- Email: [email protected]
Interleukin-6, but not relaxin, predicts outcome of rescue cerclage in women with cervical incompetence
American Journal of Obstetrics and Gynecology (2004) 191, 784e9
www.elsevier.com/locate/ajog
Interleukin-6, but not relaxin, predicts outcome of rescue cerclage in women with cervical incompetence Keun-Young Lee, MD,* Hyun-Ah Jun, MD, Hong-Bae Kim, MD, Sung-Won Kang, MD Department of Obstetrics and Gynecology, Hallym University, Seoul, Korea Received for publication December 4, 2003; revised April 7, 2004; accepted April 23, 2004
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– KEY WORDS Cervical incompetence Interleukin-6 Relaxin
Objective: We investigated the potential roles of relaxin and subclinical intra-amniotic inflammation by quantitating amniotic fluid relaxin and interleukin-6 concentrations for the prediction of outcome of rescue cerclage in women with cervical incompetence. Study design: Cervical incompetence was diagnosed when cervical dilatation exceeded 2 cm with intact but bulging membranes and no detectable uterine activity. Each woman underwent amniocentesis to facilitate the performance of a rescue cerclage between 15 and 27 weeks of gestation (n = 40 women). Forty-five additional women who underwent amniocentesis for chromosomal testing between 16 and 27 weeks of gestation served as a control group. All control patients were delivered of chromosomally normal infants atO37 weeks of gestation. All cases and control patients were singleton gestations. Interleukin-6 and relaxin were determined in all amniotic fluid samples by enzyme-linked immunosorbent assay. Results: Amniotic fluid interleukin-6 levels were significantly higher in women with cervical incompetence than in control patients (control patients, 50.4 pg/mL [range, 19.4-97.4 pg/mL] vs cervical incompetence patients, 5459.1 pg/mL [range, 1131.4-14425.7 pg/mL] ; P ! .001). In contrast to interleukin-6, relaxin levels did not differ between the 2 groups (control patients, 67.5 pg/ mL [range, 35.1-153.5 pg/mL] vs cervical incompetence patients, 45.6 pg/mL [range, 30.1-75.5 pg/ mL]; P = .061). There was a significant difference in interleukin-6 levels in women with shorter latencies (P ! .01 for all latency intervals that were examined: delivery within 24 hours, 3 days, 7 days, before 33 and 37 completed weeks of gestation). Linear regression analysis with the use of the latency interval from cerclage to delivery as the dependent and with interleukin-6 as the independent variable revealed a significant inverse relationship (r = ÿ0.62; P ! .001 after log transformation of interleukin-6). There was no relationship on regression analysis between relaxin and the latency interval. Conclusion: Amniotic fluid interleukin-6 is increased in patients with cervical incompetence, which suggests that subclinical inflammation may contribute to cervical incompetence. Further,
Supported by a grant from Hallym University Foundation Medical Research. * Reprint requests: Keun-Young Lee, MD, Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, Hallym University College of Medicine, Kangnam Sacred Heart Hospital, Youngdeungpo-Ku, Daerim-1 Dong, 948-1, Seoul, 150-071, Korea. E-mail: [email protected]
0002-9378/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ajog.2004.04.019
785
Lee et al
an elevated interleukin-6 level predicts a cerclage short-latency interval between cerclage and delivery. In contrast with interleukin-6, amniotic fluid relaxin does not appear to contribute to cervical incompetenceeinduced cervical dilation. Ó 2004 Elsevier Inc. All rights reserved.
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cervical incompetence, which is defined as painless dilation of the cervix in the second trimester, is an established cause of pregnancy loss. However, its cause is poorly understood and has been attributed variously to congenital factors, physical injury, or trauma and/ or biochemical influences.1 Hyper relaxinemia is one putative endocrine cause of cervical incompetence.1 The role of relaxin in human pregnancy and parturition is unclear, but its action of interest is an effect on cervical ripening before labor.2 Relaxin-induced cervical ripening has been studied extensively in rats where they are similar to the human collagen fibrils spilled, the concentration of collagen decreases, and the concentration of dermatan sulfate increases (ie, the dermatan sulfate to collagen ratio increases).3 Hyper relaxinemia is postulated to cause premature labor, preterm delivery, or cervical incompetence.1,4 There have been several studies on the relationship between serum relaxin and preterm birth1,4 but not on the relationship between serum relaxin and cervical length.5 Eddie et al6 suggested ovarian relaxin was not essential for dilatation of the cervix. Several investigators observed that purified porcine relaxin advances cervical softening, effacement, and dilation and delivery when administered either intravaginally or intracervically.7,8 These observations suggest that local relaxin might contribute to cervical ripening before labor. Other studies indicate that human relaxin that originates from the decidua and placenta is involved in controlled collagen remodeling in the fetal membrane.9,10 It is also suggested that relaxin in the amniotic fluid may be derived from the decidualized endometrium rather than from maternal circulation.11 We therefore investigated the amniotic fluid relaxin concentrations in women with cervical incompetence and the relationship between the relaxin concentrations and the success of rescue cerclage. In addition to relaxin, cervical ripening involves cellular and humoral components that are normally associated with inflammation12 (such as the invasion of inflammatory cells, which includes macrophages, neutrophils,13 mast cells,14 and eosinophils15) into the cervical stroma. Neutrophils and eosinophils degranulate during cervical ripening and are surrounded typically by a halo of degraded collagen. The eosinophils release well-defined cell toxic substances; the neutrophils release leukocyte collagenase and elastase.16 Further, drugs that block mast cell degranulation blunt the normal process
of cervical ripening in rats.13 In addition, the activation of a number of autocoids and hormones also resembles an inflammatory reaction.17 Considerable evidence supports a role for proinflammatory cytokines and chemokines in cervical ripening.15 Increases in interleukin-6 (IL-6) is reported in the amniotic fluid of women with chorioamnionitis18 but has not been studied apparently in women with incompetent cervix. Thus, we investigated the potential roles of intra-amniotic relaxin and IL-6 in women with the diagnosis of cervical incompetence before rescue cerclage by quantitating amniotic fluid relaxin IL-6 concentrations for the prediction of rescue cerclage outcome.
