A new rapid bedside test to diagnose and monitor intra-amniotic inflammation in preterm PROM using transcervically collected fluid

Abstract

Background:

Microbial invasion of the amniotic cavity, present in approximately 50% of patients with preterm prelabor rupture of membranes (PROM), is often associated with intra-amniotic inflammation, a risk factor for a short admission-to-delivery interval, early preterm delivery, and neonatal complications. We previously developed a cervical amniotic fluid collector, a device that allows collection of fluid noninvasively from the cervical canal when there is membrane rupture.

Objective:

This study was designed to determine whether rapid analysis of an interleukin-8 (IL-8) concentration in fluid obtained non-invasively by the cervical amniotic fluid collector can be utilized to assess the risk of intra-amniotic inflammation. We also compared the diagnostic performance of this point-of-care test (POCT) for IL-8 in transcervically obtained fluid to that of a white blood cell (WBC) count determined in amniotic fluid retrieved by amniocentesis.

Study Design:

This was a prospective cohort study performed between October 2011 and April 2017. Fluid was retrieved through both transabdominal amniocentesis and using a cervical amniotic fluid collector within 24 hours of amniocentesis in singleton pregnant patients with preterm PROM (16–35 weeks of gestation). Amniotic fluid obtained via amniocentesis was cultured for aerobic and anaerobic bacteria as well as genital mycoplasmas, and a white blood cell (WBC) count was determined. Intra-amniotic infection was diagnosed when microorganisms were identified by the culture of amniotic fluid. Intra-amniotic inflammation was defined as an elevated amniotic fluid matrix metalloproteinase-8 (MMP-8) concentration (>23 ng/ml) assayed by ELISA. IL-8 in cervical fluid obtained by the collector was measured by the POCT which used a test strip and scanner based on the fluorescence immunochromatographic analysis in 2019. The diagnostic indices, predictive values, and likelihood ratios of the two different tests were calculated.

Results:

1) IL-8 concentrations ≥9.5 ng/ml in cervical fluid, determined by the POCT, was at the knee of the receiver operating characteristic (ROC) curve analysis and had a sensitivity of 98% (56/57, 95% confidence interval [CI] 91–99.96%), specificity of 74% (40/54, 95% CI 60–85%), positive predictive value of 80% (56/70, 95% CI 72–86%), negative predictive value of 98% (40/41, 95% CI 85–99.6%), positive likelihood ratio of 3.79 (95% CI 2.41–5.96), and negative likelihood ratio of 0.02 (95% CI 0.003–0.17) in the identification of intra-amniotic inflammation; a concentration of MMP-8 >23 ng/mL by ELISA had a prevalence of 51% (57/111). 2) Cervical fluid IL-8 concentrations ≥9.5 ng/ml had significantly higher sensitivity than a transabdominally obtained amniotic fluid WBC count in the identification of intra-amniotic inflammation (sensitivity: 98% [95% CI 91–99.96%] vs 84% [95% CI 72–93%]; p<0.05) (specificity: 74% [95% CI 60–85%] vs. 76% [95% CI 62–87%); positive and negative predictive values: 80% [95% CI 72–86%] and 98% [95% CI 85–99.6%] vs. 79% [95% CI 69–86%] and 82% [95% CI 71–89%]) and in the identification of intra-amniotic inflammation/infection (gold standard: positive culture for bacteria or an MMP-8 >23ng/mL; 91% [95% CI 82–97%] vs 75% [95% CI 63–85%]; p<0.05).

Conclusions:

The POCT based on IL-8 determination in fluid retrieved by a cervical amniotic fluid collector was predictive of intra-amniotic inflammation. Therefore, analysis of cervically obtained fluid by such POCT may be used to non-invasively monitor intra-amniotic inflammation in patients with preterm PROM.

Introduction

Preterm prelabor rupture of membranes (PROM) accounts for approximately one-third of preterm births(1) and is a leading cause of perinatal morbidity and mortality.(2–11) Intra-amniotic infection/inflammation is present in 17–75% of patients (12–38), is largely subclinical, and is a major risk factor for impending preterm delivery and adverse neonatal outcome.(21, 23, 26, 37, 39–55)

Currently, amniotic fluid analysis is the most sensitive and specific method to identify intra-amniotic infection/inflammation and predict adverse pregnancy outcome in preterm PROM.(12–17, 21, 39, 56–60) However, amniotic fluid analysis requires an amniocentesis, which is invasive and may not be feasible in cases with severe oligohydramnios.(61, 62) Other methods to assess the likelihood of intra-amniotic infection or inflammation include the determination of inflammatory biomarkers in maternal serum(39, 63–68) (such as C-reactive protein) or cervicovaginal fluid.(24, 68–84) However, these methods have not performed well enough to replace the analysis of amniotic fluid obtained by amniocentesis.(39, 64, 85, 86)

A non-invasive method that would allow serial evaluation for the presence of intra-amniotic inflammation in patients with preterm PROM would be of considerable value to reduce the need for an amniocentesis in the evaluation of the inflammatory status of the amniotic cavity. We previously reported the development of a new device, the “transcervical amniotic fluid collector,” which allows retrieval of fluid from the endocervical canal.(87) This fluid has a similar concentration of alpha-fetoprotein (AFP), human chorionic gonadotrophin (hCG), and prolactin as that obtained by transabdominal amniocentesis and, therefore, may reflect the composition of fluid within the amniotic cavity.(87)

This study was designed to determine whether rapid analysis of an interleukin-8 (IL-8) concentration in fluid obtained non-invasively by the cervical amniotic fluid collector can be utilized to assess the risk of intra-amniotic inflammation. We also compared the diagnostic performance of this point-of-care test (POCT) for IL-8 in transcervically obtained fluid to that of a white blood cell (WBC) count determined in amniotic fluid retrieved by amniocentesis.

