.7)	27 (25.4–29.1)	<0.001

Data are presented as median (interquartile range) or % (n).

*Comparison between patients with a diagnosis of short cervix before and after 24 weeks of gestation. GA, gestational age.

The prevalence of microbial invasion of the amniotic cavity

The frequency of MIAC in asymptomatic patients with a sonographic short cervix was 2.2% (5/231) using cultivation techniques. However, the frequency was 12.6% (29/231) when using PCR/ESI-MS to amplify microbial DNA/RNA in AF. Table 2 shows the microorganisms identified, inflammatory markers in AF (WBC count and IL-6 concentrations), and gestational age at delivery, as well as the presence or absence of acute placental inflammation for each patient with detectable genomic material by PCR/ESI-MS or positive AF culture. Only one of the 29 patients had both a positive PCR/ESI-MS and AF culture. The most frequent microorganism identified by PCR/ESI-MS was Staphylococcus aureus (n = 6). However, in the majority of cases, this was not associated with an intra-amniotic inflammatory response, and the microbial burden or inoculum size expressed as GE/well was uniformly low (Table 2). Fusobacterium nucleatum, Gardnerella vaginalis, Mycoplasma hominis and Acinetobacter junii were detected only by PCR/ESI-MS, and not by AF culture. Seven patients (3%) had a positive viral assay using PCR/ESI-MS. Herpes simplex virus 2 (HHV2) was identified in four patients; whereas in two patients, genomic material of Cytomegalovirus was detected. One patient demonstrated Parvovirus B19 along with detection of genomic material for other microorganisms (Table 2).

Table 2. Inflammatory markers in amniotic fluid, pregnancy outcomes and placental pathology results of patients in whom microorganisms were detected in amniotic fluid using standard cultivation techniques vs. PCR/ESI-MS.

Group	AF culture	PCR/ESI-MS	GE/well	AF IL-6 (ng/mL)	AF WBC	GA at amniocentesis (weeks)	Interval amniocentesis- to-delivery	GA at delivery (weeks)	Acute placental inflammation
Microbial-associated intra-amniotic inflammation (n = 5)	Negative	Fusobacterium nucleatum	97	359.10	605	23.4	6	24.3	Yes
		Gardnerella vaginalis	168	6.63	2	23.6	18	26.1	Yes
		Herpes simplex virus 2	88	3.02	5	21.9	4	22.4	Yes
		Staphylococcus epidermidis	25	98.11	1265	24.1	20	27.0	Yes
		Staphylococcus aureus	8	8.40	26	27.3	100	41.6	No
Microbial invasion of the amniotic cavity (MIAC) (n = 28)	Negative	Acidovorax temperans	29
		Acinetobacter junii	25
		Candida tropicalis	35	0.69	4	27.9	53	35.4	No
		S. aureus	8
		Parvovirus B19	36
		Viridans/Mitis Group Streptococcus	11
		Viridans/Salivarius Group Streptococcus	3	0.44	24	23.9	59	32.3	No
		A. temperans	54
		A. junii	17	0.98	1	29.4	7	30.4	No
		Actinomyces odontolyticus	3
		Streptococcus infantis/peroris	6	0.16	0	23.1	33	27.9	No
		Herpes simplex virus 2	90	0.16	0	20.9	8	22.0	No
		S. aureus	3	0.43	0	24.6	25	28.1	No information
		Bacteria detected - No ID could be assigned	11	2.10	0	23.6	117	40.3	No
		S. aureus	6	1.11	1	25.4	92	38.6	No
		A. junii	26
		Bacteria detected - No ID could be assigned	10	0.37	0	26.6	85	38.7	Yes
		S. aureus	4	0.25	0	26.0	63	35.0	No
		A. junii	15
		Cytomegalovirus	10	0.92	0	29.1	51	36.4	No information
		Bacteroides fragilis	26
		Sneathia species	20	0.72	20	29.4	84	41.4	No
		Bacteria detected - No ID could be assigned	12	0.49	0	25.7	93	39.0	No
		Herpes simplex virus 2	24	0.98	1	30.0	63	39.0	No information
		Rhodococcus sp. RHA1	119
		A. temperans	32	0.04	2	20.3	125	38.1	No information
		A. junii	41
		Cytomegalovirus	56	0.26	0	30.1	51	37.4	Yes
		S. aureus	5	1.04	2	25.3	86	37.6	No
		G. vaginalis	15	0.65	1	21.7	99	35.9	No
		Moraxella osloensis	5	0.55	10	21.6	123	39.1	No information
		A. junii	35	0.22	1	19.1	138	38.9	No
		Kytococcus sedentarius	93	0.22	0	20.1	126	38.1	Yes
		Herpes simplex virus 2	199	0.46	2	23.3	103	38.0	Yes
		A. temperans	104	0.69	2	26.3	79	37.6	No information
	Enterococus faecalis	Corynebacterium tuberculostearicum	21	0.99	0	25.1	75	35.9	No
		Staphylococcus caprae	11
	Fusobacterium	Negative	N/A	0.66	2	22.3	27	26.1	Yes
	S. aureus	Negative	N/A	1.02	4	19.9	24	23.3	Yes
	Ureaplasma Urealyticum	Negative	N/A	0.73	4	25.1	90	38.0	No information
	Ureaplasma Urealyticum	Negative	N/A	0.65	1	22.9	105	37.9	No

