Antibiotic administration can eradicate intra-amniotic infection or inflammation in a subset of patients with preterm labor and intact membranes

Abstract

Background:

Intra-amniotic infection is present in 10% of patients with an episode of preterm labor, and is a risk factor for impending preterm delivery and neonatal morbidity/mortality. Intra-amniotic inflammation is often associated with intra-amniotic infection, but sometimes is present in the absence of detectable microorganisms. Antibiotic treatment of intra-amniotic infection has traditionally been considered to be ineffective. Intra-amniotic inflammation without microorganisms has a similar prognosis to intra-amniotic infection.

Objective:

To determine whether antibiotics can eradicate intra-amniotic infection or intra-amniotic inflammation in patients with preterm labor and intact membranes.

Study design:

The study population consisted of women who met the following criteria: 1) singleton gestation between 20–34 weeks; 2) preterm labor and intact membranes; 3) transabdominal amniocentesis performed for the evaluation of the microbiologic/inflammatory status of the amniotic cavity; 4) intra-amniotic infection and/or inflammation; and 5) received antibiotic treatment which consisted of ceftriaxone, clarithromycin, and metronidazole. Follow-up amniocentesis was performed in a subset of patients. Amniotic fluid was cultured for aerobic and anaerobic bacteria and genital mycoplasmas, and polymerase chain reaction (PCR) was performed for Ureaplasma spp. Intra-amniotic infection was defined as a positive amniotic fluid culture or positive PCR, and intra-amniotic inflammation was suspected when there was an elevated amniotic fluid white blood cell count or a positive rapid test for matrix metalloproteinase-8. For this study, the final diagnosis of intra-amniotic inflammation was made by measuring the interleukin-6 concentration in stored amniotic fluid (>2.6 ng/mL). These results were not available to managing clinicians. Treatment success was defined as eradication of intra-amniotic infection and/or inflammation or delivery ≥37 weeks.

Results:

1) Of 62 patients with intra-amniotic infection and/or inflammation, 50 received the antibiotic regimen. Of those, 29 were undelivered for ≥ 7 days and 19 underwent follow-up amniocenteses; 2) microorganisms were identified by culture or PCR of amniotic fluid obtained at admission in 21% (4/19) of patients who had follow-up amniocentesis, and were eradicated in 3 out of 4; 3) resolution of intra-amniotic inflammation was confirmed in 79% (15/19) of patients, and one other patient delivered at term, although the resolution of intra-amniotic inflammation could not be confirmed because further follow-up amniocentesis was not performed; thus, resolution of intra-amniotic inflammation/infection or term delivery (treatment success) occurred in 84% (16/19) of patients who had a follow-up amniocentesis; 4) treatment success occurred in 32% (16/50) of patients with intra-amniotic infection/inflammation who received antibiotics; and 5) the median amniocentesis-to-delivery interval was significantly longer among women who received the combination of antibiotics than among those who did not (11.4 days versus 3.1 days: p=0.04).

Conclusion:

Eradication of intra-amniotic infection/inflammation occurred in 79% of patients with preterm labor, intact membranes, and intra-amniotic infection/inflammation. Treatment success occurred in 84% of patients who underwent follow-up amniocentesis and in 32% of women who received the antibiotic regimen.

Introduction

Preterm labor is a syndrome caused by multiple pathologic processes.¹ The following mechanisms of disease have been implicated: intra-amniotic infection,^2–25 “sterile” intra-amniotic inflammation,^26–39 uterine overdistention,^40,41 maternal anti-fetal rejection,^42–46 decidual senescence,^47–50 and possibly other mechanisms that are yet to be identified.

One of every ten patients with preterm labor and intact membranes will have intra-amniotic infection^{2–7,12,16,17,22–25,33,35,51} which is largely subclinical,^{2,6,22,26,33,38,39,52,53} and these patients are at increased risk for early preterm delivery,^{3,6–8,22,33,53} neonatal complications,^{6,8,21,26,33,54–66} and maternal morbidity (such as acute pulmonary edema, when treated with tocolytics and steroids^67–69) or maternal sepsis.⁷⁰ Similar risks occur in patients with preterm PROM and intra-amniotic infection.^{4,13,17,71,72}

Given the frequency and importance of intra-amniotic infection in the pathogenesis of preterm labor with intact membranes, several randomized clinical trials have tested the efficacy and safety of antibiotic administration.^73–76 Despite initial enthusiasm,^10,73,75,77 subsequent trials have not shown beneficial effects,^74,78–83 and currently, antibiotic administration is restricted to patients with an episode of premature labor who are carriers of group B streptococcus (GBS)^84,85 or have unknown GBS status⁸⁶ to prevent vertical transmission and neonatal sepsis.^87,88

Intra-amniotic inflammation, defined as an elevated concentration of interleukin-6 or matrix metalloproteinase-8 (MMP-8) in amniotic fluid in the absence of demonstrable microorganisms detected with culture or molecular methods (“sterile” intra-amniotic inflammation), has also been associated with adverse pregnancy outcomes, including acute histologic chorioamnionitis and funisitis.^{25–28,33,71,89} Activation of the inflammasome has been implicated in the mechanisms responsible for preterm labor induced by “sterile” intra-amniotic inflammation.^{31,32,90–92}

Important advances have been made in the identification of patients at risk of spontaneous preterm delivery by assessing cervical length in the midtrimester,^93–103 as well as in the treatment of patients with a sonographic short cervix with vaginal progesterone.^104–115 However, the optimal treatment of patients with an episode of preterm labor, intact membranes, and intra-amniotic infection or intra-amniotic inflammation has not been determined. Previous reports demonstrated the eradication of microorganisms in the amniotic cavity of patients with a short cervix^116,117 and preterm PROM.^27,118 A recent report suggests that a subset of patients with preterm labor and intra-amniotic infection may benefit from antibiotic administration.¹¹⁹

We have recently reported that the antibiotic treatment of patients with preterm PROM can reduce the rate of intra-amniotic infection and intra-amniotic inflammation, as well as funisitis and the fetal systemic inflammatory response, using a combination of antibiotics (ceftriaxone, clarithromycin, and metronidazole) which target microorganisms frequently isolated from the amniotic cavity in these cases.^118,120

The purpose of this study was to determine whether antibiotics could eradicate intra-amniotic infection or intra-amniotic inflammation without demonstrable microorganisms in patients with preterm labor and intact membranes.