Material and methods The subjects were sampled prospectively before emergency cervical cerclage that was performed May 1998 to September 2001. The study was approved by our local ethics committee and institutional review board; informed consent was obtained from each women. Cervical incompetence was diagnosed after painless cervical dilatation (O2 cm) and bulging but intact membranes at the level of the external cervical os. In each instance, there was no detectable uterine activity for a mean of 4 hours (range, 1-24 hours) before cerclage. Each woman had a singleton pregnancy and no vaginal bleeding and was afebrile at surgery. Their gestational ages ranged from 15 and 27 weeks (median gestational age, 22.9 weeks). An emergency cervical cerclage (McDonald-type purse string by 5-mm Mersilene tape) under general anesthesia was performed after amnioreduction. The suture was removed, and vaginal birth was accomplished in cases in which labor or clinical infection developed. Rupture of the membranes per se was not an indication for the removal of the cerclage. A cohort of 45 women with sonographically normal pregnancy who underwent amniocentesis for chromosomal testing between 16 and 27 weeks of gestation (median, 20.3 weeks) provided a control cohort during the same period. All control patients were delivered of chromosomally normal infants at O37 weeks of gestation. The aliquots of amniotic fluid from all women were centrifuged and stored in polypropylene tubes at ÿ70(C until the batch was assayed to reduce interassay variation.
786
Lee et al value of !.05 was considered to indicate statistically significant differences among medians. Linear regression analyses were performed with the time from cerclage to delivery (latency interval) as the dependent variable and with IL-6 and relaxin concentrations as independent variables. Receiver operator curves were generated to determine the optimal cutoff point for each analyte, the curves of the 2 analytes compared, and the survival analyses conducted. Survival analysis was performed using Prism software (GraphPad, San Diego, Calif).
Figure 1 The extents of the box represent the 25th and 75th percentile, and the line inside the box is the 50th percentile. The extents of the whiskers represent the 5th and 95th percentiles. The circles are data outside 5th or 95th percentiles. CRL, Patients who underwent genetic amniocentesis; CI, patients who underwent emergency cerclage. R = 0.62; P ! .001.
Table I
Demographic and obstetric characteristics Group
Characteristic
Cervical incompetence (n = 40)
Maternal age (y)* Nulliparity (n) Gestational age (wk)*
31.5 G 8.5 11 (27%) 22.9 G 7.9
Control (n = 45)
P value
33.1 G 4.8 19 (33%) 20.3 G 3.2
.82 .51 .75
* Data are given as mean G SD.
Amniotic fluid assay of relaxin and IL-6 Amniotic fluid immunoreactive relaxin concentrations were determined with a specific human relaxin enzymelinked immunosorbent assay. Purified human H2 relaxin standard were provided by ImmunDiagnostik AG (Bensheim, Germany). We used goat anti-human relaxin antibody and a horseradish peroxidase-conjugated rabbit antihuman relaxin antibody as a detection system. The lower limit of sensitivity of the assay was 20 pg/mL. The intra-assay and interassay coefficients of variation were both !10%. IL-6 was quantitated in amniotic fluid with a commercially available enzyme-linked immunosorbent assay (R&D Systems, Minneapolis, Minn). The sensitivity of the IL-6 assay was !1.0 pg/ mL. Interassay and intra-assay coefficients of variation were !10% for each analyte.
Statistical analysis After first being tested for normal distribution with a Kolmogorov-Smirnov test, statistical comparisons were performed using the Mann-Whitney rank sum test. The laboratory results are presented as their median, followed by the 25th/75th percentiles. A probability
Results All demographic and laboratory data were available in 40 women with cervical incompetence and 45 control patients. The 2 groups were similar in selected demographic characteristics: Korean, gestational age at amniocentesis, percent nulliparity, and maternal age. The demographic characteristics of the study population are shown in Table I. Women who fulfilled the cerclage criteria had higher amniotic fluid IL-6 levels compared women who underwent genetic amniocentesis (control [50.4 pg/mL (range, 19.4-97.4 pg/mL)] vs cervical incompetence [5459.1 pg/ mL (range, 1131.4-14425.7 pg/mL)]; P ! .001). In contrast with IL-6, levels of relaxin did not show significant difference between the 2 groups (control [67.5 pg/mL (range, 35.1-153.5 pg/mL)] vs cervical incompetence [45.6 pg/mL (range, 30.1-75.5 pg/mL)]; P = .061; Figure 1). We next investigated whether the IL-6 level correlated to clinical outcome as reflected in the latency interval. We defined several clinically relevant latency intervals: 24 hours, 3 days, 7 days, before 33 and 37 completed weeks of gestation. The median amniotic fluid IL-6 and relaxin levels by latency group are shown in Table II. There was a significant difference in IL-6 levels in women with shorter latencies and preterm delivery (P ! .01 for all cut-offs) compared with those women with longer latency intervals. There was no significant difference in relaxin levels in women with shorter latencies and preterm delivery compared with those women with longer latency intervals, except in the delivery before 33 weeks of gestation group. The median relaxin level in women who were delivered before 33 weeks of gestation (55.1 pg/mL [range, 31.3-109.2 pg/mL]) was significantly higher than in those women who were delivered after 33 weeks of gestation (33.9 pg/mL [range, 18.0-45.1 pg/mL]; P = .049). Linear regression analysis (Figure 2) with latency as the dependent variable and IL-6 as the independent variable revealed a significant inverse relationship (r = ÿ0.36; P = .021 for linear data; and r = ÿ0.62; P ! .001 after log transformation of IL-6). There was no relationship between relaxin and latency interval
787
Lee et al Table II
The relationship between the median amniotic fluid IL-6 and relaxin levels and several outcome relevant latency intervals
Intervals
Outcome
IL-6
Delivery within 24 hr
No (n = 28) Yes (n = 12) No (n = 22) Yes (n = 18) No (n = 20) Yes (n = 20) No (n = 11) Yes (n = 29) No (n = 4) Yes (n = 36)
2300.0 28218.1 1511.8 14425.7 1511.8 13793.0 810.1 10009.1 235.4 7523.8
Delivery within 3 d Delivery within 7 d Delivery !33 wk Delivery !37 wk
P value (444.9-9002.2) (7737.5-44489.4) (356.1-7463.5) (5276.9-43690.2) (394.0-6552.4) (4146.8-38766.4) (243.3-4654.2) (2603.0-27059.7) (33.1-558.7) (1950.9-20396.5)
!.001 !.001 !.001 .005 .005
Relaxin 43.5 48.8 43.5 47.3 41.9 48.8 33.9 55.1 24.0 46.5
(25.8-67.9) (33.0-91.4) (29.1-66.5) (31.1-105.5) (21.7-60.2) (31.3-112.9) (18.0-45.1) (31.3-109.2) (10.5-58.4) (31.3-75.5)
P value .452 .523 .190 .049 .199
The amniotic fluid interleukin-6 level, but not the relaxin level, predicts the success of rescue cerclage.