Materials and Methods

Study design

This was a prospective cohort study which included patients admitted to Seoul National University Hospital, Seoul National University Bundang Hospital, or Dongguk University Hospital between October 2011 and April 2017 who met the following criteria: 1) preterm PROM with or without preterm labor (≤35 weeks of gestation), 2) singleton gestation, and 3) fluid collected using the cervical amniotic fluid collector device within 24 hours of amniocentesis, which was performed to assess the microbial and inflammatory status of the amniotic cavity. Rupture of membranes was diagnosed by a previous history of watery vaginal discharge and a combination of the following tests: confirming leakage of amniotic fluid from the cervical os, vaginal pooling of amniotic fluid, and a positive nitrazine test through a sterile speculum examination. Antibiotics were administered as soon as possible when the diagnosis of preterm PROM was made. All patients provided written informed consent to donate fluid for research purposes. The Institutional Review Boards of the Seoul National University Hospital, Seoul National University Bundang Hospital, and Dongguk University Hospital approved the collection and use of these samples and information for research purposes. The Seoul National University has a Federal Wide Assurance with the Office for Human Research Protections (OHRP) of the Department of Health and Human Services (DHHS) of the United States.

The transcervical amniotic fluid collector

Fluid was collected transcervically using a device previously described in detail (Figure 1).(87) Using a sterile speculum, the cervical amniotic fluid collector was placed against the external cervical os, and the balloon of the device was inflated. The fluid flowing freely through the collector device without the use of suction was collected and stored in polypropylene tubes at −70°C until assay. The goal was to obtain more than 3 cc of fluid. If the patient preferred or felt discomfort, the device was removed. The median time the device remained in place was 180 min (interquartile range [IQR]: 90–394 min). The median amount of fluid collected was 13.5 mL (IQR range: 6.5–25 mL). The device was placed by a resident, fellow or staff physician.

Figure 1. Transcervical amniotic fluid collector.

The transcervical amniotic fluid collector is shown after ballooning: length 25 cm.

Amniotic fluid analysis

Amniotic fluid was obtained by transabdominal amniocentesis under ultrasonographic guidance and was cultured for aerobic and anaerobic bacteria as well as for genital mycoplasmas (Ureaplasma spp. and Mycoplasma hominis). Intra-amniotic infection was diagnosed when microorganisms were identified by the culture of amniotic fluid. An aliquot of amniotic fluid was examined in a hemocytometer chamber to determine the WBC count. The remaining amniotic fluid was centrifuged and stored in polypropylene tubes at −70°C until assay. The results of culture for bacteria and genital mycoplasmas and WBC count of amniotic fluid obtained by transabdominal amniocentesis were used in the clinical management.

The matrix metalloproteinase-8 (MMP-8) and IL-8 concentrations were measured in the samples of stored fluid (supernatant of centrifuged amniotic fluid obtained by amniocentesis) with a commercially available enzyme-linked immunosorbent assay (R&D Systems, Minneapolis, MN). Intra-amniotic inflammation was defined when MMP-8 concentration was higher than 23 ng/ml based on prior reports.(26,78, 88–101) The sensitivity of the test of MMP-8 was 0.058 ng/mL and that of IL-8 was 0.0075 ng/mL. The intra- and inter-assay coefficients of variation were <10% for both assays.

Point-of-care test for interleukin-8

To measure IL-8 in the sample of fluid obtained by the cervical device, the POCT was performed with a scanner based on fluorescence immunonochromatographic analysis. The assay test strips were prepared and fit into a disposable cartridge. Two monoclonal antibodies to IL-8 (clone #1-1-7-3 and #KA2-16-1) were developed (by BH Yoon), and immobilized on the test and control lines (CervicAF^™ IL-8 Check, Obmed Co., Ltd., Seoul, Korea). A total of 100 μl of fluid, obtained using the cervical amniotic fluid collector, was mixed with 100 ul of detector buffer containing the two IL-8 antibodies; then, 100 ul of this mixture was loaded onto the sample well of the cartridge. After waiting 15 minutes for the immune reaction, the cartridge was inserted, and the intensity of fluorescence was scanned with the scanner for 30 seconds. The IL-8 concentration of fluid was displayed as ng/ml on the scanner screen. The sensitivity of the test was 0.001 ng/ml, and the range of detection was between 0 and 50 ng/ml and both intra- and inter-assay coefficients of variation were <10%. The POCT for IL-8 was performed in stored fluid obtained by the cervical collector in the research laboratory in 2019. We measured IL-8 concentration by the POCT in all fluids (bloody, turbid, etc) obtained by the cervical collector device. In one patient with intra-amniotic inflammation but a negative amniotic fluid culture, a small amount of highly sticky fluid (less than 0.1 cc) was obtained through the cervical collector. As the scanner did not give the result of IL-8 concentration of this fluid, this case was excluded from the analysis.