Materials and Methods

Study design

This is a retrospective case series study of pregnant women admitted to Seoul National University Hospital between January 2004 and March 2014, who met the following criteria; 1) singleton gestations between 20–34 weeks; 2) preterm labor and intact amniotic membranes determined by sterile speculum examination; 3) transabdominal amniocentesis performed for the evaluation of the microbiologic and inflammatory status of the amniotic cavity; 4) positive amniotic fluid culture or intra-amniotic inflammation; and 5) antibiotic treatment (regimen consisted of ceftriaxone, clarithromycin and metronidazole). Follow-up amniocentesis was performed in a subset of patients at the discretion of the managing physician.

At the Seoul National University Hospital, a transabdominal amniocentesis is routinely offered to all patients admitted with the diagnosis of preterm labor to assess the microbiologic status of the amniotic cavity and fetal lung maturity. Retrieval of amniotic fluid was performed after written informed consent was obtained. Preterm labor was diagnosed as the presence of regular uterine contractions (four or more contractions in 20 minutes or eight or more in 60 minutes). The Institutional Review Board of the Seoul National University Hospital approved the collection and use of these samples and clinical information for research purposes. The Seoul National University has a Federal Wide Assurance with the Office for Human Research Protection (OHRP) of the Department of Health and Human Services (DHHS) of the United States.

Amniotic fluid analysis

Amniotic fluid was cultured for aerobic and anaerobic bacteria, as well as genital mycoplasmas. Beginning in 2007, samples were also assayed for Ureaplasma spp. using polymerase chain reaction (PCR) with specific primers using methods previously described⁸. An aliquot of amniotic fluid was examined in a hemocytometer chamber to determine the white blood cell count.^33,121 In a subset of patients, MMP-8 concentration in amniotic fluid was measured using a commercially-available enzyme-linked immunosorbent assay (ELISA) (Amersham Pharmcia Biotech, Inc., Bucks, UK) and the results were available to clinicians. Intra-amniotic inflammation was suspected when the concentration of MMP-8 in the amniotic fluid was higher than 23 ng/mL, as previously reported.^{71,89,122–125}

Between March 2005 and December 2010, a rapid MMP-8 bedside test (MMP-8 PTD Check test, SK Pharma Co, Ltd, Kyunggi-do, Korea) was performed and used in patient management. Details of the MMP-8 rapid test have been previously described.^{53,60,126,127} Amniotic fluid not used for diagnostic tests was centrifuged at 800g and stored at −80C.

Intra-amniotic infection was defined as a positive amniotic fluid culture or positive PCR for Ureaplasma spp. For the purposes of this study, a definitive diagnosis of intra-amniotic inflammation was made when the interleukin-6 concentration of stored amniotic fluid was higher than 2.6 ng/mL.³³ The amniotic fluid interleukin-6 concentration was measured with a commercially available enzyme-linked immunoassay (R&D Systems, Minneapolis, MN) in 2017 and 2018. The sensitivity of the assay was 0.7 pg/mL. The intra- and inter-assay coefficients of variation were <10%. These results were not available to managing clinicians.

Clinical management

Intra-amniotic inflammation was suspected when there was an elevated amniotic fluid white blood cell count (defined as ≥ 19 cells/mm),¹²² a positive MMP-8 rapid test result,^53,126,127 or an elevated concentration of amniotic fluid MMP-8 (>23 ng/mL) measured by ELISA.^71,89,122 Suspicion of intra-amniotic inflammation, isolation of microorganisms by amniotic fluid culture, or the detection of Ureaplasma nucleic acids was an indication for the administration of antibiotics. We used a combination of antimicrobial agents previously prescribed in the management of patients with preterm PROM,^118,120 including ceftriaxone 1g (intravenous) every 24 hours, clarithromycin 500mg (oral) every 12 hours, and metronidazole 500mg (intravenous) every 8 hours. Metronidazole was administered for a maximum of 4 weeks. A follow-up amniocentesis was offered to monitor the microbiologic and inflammatory status of amniotic cavity and fetal lung maturity. The use, discontinuation, or change of antibiotic regimen or tocolytics, or interval to follow-up amniocentesis, were left to the discretion of the treating clinicians because there was no uniformity among attending physicians about these issues. Tocolytics used were ritodrine, magnesium sulphate, or atosiban. Nonsteroidal anti-inflammatory agents, such as indomethacin, were not used as tocolytic agents. Group B streptococcus (GBS) screening and intrapartum treatment are not used in our institution because neonatal GBS sepsis is extremely rare in our patient population.^128,129

Definition of treatment success in this study

Treatment success was defined as: a) eradication of intra-amniotic infection or intra-amniotic inflammation; or b) delivery at or after 37 weeks of gestation.

Diagnosis of acute histologic chorioamnionitis and clinical chorioamnionitis

Acute histologic chorioamnionitis was diagnosed in the presence of acute inflammatory changes in tissue samples including amnion and chorion-decidua.¹³⁰ Funisitis was diagnosed in the presence of neutrophil infiltration into the umbilical vessel walls or Wharton’s jelly.^{71,89,122,131-134}

Clinical chorioamnionitis was diagnosed by the presence of maternal fever (temperature >37.8°C) accompanied by two or more of the following criteria: 1) maternal tachycardia (heart rate >100 beats/min); 2) uterine tenderness; 3) foul-smelling amniotic fluid; 4) fetal tachycardia (heart rate >160 beats/min); and 5) maternal leukocytosis (leukocyte count >15,000 cells/mm³).¹³⁵ The limitations of these criteria in the identification of intra-amniotic infection have been recently described.^{118,125,136–138} The criteria for the diagnosis of neonatal morbidity can be found in Supplementary Material S1.

Statistical analysis

Continuous variables were compared between two groups with the Mann-Whitney U test. Proportions were compared with the Fisher’s exact test. The amniocentesis-to-delivery interval was compared by using the generalized Wilcoxon test for survival analysis. A p-value <0.05 was considered statistically significant. Statistical analyses were performed using SPSS (Version 22; SPSS Inc., Chicago, IL, USA).