either before (r = 0.15, P = .337) or after log transformation (r = 0.269, P = .093). A receiver operator curve analysis (Figure 3) indicated that a preoperative IL-6 level ofO1700 pg/mL had a sensitivity of 100% (95% CI, 100-100) and a specificity of 59.1% (95% CI, 36.4-79.3) for the prediction of a latency of !3 days. An IL-6 value of O12326 pg/mL predicted delivery within 24 hours with 75% sensitivity (95% CI, 42.8-94.2) and 82% specificity (95% CI, 63.1-93.9). An IL-6 value !748 pg/mL predicted a delivery delay until 33 weeks of gestation with 85% sensitivity and 67% specificity (95% CI, 71.5-100). Survival analysis (Figure 4) was conducted that compared latency intervals of patients with preoperative IL6 levels of O1700 pg/mL to those women with preoperative IL-6 levels !1700 pg/mL. The median latency interval in patients in the first group was 2 days, which was significantly shorter than the second group whose median latency interval was 35 days (log-rank test chisquared test, 13.17; P = .0003). There were no differences between groups for gestational age at amniocentesis (P = .33), patient age (P = .23), and percent nulliparity (P = .85).
Comment We investigated several factors that might be associated with cervical incompetence and the ultimate success of a rescue cerclage. Although the study design prevents any comments on the chronology of the detected changes, the results suggest that the measurement of IL-6 may be of clinical value in the treatment of cervical incompetence that requires a rescue cerclage. In 2000, Mays et al19 observed improved outcomes with the use of amniocentesis to exclude women with subclinical intra-amniotic infection, by the measurement of amniotic fluid glucose and lactic dehydrogenase levels as predictors of histologic chorioamnionitis. The mechanism underlying cervical dilation in women with cervical incompetence is unknown. Hyper
Figure 2 Regression analysis of the duration from the cervical incompetence procedure and delivery (latency) versus the amniotic fluid IL-6 levels. R = 0.62; P ! .001.
relaxinemia is one putative endocrine cause of cervical incompetence.1 Relaxin is a peptide hormone that is secreted by the corpus luteum and that is known to trigger connective tissue remodeling in several mammalian species, including human.20 Relaxin increases collagen turnover in human dermal fibroblast by stimulating collagenase expression and modulating collagen synthesis.3 Iams et al5 measured serum relaxin and the sonographic cervical length at 24 and 28 weeks of gestation in spontaneous twin pregnancies. There was a significant relationship between serum relaxin and preterm birth but not between relaxin and cervical length. This suggests
788
Lee et al
Figure 4 Survival analysis of the duration from the cervical incompetence procedure and delivery (latency).
Figure 3 Receiver operating curve analysis of amniotic fluid IL-6 levels and the duration from the cervical incompetence procedure and delivery: Delivery within 24 hours (area under the curve, 0.836; standard error, 0.68; 95% CI, 0.703-0.969; P = .001); delivery within 3 days (area under the curve, 0.871; standard error, 0.054; 95% CI, 0.765-0.977; P ! .001); delivery before 33 weeks of gestation (area under the curve, 0.858; standard error, 0.073; 95% CI, 0.715-1.00; P = .006).
that increased relaxin levels cannot explain the inverse correlation between cervical length and the prevalence of spontaneous preterm birth in women with spontaneous twin pregnancies. Further, several investigators have reported that purified porcine relaxin advances cervical softening, effacement, and dilation and shortens the interval to delivery when administered either intravaginally or intracervically.7,8 These observations suggest local relaxin may contribute to cervical ripening before labor. Indeed, human relaxin of decidual and placental origin is involved in controlled collagen remodeling in the fetal membrane.9,10 Eddie et al6 suggested that either very low levels of circulating relaxin are sufficient for its biologic effects or that decidual and placental relaxins are of greater importance to uterine and cervical function at term than relaxin from corpus luteum. Amniotic fluid relaxin may be derived from the decidualized endometrium rather than the maternal circulation.11 We therefore investigated relaxin in the amniotic fluid concentration in women with cervical incompetence and the relationship between relaxin concentration and the success of rescue cerclage. We found no significant increase in amniotic fluid relaxin in women with cervical incompetence. This suggests that relaxin is unlikely to contribute to the mechanism of cervical incompetence, at least as reflected in the amniotic fluid.
We also had expected that relaxin levels in women with cervical incompetence might be higher than those levels in the control group but were surprised to find that the differences in relaxin concentration between the 2 groups was not significant (P = .061). In point, the amniotic fluid relaxin level of the cervical incompetence group (45.6 pg/mL [range, 30.1-75.5 pg/mL]) was lower than that of the control group (67.5 pg/mL [range, 35.1-153.5 pg/mL]; P = not significant). There was no significant difference in relaxin between women with shorter latencies and preterm delivery compared with those women with longer latency intervals, except in the group that were delivered before 33 weeks of gestation. The median relaxin level in women who were delivered !33 weeks of gestation was significantly higher than in those who were deliveredO33 weeks of gestation. Interestingly, we also found the amniotic fluid relaxin level of women who were delivered !37 weeks of gestation was higher than those women who were deliveredO37 weeks of gestation (P = not significant). This observation suggests that further study of amniotic fluid relaxin in the future might be revealing. The biochemical events of cervical ripening are complex and include degradation and/or remodeling of extracellular matrix proteins and glycoproteins, disruption of tightly aligned collagen fibrils, and increased hydration caused by hyaluronan. The process has been likened to an inflammatory reaction12 that is associated with the catabolism of cervical extracellular matrix by enzymes that are released from infiltrating leukocytes. Indeed, there is an influx of inflammatory cells during cervical ripening into the cervical stroma, which includes macrophages, neutrophils,13 mast cells,14 and eosinophils.15 Romero et al21 compared the values of amniotic fluid tests in the detection of microbial invasion of the amniotic cavity and concluded that the IL-6 concentrations
Lee et al in amniotic fluid are better indicators of microbial invasion of the amniotic cavity than the amniotic fluid Gram stains, glucose concentrations, or white blood cell counts. We investigated IL-6 and demonstrated that it was increased significantly in women with cervical incompetence. We cannot determine presently whether the elevation is part of the pathophysiologic nature of cervical incompetence or the end result of it. But, we can say that the higher the IL-6 level before cerclage, the shorter the latency interval. Few women with a grossly elevated IL6 level have any likelihood of reaching viability should the cerclage be placed at !22 weeks of gestation. The mechanism underlying the increase in IL-6 level may be bacterial in origin, but antibiotics have failed to improve the outcome of emergency cerclage in the past. The identification of women with an abnormally elevated IL-6 level could lead to future therapies that would seek to interfere with either the initiation of or the continuation of the inflammatory process. Although the exact source of the cytokines in the amniotic fluid of women with cervical incompetence is unclear, a growing body of evidence suggests that ascending genital tract infection followed by deciduitis may be part of the sequence.22 The observed relationship between IL-6 level and the outcome of pregnancy is consistent with them having a pathogenic role in cervical incompetence.