Diagnosis of acute histologic chorioamnionitis and funisitis

Acute histologic chorioamnionitis was defined as the presence of acute inflammatory changes in the choriodecidua and amnion, respectively; acute funisitis was diagnosed by the presence of neutrophil infiltration into the umbilical vessel walls or Wharton’s jelly using criteria previously published.(47, 57, 88, 102, 103)

Statistical analysis

Proportions were compared using the Fisher’s exact test, and comparisons of continuous variables between groups were performed with Mann-Whitney U test. A modified t-test for correlated samples, as described by Gallen and Gambino,(104) was used to compare sensitivity and specificity. A p-value <0.05 was considered significant.

Results

Characteristics of study population

The study population consisted of 111 pregnant women who met the inclusion criteria. The rate of intra-amniotic inflammation was 51.4% (57/111), and that of a positive amniotic fluid culture was 40.5% (45/111). Microorganisms isolated in the amniotic fluid retrieved by transabdominal amniocentesis consisted of Ureaplasma species, including Ureaplasma urealyticum (n=33), Mycoplasma hominis (n=10), Lactobacillus species (n=4), Enterococcus faecalis (n=2), and one isolate each of Anaerococcus prevotii, Bifidobacterium species, Candida albicans, Escherichia coli, Gram positive bacteria, Gram positive cocci, Gram positive rod, Peptostreptococcus species, Streptococcus species, and Veillonella species. Intra-amniotic inflammation and/or infection was present in 60.4% (67/111) of the women. Table 1 describes characteristics of the study population according to the presence or absence of intra-amniotic inflammation. Patients with intra-amniotic inflammation had a significantly lower gestational age at amniocentesis and at delivery than those without intra-amniotic inflammation (p<0.001 for each).

Table 1.

Clinical characteristics and pregnancy outcomes of study population according to presence or absence of intra-amniotic inflammation

	Intra-amniotic inflammation (−) (N=54)	Intra-amniotic inflammation (+) (N=57)	P value	Adjusted P-valuea
Age, years	32 (30–35)	33 (31–35)	0.088	-
Nulliparity	64.8% (35/54)	56.1% (32/57)	0.438	-
Gestational age at amniocentesis, weeks	32.2 (30.3–33.2)	25.7 (22.7–31.0)	<.001	-
Gestational age at delivery, weeksb	33.5 (31.9–34.3)	28.6 (24.1–32.3)	<.001	-
Status of amniotic fluid obtained by amniocentesisc			0.012
clear	88.9% (49/51)	63.2% (37/55)	0.002
turbid	1.9% (1/51)	12.3% (7/55)	0.062
turbid and bloody	0	3.5% (2/55)	0.496
bloody	1.9% (1/51)	15.8% (9/55)	0.017
Interval between amniocentesis to delivery <7 days b,d,	32.4% (11/34)	56.9% (29/51)	.030	.002
Use of tocolytics	59.3% (32/54)	50.9% (29/57)	0.446	-
Use of antibiotics	100% (54/54)	100% (57/57)	1.000	-
Use of corticosteroid	81.5% (44/54)	70.2% (40/57)	0.189	-
Clinical chorioamnionitis	1.9% (1/54)	1.8% (1/57)	1.000	-
Positive amniotic fluid culture	18.5% (10/54)	61.4% (35/57)	<.001	<.001
Amniotic fluid interleukin-8 concentration (by ELISA), ng/ml	4.9 (2.8–9.5)	51.7 (23.4–73.1)	<0.01
Amniotic fluid white blood cell count, cells/mm3	4 (1–16)	290 (40–1423)	<.001	-
Amniotic fluid white blood cell count ≥ 19 cells/mm3	24.1% (13/54)	84.2% (48/57)	<.001	<.001
Status of fluid obtained by cervical fluid collector			0.012
clear	50% (27/54)	21.1% (12/57)	0.002
turbid	29.6% (16/54)	52.6% (30/57)	0.02
turbid and bloody	11.1% (6/54)	7.0% (4/57)	0.520
bloody	7.4% (4/54)	17.5% (10/57)	0.153
pinkish	1.9% (1/54)	1.8% (1/57)	1.0
Cervical fluid interleukin-8 concentration (by POCT), ng/mL	4.9 (2.8–9.5)	51.7 (23.4–73.1)	<.001
Cervical fluid interleukin-8 ≥ 9.5 ng/mL (by POCT)	25.9% (14/54)	98.2% (56/57)	<.001	<.001
Acute histologic chorioamnionitise	43.5% (10/23)	83.3% (25/30)	.004	.113
Funisitis e	8.7% (2/23)	36.7% (11/30)	.025	.020

POCT, point-of-care test

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

^aadjusted for gestational age at amniocentesis

^bTwo patients were excluded because of unavailable data

^cFive cases were excluded because of unavailable data

^dTwenty-four patients with gestational age at amniocentesis ≥ 33 weeks’ gestation were excluded from the analysis because labor was induced in most patients at 34 weeks’ gestation.