Results

Characteristics of study population

Figure 1 shows a flow diagram of patients included in this study. Sixty-two patients with intra-amniotic infection and/or inflammation were identified. A positive amniotic fluid culture was present in 11 patients; Ureaplasma spp. was detected by PCR method in 8 patients; and intra-amniotic inflammation was identified in 51 patients with a negative amniotic fluid culture for microorganisms. Bacteria identified by culture included Ureaplasma urealyticum (n=9), Mycoplasma hominis (n=2), and one isolate each of Streptococcus anginosus and Gardnerella vaginalis.

Figure 1.

Flow diagram of the study population

Of 62 patients with intra-amniotic infection and/or inflammation, 50 received the combination of ceftriaxone, clarithromycin, and metronidazole. The remaining 12 patients did not receive this antibiotic regimen (11 patients did not receive any antibiotics; one received an alternative regimen, consisting of ceftriaxone, azithromycin, and metronidazole). Of the 11 patients who did not receive any antibiotics, one had a positive amniotic fluid culture for Ureaplasma urealyticum, and antibiotics were not administered because of rapid progression of preterm labor to delivery.

The lack of antibiotic administration in the other 10 patients was because: 1) intra-amniotic infection/inflammation was not suspected because the patients had a low amniotic fluid white blood cell count when the MMP-8 rapid test was not available (n=4); however, intra-amniotic inflammation was diagnosed by an elevated interleukin-6 retrospectively; 2) the managing clinician preferred to rely on the results of the amniotic fluid white blood cell count rather than on those of the rapid MMP-8 test (n=2); 3) rapid progression of labor (n=2); 4) declined antibiotic treatment (n=1); and 5) transfer to another hospital because of unavailability of a neonatal intensive care bed (n=1).

Table 1 compares the characteristics and outcomes of patients who received the antibiotic regimen with those of patients who did not. There were no significant differences between the study groups in maternal age, cerclage use, gestational age at amniocentesis, interleukin-6 concentration, frequency of a positive amniotic fluid culture, use of tocolytics and antenatal corticosteroids, delivery within 7 and 14 days of amniocentesis, delivery <30, <34 and ≥37 weeks, clinical and acute histologic chorioamnionitis, and funisitis (p>0.1 for each). Patients who received the antibiotic regimen had a significantly higher median amniotic fluid white blood cell count (79 cells/mm³ vs. 3 cells/mm³), longer median amniocentesis-to-delivery interval (11.4 days vs. 3.1 days) and lower rate of delivery within 4 weeks of amniocentesis (58% vs. 91.7%) than those in whom the antibiotic regimen was not used (p<0.05 for each).

Table 1.

Clinical characteristics and outcomes of patients who did vs. did not use the regimen of antibiotics consisting of ceftriaxone, clarithromycin, and metronidazole

	Use of ceftriaxone, clarithromycin, and metronidazole (n=50)	No antibiotics or use of other antibiotics (n=12)	p-value
Maternal age (years)	31 (29–34)	34 (31–36)	0.12
Nulliparity (%)	62.0% (31/50)	25.0% (3/12)	0.027
Cerclage before the onset of preterm labor	12.0% (6/50)	8.3% (1/12)	0.99
Cerclage after the onset of preterm labor and preterm labor stopped	4.0% (2/50)	8.3% (1/12)	0.48
Initial amniocentesis
Gestational age at amniocentesis (weeks)	25.4 (22.1–27.5)	25.7 (22.6–28.6)	0.63
Positive amniotic fluid culture (%)	20.0% (10/50)	9.1% (1/12)	0.68
Positive amniotic fluid PCR for Ureaplasma spp.	21.2% (7/33)	11.1% (1/9)	0.66
Amniotic fluid WBC count (cells/mm3)	79 (2–860)	3 (0–65)	0.048
Amniotic fluid WBC count (≥ 19 cells/mm3)	58.3% (28/48)	25.0% (3/12)	0.054
Amniotic fluid interleukin-6 (ng/mL)	18.2 (4.1–43.0)	7.8 (3.2–16.9)	0.12
Amniotic fluid interleukin-6 (>2.6 ng/mL)	100% (49/49)	100% (12/12)	>0.99
Cervical dilatation > 3cm (%)	10.0% (5/50)	16.7% (2/12)	0.61
Use of tocolytics (%)	98.0% (49/50)	91.7% (11/12)	0.35
Antenatal corticosteroids administration (%)	62.0% (31/50)	58.3% (7/12)	>0.99
Gestational age at delivery (weeks)	28.9 (25.5–33.9)	27.3 (23.4–31.7)	0.29
Interval between amniocentesis to delivery (days)	11.4 (2.8–57.0)	3.1 (0.3–17.8)	0.04 b
Delivery within 7 days of amniocentesis	42.0% (21/50)	58.3% (7/12)	0.35
Delivery within 14 days of amniocentesis	52.0% (26/50)	67% (8/12)	0.52
Delivery within 4 weeks of amniocentesis	58.0% (29/50)	91.7% (11/12)	0.042
Delivery before 30 weeks a	57.4% (27/47)	81.8% (9/11)	0.18
Delivery before 34 weeks	76.0% (38/50)	91.7% (11/12)	0.43
Delivery at term (>=37 weeks)	8.0% (4/50)	8.3% (1/12)	>0.99
Clinical chorioamnionitis	12.0% (6/50)	0% (0/12)	0.59
Acute histologic chorioamnionitis	69.2% (27/39)	88.9% (8/9)	0.41
Funisitis	30.8% (12/39)	22.2% (2/9)	>0.99

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

PCR, polymerase chain reaction; WBC, white blood cell.

^aCases who underwent amniocentesis at or beyond 30 weeks were excluded from the analysis.

^bThe amniocentesis-to-delivery interval was compared by using the generalized Wilcoxon test for survival analysis.

Of 50 patients treated with the antibiotic regimen, 29 remained undelivered for at least one week (figure 1). Microorganisms identified in the amniotic fluid of 29 patients undelivered for at least one week included Ureaplasma urealyticum (n=4) and Mycoplasma hominis (n=1). One patient had a mixed infection of Ureaplasma urealyticum and Mycoplasma hominis. Microorganisms identified in the amniotic fluid of 21 patients delivered within 7 days of amniocentesis included Ureaplasma urealyticum (n=4), and one isolate each of Mycoplasma hominis, Streptococcus anginosus and Gardnerella vaginalis. One patient had a mixed infection of Ureaplasma urealyticum and Mycoplasma hominis.