References 1. Iams JD. Cervical incompetence. In: Creasy R, Resnik R, editors. Maternal fetal medicine. 4th ed. Philadelphia: Saunders; 1999. p. 445-64. 2. MacLennan AH. The role of relaxin in human reproduction. Clin Reprod Fertil 1983;2:77-95. 3. Downing SJ, Sherwood OD. The physiological role of relaxin in the pregnant rat: IV, the influence of relaxin on cervical collagen and glucosaminoglycans. Endocrinology 1986;118:471-9. 4. Weiss G, Goldsmith LT, Sachdev R, Hagen SV, Lederer K. Elevated first-trimester serum relaxin concentrations in pregnant women after ovarian stimulation predict prematurity risk and preterm delivery. Obstet Gynecol 1993;82:821-8. 5. Iams JD, Goldsmith LT, Weiss G. The preterm prediction study: maternal serum relaxin, sonographic cervical length, and spontaneous preterm birth in twins. J Soc Gynecol Investig 2001;8:39-42. 6. Eddie LW, Cameron IT, Leeton JF, Healy DL, Renou P. Ovarian relaxin is not essential for dilatation of cervix. Lancet 1990;336: 243.
789 7. Evans MI, Dougan MB, Moawad AH, Evans WJ, Bryant-Greenwood GD, Greenwood FC. Ripening of the human cervix with porcine ovarian relaxin. Am J Obstet Gynecol 1983;147:410-4. 8. Bell RJ, Permezel M, McLennan A, Hughes C, Healy D, Brennecke S. A randomized, double-blind, placebo-controlled trial of the safety of vaginal recombinant human relaxin for cervical ripening. Obstet Gynecol 1993;82:328-33. 9. Bogic LV, Mandel M, Bryant-Greenwood GD. Relaxin gene expression in human reproductive tissues by in situ hybridization. J Clin Endocrinol Metab 1995;80:130-7. 10. Bryant-Greenwood GD, Schwabe C. Human relaxins: chemistry and biology. Endocr Rev 1994;15:5-26. 11. Johnson MR, Abbas A, Nicolaides KH, Ligheman SL. Distribution of relaxin between human maternal and fetal circulation and amniotic fluid. J Endocr 1992;134:313-7. 12. Liggins G. Cervical ripening as an inflammatory reaction. In: Ellwood DA, Anderson ABM, editors. The cervix in pregnancy and labour: clinical and biochemical investigations. Edinburgh, Scotland: Churchill-Livingstone; 1981. p. 1-12. 13. Spanggaard H, Knudsen UB, Uldbjerg N, Jeziorska M, Woolley DE, Danielsen CC. Mast cells in cervical ripening-an immunohistochemical and biochemical study in rats. Eur J Obstet Gynecol Reprod Biol 1997;73:91-7. 14. Tschesche H. Human neutrophil collagenase. Methods Enzymol 1995;248:431-49. 15. Osmers RG, Adelmann-Grill BC, Rath W, Stuhlsatz HW, Tschesche H, Kuhn W. Biochemical events in cervical ripening dilatation during pregnancy and parturition. J Obstet Gynecol 1995;21:185-94. 16. Uldbjerg N, Forman A. Biomechanical and biochemical changer of the uterus and cervix during pregnancy. In: Reece AE, Hobbins JC, editors. Medicine of the fetus and mother. 2nd ed. Philadelphia: Lippincott-Raven; 1999. p. 921-31. 17. Uldbjerg N, Ulmsten U, Ekman G. The physiological role of eicosanoids in controlling the form and function of the cervix. In: Hiller K, editor. Advances in eicosanoid research. Lancaster (PA): MTP Press; 1987. p. 163-83. 18. Yoon BH, Romero R, Kim JC, Jun JK, Gomez R, Choi JH, et al. Amniotic fluid interleukin-6: a sensitive test for antenatal diagnosis of acute inflammatory lesions of preterm placenta and prediction of perinatal morbidity. Am J Obstet Gynecol 1995;172:960-70. 19. Mays JK, Figueroa R, Shah J, Khakoo H, Kaminsky S, Thjani N. Amniocentesis for selection before rescue cerclage. Obstet Gynecol 2000;95:652-5. 20. Sherwood OD. Relaxin. In: Knobil E, Neill JD, Greenwald GS, Markert C, Pfaff DW, editors. The physiology of reproduction. 2nd ed. New York: Raven Press; 1994. p. 861-1009. 21. Romero R, Yoon BH, Mazor M, Gomez R, Diamond MP, Kenny JS, et al. The diagnostic and prognostic value of amniotic fluid white blood cell count, glucose, interleukin- 6, and Gram stain in patients with preterm labor and intact membranes. Am J Obstet Gynecol 1993;169:805-16. 22. Romero R, Gonzalez R, Sepulveda W, Brandt F, Ramirez M, Sorokin Y, et al. VIII, Microbial invasion of the amniotic cavity in patients with suspected cervical incompetence: prevalence and clinical significance. Am J Obstet Gynecol 1992;167:1086-91.