AF, amniotic fluid; PCR, polymerase chain reaction; ESI-MS, electrospray ionization mass spectrometry; GE/well, genome equivalents per PCR reaction well; IL-6, interleukin 6; N/A, not applicable; WBC, white blood cell count.

The prevalence of intra-amniotic inflammation

Intra-amniotic inflammation (defined as AF IL-6 ≥2.6 ng/mL) was present in 12.6% (29/231) of patients. When combining the results of AF cultures, PCR/ESI-MS and AF IL-6 concentrations, 2.2% (5/231) of patients had microbial-associated intra-amniotic inflammation. Ten percent (24/231) of patients had intra-amniotic inflammation without detection of either bacteria or viruses, and were thus categorized as having sterile intra-amniotic inflammation. Therefore, in asymptomatic patients with a sonographic short cervix, sterile intra-amniotic inflammation was significantly more frequent than microbial-associated intra-amniotic inflammation [10.4% (24/231) versus 2.2% (5/231); p < 0.001]. Four of the five patients with microbial-associated intra-amniotic inflammation had a sPTD <32 weeks of gestation, and all were associated with acute histologic placental inflammation (Table 2).

Clinical characteristics and outcomes of patients with microbial-associated and sterile intra-amniotic inflammation

Table 3 displays the clinical characteristics of patients in whom the diagnosis of a sonographic short cervix was made before 24 weeks of gestation, according to the results of AF concentrations of IL-6, AF cultures and PCR/ESI-MS. Although there were no differences in the median (IQR) gestational age at amniocentesis among different clinical groups (p = 0.09), the median (IQR) CL at diagnosis was significantly shorter in patients with sterile and microbial-associated intra-amniotic inflammation than in women without intra-amniotic inflammation [0 (IQR: 0–9) and 0 mm (IQR: 0–1) versus 13 mm (0.7–13.2), respectively (each p < 0.001); Table 3]. Importantly, the median (IQR) gestational age at delivery was similar between patients with microbial-associated intra-amniotic inflammation and those with sterile intra-amniotic inflammation [24 (22.4–26.1) versus 25 (20.8–30.5) weeks; p = 0.6]. The rate of sPTD <34 weeks of gestation was also significantly higher in patients with sterile intra-amniotic inflammation than in those without intra-amniotic inflammation [80% (12/15) versus 46.8% (36/77); p < 0.001]. This difference remained significant when patients with a CL ≤25 mm at or after 24 weeks of gestation were included in the analysis [70.8% (17/24) versus 31.6% (55/174); p < 0.001]. No significant differences were detected in the prevalence of acute placental inflammation (acute histologic chorioamnionitis and/or funisitis) between patients with sterile intra-amniotic inflammation and those without intra-amniotic inflammation [41.7% (5/12) versus 46.9% (30/77); p = 0.71] (Table 3).

