Premature rupture of membranes (PROM), defined as spontaneous rupture of the fetal membranes before the onset of uterine contractions, is one of the most common diagnostic dilemmas in contemporary obstetric practice. Premature rupture of membranes can occur at any gestational age, and preterm PROM (defined as PROM before 37 weeks) is responsible for 20–40% of preterm births.1 Early and accurate diagnosis of PROM would allow for gestational age-specific obstetric interventions designed to optimize perinatal outcome and minimize serious complications such as cord prolapse and infectious morbidity (chorioamnionitis, neonatal sepsis).2–6 Conversely, a false-positive diagnosis of PROM may lead to unnecessary obstetric interventions, including hospitalization, administration of antibiotics and corticosteroids, and even induction of labor.7–9
Conventional clinical methods for diagnosing ROM are seriously flawed.10–17 The most commonly used method, the nitrazine test, is designed to confirm only an alkaline pH in the cervicovaginal secretions and is associated with high false-positive rates related to cervicitis, vaginitis, alkaline urine, and contamination with blood, semen, or antiseptic agents.11,12 As such, the sensitivity and specificity of this test range from 90% to 97% and from 16% to 70%, respectively.13,14 The “fern test” (crystallization of amniotic fluid on drying) may give false-positive results due to fingerprints or contamination with semen and cervical mucus as well as false-negative results due to technical error (dry swab) or contamination with blood.11,15,16 Reported sensitivity and specificity for the fern test are 51% and 70%, respectively, in patients without labor and 98% and 88%, respectively, in patients in labor.17
A bedside immunoassay designed to measure levels of the 34 kd fetal glycoprotein, placental alpha-microglobulin-1, in cervicovaginal secretions18 was recently introduced into clinical practice in an attempt to improve on conventional clinical tests.19 Placental alpha-microglobulin-1 is an ideal candidate marker for ROM because its concentration in amniotic fluid is 1,000- to 10,000-fold higher than that in the cervicovaginal discharge with intact membranes (2,000–25,000 ng/mL versus 0.05–2.0 ng/mL, respectively).18–20 Despite the importance of an accurate and timely diagnosis of ROM and the theoretic advantages of using placental alpha-microglobulin-1 as a marker of ROM, there is little prospective data on the clinical performance of the placental alpha-microglobulin-1 immunoassay. To this end, the current study compares the accuracy of the immunoassay with that of conventional clinical assessment for the diagnosis of ROM.
MATERIALS AND METHODS
A prospective observational study was performed involving consecutive patients presenting to Seoul National University Hospital with signs or symptoms of ROM from March 2005 to February 2006. The study was approved by the hospital’s Institutional Review Board and written consent was obtained from all patients for the collection and analysis of specimens and clinical data.
All patients were evaluated for ROM with a detailed history, physical examination (including sterile speculum examinations), and transabdominal ultrasound examination. At initial speculum examination without the use of antiseptic solutions, all patients were assessed using both conventional clinical tests (visual evaluation of leakage or pooling of amniotic fluid in the posterior fornix, nitrazine testing, and ferning), and the placental alpha-microglobulin-1 immunoassay. Patients with active vaginal bleeding were excluded from the study. Women were included in the study only at their initial presentation (ie, no woman was included twice).
A diagnosis of ROM was made at initial examination using conventional clinical criteria: 1) if amniotic fluid was seen leaking from the cervical os on speculum examination, or 2) if two of the following three clinical signs were present: visual pooling of fluid in the posterior fornix, positive nitrazine test, and microscopic evidence of ferning. The presence of oligohydramnios on transabdominal ultrasound examination was not one of the diagnostic criteria and was not sufficient to make the diagnosis of ROM.
The placental alpha-microglobulin-1 immunoassay was performed according to the manufacturer’s instructions (AmniSure ROM test, N-Dia Inc, New York, NY). Briefly, a sterile Dacron swab supplied by the manufacturer was placed in the posterior fornix of the vagina. After a period of 1 minute to ensure saturation, the swab was removed and agitated in a vial containing solvent for 1 minute. The test strip was then placed in the solvent, and the sample in the vial was allowed to migrate through the membranes by capillary reaction. The test strip indicated a negative or positive result within 5 minutes. Results were determined by the presence or absence of the control and test lines. Immunoassay results were reported as positive or negative. Obstetric care providers were blinded to the results of the placental immunoassay.
After an initial assessment for ROM, all patients were managed by standard gestational age–specific clinical algorithms. Briefly, women in whom ROM could not be confirmed clinically were admitted to the hospital, and a follow-up evaluation was performed in 1–3 days. In a small subset of high-risk women remote from term, clinically indicated amnio-dye testing (intra-amniotic injection of indigo carmine with staining of a vaginal tampon within 20–30 minutes) was performed so that a definitive diagnosis could be made. Women with confirmed ROM after 35 weeks were admitted for delivery. Women with PROM from 24 to 35 weeks were managed expectantly as inpatients. Expectant management included serial physical examinations to exclude intra-amniotic infection, regular fetal testing, neonatal consultation, and the administration of broad-spectrum antibiotics (to prolong latency) or antenatal corticosteroids (to optimize fetal outcome), or both if indicated. Repeat sterile speculum examinations were performed as clinically indicated. Indications for delivery included intrauterine infection, nonreassuring fetal testing, excessive bleeding, unstoppable preterm labor, and 35 weeks of gestation.