www.elsevier.com/locate/ajog
Interleukin-6, but not relaxin, predicts outcome of rescue cerclage in women with cervical incompetence Keun-Young Lee, MD,* Hyun-Ah Jun, MD, Hong-Bae Kim, MD, Sung-Won Kang, MD Department of Obstetrics and Gynecology, Hallym University, Seoul, Korea Received for publication December 4, 2003; revised April 7, 2004; accepted April 23, 2004
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––– KEY WORDS Cervical incompetence Interleukin-6 Relaxin
Objective: We investigated the potential roles of relaxin and subclinical intra-amniotic inflammation by quantitating amniotic fluid relaxin and interleukin-6 concentrations for the prediction of outcome of rescue cerclage in women with cervical incompetence. Study design: Cervical incompetence was diagnosed when cervical dilatation exceeded 2 cm with intact but bulging membranes and no detectable uterine activity. Each woman underwent amniocentesis to facilitate the performance of a rescue cerclage between 15 and 27 weeks of gestation (n = 40 women). Forty-five additional women who underwent amniocentesis for chromosomal testing between 16 and 27 weeks of gestation served as a control group. All control patients were delivered of chromosomally normal infants atO37 weeks of gestation. All cases and control patients were singleton gestations. Interleukin-6 and relaxin were determined in all amniotic fluid samples by enzyme-linked immunosorbent assay. Results: Amniotic fluid interleukin-6 levels were significantly higher in women with cervical incompetence than in control patients (control patients, 50.4 pg/mL [range, 19.4-97.4 pg/mL] vs cervical incompetence patients, 5459.1 pg/mL [range, 1131.4-14425.7 pg/mL] ; P ! .001). In contrast to interleukin-6, relaxin levels did not differ between the 2 groups (control patients, 67.5 pg/ mL [range, 35.1-153.5 pg/mL] vs cervical incompetence patients, 45.6 pg/mL [range, 30.1-75.5 pg/ mL]; P = .061). There was a significant difference in interleukin-6 levels in women with shorter latencies (P ! .01 for all latency intervals that were examined: delivery within 24 hours, 3 days, 7 days, before 33 and 37 completed weeks of gestation). Linear regression analysis with the use of the latency interval from cerclage to delivery as the dependent and with interleukin-6 as the independent variable revealed a significant inverse relationship (r = ÿ0.62; P ! .001 after log transformation of interleukin-6). There was no relationship on regression analysis between relaxin and the latency interval. Conclusion: Amniotic fluid interleukin-6 is increased in patients with cervical incompetence, which suggests that subclinical inflammation may contribute to cervical incompetence. Further,
Supported by a grant from Hallym University Foundation Medical Research. * Reprint requests: Keun-Young Lee, MD, Division of Maternal-Fetal Medicine, Department of Obstetrics and Gynecology, Hallym University College of Medicine, Kangnam Sacred Heart Hospital, Youngdeungpo-Ku, Daerim-1 Dong, 948-1, Seoul, 150-071, Korea. E-mail: [email protected]
0002-9378/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.ajog.2004.04.019
785
Lee et al
an elevated interleukin-6 level predicts a cerclage short-latency interval between cerclage and delivery. In contrast with interleukin-6, amniotic fluid relaxin does not appear to contribute to cervical incompetenceeinduced cervical dilation. Ó 2004 Elsevier Inc. All rights reserved.
–––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––––
Cervical incompetence, which is defined as painless dilation of the cervix in the second trimester, is an established cause of pregnancy loss. However, its cause is poorly understood and has been attributed variously to congenital factors, physical injury, or trauma and/ or biochemical influences.1 Hyper relaxinemia is one putative endocrine cause of cervical incompetence.1 The role of relaxin in human pregnancy and parturition is unclear, but its action of interest is an effect on cervical ripening before labor.2 Relaxin-induced cervical ripening has been studied extensively in rats where they are similar to the human collagen fibrils spilled, the concentration of collagen decreases, and the concentration of dermatan sulfate increases (ie, the dermatan sulfate to collagen ratio increases).3 Hyper relaxinemia is postulated to cause premature labor, preterm delivery, or cervical incompetence.1,4 There have been several studies on the relationship between serum relaxin and preterm birth1,4 but not on the relationship between serum relaxin and cervical length.5 Eddie et al6 suggested ovarian relaxin was not essential for dilatation of the cervix. Several investigators observed that purified porcine relaxin advances cervical softening, effacement, and dilation and delivery when administered either intravaginally or intracervically.7,8 These observations suggest that local relaxin might contribute to cervical ripening before labor. Other studies indicate that human relaxin that originates from the decidua and placenta is involved in controlled collagen remodeling in the fetal membrane.9,10 It is also suggested that relaxin in the amniotic fluid may be derived from the decidualized endometrium rather than from maternal circulation.11 We therefore investigated the amniotic fluid relaxin concentrations in women with cervical incompetence and the relationship between the relaxin concentrations and the success of rescue cerclage. In addition to relaxin, cervical ripening involves cellular and humoral components that are normally associated with inflammation12 (such as the invasion of inflammatory cells, which includes macrophages, neutrophils,13 mast cells,14 and eosinophils15) into the cervical stroma. Neutrophils and eosinophils degranulate during cervical ripening and are surrounded typically by a halo of degraded collagen. The eosinophils release well-defined cell toxic substances; the neutrophils release leukocyte collagenase and elastase.16 Further, drugs that block mast cell degranulation blunt the normal process
of cervical ripening in rats.13 In addition, the activation of a number of autocoids and hormones also resembles an inflammatory reaction.17 Considerable evidence supports a role for proinflammatory cytokines and chemokines in cervical ripening.15 Increases in interleukin-6 (IL-6) is reported in the amniotic fluid of women with chorioamnionitis18 but has not been studied apparently in women with incompetent cervix. Thus, we investigated the potential roles of intra-amniotic relaxin and IL-6 in women with the diagnosis of cervical incompetence before rescue cerclage by quantitating amniotic fluid relaxin IL-6 concentrations for the prediction of rescue cerclage outcome.
