OBJECTIVE: Extra-amniotic saline infusion has been shown to be equal to or more efficacious than misoprostol, laminaria, dinoprostone, and prostaglandin estradiol for cervical ripening and labor induction. Because of the introduction of a foreign body into the uterus, extra-amniotic saline infusion may potentially cause increased rates of chorioamnionitis. This study examines the risk of chorioamnionitis with extra-amniotic saline infusion compared with other methods of induction and spontaneous labor.
METHODS: A retrospective analysis was performed based on deliveries at Bellevue Hospital Center, a tertiary-care facility, from August 2000 to December 2002. Three groups were identified: extra-amniotic saline infusion, other methods of induction, and spontaneous labor. Differences in chorioamnionitis rates were analyzed by using analysis of variance and multivariable logistic regression as appropriate.
RESULTS: There were 625 charts evaluated: 171 extra-amniotic saline infusion, 190 other, and 264 with spontaneous labor. The rates of chorioamnionitis were 26.9%, 17.9%, and 13.3%, respectively. After adjusting for confounding variables, such as instrumentation, length of rupture, and number of exams, subjects who were induced with extra-amniotic saline infusion were significantly more likely to develop chorioamnionitis (relative risk = 2.2; 95% confidence interval 1.4, 4.0; P = .006).
CONCLUSION: Extra-amniotic saline infusion may be associated with a greater risk of chorioamnionitis when compared with other methods of labor induction. Given the increased risk of chorioamnionitis associated with extra-amniotic saline infusion, its use should be in the context of a careful assessment of the risks and benefits of various methods of labor induction.
LEVEL OF EVIDENCE: II-2
Extra amniotic saline infusion increases the risk of chorioamnionitis compared with other methods of cervical ripening or labor induction.
From the *Department of Obstetrics and Gynecology, New York University School of Medicine, New York, New York; and †Department of Obstetrics and Gynecology, Yale University School of Medicine, New Haven, Connecticut.
Address reprint requests to: Kenneth A. Levey, MD, MPH, NYU School of Medicine, Department of Obstetrics and Gynecology, 550 First Avenue, Suite 9E2, New York, NY 10016; e-mail: firstname.lastname@example.org.
Received August 9, 2003. Received in revised form December 21, 2003. Accepted January 8, 2004.
Several different methods have been developed for cervical ripening and induction of labor. These methods fall into 2 general categories: mechanical and biochemical. Some methods serve both purposes. The primarily mechanical methods include transcervical placement of a Foley catheter (with or without saline infusion) and laminaria. The biochemical methods include misoprostol, prostaglandin estradiol (PGE2) gel, oxytocin infusion, and dinoprostone.
Of the primarily mechanical methods, extra-amniotic saline infusion has received increasing interest in the last 10 years, as evidenced by the number of studies examining its efficacy. At our institution, extra-amniotic saline infusion has been used routinely since 1999 as a primary method of both cervical ripening and labor induction.
Extra-amniotic saline infusion was first described in 1974 by Halbrecht and Blum,1 who used it for cervical ripening in performing second-trimester abortions. Since then, multiple studies have shown its efficacy and safety to be either superior or equal to misoprostol,2,3 laminaria,4 dinoprostone gel,5 and PGE2 gel6,7 for cervical ripening and labor induction in term gravidas.
Several of these studies2,8 have shown an increased risk of intrapartum chorioamnionitis, but none had adequate sample size and power to demonstrate a statistically significant difference. This is most likely secondary to the fact that these studies were powered to examine the efficacy of the induction methods but not the potential complications. In the current study, we sought to estimate the risk of infectious morbidity among patients who underwent cervical ripening and induction of labor by the extra-amniotic saline infusion method and compare these results to other methods.
MATERIALS AND METHODS
This study was conducted at Bellevue Hospital Center, which is a tertiary-care facility affiliated with the New York University School of Medicine. After approval by the institutional review board, a retrospective cohort analysis was performed on records of patients who delivered between August 2000 and December 2002. Three groups were identified: extra-amniotic saline infusion, “other” methods of induction, and spontaneous labor. The charts and the delivery room logbook were reviewed. Patients were included if they had a term delivery at Bellevue Hospital Center during the study period, whether spontaneous or induced. Patients with ruptured membranes before induction who were at less than 37 weeks of gestation or who had an elective cesarean delivery were excluded.
