Share this article on:

A Retrospective Comparison of Antibiotic Regimens for Preterm Premature Rupture of Membranes

Pierson, Rebecca C. MD; Gordon, Sashana S. MS; Haas, David M. MD, MS

doi: 10.1097/AOG.0000000000000426
Contents: Original Research

OBJECTIVE: To evaluate whether the use of ampicillin and azithromycin leads to a similar latency period in preterm premature rupture of membranes as ampicillin and erythromycin and whether the substitution of azithromycin for erythromycin effects rates of other outcomes.

METHODS: We performed a retrospective cohort study of women with preterm premature rupture of membranes between 24 and 34 completed weeks of gestation and compared two groups: those who received ampicillin and erythromycin and those who received ampicillin and azithromycin. Primary outcome was length of latency (defined as time from first antibiotic dose to delivery) and secondary outcomes were rates of chorioamnionitis, cesarean delivery, Apgar scores, birth weight, neonatal death, neonatal sepsis, and neonatal respiratory distress syndrome.

RESULTS: Of 168 women who met inclusion criteria, 75 received ampicillin and erythromycin and 93 received ampicillin and azithromycin. There was no difference in latency between groups: 9.6±13.2 days (erythromycin) compared with 9.4±10.0 (azithromycin) days (P=.40). Secondary outcomes did not differ between groups. We had 80% power to detect a difference of 5 days.

CONCLUSION: Among women with preterm premature rupture of membranes between 24 and 34 completed weeks of gestation, substitution of azithromycin for erythromycin in the recommended antibiotic regimen did not affect latency or any other measured maternal or fetal outcomes.


There is no difference in latency when azithromycin and ampicillin are substituted for erythromycin and ampicillin in patients with preterm premature rupture of membranes.

Indiana University School of Medicine and Department of Medicine, Division of Clinical Pharmacology, and Department of Obstetrics and Gynecology, Indiana University School of Medicine, Indianapolis, Indiana.

Corresponding author: David M. Haas, MD, MS, Department of Obstetrics and Gynecology, 550 N University Boulevard, UH 2440, Indianapolis, IN 46202; e-mail:

Supported by PREGMED: The Indiana University Center for Pharmacogenetics and Therapeutics Research in Maternal and Child Health (5U01HD063094) (D.M.H.) and National Institutes of Health–National Institute of General Medical Sciences: Indiana University Comprehensive Training in Clinical Pharmacology (2T32GM008425-21) (R.C.P.).

Financial Disclosure The authors did not report any potential conflicts of interest.

Preterm premature rupture of membranes (PROM) complicates 3% of pregnancies1 and is responsible for one-fourth to one-third of preterm births.2 The etiology is often unknown but may be related to infection,3–5 membrane dysfunction on a molecular level,6,7 genetic variation,8,9 or a combination of factors.10 Treatment with antibiotics in the absence of labor has been shown to decrease the frequency of chorioamnionitis,11 prolong latency, and decrease fetal and neonatal complications.12–14 Given the wide range of possible causative organisms and the often polymicrobial nature of infectious preterm PROM,15 the American College of Obstetricians and Gynecologists (the College) recommends broad-spectrum coverage in latency antibiotic regimens. Many antibiotic regimens have been evaluated in both a prospective and retrospective fashion and no treatment regimen has been found superior; all regimens are superior to placebo.16

In the absence of major infectious complications such as chorioamnionitis, prolongation of the pregnancy is recommended to reduce the risks associated with premature delivery. According to recommendations from the College, women with preterm PROM should receive therapy with ampicillin and erythromycin (48 hours of parenteral antibiotics followed by 5 days of oral amoxicillin and erythromycin).17 Although azithromycin is often substituted for erythromycin for ease of administration, presumed equivalency, and possible decreased cost, to our knowledge, an azithromycin-containing regimen has never been compared with the recommended regimen for preterm PROM. In a search of Medline (1946 to present; search date May 2014), using the terms “fetal membranes, premature rupture or preterm rupture membranes” and “azithromycin,” no studies comparing azithromycin with any other antibiotic for preterm PROM were found.

The objective of this study was to evaluate whether the use of ampicillin and azithromycin led to a similar latency period compared with the recommended regimen of ampicillin and erythromycin. Our secondary objective was to determine whether the substitution of azithromycin for erythromycin affected rates of other maternal and neonatal outcomes.

Back to Top | Article Outline


This study was approved by the institutional review board at Indiana University. All abstracted records were reviewed for accuracy by a second member of the study team.