Material and methods The subjects were sampled prospectively before emergency cervical cerclage that was performed May 1998 to September 2001. The study was approved by our local ethics committee and institutional review board; informed consent was obtained from each women. Cervical incompetence was diagnosed after painless cervical dilatation (O2 cm) and bulging but intact membranes at the level of the external cervical os. In each instance, there was no detectable uterine activity for a mean of 4 hours (range, 1-24 hours) before cerclage. Each woman had a singleton pregnancy and no vaginal bleeding and was afebrile at surgery. Their gestational ages ranged from 15 and 27 weeks (median gestational age, 22.9 weeks). An emergency cervical cerclage (McDonald-type purse string by 5-mm Mersilene tape) under general anesthesia was performed after amnioreduction. The suture was removed, and vaginal birth was accomplished in cases in which labor or clinical infection developed. Rupture of the membranes per se was not an indication for the removal of the cerclage. A cohort of 45 women with sonographically normal pregnancy who underwent amniocentesis for chromosomal testing between 16 and 27 weeks of gestation (median, 20.3 weeks) provided a control cohort during the same period. All control patients were delivered of chromosomally normal infants at O37 weeks of gestation. The aliquots of amniotic fluid from all women were centrifuged and stored in polypropylene tubes at ÿ70(C until the batch was assayed to reduce interassay variation.
786
Lee et al value of !.05 was considered to indicate statistically significant differences among medians. Linear regression analyses were performed with the time from cerclage to delivery (latency interval) as the dependent variable and with IL-6 and relaxin concentrations as independent variables. Receiver operator curves were generated to determine the optimal cutoff point for each analyte, the curves of the 2 analytes compared, and the survival analyses conducted. Survival analysis was performed using Prism software (GraphPad, San Diego, Calif).
Figure 1 The extents of the box represent the 25th and 75th percentile, and the line inside the box is the 50th percentile. The extents of the whiskers represent the 5th and 95th percentiles. The circles are data outside 5th or 95th percentiles. CRL, Patients who underwent genetic amniocentesis; CI, patients who underwent emergency cerclage. R = 0.62; P ! .001.
Table I
Demographic and obstetric characteristics Group
Characteristic
Cervical incompetence (n = 40)
Maternal age (y)* Nulliparity (n) Gestational age (wk)*
31.5 G 8.5 11 (27%) 22.9 G 7.9
Control (n = 45)
P value
33.1 G 4.8 19 (33%) 20.3 G 3.2
.82 .51 .75
* Data are given as mean G SD.
Amniotic fluid assay of relaxin and IL-6 Amniotic fluid immunoreactive relaxin concentrations were determined with a specific human relaxin enzymelinked immunosorbent assay. Purified human H2 relaxin standard were provided by ImmunDiagnostik AG (Bensheim, Germany). We used goat anti-human relaxin antibody and a horseradish peroxidase-conjugated rabbit antihuman relaxin antibody as a detection system. The lower limit of sensitivity of the assay was 20 pg/mL. The intra-assay and interassay coefficients of variation were both !10%. IL-6 was quantitated in amniotic fluid with a commercially available enzyme-linked immunosorbent assay (R&D Systems, Minneapolis, Minn). The sensitivity of the IL-6 assay was !1.0 pg/ mL. Interassay and intra-assay coefficients of variation were !10% for each analyte.
Statistical analysis After first being tested for normal distribution with a Kolmogorov-Smirnov test, statistical comparisons were performed using the Mann-Whitney rank sum test. The laboratory results are presented as their median, followed by the 25th/75th percentiles. A probability
Results All demographic and laboratory data were available in 40 women with cervical incompetence and 45 control patients. The 2 groups were similar in selected demographic characteristics: Korean, gestational age at amniocentesis, percent nulliparity, and maternal age. The demographic characteristics of the study population are shown in Table I. Women who fulfilled the cerclage criteria had higher amniotic fluid IL-6 levels compared women who underwent genetic amniocentesis (control [50.4 pg/mL (range, 19.4-97.4 pg/mL)] vs cervical incompetence [5459.1 pg/ mL (range, 1131.4-14425.7 pg/mL)]; P ! .001). In contrast with IL-6, levels of relaxin did not show significant difference between the 2 groups (control [67.5 pg/mL (range, 35.1-153.5 pg/mL)] vs cervical incompetence [45.6 pg/mL (range, 30.1-75.5 pg/mL)]; P = .061; Figure 1). We next investigated whether the IL-6 level correlated to clinical outcome as reflected in the latency interval. We defined several clinically relevant latency intervals: 24 hours, 3 days, 7 days, before 33 and 37 completed weeks of gestation. The median amniotic fluid IL-6 and relaxin levels by latency group are shown in Table II. There was a significant difference in IL-6 levels in women with shorter latencies and preterm delivery (P ! .01 for all cut-offs) compared with those women with longer latency intervals. There was no significant difference in relaxin levels in women with shorter latencies and preterm delivery compared with those women with longer latency intervals, except in the delivery before 33 weeks of gestation group. The median relaxin level in women who were delivered before 33 weeks of gestation (55.1 pg/mL [range, 31.3-109.2 pg/mL]) was significantly higher than in those women who were delivered after 33 weeks of gestation (33.9 pg/mL [range, 18.0-45.1 pg/mL]; P = .049). Linear regression analysis (Figure 2) with latency as the dependent variable and IL-6 as the independent variable revealed a significant inverse relationship (r = ÿ0.36; P = .021 for linear data; and r = ÿ0.62; P ! .001 after log transformation of IL-6). There was no relationship between relaxin and latency interval
787
Lee et al Table II
The relationship between the median amniotic fluid IL-6 and relaxin levels and several outcome relevant latency intervals
Intervals
Outcome
IL-6
Delivery within 24 hr
No (n = 28) Yes (n = 12) No (n = 22) Yes (n = 18) No (n = 20) Yes (n = 20) No (n = 11) Yes (n = 29) No (n = 4) Yes (n = 36)
2300.0 28218.1 1511.8 14425.7 1511.8 13793.0 810.1 10009.1 235.4 7523.8
Delivery within 3 d Delivery within 7 d Delivery !33 wk Delivery !37 wk
P value (444.9-9002.2) (7737.5-44489.4) (356.1-7463.5) (5276.9-43690.2) (394.0-6552.4) (4146.8-38766.4) (243.3-4654.2) (2603.0-27059.7) (33.1-558.7) (1950.9-20396.5)
!.001 !.001 !.001 .005 .005
Relaxin 43.5 48.8 43.5 47.3 41.9 48.8 33.9 55.1 24.0 46.5
(25.8-67.9) (33.0-91.4) (29.1-66.5) (31.1-105.5) (21.7-60.2) (31.3-112.9) (18.0-45.1) (31.3-109.2) (10.5-58.4) (31.3-75.5)
P value .452 .523 .190 .049 .199
The amniotic fluid interleukin-6 level, but not the relaxin level, predicts the success of rescue cerclage.