Independent variables included were age, parity, race, maternal weight, delivery mode, neonatal weight, Apgar scores at 1 and 5 minutes, gestational age, intrapartum and postpartum temperatures, change in white blood cells from admission to postpartum day 1, number of exams after rupture of membranes, time from membrane rupture to delivery, and the presence or absence of intrapartum instrumentation (use of either intrauterine pressure catheter or fetal scalp electrode).
Dependent variables included presence or absence of chorioamnionitis (defined as intrapartum maternal temperature elevation of 38°C or greater with or without uterine tenderness, with maternal [more than 100 bpm] or fetal [more than 160 bpm] tachycardia) and postpartum endomyometritis (defined as 2 temperature elevations of 38°C or greater accompanied by uterine tenderness, with or without malodorous lochia, without another source of fever in a patient who did not have chorioamnionitis). Temperatures were measured orally with a TurboTemp® electronic thermometer (IVAC, San Diego, CA).
Extra-amniotic saline infusion subjects were defined as those who had an induction of labor with extra-amniotic saline infusion, irrespective of the indication, performed in accordance with a standardized departmental protocol. The spontaneous labor group included patients who presented to the labor unit in active phase. Active phase was defined as regular contractions and cervical dilatation of 3 cm or greater. Patients were included in the spontaneous group regardless of the subsequent use of oxytocin augmentation, which was provided when clinically indicated. The “other methods” group was defined by subjects who received either oxytocin or misoprostol as the primary method of induction. Patients were included in this group regardless of the subsequent use of clinically indicated oxytocin augmentation once in active phase.
A sample size calculation was performed prospectively to detect a 10% difference in chorioamnionitis using a β = .2 and α = .05 (two-tailed). A baseline chorioamnionitis rate of 13% was presumed based on known rates of infection determined by routine hospital surveillance. Thus, 182 subjects would be required in each group. Descriptive statistics were performed by using one-way analysis of variance for continuous variables and Fisher exact test for nominal variables. Pair-wise comparisons were performed after analysis of variance by using the χ2 test. Multivariable logistic regression analysis was used to control for confounding variables. Significance was defined as a P < .05. Statistical analysis was performed on SPSS for Windows (SPSS Inc, Chicago, IL).
There were 625 charts evaluated. Of these, 171 patients were induced with the extra-amniotic saline infusion method, 190 patients were induced with other methods, and 264 patients presented in spontaneous labor. The mean maternal age, maternal weight, parity, gestational age, neonatal weight, and Apgar scores at both 1 and 5 minutes were similar among all 3 groups (Table 1). There were nonsignificant differences in the racial makeup of the 3 groups.
Table 2 describes differences in mode of delivery. When compared with extra-amniotic saline infusion, there was a lower rate of cesarean delivery for both spontaneously laboring patients (7.2% versus 34.0%; P < .001) and for all other methods of induction (17.9% versus 34.0%; P < .001). There were no differences in the rates of operative vaginal deliveries when comparing extra-amniotic saline infusion with spontaneous labor (P = .29) and extra-amniotic saline infusion with other methods (P = .91).
The time from rupture of membranes to delivery was similar among the 3 groups (Table 3), but pair-wise comparison (not shown) revealed significant differences when comparing extra-amniotic saline infusion to both spontaneously laboring patients (P = .03) and other methods of induction (P < .001). There were no significant differences in the percentage of patients that had internal instrumentation (intrauterine pressure catheter and/or fetal scalp electrode) among the 3 groups. Comparing extra-amniotic saline infusion, spontaneous labor, and other, there were 41.0%, 23.4%, and 48.1% of patients instrumented, respectively. Only extra-amniotic saline infusion and spontaneous labor differed significantly (P = .03) on pair-wise analysis. There were no differences in the number of postrupture exams among the 3 groups. Additionally, intrapartum temperature differed significantly when comparing extra-amniotic saline infusion with spontaneous (37.5°C versus 37.2°C; P = .025) and to other methods (37.5°C versus 37.3°C; P = .02). The rates of group B Streptococcus prophylaxis were similar for all 3 groups.