Table 2 compares the characteristics and outcomes of patients delivered within seven days of amniocentesis and those who were undelivered for at least seven days. There were no significant differences in the median gestational age at amniocentesis and the frequency of a positive amniotic fluid culture for microorganisms between the two groups (p>0.1 for each). Patients who remained undelivered for at least one week had a significantly lower median concentration of amniotic fluid interleukin-6 and white blood cell count than those who delivered before one week (p<0.005 for both).

Table 2.

Characteristics and outcomes of patients delivered within 7 days of amniocentesis and those who were undelivered for at least 7 days

	Delivery before 1 week (n=21)	Undelivered for ≥ 1 week (n=29)	p-value
Maternal age (years)	33 (30–36)	30 (28–33)	0.07
Nulliparity (%)	47.6% (10/21)	72.4% (21/29)	0.09
Initial amniocentesis
Gestational age at amniocentesis (weeks)	26.4 (22.6–28.4)	24.3 (21.9–26.9)	0.13
Positive amniotic fluid culture (%)	28.6% (6/21)	13.8% (4/29)	0.29
Positive amniotic fluid PCR for Ureaplasma spp.	14.3% (2/14)	26.3% (5/19)	0.67
Amniotic fluid WBC count (cells/mm3)	725 (94->1000)	5 (1–100)	0.002
Amniotic fluid WBC count (≥ 19 cells/mm3)	81.0% (17/21)	40.7% (11/27)	0.008
Amniotic fluid interleukin-6 (ng/mL)	28.2 (14.0–46.5)	10.3 (3.4–21.8)	0.001
Amniotic fluid interleukin-6 (>2.6 ng/mL)	100% (21/21)	100% (28/28)	>0.99
Cervical dilatation > 3cm (%)	14.3% (3/21)	6.9% (2/29)	0.64
Use of tocolytics (%)	95.2% (20/21)	100% (29/29)	0.42
Antenatal corticosteroids administration (%)	52.4% (11/21)	69.0% (20/29)	0.26
Gestational age at delivery (weeks)	26.6 (22.9–28.7)	33.1 (27.3–34.9)	<0.001
Delivery within 14 days of amniocentesis	100% (21/21)	17.2% (5/29)	<0.001
Delivery within 4 weeks of amniocentesis	100% (21/21)	27.6% (8/29)	<0.001
Delivery before 30 weeks a	94.7% (18/19)	32.1% (9/28)	<0.001
Delivery before 34 weeks	100% (21/21)	58.6% (17/29)	0.001
Delivery at term (>=37 weeks)	0% (0/21)	13.8% (4/29)	0.13
Clinical chorioamnionitis	23.8% (5/21)	3.4% (1/29)	0.07
Acute histologic chorioamnionitis	81.3% (13/16)	60.9% (14/12)	0.29
Funisitis	37.5% (6/16)	26.1% (6/23)	0.50

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

PCR, polymerase chain reaction; WBC, white blood cell.

^aCases who underwent amniocentesis at or beyond 30 weeks were excluded from the analysis.

Of 29 patients undelivered for ≥ 7 days, 10 did not have a follow-up amniocentesis (five declined the procedure, 2 had severe oligohydramnios due to rupture of membranes, 2 were transferred to another hospital, and in one the treating physician did not recommend the procedure). The remaining 19 patients had a follow-up amniocentesis to determine if intra-amniotic infection had been eradicated, if intra-amniotic inflammation was being treated, and to determine whether antibiotic treatment should be continued or stopped. Generally, antibiotics were discontinued if patients had a negative MMP-8 test result or if the amniotic fluid white blood cell count became normal. However, the final decision was made by the attending obstetrician.

There were no significant differences in the median gestational age at amniocentesis, amniotic fluid interleukin-6 concentration and white blood cell count and the frequency of a positive amniotic fluid culture between patients who were undelivered for at least one week and had follow-up amniocentesis and those who had not (p>0.1 for each). Patients who did not have a follow-up amniocentesis delivered significantly earlier than those who had follow-up amniocenteses (27.3 weeks [interquartile range, 25.0–33.9 weeks] vs 34.1 weeks [interquartile range, 31.7–35.6 weeks]; p<0.05).

Treatment success with antibiotics in this study

Of the 23 patients who had follow-up amniocentesis, intra-amniotic inflammation was determined to be successfully eradicated in 15, and intra-amniotic infection was eradicated in three (one with a positive culture and a positive PCR for Ureaplasma spp., and two with a negative culture but a positive PCR for Ureaplasma spp.). All patients with intra-amniotic infection also had intra-amniotic inflammation.

Microbiologic or biochemical evidence of successful treatment was demonstrated in 79% (15/19). One patient who did not have confirmation of eradication of intra-amniotic infection/inflammation at follow-up amniocentesis delivered at term. None of the 10 patients who did not have a follow-up amniocentesis delivered at term. Thus, treatment success of antibiotics (defined as eradication of intra-amniotic infection/inflammation or delivery ≥ 37 weeks of gestation) occurred in 84% (16/19) of patients who had follow-up amniocentesis and was possible in at least 32% (16/50) of patients with intra-amniotic infection/inflammation who received the antimicrobial agents.

Clinical outcome of patients treated with antimicrobial agents and who had a follow-up amniocentesis

Table 3 shows the details and Table 4 summarizes the characteristics and outcomes of 19 patients who were treated with the antimicrobial agents and had a follow-up amniocentesis. A detailed description of each patient can be found in the Supplementary Material S2.

Table 3.