^ePlacental histologic findings were analyzed in 53 women whose interval between sampling and delivery was within 7 days to preserve a meaningful temporal relationship.

Transcervically collected fluid interleukin-8 concentrations determined by POCT

Figure 2 shows the IL-8 concentrations determined by the POCT in fluid obtained with the transcervical collector according to the presence or absence of intra-amniotic infection/inflammation (comparison A: patients with intra-amniotic inflammation [amniotic fluid MMP-8 >23 ng/ml] vs. those without intra-amniotic inflammation [amniotic fluid MMP-8 ≤23ng/ml]; comparison B: patients with a positive amniotic fluid culture for bacteria [intra-amniotic infection] vs. those with a negative amniotic fluid culture; comparison C: patients with intra-amniotic infection and/or inflammation vs. those without intra-amniotic infection and inflammation). Patients in these three groups had a significantly higher median transcervically collected fluid IL-8 concentration than those without each of these specific outcomes (comparison A: median IL-8 concentration [interquartile range], 51.7 ng/mL [23.4–73.1 ng/mL] vs. 4.9 ng/mL [2.8–9.5 ng/mL], p<0.001; comparison B: median IL-8 concentration [interquartile range], 51.4 ng/mL [15.7–69.6 ng/mL] vs. 7.8 ng/mL [3.3–24.1 ng/mL], p<0.001): comparison C: median IL-8 concentration [interquartile range], 48.1 ng/mL [17.0–63.5 ng/mL] vs. 4.5 ng/mL [2.7–7.8 ng/mL], p<0.001) (Figure 2).

Figure 2. Interleukin-8 concentration determined by the point-of-care test of fluid.

Interleukin-8 concentration determined by the point-of-care test of fluid obtained with the transcervical collector device. (A) according to the presence or absence of intra-amniotic inflammation (amniotic fluid MMP-8 >23 ng/ml) (median IL-8 concentration [interquartile range], 51.7 ng/mL [23.4–73.1 ng/mL] vs. 4.9 ng/mL [2.8–9.5 ng/mL], p<0.001); (B) according to the presence or absence of intra-amniotic infection (a positive amniotic fluid culture for bacteria) (median IL-8 concentration [interquartile range], 51.4 ng/mL [15.7–69.6 ng/mL] vs. 7.8 ng/mL [3.3–24.1 ng/mL], p<0.001); and (C) according to the presence or absence of intra-amniotic infection and/or inflammation (median IL-8 concentration [interquartile range], 48.1 ng/mL [17.0–63.5 ng/mL] vs. 4.5 ng/mL [2.7–7.8 ng/mL], p<.001),

After inspecting the receiver operating characteristic (ROC) curve shown in Figure 3, a transcervically collected fluid IL-8 cutoff value of ≥ 9.5 ng/mL was selected for the identification of intra-amniotic inflammation.

Figure 3. Receiver operating characteristic curves.

Receiver operating characteristic curves demonstrating the performance of interleukin-8 concentration determined by the POCT of fluid obtained using the transcervical collector device. (A) in the identification of intra-amniotic inflammation (area under the curve, 0.949 [95% confidence interval, 0.890–0.982]; p<.0.001), and (B) in the identification of intra-amniotic infection/inflammation (area under the curve, 0.914 [95% confidence interval, 0.845–0.958]; p<.001). AUC, area under the curve.

Diagnostic performance of transcervically obtained fluid interleukin-8 determined by POCT and WBC count of amniotic fluid obtained by amniocentesis

Table 2 shows the diagnostic indices and predictive values for the analyte thresholds and for transabdominal amniotic fluid WBC counts using previously reported thresholds (amniotic fluid WBC count ≥ 19 cells/mm³,(88, 105–108) and MMP-8 > 23ng/mL(26, 109)). Transcervically collected fluid IL-8 concentrations ≥9.5 ng/ml determined by the POCT had significantly higher sensitivity than the transabdominally obtained amniotic fluid WBC counts (≥19 cells/mm³) in the identification of intra-amniotic inflammation, as well as in the identification of intra-amniotic infection and/or inflammation (gold standard: positive culture for bacteria or an MMP-8 > 23ng/mL; p<0.05 for each).

Table 2.