Table 3. Frequency of intra-amniotic inflammation and clinical outcome in asymptomatic patients with a sonographic short cervix before 24 weeks of gestation according to the results of PCR/ESI-MS, AF cultures and AF interleukin-6 concentrations.

	No MIAC or intra-amniotic inflammation (n = 77)	MIAC (n = 13)	Microbial associated intra-amniotic inflammation (n = 3)	Sterile intra-amniotic inflammation (n = 15)	p value
Gestational age at amniocentesis (weeks)	20.7 (19.3–22.5)	21.7 (20.2–23.2)	23.4 (21.8–23.6)	21.1 (20.1–22.1)	0.09
Cervical length (mm)	13 (8.5–17)	12 (6.5–18)	0 (0–1)	0 (0–9)	<0.001
AF white blood cells count (cells/mm^3)	3.5 (0.7–13.2)	1 (0–3)	2 (5–605)	2 (2–6)	0.26
AF glucose (mg/dL)	32 (27–36.7)	37.8 (27–38.5)	22.4 (24.3–36)	26 (20–31)	0.02
AF interleukin-6 (ng/mL)	0.4 (0.23–0.93)	0.46 (0.19–0.66)	3.0 (6.6–359)	7.6 (4.4–22.3)	<0.001
Amniocentesis-to-delivery interval (days)	92 (38–121)	103 (30–124)	6 (4–18)	16 (6–62)	<0.001
Gestational age at delivery (weeks)	34.5 (26.5–38.7)	37.8 (27–38.5)	24.2 (22.4–26.1)	25 (20–30)	0.001
sPTD before 32 weeks	35.1% (27)	30.8% (4)	100% (3)	80% (12)	0.02
sPTD before 34 weeks	46.8% (36)	38.5% (5)	100% (3)	80% (12)	0.03
sPTD before 37 weeks	53.2% (41)	46.2% (6)	100% (3)	80% (12)	0.09
Acute placental inflammation*	46.9% (30/64)	36.4% (4/11)	100% (3/3)	41.7% (5/12)	0.26
Acute histologic chorioamnionitis	46.9% (30)	27.3% (3/11)	100% (3/3)	41.7% (5/12)
Funisitis	32.8% (21/64)	27.3% (3/11)	100% (3/3)	25% (3/12)

Data are presented as % (n/N).

AF, amniotic fluid; MIAC, microbial invasion of the amniotic cavity; sPTD, spontaneous preterm delivery.

*Data were calculated over 90 placenta samples available for pathology examination in patients with a sonographic short cervix <24 weeks.

Outcome in patients with and without a history of preterm birth

Among the study population, 74 patients (32%) had a history of sPTD. Of this population, 8 patients had a cervical cerclage and 16 received 17 α-hydroxyprogesterone caproate in the index pregnancy. There were no significant differences in the demographic characteristics (age, race, pre-pregnancy BMI, smoke during pregnancy) between low and high risk patients (data not shown). The rate of sterile intra-amniotic inflammation was not different between patients with without a history of sPTD [10.8% (8/74) versus 10.2% (16/157); p = 0.8]. Of the patients who had a cervical cerclage in the index pregnancy, three had intra-amniotic inflammation and all delivered before 34 weeks of gestation. The rate of sPTD <37 weeks of gestation was significantly higher in patients with a history of sPTD than in those without [63.5% (47/74) versus 40.8% (64/157); p = 0.001].