either before (r = 0.15, P = .337) or after log transformation (r = 0.269, P = .093). A receiver operator curve analysis (Figure 3) indicated that a preoperative IL-6 level ofO1700 pg/mL had a sensitivity of 100% (95% CI, 100-100) and a specificity of 59.1% (95% CI, 36.4-79.3) for the prediction of a latency of !3 days. An IL-6 value of O12326 pg/mL predicted delivery within 24 hours with 75% sensitivity (95% CI, 42.8-94.2) and 82% specificity (95% CI, 63.1-93.9). An IL-6 value !748 pg/mL predicted a delivery delay until 33 weeks of gestation with 85% sensitivity and 67% specificity (95% CI, 71.5-100). Survival analysis (Figure 4) was conducted that compared latency intervals of patients with preoperative IL6 levels of O1700 pg/mL to those women with preoperative IL-6 levels !1700 pg/mL. The median latency interval in patients in the first group was 2 days, which was significantly shorter than the second group whose median latency interval was 35 days (log-rank test chisquared test, 13.17; P = .0003). There were no differences between groups for gestational age at amniocentesis (P = .33), patient age (P = .23), and percent nulliparity (P = .85).
Comment We investigated several factors that might be associated with cervical incompetence and the ultimate success of a rescue cerclage. Although the study design prevents any comments on the chronology of the detected changes, the results suggest that the measurement of IL-6 may be of clinical value in the treatment of cervical incompetence that requires a rescue cerclage. In 2000, Mays et al19 observed improved outcomes with the use of amniocentesis to exclude women with subclinical intra-amniotic infection, by the measurement of amniotic fluid glucose and lactic dehydrogenase levels as predictors of histologic chorioamnionitis. The mechanism underlying cervical dilation in women with cervical incompetence is unknown. Hyper
Figure 2 Regression analysis of the duration from the cervical incompetence procedure and delivery (latency) versus the amniotic fluid IL-6 levels. R = 0.62; P ! .001.
relaxinemia is one putative endocrine cause of cervical incompetence.1 Relaxin is a peptide hormone that is secreted by the corpus luteum and that is known to trigger connective tissue remodeling in several mammalian species, including human.20 Relaxin increases collagen turnover in human dermal fibroblast by stimulating collagenase expression and modulating collagen synthesis.3 Iams et al5 measured serum relaxin and the sonographic cervical length at 24 and 28 weeks of gestation in spontaneous twin pregnancies. There was a significant relationship between serum relaxin and preterm birth but not between relaxin and cervical length. This suggests
788
Lee et al
Figure 4 Survival analysis of the duration from the cervical incompetence procedure and delivery (latency).
Figure 3 Receiver operating curve analysis of amniotic fluid IL-6 levels and the duration from the cervical incompetence procedure and delivery: Delivery within 24 hours (area under the curve, 0.836; standard error, 0.68; 95% CI, 0.703-0.969; P = .001); delivery within 3 days (area under the curve, 0.871; standard error, 0.054; 95% CI, 0.765-0.977; P ! .001); delivery before 33 weeks of gestation (area under the curve, 0.858; standard error, 0.073; 95% CI, 0.715-1.00; P = .006).
that increased relaxin levels cannot explain the inverse correlation between cervical length and the prevalence of spontaneous preterm birth in women with spontaneous twin pregnancies. Further, several investigators have reported that purified porcine relaxin advances cervical softening, effacement, and dilation and shortens the interval to delivery when administered either intravaginally or intracervically.7,8 These observations suggest local relaxin may contribute to cervical ripening before labor. Indeed, human relaxin of decidual and placental origin is involved in controlled collagen remodeling in the fetal membrane.9,10 Eddie et al6 suggested that either very low levels of circulating relaxin are sufficient for its biologic effects or that decidual and placental relaxins are of greater importance to uterine and cervical function at term than relaxin from corpus luteum. Amniotic fluid relaxin may be derived from the decidualized endometrium rather than the maternal circulation.11 We therefore investigated relaxin in the amniotic fluid concentration in women with cervical incompetence and the relationship between relaxin concentration and the success of rescue cerclage. We found no significant increase in amniotic fluid relaxin in women with cervical incompetence. This suggests that relaxin is unlikely to contribute to the mechanism of cervical incompetence, at least as reflected in the amniotic fluid.
We also had expected that relaxin levels in women with cervical incompetence might be higher than those levels in the control group but were surprised to find that the differences in relaxin concentration between the 2 groups was not significant (P = .061). In point, the amniotic fluid relaxin level of the cervical incompetence group (45.6 pg/mL [range, 30.1-75.5 pg/mL]) was lower than that of the control group (67.5 pg/mL [range, 35.1-153.5 pg/mL]; P = not significant). There was no significant difference in relaxin between women with shorter latencies and preterm delivery compared with those women with longer latency intervals, except in the group that were delivered before 33 weeks of gestation. The median relaxin level in women who were delivered !33 weeks of gestation was significantly higher than in those who were deliveredO33 weeks of gestation. Interestingly, we also found the amniotic fluid relaxin level of women who were delivered !37 weeks of gestation was higher than those women who were deliveredO37 weeks of gestation (P = not significant). This observation suggests that further study of amniotic fluid relaxin in the future might be revealing. The biochemical events of cervical ripening are complex and include degradation and/or remodeling of extracellular matrix proteins and glycoproteins, disruption of tightly aligned collagen fibrils, and increased hydration caused by hyaluronan. The process has been likened to an inflammatory reaction12 that is associated with the catabolism of cervical extracellular matrix by enzymes that are released from infiltrating leukocytes. Indeed, there is an influx of inflammatory cells during cervical ripening into the cervical stroma, which includes macrophages, neutrophils,13 mast cells,14 and eosinophils.15 Romero et al21 compared the values of amniotic fluid tests in the detection of microbial invasion of the amniotic cavity and concluded that the IL-6 concentrations
Lee et al in amniotic fluid are better indicators of microbial invasion of the amniotic cavity than the amniotic fluid Gram stains, glucose concentrations, or white blood cell counts. We investigated IL-6 and demonstrated that it was increased significantly in women with cervical incompetence. We cannot determine presently whether the elevation is part of the pathophysiologic nature of cervical incompetence or the end result of it. But, we can say that the higher the IL-6 level before cerclage, the shorter the latency interval. Few women with a grossly elevated IL6 level have any likelihood of reaching viability should the cerclage be placed at !22 weeks of gestation. The mechanism underlying the increase in IL-6 level may be bacterial in origin, but antibiotics have failed to improve the outcome of emergency cerclage in the past. The identification of women with an abnormally elevated IL-6 level could lead to future therapies that would seek to interfere with either the initiation of or the continuation of the inflammatory process. Although the exact source of the cytokines in the amniotic fluid of women with cervical incompetence is unclear, a growing body of evidence suggests that ascending genital tract infection followed by deciduitis may be part of the sequence.22 The observed relationship between IL-6 level and the outcome of pregnancy is consistent with them having a pathogenic role in cervical incompetence.