The multivariable logistic regression model accounted for method of induction, number of exams performed after rupture of membranes, time from rupture to delivery, and instrumentation. Chorioamnionitis was independently predicted by number of exams more than 4 (relative risk [RR] 2.1; 95% confidence interval [CI] 1.3, 3.5; P = .003), time from rupture to delivery more than 12 hours (RR 1.8; 95% CI 1.2, 2.8; P = .008), instrumentation (RR 2.3; 95% CI 1.5, 3.5; P < .001). When adjusted for number of exams, length of rupture, and instrumentation, subjects who were induced using extra-amniotic saline infusion were significantly more likely to develop chorioamnionitis (RR = 2.2; 95% CI 1.4, 4.0; P = .006). Other methods of labor induction (RR 0.8; 95% CI 0.5, 1.5; P = .56) and spontaneous labor (RR 0.9; 95% CI 0.6, 1.4; P = .71) did not independently predict chorioamnionitis in this model.
There were no significant differences in postpartum endomyometritis among the 3 groups (Table 4). The extra-amniotic saline infusion group had nonsignificant trends toward increased rates of postpartum infection versus the spontaneous labor group (RR 2.1; 95% CI 0.7, 5.8) and toward decreased rates versus other methods (RR 0.8; 95% CI 0.7, 5.8).
At our institution, extra-amniotic saline infusion is a commonly used method for both cervical ripening and labor induction. Multiple published studies have shown that extra-amniotic saline infusion is similar in efficacy to other methods of cervical ripening and labor induction.3–7 Our results show a clear association between extra-amniotic saline infusion and rates of chorioamnionitis. These results accounted for important covariates known to be associated with chorioamnionitis, such as length of time from rupture of membranes to delivery9 and number of vaginal exams.10,11
These results are concordant with other studies, which were not designed to adequately evaluate potential complications of extra-amniotic saline infusion. The largest of the available studies, by Guinn et al 2000,7 prospectively evaluated extra-amniotic saline infusion versus laminaria and PGE2 in 444 patients. Of those, 169 subjects were exposed to extra-amniotic saline infusion and 165 to laminaria. There was an intrapartum infection rate of 17.8% for extra-amniotic saline infusion (30 of 107) versus 11.5% (19 of 104) in laminaria patients. The chorioamnionitis rates were not reported for PGE2 gel in this report. The increased rates of chorioamnionitis for extra-amniotic saline infusion versus laminaria did not reach statistical significance (P = .11). In another relatively large study, reported by Vengalil et al3 in 1998, 248 patients were evaluated prospectively for efficacy of extra-amniotic saline infusion versus misoprostol. The chorioamnionitis rates were 12.5% (16 of 128) and 6.7% (8 of 120) for extra-amniotic saline infusion and misoprostol, respectively, with a P value of .12. Unlike these reports, our study was specifically designed to evaluate the potential risk of infectious morbidity with extra-amniotic saline infusion.
The main limitation of this study is its retrospective design. Although we were able to control for many confounding variables, we could not, however, evaluate baseline cervical status by Bishop score because relatively few charts adequately documented all of the elements of this scoring system. At present, we do not have detailed information on the frequency of other methods of induction, although the vast majority of inductions at our institution not performed with extra-amniotic saline infusion are performed with either oxytocin or misoprostol. We acknowledge our institutional rate of chorioamnionitis is quite high and may limit external validity. However, our high baseline institutional rate of chorioamnionitis in spontaneously laboring subjects (13%) should have blunted the effects of elevated infection risk in extra-amniotic saline infusion subjects. Furthermore, there was a significantly shorter time period between rupture of membranes and delivery in extra-amniotic saline infusion subjects versus both spontaneously laboring subjects and those induced with other methods. Although length of time membranes are ruptured does independently predict risk of chorioamnionitis, extra-amniotic saline infusion subjects had increased rates of infection despite decreased exposure time. Finally, when extra-amniotic saline infusion was compared with other methods of induction, there was no difference in the number of postrupture exams or in the exposure to intrauterine pressure catheters of fetal scalp electrodes. As a result, we believe extra-amniotic saline infusion is a direct cause of chorioamnionitis.