The relationship between intra-amniotic inflammation and time-to-delivery

Figure 1 displays the amniocentesis-to-delivery interval, according to the presence of intra-amniotic inflammation, MIAC and sterile intra-amniotic inflammation among asymptomatic patients with a sonographic short cervix. Patients with sterile intra-amniotic inflammation had a significantly shorter amniocentesis-to-delivery interval than women without intra-amniotic inflammation [median 35, (IQR: 10–70) versus median 71, (IQR: 47–98) days; p < 0.0001 (Figure 1)]. There was no significant difference in the amniocentesis-to-delivery interval between patients with microbial-associated intra-amniotic inflammation and those with sterile intra-amniotic inflammation [median 18 (IQR: 6–20) versus median 35 (IQR: 10–70) days; p = 0.7]. Importantly, the amniocentesis-to-delivery interval was similar in patients with MIAC and those without any evidence of intra-amniotic inflammation [median 79, (IQR: 51–99) versus median 71, (IQR: 47–98) days; p = 0.9, (Figure 1)]. Therefore, we conclude that the mere presence of microorganisms in the amniotic cavity without a detectable inflammatory response is unlikely to have pathologic consequences that can be discernable based on obstetrical outcomes.

Figure 1. Kaplan–Meier survival curves of amniocentesis-to-delivery interval (days) among asymptomatic patients diagnosed with a CL ≤25 mm, according to the presence of microbial-associated or sterile intra-amniotic inflammation. Patients in whom labor was induced were censored and are represented by crosses. The amniocentesis-to-delivery interval among women with sterile intra-amniotic inflammation was significantly shorter than that of: (1) women without intra-amniotic inflammation; and (2) women with MIAC [median 35, (IQR: 10–70) versus median 71, (IQR: 47–98) days, and median 79, (IQR: 51–99) days (p < 0.001 and p = 0.02), respectively]. There was no significant difference in the amniocentesis-to-delivery interval between patients with sterile intra-amniotic inflammation and those with microbial-associated intra-amniotic inflammation (p > 5).

Multivariable survival analysis was used to examine the amniocentesis-to-delivery interval among different clinical groups, adjusting for maternal age, nulliparity, race, smoking status and pre-pregnancy BMI using Cox proportional hazard modeling. Patients with sterile intra-amniotic inflammation were at nearly four times greater risk of delivery per unit of time as those without intra-amniotic inflammation [hazard ratio of 3.6 (95% confidence interval (CI), 1.6–8.0)] (Table 4). Moreover, when considering only patients in whom a sonographic short cervix was diagnosed before 24 weeks of gestation, those with sterile intra-amniotic inflammation were at five times greater risk of delivery per unit of time as those without intra-amniotic inflammation [hazard ratio of 5.4 (95% CI%, 1.5–19.0)]. However, this was not the case in patients with microorganisms detected with molecular techniques and without intra-amniotic inflammation (Table 4).

Table 4. Magnitude of association between type of intra-amniotic inflammation and risk of spontaneous preterm delivery at <34 weeks of gestation both unrestricted and restricted to patients whose diagnosis of a sonographic short cervix was performed <24 weeks of gestation.

Study group			Unadjusted			Adjusted*
	n	%	HR	95%CI		HR	95%CI
All the study population
No MIAC or intra-amniotic inflammation	174	75	1	Reference		1	Reference
MIAC	24	11	1.0	0.3	3.5	1.0	0.3	3.1
Microbial-associated intra-amniotic inflammation*	5	2	4.8	0.7	31.2	3.9	0.6	26.0
Sterile intra-amniotic inflammation	28	12	3.9	1.9	7.8	3.6	1.6	8.0
Sonographic short cervix <24 weeks
No MIAC or intra-amniotic inflammation	77	71	1	Reference		1	Reference
MIAC	15	14	1.6	0.4	5.9	1.6	0.5	5.0
Microbial-associated intra-amniotic inflammation*	3	3	13.6	4.7	39.3	14.7	2.8	75.8
Sterile intra-amniotic inflammation	13	12	4.3	1.8	10.5	5.4	1.5	19.0

HR, hazard ratio. Bold and shaded values correspond to the adjusted and unadjusted HR for sterile and microbial intra-amniotic inflammation, respectively.