We performed a retrospective chart review of patient records of all women admitted with a possible diagnosis of preterm PROM from January 1, 2009, to March 31, 2013. Patients were identified using pharmacy records, census lists, and hospital delivery records in an effort to capture all possible preterm PROM diagnoses. Women with pregnancies at less than 24 completed weeks of gestation were excluded given the highly variable management options for these individuals. Women with pregnancies at greater than 34 completed weeks of gestation were excluded because latency antibiotics are not indicated for this group. Patients admitted in preterm labor with concomitant rupture of membranes were excluded. Patients with a cerclage, a multiple gestation, a history of amniocentesis or fetal surgery, a history of abdominal trauma, or who were carrying a fetus with lethal anomalies were also excluded. Demographic information was recorded for each patient as well as antibiotic information (regimen, timing of antibiotic administration) and obstetric course (time of delivery, route of delivery, and presence or absence of chorioamnionitis). The choice of antibiotic regimen had been made by the obstetric provider based on personal preference at the time of admission. No standard protocol guided therapy choice. The diagnosis of chorioamnionitis was assigned based on documentation in the medical record of a clinical diagnosis or on a placental pathology result of chorioamnionitis. Matching infant medical records were reviewed, and the outcomes of death, neonatal respiratory distress syndrome, and neonatal sepsis were recorded.

The primary outcome of this study was the latency time (defined as the time from first antibiotic dose administration to the time of delivery). Secondary outcomes were gestational age at delivery, any noted adverse drug side effects requiring discontinuation of the drug, birth weight, Apgar scores, and neonatal outcomes (death, neonatal respiratory distress syndrome, sepsis).

All analyses were performed using IBM SPSS Statistics 21.0 on an intent-to-treat basis. The mean latency times in each antibiotic regimen were compared using the nonparametric Mann–Whitney U test. Unpaired t test was also used to evaluate the secondary outcomes of gestational age at delivery, birth weight, and Apgar scores if normally distributed. Nonparametric data were compared using Mann–Whitney U tests. The χ2 test was used to compare all categorical variables. A logistic regression model was created to control for the effect of gestational age on latency. Survival curves for latency were also compared between the two groups using Mantel–Cox testing. We did not have an a priori sample size calculation. A post hoc power analysis demonstrated that we had 80% power to detect a difference in latency period of 5 days.

Back to Top | Article Outline


From January 1, 2009, through March 31, 2013, 1,224 potential patients were identified. A significant portion of these were identified through medication administration records but were in fact ineligible. The remaining 526 patients had preterm rupture of membranes based on documentation in the medical record of the clinical criteria for rupture. Of these, 168 patients met inclusion criteria for this study and were included in the analysis (Fig. 1). Baseline characteristics and demographics were similar between groups (Table 1).

Fig. 1

Fig. 1

Table 1

Table 1

There was no statistically significant difference in time of latency between women who received ampicillin and azithromycin (9.4±10.4 days) and women who received ampicillin and erythromycin (9.6±13.2 days) (P=.40), as shown in Table 2 and Figure 2.

Table 2

Table 2

Fig. 2

Fig. 2

The latency time from rupture of membranes to delivery also was not different between the groups (9.9±10.0 for azithromycin compared with 10.1±13.2 days for erythromycin group, P=.40). Length of latency also did not differ when controlling for gestational age at the time of rupture in the regression analysis.

There were similar rates of chorioamnionitis, similar birth weight, Apgar scores, and neonatal complications between the two groups (Table 2). Of the 95 cases of chorioamnionitis diagnosed, five were diagnosed by clinical criteria alone, 78 by pathology examination of the placenta alone, and 12 by both clinical and pathology confirmation. If only the clinical diagnosis of chorioamnionitis were used, only eight of 93 (8.6%) of the women in the ampicillin and azithromycin group and nine of 75 (12%) women in the ampicillin and erythromycin group would have been diagnosed, respectively (P=.47). There were no differences in the rates of neonatal sepsis, neonatal respiratory distress syndrome, or death. There was no difference in the frequency of cesarean delivery between groups. We also recorded the reason for delivery for each patient (spontaneous labor, chorioamnionitis, scheduled delivery after 34 completed weeks of gestation, fetal indication, maternal indication other than chorioamnionitis); there were no differences between groups (data not shown, P=.42). One woman receiving erythromycin developed chest pain after 2 days of therapy; although this symptom was not felt to be an adverse drug reaction, the health care providers elected to discontinue erythromycin. No women receiving azithromycin had severe side effects noted.