The diagnostic indices and predictive values in the identification of intra-amniotic infection/inflammation

Predictors	Sensitivity*	Specificity*	Positive predictive value*	Negative predictive value*	Positive likelihood ratio**	Negative likelihood ratio**
Identification of Intra-amniotic inflammation (defined by AF MMP-8 >23ng/mL)
Cervical fluid IL-8 ≥ 9.5 ng/ml	98.2% (90.9–99.96%) (56/57) a	74.1% (60.3–85.0%) (40/54)	80% (71.8–86.3%) (56/70)	97.6% (85.1–99.6%) (40/41)	3.79 (2.41–5.96)	0.02 (0.003–0.17)
AF WBC count ≥ 19 cells/mm3	84.2% (72.1–92.5%) (48/57) a	75.9% (62.4–86.5%) (41/54)	78.7% (69.4–85.7%) (48/61)	82% (71.1–89.4%) (41/50)	3.50 (2.15–5.69)	0.21 (0.11–0.39)
Identification of Intra-amniotic infection
Cervical fluid IL-8 ≥ 9.5 ng/ml	86.7% (73.2–94.9%) (39/45)	53.0% (40.3–65.4%) (35/66)	55.7% (48.7–62.5%) (39/70)	85.4% (72.8–92.7%) (35/41)	1.85 (1.39–2.44)	0.25 (0.12–0.55)
AF WBC count ≥ 19 cells/mm3	75.6% (60.5–87.1%) (34/45)	59.1% (46.3–71.0%) (39/66)	55.7% (47.4–63.8%) (34/61)	78.0% (67.1–86.0%) (39/66)	1.85 (1.32–2.58)	0.41 (0.24–0.72)
Identification of Intra-amniotic infection and/or intra-amniotic inflammation
Cervical fluid IL-8 ≥ 9.5 ng/ml	91.0% (81.5–96.6%) (61/67)a	79.5% (64.7–90.2%) (35/44)	87.1% (79.0–92.4%) (61/70)	85.4% (72.8–92.7%) (35/41)	4.45 (2.48–8.01)	0.11 (0.05–0.25)
AF WBC count ≥ 19 cells/mm3	74.6% (62.5–84.5%) (50/67) a	75% (59.7–86.8%) (33/44)	82.0% (72.8–88.5%) (50/61)	66% (55.4–75.2%) (33/50)	2.99 (1.76–5.07)	0.34 (0.22–0.53)

^*,Data are given as % (95% confidence interval) (n/N).

^**,Data are given as likelihood ratio (95% confidence interval).

^aP < 0.05 by Galen and Gambino analysis

AF, amniotic fluid; IL-8, interleukin-8; MMP-8, matrix metalloproteinase-8; WBC, white blood cell.

Interleukin-8 determined by ELISA in amniotic fluid obtained by amniocentesis

There was a significant relationship between the transcervically obtained fluid (stored fluid without centrifugation) IL-8 concentration determined by the POCT and the concentration of transabdominally obtained amniotic fluid (stored supernatant after centrifugation) IL-8 determined by ELISA (p<0.001, γ=0.681; Figure 4).

Figure 4. Relationship between transcervical and transabdominal fluids.

There was a significant relationship between the transcervically obtained fluid (stored fluid without centrifugation) interleukin-8 concentration determined by the point-of-care test and the concentration of transabdominally obtained amniotic fluid (stored supernatant after centrifugation) interleukin-8 determined by ELISA (p<.001, γ=0.681). IL, interleukin.

The diagnostic performance of transcervically obtained fluid IL-8 concentration determined by the POCT was also examined in the identification of intra-amniotic inflammation defined by the concentration of trans-abdominally obtained amniotic fluid IL-8. As we were not able to find the cut-off value of transabdominally obtained amniotic fluid IL-8 in the diagnosis of intra-amniotic inflammation in the literature, we performed an ROC curve analysis of amniotic fluid IL-8 in the identification of intra-amniotic inflammation (defined by elevated amniotic fluid MMP-8 >23 ng/ml), found that amniotic fluid IL-8 concentration of 1.8 ng/ml was at the knee of the ROC curve, and selected this as a cut-off value in the diagnosis of intra-amniotic inflammation. Table 3 shows diagnostic indices and predictive values of transcervically obtained fluid IL-8 concentration determined by the POCT in the identification of intra-amniotic inflammation (defined by an elevated concentration of transabdominally obtained amniotic fluid IL-8 >1.8 ng/ml) and/or intra-amniotic infection. There were no statistically significant differences in the diagnostic indices between transcervically collected fluid IL-8 concentrations ≥9.5 ng/ml determined by the POCT and the transabdominally obtained amniotic fluid WBC count (≥19 cells/mm³) in the identification of intra-amniotic inflammation (defined by elevated concentration of transabdominally obtained amniotic fluid IL-8 >1.8 ng/mL).

Table 3.

Diagnostic indices and predictive values in the identification of intra-amniotic inflammation (defined by amniotic fluid interleukin-8 concentration >1.8 ng/ml) and/or intra-amniotic infection