Multivariable adjustments were made for maternal age, nulliparity, race, 17 OHPC treatment, smoking status and pre-pregnancy body mass index.

*Interpretation of these hazard ratios should take into account the small number of subjects.

Discussion

Principal findings of the study

(1) Sterile intra-amniotic inflammation was present in 10% of asymptomatic women with a sonographic short cervix (CL ≤25 mm) and more common than microbial-associated intra-amniotic inflammation; and (2) patients with a short cervix and sterile intra-amniotic inflammation had a higher rate of sPTD and a shorter amniocentesis-to-delivery interval than those without intra-amniotic inflammation.

The prevalence and significance of microbial invasion of the amniotic cavity in patients with a sonographic short cervix

AF is normally sterile [163,165–167]. Intra-amniotic infection [168–176] and inflammation [151,158,175,177–195] are risk factors for adverse outcomes in spontaneous preterm labor with intact membranes [140,157,161,196–209], preterm PROM [141,210–217], cervical insufficiency [63,64,69,218,219], a sonographic short cervix [65,67,220], patients with idiopathic vaginal bleeding [221], placenta previa [162,222], and those with an intrauterine device (IUD) [223]. The presence of viable microorganisms detectable with culture has been observed in asymptomatic patients with a sonographic short cervix [65,220]; 9% of patients with a CL ≤ 25 mm (14–24 weeks) have a positive AF culture [65]. Moreover, in patients with preterm labor and intact membranes, the shorter the cervix and the earlier the gestational age at diagnosis, the higher the risk of infection/inflammation [220]. Patients with a CL ≤ 15 mm had a higher rate of positive AF culture [26% (15/57)] than patients with a CL ≥15 mm [3.8% (13/344)] [220]. It is unclear if a short cervix leads to compromised innate immunity following the loss of the anti-microbial properties of the cervical mucus plug [224–229] and predisposes to ascending intra-amniotic infection, or alternatively, if intra-amniotic infection leads to a short cervix.

Molecular microbiologic techniques for the detection of microbial invasion of the amniotic cavity

Given the well-known limitations of cultivation methods [230], molecular techniques have been used for the detection of microorganisms in the amniotic cavity [140–143,208,209,217,219,231–249]. The advantage of PCR/ESI-MS technology over conventional PCR methods using broad primers is that this method uses a “closed system” designed to minimize the possibility of contamination. Mass spectrometry to quantitate the molecular weight of the amplicons allows identification of the organism [111,117], within 8 h [106,120,128].

To determine the clinical significance of a positive PCR/ESI-MS signal, we evaluated the presence of intensity of the intra-amniotic inflammatory response, the interval to delivery and the presence of inflammatory lesions of the placenta. PCR/ESI-MS was positive in 29 cases; however, in most, the organisms were those found in skin (e.g. S. aureus) (Table 2). Patients with a low microbial burden without intra-amniotic inflammation are at low risk for adverse pregnancy outcome. We previously observed that patients with a positive PCR/ESI-MS and a high microbial burden >17 GE/well were likely to have an inflammatory response and adverse pregnancy outcome, and that those with a positive PCR and a low inoculum size did not have demonstrable evidence of adverse pregnancy outcome or complications [135]. Therefore, we propose that patients with a positive PCR/ESI-MS with a low inoculum size, negative culture, and without evidence of intra-amniotic inflammation probably represent false-positive results attributable to contamination of the specimen in the clinical setting or laboratory. Thus, we consider that interpretation of the clinical significance of a positive PCR result needs to be informed by the microbial burden (number of genome equivalents) and the magnitude of the intra-amniotic inflammatory response.