Back to Top | Article Outline


In this retrospective study of two commonly used antibiotic regimens in the treatment of preterm PROM, there were no differences in outcomes between patients who received ampicillin and erythromycin and patients who received ampicillin and azithromycin. Both groups had a latency of slightly greater than 9 days. This is in agreement with other published studies of antibiotic use in preterm PROM in which the latency period ranged from 1.4 to 12.0 days.18

The spectrum of microbial coverage of azithromycin is similar to erythromycin because they are very similar in structure; however, their pharmacokinetic properties are different: the half-life of azithromycin is approximately 68 hours (Zithromax [package insert]. New York, NY: Pfizer; 2013:1–35) and the half-life of erythromycin is 1.6 hours.19 The dosing regimens commonly used for latency in preterm PROM are also different (7 days of erythromycin compared with a single dose of azithromycin), which contributes to a different pharmacokinetic profile. If this protocol were shown to be equivalent to the standard ampicillin and erythromycin protocol, this would be a treatment regimen with potentially decreased cost and increased ease of administration.

A College Practice Bulletin discusses the use of ampicillin and erythromycin as the recommended regimen for preterm PROM.17 As a result of concerns about the maternal side effects of erythromycin and the ease of administration of azithromycin, many health care providers have changed their preferred antibiotic regimen to the combination of ampicillin and azithromycin. Given the essential equivalence of our latency findings in this study between the two regimens, there is reason to believe that the regimens may be potentially interchangeable with regard to outcomes. However, this finding should be evaluated in a randomized controlled trial to rigorously test the equivalence hypothesis, particularly for important secondary neonatal outcomes.

We also found no differences in any of our secondary maternal or neonatal outcomes. We were surprised that our rates of chorioamnionitis were higher than other studies, which have reported rates from 12% (18) to 23% (14) compared with our rates of 49%. The higher rates in our study reflect the fact that we accepted either a clinical or a pathologic diagnosis of chorioamnionitis and that we send nearly all placentas from preterm deliveries to pathology. Thus, a higher component of pathology-diagnosed chorioamnionitis contributed to our higher rate. Also, the inner-city catchment patient population served in the hospital may have a higher risk of chorioamnionitis or of predisposing genital infections.

Because we wished to evaluate the two treatments in a specific subset of patients with preterm birth (preterm PROM in the absence of cervical incompetence, preterm labor, or iatrogenic risk factors), our study excluded any patients with complications of pregnancy likely to contribute to an increased risk of preterm PROM (multiple gestation, preterm labor, cervical insufficiency, cerclage, intrauterine device in situ, history of amniocentesis or fetal surgery). Although limiting the generalizability, these exclusions likely also limited some confounding factors. We analyzed all data on an intent-to-treat basis; we did not exclude patients who did not complete the entire course of therapy.

As a retrospective study, our results have some limitations. Despite using three different methodologies to identify patients, we could have failed to identify some women with preterm PROM. We attempted to limit data acquisition errors by having data doubly extracted with quality control checks. Although studies have shown an increased risk of preterm PROM with subsequent pregnancies after a prior pregnancy complicated by preterm PROM,20–22 we did not analyze or control for prior pregnancy outcome because the etiology of this effect is unknown and the increased risk is not universal.20–22 We also did not exclude patients who received tocolytic therapy or magnesium sulfate for its neuroprotective effects; although the presence of these agents is potentially a confounding factor that may well influence the length of latency, the subset of patients in our study who received tocolytic agents or magnesium sulfate was a minority and the use was generally only long enough to administer antenatal corticosteroids. In addition, the rates of tocolytic therapy were similar between groups. We did not analyze the rate of additional maternal antibiotics received for reasons other than latency and we did not exclude patients with genitourinary infections. Concomitant antibiotic receipt may complicate the clinical picture, especially in the patient with cervical or genitourinary infection at the time of rupture. However, the rates of women with genitourinary infections were similar between the two groups (P=.10). In practice, health care providers may have been more likely to use azithromycin for women with concomitant genitourinary infections. Our sample size would have had 80% power to detect a latency difference of 5 days. To power an equivalency trial using the mean latency we found and ensuring one antibiotic regimen was no worse than 2 days different would require 610 women in each group.

In summary, within the limitations of this nonrandomized study of modest sample size, we have found equivalence of latency of two commonly used antibiotic regimens in the treatment of preterm PROM: ampicillin and erythromycin and ampicillin and azithromycin.