Details of presentations of initial amniocentesis, and outcomes of patients who were treated with antibiotics (ceftriaxone, clarithromycin and metronidazole) and had follow-up amniocentesis

	Gestational age, weeks		Initial amniocentesis					Interval between initial amniocentesis and resolution (days)a
Case No.	Amniocentesis	Delivery	Culture	PCR for Ureaplasm a spp.	interleukin-6 (ng/mL)	WBC count (cells/mm3)	MMP-8 rapid test		Birth weight (gm)	Steroid for fetal lung maturity (weeks)	Acute histologic chorioamnionitis/funisitis	Neonatal outcomes
Group A :			Resolution confirmed and delivered after 34 weeks of gestation
1	25.1	35.6	Neg.	Pos.	5.0	11	Pos.	28	unknown	25.3	N/A	Survival without morbidity
2	22.7	38.3	Neg.	N/A	33.4	100	Pos.	14	2620	22.9	−/−	Survival without morbidity
3	26.0	40.1	Neg.	N/A	2.7	5	Pos.	7	3860	26.0	−/−	Survival without morbidity
4	29.6	34.1	Neg.	N/A	2.7	0	Pos.	7	2200	33.3	−/−	Survival without morbidity
5	22.9	38.0	Neg.	Neg.	2.8	5	Pos.	38	2850		−/−	Survival without morbidity
6	26.7	34.7	Neg.	N/A	4.1	0	Pos.	7	2610	26.3	+/−	Survival without morbidity
7	27.6	35.4	Neg.	Neg.	4.0	1	Pos.	10	unknown		N/A	Survival without morbidity
8	20.9	34.7	Neg.	Neg.	3.3	1	Pos.	67	2040		+/−	Survival without morbidity
9	25.6	35.0	Neg.	Neg.	3.6	25	N/A	48	2290		+/−	Survival without morbidity
Group B :			Resolution confirmed but delivered before 34 weeks of gestation
10	28.4	31.7	Neg.	Pos.	42.6	720	N/A	13	1840	28.4	+/−	Survival without morbidity
11	24.0	32.6	Neg.	Neg.	2.9	2	N/A	30	1990	29.7	+/−	Survival without morbidity
12	25.4	30.4	Neg.	N/A	2.6	0	Pos.	15	1400	24.0	+/+	Survival without morbidity
13	21.0	33.3	Neg.	Neg.	51.5	105	Pos.	22	1800	27.9	−/−	Survived and diagnosed as BPD
14	21.0	29.6	Pos.	Pos.	4.8	0	N/A	30	1640	29.4	+/−	Survival without morbidity
15	20.1	25.4	Neg.	Neg.	11.1	16	Pos.	15	700	25.4	+/+	Shortly died (5 hours from birth)
Group C :			Resolution not confirmed but delivered after 37 weeks of gestation
16	21.6	38.0	Neg.	Neg.	19.4	2	N/A	-	2760		−/−	Survival without morbidity
Group D :			Resolution not confirmed and delivered before 34 weeks of gestation
17	31.4	32.9	Neg.	Neg.	22.0	100	Pos.	-	1710	30.9	+/+	Survival without morbidity
18	26.4	32.6	Pos.	N/A	18.2	50	Pos.	-	2060		+/−	Survival without morbidity
19	22.1	23.6	Neg.	N/A	23.8	54	Pos.	-	620		+/−	Survived and diagnosed as RDS, BPD, IVH, PVL

PCR, polymerase chain reaction; MMP-8, matrix metalloproteinase-8; Pos., positive result; Neg., negative result; N/A, not assessed; BPD, bronchopulmonary dysplasia; RDS, respiratory distress syndrome; IVH, intraventricular hemorrhage; PVL, periventricular leukomalacia

^athe first amniocentesis without evidence of intra-amniotic infection/inflammation

Table 4.

Characteristics and outcomes of 19 patients who were treated with antibiotics (ceftriaxone, clarithromycin and metronidazole) and had follow-up amniocentesis

	Resolution of intra-amniotic inflammation
	Confirmed (n=15)		Not confirmed (n=4)
	Group A: Delivery at or after 34 weeks (n=9)	Group B: Delivery before 34 weeks (n=6)	Group C: Delivery at or after 34 weeks (n=1)	Group D: Delivery before 34 weeks (n=3)
Nulliparity	88.9% (8/9)	66.7% (4/6)	0% (0/1)	66.6% (2/3)
History of preterm delivery	0% (0/9)	16.7% (1/6)	0% (0/1)	0% (0/3)
Progesterone treatment	0	0	0	0
Cerclage before the onset of preterm Labor	11.1% (1/9)	16.7% (1/6)	0% (0/1)	33.3% (1/3)
Cerclage after the onset of preterm labor and labor stopped	0% (0/9)	16.7% (1/6)	0% (0/1)	0% (0/1)
Initial amniocentesis
Gestational age at amniocentesis	25.6 (22.9–26.7)	22.5 (21.0–25.4)	21.6	26.4 (22.1–31.4)
Positive amniotic fluid culture (%)	0% (0/9)	16.7% (1/6)	0% (0/1)	33.3% (1/3)
Positive amniotic fluid PCR for Ureaplasma spp.	20% (1/5)	40% (2/5)	0% (0/1)	0% (0/1)
Amniotic fluid WBC count (cells/mm3)	5 (1–11)	16 (2–105)	2	54 (50–100) a
Positive MMP-8 Rapid test	87.5% (7/8)	100% (3/3)	0% (0/0)	100% (3/3)
Amniotic fluid interleukin-6 (ng/mL)	3.56 (2.79–4.09)	7.02 (2.6–42.6)	19.40	21.97 (18.22–23.84)
Days from initial amniocentesis to resolution	14 (7–38)	18 (13–22)	N/A	N/A
Duration of new antibiotic regimen use (days)b	21 (14–25)	25.5 (21–31)	14	10 (10–33)
Number of amniocentesis	4 (4–4)	3.5 (3–4)	2	2 (2–3)
Gestational age at delivery (weeks)	35.4 (34.7–38.0)	31.1 (29.6–32.6) a	38.0	32.6 (23.6–32.9) a
Days from initial amniocentesis to delivery	73 (56–103)	48 (32–66)	115	10 (10–43) a
Delivery within 14 days of amniocentesis	0% (0/9)	0% (0/6)	0% (0/1)	66.7% (2/3) a
Delivery within 4 weeks of amniocentesis	0% (0/9)	16.7% (1/6)	0% (0/1)	66.7% (2/3) a
Delivery before 30 weeks	0% (0/9)	33.3% (2/6)	0% (0/1)	50.0% (1/2) c
Delivery before 34 weeks	0% (0/9)	100% (6/6) a	0% (0/1)	100% (3/3) a
Delivery at term (>=37 weeks)	66.7% (6/9)	0% (0/6)	100% (1/1)	0% (0/3)
Birth weight (grams)	2610 (2200–2850)	1720 (1225–1878) a	2760	1710 (620–2060) a
Clinical chorioamnionitis	0% (0/9)	0% (0/6)	0% (0/1)	0% (0/3)
Acute histologic chorioamnionitis	28.6% (2/7)	83.3% (5/6)	0% (0/1)	100% (3/3)
Funisitis	0% (0/7)	50% (3/6)	0% (0/1)	33.3% (1/3)
Significant neonatal morbidity	0% (0/9)	33.3% (2/6)	0% (0/1)	33.3% (1/3)

Data are median (interquartile ranges for group A and B, range for group D) or % (n/N).