Sterile inflammation as a determinant of adverse pregnancy outcome in patients with a sonographic short cervix in the midtrimester

Sterile intra-amniotic inflammation (defined as the presence of intra-amniotic inflammation as determined by an elevation in cytokines, chemokines, matrix degrading enzymes or other inflammatory markers [89,91–95,97,100,157,158,160,162,178,179,189,250–259] etc.) in the absence of detectable microorganisms has long been suspected to play a role in pregnancy complications. However, a major obstacle has been the reliance on cultivation techniques to detect the presence of microorganisms. Ever since the description of the “great plate count anomaly”, most of the microbial world has been considered non-culturable [260–264], and therefore, intra-amniotic inflammation associated with a negative AF culture was often attributed to organisms that escaped detection with standard microbiologic techniques. However, molecular microbiologic techniques have been used to address the issue of the limited sensitivity of cultivation techniques. Our studies provide evidence that microorganisms have not been found in most cases with intra-amniotic inflammation (both bacteria and several viruses implicated in human disease). Therefore, we believe it is prudent to consider that many intra-amniotic inflammatory processes are likely due to danger signals of non-microbial origin.

The precise nature of the non-microbial danger signals that induce intra-amniotic inflammation are unknown. Similarly, the mechanisms whereby danger signals would lead to the onset of preterm labor, preterm PROM, a short cervix, and other complications of pregnancy need investigation. We have previously reported that patients with intra-amniotic inflammation without detectable microorganisms have an elevation of caspase-1, an enzyme implicated in the activation of the inflammasome complex [265]. Therefore, we propose that danger signals can induce inflammation through the activation of the inflammasome in preterm labor not associated with infection, and also in spontaneous labor at term [265]. Parturition at term is an example of physiologic sterile inflammation. Indeed, unbiased approaches such as transcriptomics have demonstrated a molecular inflammatory signature in the myometrium and chorioamnionitic membranes, even in the absence of histologic evidence of inflammation [82,83,85,266,267]. In addition, AF concentrations of IL-1, IL-6, IL-8, MCP-1 and others are higher in term gestations during labor than in those without labor [178,257,268–270].

The findings reported herein are consistent with those of Kiefer et al., who reported an association between a sonographic short cervix (CL ≤ 5 mm) in the midtrimester and high AF concentrations of chemokines/cytokines such as monocyte chemotactic protein (MCP)-1 and IL-6 [66]. Subsequently, using a comprehensive panel of 25 cytokines in asymptomatic patients with a sonographic short cervix between 14 and 25 weeks of gestation, Keeler et al. reported that 9 cytokines were significantly correlated with the amniocentesis-to-delivery interval [104]; specifically, AF concentrations of MCP-1 >1320 pg/mL were predictive of preterm delivery within one week of amniocentesis [104]. The same group recently reported that an inflammatory score based on 14 cytokines can predict sPTD <34 weeks of gestation in patients with a sonographic short cervix (sensitivity of 87%, specificity of 100%, positive predictive value of 100%, and negative predictive value of 87.5%) [105]. Collectively, this information suggests that patients with a sonographic short cervix and intra-amniotic inflammation are at risk for preterm delivery.

The diagnosis of sterile inflammation in patients with a short cervix

The standard approach for the rapid evaluation of AF includes the performance of a Gram stain [155,271–275], AF WBC count [155,272,273,275–277], and AF glucose concentrations [155,272–275,278–284]. The Gram stain of AF detects bacteria, and has a high specificity but low sensitivity [271–274]. A negative Gram stain does not exclude the presence of microorganisms (the most frequent being Mycoplasma or Ureaplasma species, which are not detected by Gram stain). A low AF glucose concentration and an elevated AF WBC count have been considered evidence of intra-amniotic inflammation in response to microbial invasion. However, we previously reported that patients with sterile intra-amniotic inflammation do not have an elevated WBC count or low AF glucose concentration [285]. Such observations are consistent with those reported herein (see Table 3). Consequently, reliance on the standard methods of AF analysis is not adequate for the identification of the most frequent form of intra-amniotic inflammation. Detection of sterile inflammation requires assessment for cytokines and/or chemokines, or MMPs, such as IL-6 or MMP-8 [139,286]. Until recently, these tests were research tools. However, point of care tests have now been developed which can identify intra-amniotic inflammation in less than 30 min and do not require sophisticated equipment [287].