Back to Top | Article Outline


1. Waters TP, Mercer B. Preterm PROM: prediction, prevention, principles. Clin Obstet Gynecol 2011;54:307–12.
2. Meis PJ, Ernest JM, Moore ML. Causes of low birth weight births in public and private patients. Am J Obstet Gynecol 1987;156:1165–8.
3. Averbuch B, Mazor M, Shoham-Vardi I, Chaim W, Vardi H, Horowitz S, et al.. Intra-uterine infection in women with preterm premature rupture of membranes: maternal and neonatal characteristics. Eur J Obstet Gynecol Reprod Biol 1995;62:25–9.
4. Garite TJ, Freeman RK. Chorioamnionitis in the preterm gestation. Obstet Gynecol 1982;59:539–45.
5. Naeye RL, Peters EC. Causes and consequences of premature rupture of fetal membranes. Lancet 1980;1:192–4.
6. Skinner SJM, Campos GA, Liggins GC. Collagen content of human amniotic membranes: effect of gestation length and premature rupture. Obstet Gynecol 1981;57:487–9.
7. Moore RM, Mansour JM, Redline RW, Mercer BM, Moore JJ. The physiology of fetal membrane rupture: insight gained from the determination of physical properties. Placenta 2006;27:1037–51.
8. Rocha FG, Slavin TP, Li D, Tiirikainen MI, Bryant-Greenwood GD. Genetic associations of relaxin: preterm birth and premature rupture of fetal membranes. Am J Obstet Gynecol 2013;209:258.e1–8.
9. Romero R, Friel LA, Edwards DRV, Kusanovic JP, Hassan SS, Mazaki-tovi S, et al.. A genetic association study of maternal and fetal candidate genes that predispose to preterm prelabor rupture of membranes (PROM). Am J Obstet Gynecol 2010;203:361.e1–30.
10. French JI, McGregor JA. The pathobiology of premature rupture of membranes. Semin Perinatol 1996;20:344–68.
11. Ovalle A, Martínez MA, Kakarieka E, Gómez R, Rubio R, Valderrama O, et al.. Antibiotic administration in patients with preterm premature rupture of membranes reduces the rate of histological chorioamnionitis: a prospective, randomized, controlled study. J Matern Neonatal Med 2002;12:35–41.
12. Kenyon S, Taylor DJ, Tarnow-Mordi WO; ORACLE Collaborative Group. ORACLE—antibiotics for preterm prelabour rupture of the membranes: short-term and long-term outcomes. Acta Paediatr 2002;91:12–5.
13. Mercer BM, Arheart KL. Antimicrobial therapy in expectant management of preterm premature rupture of the membranes. Lancet 1995;346:1271–9.
14. Mercer BM, Miodovnik M, Thurnau GR, Goldenberg RL, Das AF, Ramsey RD, et al.. Antibiotic therapy for reduction of infant morbidity after preterm premature rupture of the membranes. A randomized controlled trial. National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network. JAMA 1997;278:989–95.
15. Czikk MJ, McCarthy FP, Murphy KE. Chorioamnionitis: from pathogenesis to treatment. Clin Microbiol Infect 2011;17:1304–11.
16. Kenyon S, Boulvain M, Neilson J. Antibiotics for preterm rupture of membranes. The Cochrane Database of Systematic Reviews 2013, Issue 12. Art. No.: CD001058. DOI: 10.1002/14651858.CD001058.pub3.
17. Premature rupture of membranes. Practice Bulletin No. 139. American College of Obstetricians and Gynecologists. Obstet Gynecol 2013;122:918–30.
18. Mercer BM, Arheart KL. Antibiotic therapy for preterm premature rupture of the membranes. Semin Perinatol 1996;20:426–38.
19. Hardman J, Limbird L, Gilman A. Goodman and Gilman’s the pharmacological basis of therapeutics. 11th ed. New York (NY): McGraw-Hill; 2006.
20. Asrat T, Lewis DF, Garite TJ, Major CA, Nageotte MP, Towers CV, et al.. Rate of recurrence of preterm premature rupture of membranes in consecutive pregnancies. Am J Obstet Gynecol 1991;165:1111–5.
21. Harger JH, Hsing AW, Tuomala RE, Gibbs RS, Mead PB, Eschenbach DA, et al.. Risk factors for preterm premature rupture of fetal membranes: a multicenter case-control study. Am J Obstet Gynecol 1990;163:130–7.
22. Naeye RL. Factors that predispose to premature rupture of the fetal membranes. Obstet Gynecol 1982;60:93–8.
© 2014 by The American College of Obstetricians and Gynecologists.