Predictors	Sensitivity*	Specificity*	Positive predictive value*	Negative predictive value*	Positive likelihood ratio**	Negative likelihood ratio**
Identification of Intra-amniotic inflammation (AF IL-8 >1.8 ng/mL)
Cervical fluid IL-8 ≥ 9.5 ng/ml	96.3% (87.3–99.5%) (52/54)	68.4% (54.8–80.1%) (39/57)	74.3% (66.3–80.9%) (52/70)	95.1% (83.2–98.7%) (39/41)	3.05 (2.07–4.49)	0.05 (0.01–0.21)
AF WBC count ≥ 19 cells/mm3	87.0% (75.1–94.6%) (47/54)	75.4% (62.2–85.9%) (43/57)	77.0% (67.8–84.3%) (47/61)	86.0% (75.2–92.6%) (43/50)	3.544 (2.22–5.65)	0.17 (0.09–0.35)
Identification of Intra-amniotic infection
Cervical fluid IL-8 ≥ 9.5 ng/ml	86.7% (73.2–94.9%) (39/45)	53.0% (40.3–65.4%) (35/66)	55.7% (48.7–62.5%) (39/70)	85.4% (72.8–92.7%) (35/41)	1.85 (1.39–2.44)	0.25 (0.12–0.55)
AF WBC count ≥ 19 cells/mm3	75.6% (60.5–87.1%) (34/45)	59.1% (46.3–71.0%) (39/66)	55.7% (47.4–63.8%) (34/61)	78.0% (67.1–86.0%) (39/50)	1.85 (1.32–2.58)	0.41 (0.24–0.72)
Identification of Intra-amniotic infection and/or intra-amniotic inflammation
Cervical fluid IL-8 ≥ 9.5 ng/ml	89.1% (78.8–95.5%) (57/64)	72.3% (57.4–84.4%) (34/47)	81.4% (73.3–87.5%) (57/70)	82.9% (70.3–90.9%) (34/41)	3.22 (2.01–5.15)	0.15 (0.07–0.31)
AF WBC count ≥ 19 cells/mm3	76.6% (64.3–86.2%) (49/64)	74.5% (59.7–86.1%) (35/47)	80.3% (71.1–87.1%) (49/61)	70.0% (59.2–78.9%) (35/50)	3.00 (1.81–4.98)	0.32 (0.20–0.51)

^*,Data are given as % (95% confidence interval) (n/N).

^**,Data are given as likelihood ratio (95% confidence interval)

AF, amniotic fluid; IL-8, interleukin-8; WBC, white blood cell.

Acute histologic chorioamnionitis and transcervically collected fluid IL-8 determined by POCT

The relationships between acute histologic chorioamnionitis and transcervically collected fluid IL-8 concentration determined by POCT were examined in the subset of 53 patients who delivered within 7 days after cervical fluid sampling. This interval was chosen to preserve a meaningful temporal relationship between the results of transcervically collected fluid IL-8 concentration and the results of placental pathologic examination. The median transcervically collected fluid IL-8 concentration was significantly higher in patients whose placentas demonstrated acute histologic chorioamnionitis than in those without acute histologic chorioamnionitis (median 49.1 ng/mL, interquartile range 9.6–62.0 ng/mL vs median 6.4 ng/mL, interquartile range 3.5–16.1 ng/mL, p=.001).

Amniotic fluid MMP-8 and transcervically obtained fluid interleukin-8 determined by POCT

To identify intraamniotic infection (positive amniotic fluid culture), acute histologic chorioamnionitis, and funisitis, we compared ROC curves of transcervically obtained fluid IL-8 concentration determined by POCT and amniotic fluid MMP-8 concentration retrieved by trans-abdominal amniocentesis determined by ELISA in the identification of intra-amniotic infection (positive amniotic fluid culture), acute histologic chorioamnionitis, and funisitis. There were no significant differences in the area under the curve in the identification of intra-amniotic infection (0.763 vs. 0.735), acute histologic chorioamnionitis (0.786 vs. 0.784) and funisitis (0.712 vs. 0.615) (p>0.05 for each) between transcervically collected IL-8 and amniotic fluid MMP-8.

Discussion

Principal findings of the study

The IL-8 concentration determined by the POCT of fluid retrieved noninvasively using the cervical collector was predictive of intra-amniotic inflammation, which was defined by a pre-specified “gold standard” (analysis of amniotic fluid with an MMP-8 ELISA). Moreover, the POCT analysis of transcervically obtained fluid IL-8 was superior to the transabdominal amniotic fluid WBC count.

Intra-amniotic infection/inflammation and the management of preterm PROM

The management of preterm PROM remains a major challenge in clinical obstetrics. Key decisions rest on balancing the risks of preterm birth, intra-amniotic infection predisposing to neonatal sepsis, (110, 111) abruptio placentae, (112–117) and umbilical cord accidents, (118, 119)– all of which may occur during expectant management.

Delivery is universally agreed upon as the strategy of choice for patients with clinical chorioamnionitis.(120–122) However, the optimal management of patients without evidence of intra-amniotic infection is unclear. Transabdominal amniocentesis can provide additional information in these patients by allowing the detection of intra-amniotic infection and intra-amniotic inflammation and by assessing the state of fetal lung maturity.(12, 123) However, amniocentesis is an invasive procedure and can be difficult to perform in cases of oligohydramnios (although contemporary series suggest a success rate of over 80% in patients with preterm PROM).(16, 123) In these cases, assessing the risk of intra-amniotic infection is even more important, given that a reduced volume of amniotic fluid is a risk factor for intra-amniotic infection.(95, 124–127) This diagnostic challenge is also present when patients are evaluated after admission to the antepartum ward. Indeed, a recent study reported that two-thirds of those who develop clinical chorioamnionitis do not have demonstrable intra-amniotic infection by culture.(99)