Therapeutic implications of the diagnosis of sterile intra-amniotic inflammation

When intra-amniotic inflammation is detected in patients with negative AF cultures for microorganisms, most investigators have recommended treatment with antibiotics assuming that an infection has not been detected with microbiologic techniques. In light of the findings reported herein, this approach needs to be reevaluated. A rational strategy to this clinical challenge requires identification of the danger signal inducing intra-amniotic inflammation. Patients with sterile intra-amniotic inflammation are more likely to have acute inflammatory lesions of the placenta and neonatal morbidity than those without intra-amniotic inflammation, even after adjustment for gestational age [285]. Therefore, further work is necessary to characterize the frequency, type of multi-organ involvement, and the behavior of the cytokine network in the fetal systemic inflammatory response associated with sterile intra-amniotic inflammation.

It is noteworthy that the standard treatment of patients with spontaneous preterm labor includes the administration of glucocorticoids, which are powerful anti-inflammatory agents [288,289]. Moreover, magnesium sulfate is frequently administered to patients at risk of sPTD to reduce the rate of cerebral palsy in the offspring [290–293]. Magnesium sulfate has also been reported to have anti-inflammatory properties [294].

An important question is whether patients with a short cervix and sterile intra-amniotic inflammation respond to vaginal progesterone, cervical cerclage, or a cervical pessary. Our study was conducted before the publications of studies reporting that vaginal progesterone reduces the rate of preterm delivery and neonatal morbidity, and thus, none of our patients received vaginal progesterone. Insofar as cerclage, of the 19 patients who had this procedure, three had intra-amniotic inflammation and all delivered before 34 weeks of gestation. This is consistent with previous observations that patients with an elevated IL-6 concentration predict a short interval cerclage-to-delivery and have a higher rate of sPTD [295]. Therapeutic interventions for treatment of intra-amniotic inflammation require further basic and clinical studies.

Strengths and limitations

Strengths of this study include the large sample size, the use of state-of-the-art molecular microbiologic techniques for the detection of bacterial and viral footprints, the assessment of AF inflammation using well-established methods (IL-6 determination), and placental pathology. This study is the first to use molecular microbiologic techniques for bacteria and viruses in patients with a sonographic short cervix. We have used the term sterile intra-amniotic inflammation to refer to a state in which there is an elevated AF IL-6 without detectable microorganisms by culture and PCR. The absence of microorganisms is, of course, a function of the sensitivity of the technique. Is it possible that more sensitive techniques will find microorganisms in patients with intra-amniotic inflammation? Future studies using a metagenomic approach may find microbial footprints which have not been detected with the methods used in this study. We have debated the pros and cons of using the term sterile “inflammation” and we believe that its use will focus investigation on the potential non-microbial etiology of inflammation and preterm birth.

Conclusion

Sterile intra-amniotic inflammation is more common than microbial-associated intra-amniotic inflammation in asymptomatic patients with a sonographic short cervix, and its presence is a risk factor for sPTD. Elucidation of the causes of sterile inflammation and the mechanisms whereby it predisposes to sPTD are important issues. The short- and long-term consequences of sterile intra-amniotic inflammation in the fetus/neonate, and whether such inflammation is amenable to treatment, require investigation.

Declaration of interest

The authors report no conflicts of interest. This research was supported, in part, by the Perinatology Research Branch, Division of Intramural Research, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Department of Health and Human Services (NICHD/NIH); and, in part, with Federal funds from NICHD, NIH under Contract No. HHSN275201300006C.