Abramovici et al12 studied 197 subjects undergoing induction; 98 were with misoprostol and 99 with a Foley balloon without saline infusion. The 2 groups had similar rates of chorioamnionitis, 14.3% and 15.2%, respectively. It has been demonstrated that extra-amniotic infusion of saline infiltrates the entire choriodecidual space13 potentially inoculating the uterine cavity with vaginal flora. Furthermore, inflammatory mediators, such as prostaglandins, are released through stimulation of the decidua by saline.14 It is possible that saline infusion, rather than the simple presence of a foreign body, may be responsible for initiating the inflammatory cascade.
The extra-amniotic saline infusion method of induction may be associated with an increased risk of chorioamnionitis, but a prospective study is necessary to definitively establish causation. Until such results are available, the potential risk of infection should be considered when choosing this modality.
1. Halbrecht I, Blum M. Induction of mid-trimester abortion by means of extra-amniotic infusion of an isotonic saline solution combined with intravenous oxytocin drip infusion. Contraception 1974;10:637–43.
2. Buccellato CA, Stika CS, Frederiksen MC. A randomized trial of misoprostol versus extra-amniotic sodium chloride infusion with oxytocin for induction of labor. Am J Obstet Gynecol 2000;182:1039–44.
3. Vengalil SR, Guinn DA, Olabi NF, Burd LI, Owen J. A randomized trial of misoprostol and extra-amniotic saline infusion for cervical ripening and labor induction. Obstet Gynecol 1998;91:774–9.
4. Lin A, Kupferminc M, Dooley SL. A randomized trial of extra-amniotic saline infusion versus laminaria for cervical ripening. Obstet Gynecol 1995;86:545–9.
5. Goldman JB, Wigton TR. A randomized comparison of extra-amniotic saline infusion and intracervical dinoprostone gel for cervical ripening. Obstet Gynecol 1999;93:271–4.
6. Rouben D, Arias F. A randomized trial of extra-amniotic saline infusion plus intracervical Foley catheter balloon versus prostaglandin E2 vaginal gel for ripening the cervix and inducing labor in patients with unfavorable cervices. Obstet Gynecol 1993;82:290–4.
7. Guinn DA, Goepfert AR, Christine M, Owen J, Hauth JC. Extra-amniotic saline, laminaria, or prostaglandin E(2) gel for labor induction with unfavorable cervix: a randomized controlled trial. Obstet Gynecol 2000;96:106–12.
8. Lyndrup J, Nickelsen C, Weber T, Molnitz E, Guldbaek E. Induction of labour by balloon catheter with extra-amniotic saline infusion (BCEAS): a randomised comparison with PGE2 vaginal pessaries. Eur J Obstet Gynecol Reprod Biol 1994;53:189–97.
9. Daikoku NH, Kaltreider DF, Khouzami VA, Spence M, Johnson JW. Premature rupture of membranes and spontaneous preterm labor: maternal endometritis risks. Obstet Gynecol 1982;59:13–20.
10. Lewis DF, Major CA, Towers CV, Asrat T, Harding JA, Garite TJ. Effects of digital vaginal examinations on latency period in preterm premature rupture of membranes. Obstet Gynecol 1992;80:630–4.
11. Alexander JM, Mercer BM, Miodovnik M, Thurnau GR, Goldenberg RL, Das AF, et al. The impact of digital cervical examination on expectantly managed preterm rupture of membranes. Am J Obstet Gynecol 2000;183:1003–7.
12. Abramovici D, Goldwasser S, Mabie B, Mercer B, Goldwasser R, Sibai B. A randomized comparison of oral misoprostol versus Foley catheter and oxytocin for indcution of labor at term. Am J Obstet Gynecol 1999;181:1108–12.
13. Gustavii B, Gothlin J. Radiographic study of extra-amniotically injected hypertonic saline in therapeutic abortion. Acta Obstet Gynecol Scand 1971;50:315–20.
14. Gustavii B. Studies on the mode of action of intra-amniotically and extra-amniotically injected hypertonic saline in therapeutic abortion. Acta Obstet Gynecol Scand Suppl 1973;1–22.