PCR, polymerase chain reaction; WBC, white blood cell; MMP-8, matrix metalloproteinase-8.

^ap < 0.05 compared to Group A.

^bSome patients restarted antibiotic administration because they developed preterm rupture of membranes or preterm labor and intra-amniotic infection/inflammation after the discontinuation of antibiotics as intra-amniotic infection/inflammation resolved and preterm labor stopped. This duration was not included in this analysis

^cOne case who underwent amniocentesis at or beyond 30 weeks was excluded from the analysis.

Comment

Principal findings of the study:

1) antibiotics were effective in treating intra-amniotic infection/inflammation in women with preterm labor and intact membranes as demonstrated by analysis of amniotic fluid obtained before and after antibiotics were administered; 2) resolution of intra-amniotic infection/inflammation was objectively demonstrated in 79% (15/19) of patients who received the antimicrobial agents and had follow-up amniocentesis; 3) the overall treatment success (defined as resolution of intra-amniotic inflammation or infection, or delivery ≥37 weeks) rate among patients who underwent follow-up amniocentesis was 84% (16/19). The overall success rate among all women with intra-amniotic infection/inflammation who received the antimicrobial agents was 32% (16/50).

The prevalence and clinical import of intra-amniotic infection/inflammation in patients with preterm labor and intact membranes

The frequency of a positive amniotic fluid culture for microorganisms in patients presenting with an episode of preterm labor and intact membranes is approximately 10%,^{3,6,12,14,33,139} and these patients are more likely to develop maternal complications such as clinical chorioamnionitis³ and pulmonary edema while receiving tocolytics,^68,69 and deliver a preterm neonate shortly after admission.^{3,14,16,33,140,141} In addition, patients with intra-amniotic infection are more likely to show evidence of histologic chorioamnionitis (a maternal host response) or funisitis/chorionic vasculitis (pathologic hallmarks of the fetal inflammatory response syndrome, or FIRS).^{59,124,131–133} One of every four preterm neonates is born to a mother with microorganisms in the amniotic cavity.^{3,11,12,22,142–146}

When microorganisms invade the human fetus, a systemic inflammatory response can be elicited, and this condition is referred to as FIRS (diagnosed by an elevated umbilical cord blood plasma IL-6 concentration). This condition is associated with a higher rate of neonatal complications^147,148 because, before birth, these fetuses have multi-systemic involvement or dysfunction.¹⁴⁹ Examples include leukocyte activation,¹⁵⁰ leukocytosis,¹⁵¹ adrenal gland hyperactivity (elevated concentrations of cortisol in peripheral blood),¹⁵² cardiac dysfunction,^153,154 and increased concentrations of matrix-degrading enzymes in amniotic fluid and fetal blood.^155–160 FIRS is a risk factor for neonatal morbidity, as well as long-term complications such as cerebral palsy ^54,161 and chronic lung disease.^162–165

In summary, a strong body of evidence indicates that fetal exposure to microorganisms or intra-amniotic inflammation is associated with adverse outcome.^{1,5,6,8,9,39,55,64,71,131,161,166–170} Despite this overwhelming evidence, obstetricians in practice do not routinely ascertain whether patients with preterm labor have intra-amniotic infection/inflammation. The reason is two-fold: first, the best method to determine the presence of intra-amniotic infection/inflammation is analysis of amniotic fluid, which requires an invasive procedure (amniocentesis); second, the evidence that treatment with antimicrobial agents can eradicate intra-amniotic infection has been based on case reports. Therefore, in practice, clinicians rely on signs and symptoms of clinical chorioamnionitis (e.g. fever, maternal tachycardia, etc.) to exclude intra-amniotic infection. However, it is now well-established that these clinical signs are both insensitive and non-specific for the identification of intra-amniotic infection in both preterm^{6,33,121,125,171} and term gestations^{136,137,172–176}. This is also the case for maternal circulating white blood cell count and other biomarkers of the acute phase response (such as serum C-reactive protein).^{18,121,167,177,178} One argument against the analysis of amniotic fluid used to be that results were not immediately available to affect patient management, as culture for microorganisms may take several days. However, rapid tests are now available for the diagnosis of intra-amniotic inflammation (such as amniotic fluid white blood cell count, glucose, amniotic fluid MMP-8, or interleukin-6, among others),^{4,19,38,53,60,126,127,167,176,177,179–184}, and for the diagnosis of infection using PCR. ^{8,23,123,140,185,188}

Antibiotic administration to patients in preterm labor with intact membranes

The evidence that intra-amniotic infection is causally linked to spontaneous preterm labor and delivery coalesced in the 1980s:^{2,3,13,14,144,189,190} this led to the conduct of several randomized clinical trials in which patients with an episode of preterm labor were allocated to antimicrobial agents vs. placebo or no treatment.^{73,74,77–80,82,191–194} Although the initial trials reported pregnancy prolongation,^75,77,191 and, in some cases, a lower frequency in the rate of preterm delivery, ^73,77 these findings were not supported by subsequent clinical trials^78,81 or systematic reviews and meta-analyses.^195–197 This led professional organizations, including the American College of Obstetricians and Gynecologists¹⁹⁸ and Society of Maternal-Fetal Medicine,⁸⁶ to recommend that antibiotics not be administered to patients with preterm labor and intact membranes, with the objective of prolonging pregnancy or reducing the rate of preterm birth. Antibiotics have been recommended in the context of preterm labor with intact membranes when delivery is impending and the patient is a carrier of Group B streptococci or Streptococcus agalactiae.^{85,86,199,200}

Why are antibiotics considered ineffective in prolonging pregnancy and preventing preterm delivery in patients with preterm labor and intact membranes?