Analysis of transcervically obtained fluid to assess the risk of intra-amniotic inflammation

An alternative approach to amniocentesis is desirable to avoid an invasive procedure and to facilitate serial assessment of patients admitted to the antepartum ward. Investigators have explored the value of several biomarkers in maternal blood (63–66, 128–130) and cervicovaginal fluid (e.g. IL-6, IL-8, C-reactive protein, and neutrophil defensins).(24, 69, 80, 131, 132) Vaginal fluid obtained from the posterior fornix has also been used to assess the inflammatory status of the amniotic cavity. A high C-reactive protein concentration and low glucose concentration in amniotic fluid from the posterior fornix are somewhat predictive of intra-amniotic infection/inflammation in preterm PROM, as determined by amniocentesis.(73, 133) However, one limitation of cervicovaginal fluid assessment is that it is sometimes unavailable, and buffers used to preserve samples may dilute the scant volume of fluid in the posterior fornix. Kacerovsky et al (134) reported that IL-6 determination in vaginal fluid is related to the presence of microbial invasion of the amniotic cavity (MIAC), based on PCR for Ureaplasma spp., Mycoplasma hominis, or a positive culture. Of interest, IL-8 was associated with MIAC and acute histologic chorioamnionitis, but this difference was not statistically significant. In a separate study, Musilova et al. found that a POCT, based on the determination of IL-6 in fluid obtained from the vaginal fornix, was strongly correlated with that obtained by transabdominal amniocentesis; yet, the concentration of IL-6 in vaginal fluid was three-fold higher.(135) The authors considered that the vaginal fluid retrieved from the posterior fornix reflected a combination of amniotic fluid, vaginal fluid, and cervical material.

Fluid obtained by a transcervical collector: a simple and new approach to assess the inflammatory state of the intra-amniotic cavity

In the current study, we collected fluid transcervically using a collector device. To avoid possible contamination with vaginal fluid, the device was positioned at the level of the external cervical os. We have previously shown that the concentrations of proteins in fluid collected by the transcervical fluid collector correlated well with those in amniotic fluid collected by amniocentesis.(87) Herein, we found that transcervically collected amniotic fluid IL-8 concentrations measured by the rapid bedside test had better diagnostic indices than the amniotic fluid WBC count obtained transabdominally in the identification of intra-amniotic infection and/or inflammation (determined by culture or MMP-8).

Therefore, based on our results, fluid obtained using the collector device may be used as a screening test to determine the likelihood of intra-amniotic inflammation. If inflammation is likely to be present, this can be confirmed by amniocentesis to assess the state of inflammation and the presence/type of microorganisms with a combination of cultivation and molecular microbiologic techniques.(13–16, 28, 36, 44, 136–140)

Given the diagnostic accuracy of transcervically collected fluid IL-8 concentrations, future work to evaluate the potential value of serial measurements is necessary. Moreover, concentrations of other inflammatory cytokines and mediators in cervical fluid may provide complementary information to that provided by IL-8.

In this study, we used MMP-8 concentrations to diagnose intra-amniotic inflammation based on a large body of evidence showing a relationship between the concentrations of this protein and adverse pregnancy outcome, including the likelihood of acute histologic chorioamnionitis and funisitis.(26, 90–92, 97, 99, 127, 141, 142) However, when examining fluid collected transcervically, we chose IL-8, given that this cytokine has been be informative in our studies. Similarly, a group of investigators(143) found that IL-8 was informative in cervical fluid in patients with intact membranes undergoing cerclage (a different clinical condition from PROM). Nonetheless, IL-8 is dramatically elevated in cases of intra-amniotic inflammation.(144)

Clinical Implications

The management of preterm PROM largely consists of admission to the hospital,(121) the administration of antibiotics(145–152) and corticosteroids,(153) and the timely induction of labor,(121, 153) if spontaneous labor does not occur. Although the clinical significance of intra-amniotic inflammation is well established,(26, 36) clinical utilization depends upon the availability of ultrasound and skilled operators to perform amniocentesis in patients with reduced volume of amniotic fluid. This obstacle may be overcome by assessing the inflammatory status of the amniotic cavity using fluid collected from the cervical os or from accumulated fluid in the vagina.(134, 154, 155)

Therefore, when a patient presents and is admitted with preterm PROM, an initial assessment of the inflammatory state of the amniotic cavity can be performed using the collector device, and antibiotics started. Serial sampling would allow monitoring of the intra-amniotic inflammatory response. It is noteworthy that because the negative predictive value of the IL-8 concentration determined by the POCT in this study is 98% [95% CI 85.1–99.6%] (in a population with a prevalence of intra-amniotic inflammation of 51%) and the likelihood ratio of a negative test is 0.02 (95% CI 0.003–0.17), a low concentration of IL-8 in transcervically collected fluid is reassuring to exclude intra-amniotic inflammation.

In contrast, an elevated IL-8 concentration in fluid collected transcervically would require further diagnostic workup (e.g. amniocentesis) to determine whether intra-amniotic infection and/or intra-amniotic inflammation is present. Microbiologic studies will require amniotic fluid collected by amniocentesis because fluid collected transcervically is prone to contamination with local microorganisms. Only amniotic fluid obtained by amniocentesis allows the precise identification of the microorganisms and an appropriate antibiotic regimen.