References 1. Iams JD. Cervical incompetence. In: Creasy R, Resnik R, editors. Maternal fetal medicine. 4th ed. Philadelphia: Saunders; 1999. p. 445-64. 2. MacLennan AH. The role of relaxin in human reproduction. Clin Reprod Fertil 1983;2:77-95. 3. Downing SJ, Sherwood OD. The physiological role of relaxin in the pregnant rat: IV, the influence of relaxin on cervical collagen and glucosaminoglycans. Endocrinology 1986;118:471-9. 4. Weiss G, Goldsmith LT, Sachdev R, Hagen SV, Lederer K. Elevated first-trimester serum relaxin concentrations in pregnant women after ovarian stimulation predict prematurity risk and preterm delivery. Obstet Gynecol 1993;82:821-8. 5. Iams JD, Goldsmith LT, Weiss G. The preterm prediction study: maternal serum relaxin, sonographic cervical length, and spontaneous preterm birth in twins. J Soc Gynecol Investig 2001;8:39-42. 6. Eddie LW, Cameron IT, Leeton JF, Healy DL, Renou P. Ovarian relaxin is not essential for dilatation of cervix. Lancet 1990;336: 243.
789 7. Evans MI, Dougan MB, Moawad AH, Evans WJ, Bryant-Greenwood GD, Greenwood FC. Ripening of the human cervix with porcine ovarian relaxin. Am J Obstet Gynecol 1983;147:410-4. 8. Bell RJ, Permezel M, McLennan A, Hughes C, Healy D, Brennecke S. A randomized, double-blind, placebo-controlled trial of the safety of vaginal recombinant human relaxin for cervical ripening. Obstet Gynecol 1993;82:328-33. 9. Bogic LV, Mandel M, Bryant-Greenwood GD. Relaxin gene expression in human reproductive tissues by in situ hybridization. J Clin Endocrinol Metab 1995;80:130-7. 10. Bryant-Greenwood GD, Schwabe C. Human relaxins: chemistry and biology. Endocr Rev 1994;15:5-26. 11. Johnson MR, Abbas A, Nicolaides KH, Ligheman SL. Distribution of relaxin between human maternal and fetal circulation and amniotic fluid. J Endocr 1992;134:313-7. 12. Liggins G. Cervical ripening as an inflammatory reaction. In: Ellwood DA, Anderson ABM, editors. The cervix in pregnancy and labour: clinical and biochemical investigations. Edinburgh, Scotland: Churchill-Livingstone; 1981. p. 1-12. 13. Spanggaard H, Knudsen UB, Uldbjerg N, Jeziorska M, Woolley DE, Danielsen CC. Mast cells in cervical ripening-an immunohistochemical and biochemical study in rats. Eur J Obstet Gynecol Reprod Biol 1997;73:91-7. 14. Tschesche H. Human neutrophil collagenase. Methods Enzymol 1995;248:431-49. 15. Osmers RG, Adelmann-Grill BC, Rath W, Stuhlsatz HW, Tschesche H, Kuhn W. Biochemical events in cervical ripening dilatation during pregnancy and parturition. J Obstet Gynecol 1995;21:185-94. 16. Uldbjerg N, Forman A. Biomechanical and biochemical changer of the uterus and cervix during pregnancy. In: Reece AE, Hobbins JC, editors. Medicine of the fetus and mother. 2nd ed. Philadelphia: Lippincott-Raven; 1999. p. 921-31. 17. Uldbjerg N, Ulmsten U, Ekman G. The physiological role of eicosanoids in controlling the form and function of the cervix. In: Hiller K, editor. Advances in eicosanoid research. Lancaster (PA): MTP Press; 1987. p. 163-83. 18. Yoon BH, Romero R, Kim JC, Jun JK, Gomez R, Choi JH, et al. Amniotic fluid interleukin-6: a sensitive test for antenatal diagnosis of acute inflammatory lesions of preterm placenta and prediction of perinatal morbidity. Am J Obstet Gynecol 1995;172:960-70. 19. Mays JK, Figueroa R, Shah J, Khakoo H, Kaminsky S, Thjani N. Amniocentesis for selection before rescue cerclage. Obstet Gynecol 2000;95:652-5. 20. Sherwood OD. Relaxin. In: Knobil E, Neill JD, Greenwald GS, Markert C, Pfaff DW, editors. The physiology of reproduction. 2nd ed. New York: Raven Press; 1994. p. 861-1009. 21. Romero R, Yoon BH, Mazor M, Gomez R, Diamond MP, Kenny JS, et al. The diagnostic and prognostic value of amniotic fluid white blood cell count, glucose, interleukin- 6, and Gram stain in patients with preterm labor and intact membranes. Am J Obstet Gynecol 1993;169:805-16. 22. Romero R, Gonzalez R, Sepulveda W, Brandt F, Ramirez M, Sorokin Y, et al. VIII, Microbial invasion of the amniotic cavity in patients with suspected cervical incompetence: prevalence and clinical significance. Am J Obstet Gynecol 1992;167:1086-91.