Preterm labor is a syndrome defined by the presence of increased uterine contractility, cervical dilatation, and decidual membrane activation, each caused by multiple pathologic processes.^15,201,202 Intra-amniotic infection is only one of the potential mechanisms of disease responsible for this syndrome. If the frequency of intra-amniotic infection is only 10%²²; then antimicrobial agents can only be effective in that small fraction of patients.²⁶

The ORACLE II trial randomized 6,295 women with an episode of preterm labor with intact membranes to placebo or antibiotics: these patients did not have clinical evidence of infection, and amniocenteses were not performed to diagnose intra-amniotic infection.⁷⁸ Therefore, 90% of patients enrolled in the ORACLE II trial could not have benefitted from antibiotic administration, and the negative results are not surprising. The same applies to all other randomized clinical trials of antibiotics in patients with preterm labor and intact membranes.^74,79–82 However, these results should not be interpreted as indicating that antibiotics are ineffective when administered to the “right” patients: those who have proven intra-amniotic infection.

Experimental evidence that anti-microbial agents can eradicate intrauterine infection and prolong pregnancy

McDuffie et al. reported that, in pregnant rabbits, antibiotic administration (ampicillin and sulbactam) at or before inoculation with E. coli led to fewer preterm deliveries and more live pups than those whose treatment was delayed for more than four hours.²⁰³ Subsequently, Fidel et al., using the same experimental model, showed that antibiotic administration within 12 hours of inoculation – but not after 18 hours – increased the duration of pregnancy and reduced perinatal mortality.²⁰⁴ Collectively, these results suggest that antibiotics can be beneficial in cases of intrauterine infection.

These observations were subsequently confirmed in non-human primates. Investigators at the Oregon Primate Center administered Group B Streptococci to pregnant Rhesus monkeys (Macaca mulatta) on day 130 of gestation (term: 167 days), and observed an increase in uterine contractility at a median of 28 hours (range: 14–40 hours) after inoculation.⁵ This model of intra-amniotic infection has many features in common with intra-amniotic infection in women. Importantly, the onset of contractions was preceded by an increase in amniotic fluid concentrations of proinflammatory cytokines (IL-1β, TNFα, IL-6) and prostaglandins (E2 and F2a). None of the animals became febrile or had leukocytosis; yet, all delivered preterm. Subsequent studies demonstrated that dexamethasone,²⁰⁵ indomethacin,²⁰⁶ and IL-10 blocked IL1–induced uterine contractility (a model of intra-amniotic inflammation), suggesting a role for anti-inflammatory agents in the treatment of inflammation-induced preterm labor.²⁰⁷

Antibiotics used in this study to treat intra-amniotic infection/inflammation

An important principle in the treatment of infectious diseases is that antibiotic selection should be tailored to the microorganisms causing the infection. The rationale for the antibiotic regimen used in our study was described in previous studies of patients with preterm PROM.^118,120 Briefly, two macrolides, erythromycin or azithromycin, have been used to treat intra-amniotic infection in women, and there is experimental evidence in non-human primates that azithromycin can eradicate Ureaplasma from the amniotic cavity and reduces fetal lung injury.^208,209 Clarithromycin was chosen at our Institution because it has a much higher rate of transplacental passage than erythromycin or azithromycin, and this agent is effective against Ureaplasma species, the most common microorganism identified in the amniotic fluid of patients at risk for preterm delivery.²¹⁰ Ceftriaxone was included because of its enhanced coverage of aerobic bacteria and high rate of transplacental passage.^211,212 Metronidazole was selected because anaerobic organisms are frequently present in the amniotic cavity, and this drug provides optimal coverage for these microorganisms. We reported that, in patients with preterm PROM, this antibiotic combination eradicated intra-amniotic infection and/or inflammation in at least 33% of patients as demonstrated by repeat analysis of amniotic fluid.¹¹⁸ Whether other antimicrobial combinations can achieve the same result would need to be determined.

Evidence that intra-amniotic infection can be treated

Intra-amniotic infection has been successfully treated in patients with a sonographic short cervix without clinical manifestations of infection (fever, uterine tenderness, etc.).¹¹⁷ Eradication of intra-amniotic infection has also been reported in cases of preterm PROM^118,213and preterm labor.^214,215 Whether this approach is effective in patients with preterm labor with intact membranes had not been studied until recently.¹¹⁹ In patients with preterm labor and proven intra-amniotic infection, there was a shorter diagnosis-to-delivery interval.^{3,6–8,22,33,53} Indeed, it was generally believed that once patients present with preterm labor, an intra-amniotic “cytokine storm” would inevitably lead to preterm delivery.

The results reported herein represent the first objective confirmation that antibiotic treatment can eradicate intra-amniotic infection in preterm labor with intact membranes in a case series. This was demonstrated in three patients: the first had microbial invasion of the amniotic cavity with Ureaplasma spp. detected by culture; the other two had microbial nucleic acids detected by PCR.

In all three cases, a repeat amniocentesis yielded a negative amniotic fluid culture and negative PCR for microorganisms. Details of each specific case are illustrated in Table 3 (see cases 1, 10 and 14). It is interesting that in case 14, the first amniocentesis at 21 weeks was positive for Ureaplasma spp. and showed an elevated interleukin-6 (4.8 ng/mL). Antibiotic treatment eradicated both the microorganisms and evidence of the intra-amniotic inflammatory process (interleukin-6: 1.93 ng/mL). The treating physician elected to continue oral clarithromycin. Four weeks after successful treatment, the patient was suspected to have rupture of membranes and the amniotic fluid became positive for Morganella morganii, a Gram negative bacilli frequently implicated in nosocomial infections.²¹⁶ Intra-amniotic inflammation (interleukin-6: 6.89 ng/mL) recurred, labor progressed, and the patient delivered at 29.4 weeks a 1640 g neonate that had no major complications. This case indicates that patients with an intra-amniotic infection may be susceptible to recurring infection with other microorganisms. Whether this indicates a deficit in host defense or an opportunistic infection during the course of antimicrobial therapy is unclear. Chorioamnionitis caused by Morganella morganii has been reported in an immunocompetent patient.²¹⁷ Recent evidence derived from whole exome sequencing indicates that some patients may have deleterious mutations in genes encoding for proteins implicated in host defense against microbial invasion.^218–221 There is evidence that acute chorioamnionitis may be recurrent in successive pregnancies²²² and therefore, the predisposition to intra-amniotic infection may have a genetic basis.^29,223–227

Recently, a group of investigators reported that antimicrobial agents in patients with intra-amniotic infection may result in prolongation of pregnancy and a decreased rate of admission to the neonatal intensive care unit without a change in perinatal morbidity.¹¹⁹ No follow-up amniocenteses were performed in that study; therefore, there was no objective evidence to determine whether antimicrobial therapy was effective in treating intra-amniotic infection/inflammation. Nonetheless, the reports of such studies represent indirect evidence consistent with our findings.