Determining whether there is MIAC or intra-amniotic inflammation is important, given our recent report that a combination of antibiotics (ceftriaxone, clarithromycin, and metronidazole) is effective in eradicating intra-amniotic infection and intra-amniotic inflammation, prolonging the latency period, and reducing the frequency of funisitis and cerebral palsy.(60, 156, 157) Similar results have been reported in women with cervical insufficiency and preterm labor with intact membranes.(60, 107, 108)

Research Implications

The noninvasive assessment of the inflammatory status of the amniotic cavity in patients with preterm PROM creates opportunities for personalized treatment of the patient with this frequent complication of pregnancy. Thus far, the management of preterm PROM is based on data derived from population studies; however, nearly one-half of the patients with this condition do not have evidence of MIAC or intra-amniotic inflammation.(26, 36) It is unclear if these patients could benefit from treatment with antibiotics, the current standard of care. Moreover, it is well-established that the frequency of MIAC increases during the latency period, and can increase up to 75% at the time of onset of preterm labor.(16) Further studies are required to determine if antibiotics should be administered only to patients with intra-amniotic inflammation at the time of admission or when they develop intra-amniotic inflammation during the course of expectant management.

Strengths and Limitations

This study demonstrates that a noninvasive method and the POCT can be used to assess the likelihood of intra-amniotic inflammation. Such an approach requires a device that is simple to use, safe, and acceptable to patients. The selection of a cytokine (IL-8) as a biomarker is based on extensive studies of amniotic fluid in the context of preterm labor,(77, 158–160) preterm PROM,(25, 43, 58, 161–163) labor at term,(164–166) and other complications of pregnancy.(41, 43, 45) Another strength of this study is that the results of transcervically collected fluid were not available for patient management; therefore, this information did not influence clinical management.

The third strength of this study is that we used a definition of intra-amniotic inflammation based on the concentration of MMP-8 in amniotic fluid retrieved by transabdominal amniocentesis, an independent criterion from the biomarker selected for the POCT (i.e. IL-8). There is a substantial body of evidence indicating that elevated amniotic fluid MMP-8 concentrations are associated with adverse pregnancy outcome,(26, 89, 91–93, 109, 167, 168) neonatal complications,(26, 90, 92, 93) and histologic chorioamnionitis.(57, 97)

A limitation of this study is that the cutoff chosen for IL-8 requires replication in an independent set of patients. It is important to note that the assay for IL-8 in amniotic fluid retrieved by transabdominal amniocentesis uses a different antibody pair than that of the POCT. The former assay is commercially available for research purposes. The latter assay was specifically developed for this clinical indication.

Conclusion

A bedside analysis of IL-8 concentration in fluid obtained using a transcervical fluid collector is an excellent indicator of the status of intra-amniotic inflammation. The high negative predictive value of this approach (and the low likelihood ratio of a negative result) would allow identification of patients who may not require amniocentesis for the assessment of the inflammatory state in the amniotic cavity, thus reducing the need for invasive procedures. Moreover, the noninvasive nature of the cervical collector may facilitate serial studies of patients with preterm PROM, as there is evidence that the frequency of intra-amniotic infection and intra-amniotic inflammation increases with the duration of the latency period.(16, 169) Such a device can also be used in patients with PROM at term to identify those with intra-amniotic inflammation in whom an induction of labor may be desirable.

Condensation

In patients with preterm PROM, amniotic fluid can be collected using a transcervical device and the status of intra-amniotic inflammation can be assessed using a rapid point-of-care test, which measures IL-8 concentration in this fluid. This approach allows non-invasive diagnosis and monitoring of intra-amniotic inflammation.

AJOG-at-a-Glance

Why was the study performed?

To determine if a point-of-care test based on the determination of IL-8 concentration in amniotic fluid obtained noninvasively by a cervical collector can identify intra-amniotic inflammation in patients with preterm PROM.

Key findings:

• An elevated cervical fluid IL-8 concentration (≥ 9.5ng/ml) determined with a point-of-care test had a sensitivity of 98% (95% CI 91–99.96%), specificity of 74% (95% CI 60–85%), negative predictive value of 98% (95% CI 85–99.6%), and a negative likelihood ratio of 0.02 (95% CI 0.003–0.17) in the identification of intra-amniotic inflammation (prevalence 51%).

• This test had a significantly greater sensitivity than amniotic fluid white blood cell count based on analysis of amniotic fluid retrieved by amniocentesis.

• Collecting amniotic fluid non-invasively and the use of IL-8 concentrations would reduce the need for invasive procedures and allow accurate assessment of intra-amniotic inflammation, which is a risk factor for adverse pregnancy outcome.

What does the study add to what is already known?

We report for the first time that non-invasive assessment of intra-amniotic inflammation in patients with preterm PROM with the combined use of a device to collect amniotic fluid through the endocervical canal coupled with a point-of-care test to measure IL-8 in this fluid. This approach would make it possible to diagnose and monitor intra-amniotic inflammation without amniocentesis.