Successful treatment of intra-amniotic inflammation in preterm labor with intact membranes with antibiotics

Intra-amniotic inflammation in the absence of demonstrable microorganisms is more frequent than intra-amniotic infection in patients with preterm labor and intact membranes,^{24,26,33,36,39} a sonographic short cervix,^28,117 and even preterm PROM.^27,71 This type of intra-amniotic inflammation may be caused by either microorganisms which escaped detection^27,28 or by danger signals or alarmins^{30–32,91,228,229} which are released by cells under stress or during the course of cell death such as necrosis.^230–233 Examples of danger signals include high mobility group box (HMGB1), S100 calcium-binding protein B (S100B), and interleukin-1α, which can induce preterm labor by the activation of the inflammasome.^{31,90,92,234–238}

The treatment of sterile inflammation is a major challenge in medicine. The conventional approach is to use anti-inflammatory agents, such as glucocorticoids^239,240 or non-steroidal antiinflammatory agents.^241,242 In some cases, treatment is possible with a specific agent that decreases the concentration of the danger signal such as allopurinol to decrease the concentration of uric acid in gout. However, in the case of intra-amniotic inflammation without demonstrable microorganisms, the optimal treatment is uncertain. Preliminary evidence from our laboratory suggests that inhibitors of the inflammasome may have therapeutic benefits in preventing preterm labor induced by specific danger signals such as S100B.²³⁶

How can antibiotics be effective in cases of intra-amniotic inflammation without demonstrable microorganisms? The antibiotic combination used at the Seoul National University Hospital included clarithromycin, which has been shown to have immunomodulatory properties and specifically inhibits AP-1 and NF-kappa B, two transcription factors which induce production of proinflammatory cytokines and effectively act as anti-inflammatory agents.^243,244 We have previously shown that NFkB is upregulated by IL-1B.^245–249

Our study shows that intra-amniotic inflammation/infection was successfully treated in 84% (16/19) of cases in which follow-up amniocentesis was performed. It is unlikely that this therapeutic success can be attributed to glucocorticoids because these agents were not used in 31% (5/16) of patients in whom intra-amniotic inflammation improved.

The results of the current study are consistent with our previous observations in the context of preterm PROM. The antimicrobial agents used in this study were able to treat and prevent intra-amniotic infection/inflammation, prolong the latency period, reduce acute histologic chorioamnionitis and funisitis, and improve neonatal outcomes in patients with preterm PROM.¹²⁰

Strengths and limitations

Strengths of the study include: 1) this is the first case series in which women with intra-amniotic infection/inflammation were treated with antibiotics and monitored with serial amniocentesis to determine if there was therapeutic success in patients with preterm labor and intact membranes; 2) assessment of intra-amniotic infection/inflammation was performed by analysis of amniotic fluid using the amniotic fluid white blood cell count or a rapid test for MMP-8; 3) the retrospective diagnosis of intra-amniotic inflammation was performed using amniotic fluid concentrations of interleukin-6, which has been demonstrated to be a reliable marker, widely used in many reports to diagnose this condition; and 4) this study used serial evaluation of amniotic fluid. This is the only objective method to ascertain whether there is therapeutic efficacy.

Limitations of this study include its observational nature. This is not a randomized clinical trial in which there was a placebo arm. However, clinicians in our institution are unwilling to randomize patients with intra-amniotic infection/inflammation to placebo because such patients are at increased risk for clinical chorioamnionitis, maternal sepsis, and neonatal complications, such as early neonatal sepsis, among others. In a previous observational study, we reported that 91% of patients with intra-amniotic inflammation delivered within one week of amniocentesis.³³ In contrast, in the current study, only 42% (21/50) delivered within one week: this is also indirect evidence of efficacy.

We have grouped together patients with intra-amniotic infection and intra-amniotic inflammation without demonstrable microorganisms. A limitation of our study is that there were only three patients with intra-amniotic infection who were successfully treated and that most of the patients had intra-amniotic inflammation without detectable microorganisms with the methods used in our institution. It is possible that more organisms could have been detected by using assays for the conserved region of the microbial genome or sequencing of microbial cell-free DNA.^250–252 Further studies using molecular microbiologic methods are required to address this issue.

Another potential limitation to our interpretation of the results of this case series is that we used as a definition of “success” as delivery at ≥37 weeks of gestation, rather than ≥34 weeks. This may be a very stringent criterion to assess the prognosis of a patient with preterm labor and intra-amniotic infection/inflammation; however, using this definition strengthens the evidence of the effectiveness of antibiotics, as patients may benefit from antimicrobial agents without delivering at term (e.g. delivery at 36 weeks). Indeed, we performed a sensitivity analysis, and if treatment success was defined as eradication of intra-amniotic infection/inflammation or delivery at or after 32 weeks of gestation, the overall efficacy would be 44% (22/50).

Conclusion

The administration of antibiotics to patients with preterm labor and intact membranes with proven intra-amniotic infection/inflammation is associated with eradication of infection and inflammation in a subset of patients.

AJOG at a Glance:

1. What is the research question?

   Can intra-amniotic infection or intra-amniotic inflammation be treated with antibiotics in patients with preterm labor and intact membranes?

2. What are the key findings?

   Resolution of intra-amniotic inflammation or intra-amniotic infection was objectively demonstrated by analysis of amniotic fluid after treatment with antibiotics in 83% of patients.

3. What does this add to what is known?

   Contrary to what is widely believed, antimicrobial treatment of intra-amniotic infection or intra-amniotic inflammation can be successful in a subset of patients with preterm labor and intact membranes. These observations open new therapeutic alternatives and call for personalized assessment of patients with preterm labor and intact membranes to identify those who can benefit from this intervention.