A final determination of whether or not the patient had ROM at her initial presentation was made after the patient had delivered and the clinical record was reviewed. This assessment was made after review of the initial and all follow-up evaluations and subsequent clinical course. A diagnosis of ROM at the follow-up evaluation in 1–3 days was taken as evidence of ROM at the initial visit. This determination was made by investigators who were blinded to the placental immunoassay results.
Performance (sensitivity, specificity, positive predictive value, negative predictive value, and false-negative rate [defined as 1-specificity]) of the placental alpha-microglobulin-1 immunoassay in diagnosing ROM at initial presentation was compared with that of the conventional clinical tests, either alone or in combination. Analyses were performed using McNemar χ2 and Fisher exact tests. P<.05 was regarded as significant. Given the number of women included (n=184), this study had an 80% power to detect a 10% difference between the two tests.
A total of 184 patients (gestational age 35±0.5 weeks [mean±standard error of the mean], range 11–42 weeks) were recruited; 57% (104 of 184) were at term (37 weeks or later), and 43% (80 of 184) were preterm. Of these 184 patients, an initial diagnosis of ROM was made in 76% (139 of 184) using conventional clinical assessment and 88% (161 of 184) using the placental alpha-microglobulin-1 immunoassay. One patient was lost to follow-up and was therefore excluded from the final analysis. Subsequent review of the medical records confirmed that 159 of the final 183 patients (87%) had ruptured membranes at their initial presentation, whereas 24 patients (13%) had intact membranes (Fig. 1). Of the 20 patients in whom ROM was not confirmed at the initial examination but who were subsequently determined to have ROM at initial presentation, the median time to diagnosis of ROM was 6 hours (15 were diagnosed within 24 hours, three within 2 days, one in 3 days, and one in 8 days). Gestational age at delivery was 37±0.4 weeks (mean±standard error of the mean, range, 14–42 weeks), with a mean latency from initial examination to delivery of 9±2 days (range 0–182 days).
Figure 1 summarizes the results of placental alpha-microglobulin-1 immunoassay and conventional clinical tests for the diagnosis of ROM at the initial visit. The longitudinal assessment of the patient’s pregnancy (including initial examination, follow-up examinations, and subsequent clinical course), as determined by review of the medical records after delivery, was used as the clinical gold standard to determine whether or not the patient had ROM at the initial visit.
Using the final determination as the gold standard (Fig. 1), placental alpha-microglobulin-1 immunoassay confirmed ROM at initial presentation with a sensitivity of 98.7% (157 of 159), specificity of 87.5% (21 of 24), positive predictive value of 98.1% (157 of 160), negative predictive value of 91.3% (21 of 23), and false positive rate of 12.5%. By comparison, the conventional combined clinical tests (pooling, nitrazine, and ferning) had a sensitivity of 87.4% (139 of 159, 95% confidence interval [CI] 0.81–0.92), specificity of 100% (24 of 24, 95% CI 0.83–1), positive predictive value of 100% (139 of 139, 95% CI 0.97–1), and negative predictive value of 54.5% (24 of 44, 95% CI 0.39–0.69) in diagnosing ROM. Compared with the conventional combined clinical tests, the placental immunoassay was significantly more sensitive in diagnosing ROM (98.7% [157 of 159] versus 87.4% [139 of 159], P<.001, McNemar χ2 test). Although the conventional clinical test had an excellent specificity (100%), it was not statistically different from that of the placental immunoassay (P=.25).
In 25 cases, there were discrepant results between the conventional clinical assessment and placental immunoassay. Table 1 describes the clinical characteristics and outcomes of patients with discrepant results. In three cases (cases 3–5), the placental immunoassay produced a false-positive result because the patients showed uneventful clinical courses after the initial examinations. Among the 20 patients with a positive placental immunoassay result and negative clinical assessment who were subsequently determined to have PROM (cases 6–25), nine cases (case 6–14) revealed a positive result in one of the three conventional clinical tests: seven were nitrazine test positive, one showed evidence of pooling, and one had a positive fern test. However, 11 cases (case 15–25) showed only a positive placental immunoassay (Table 1).
The performance of the placental immunoassay in identifying women with ROM was compared also with the nitrazine test alone (Table 2). The sensitivity of the immunoassay was significantly greater than that of the nitrazine test alone (98.7% [157 of 159] versus 88.1% [140 of 159], P<.001).
In an attempt to design a test with improved diagnostic accuracy, we examined the performance of combined placental alpha-microglobulin-1 immunoassay and nitrazine tests in identifying women with ROM (Table 3). A positive placental immunoassay with a positive nitrazine test was 100% accurate in identifying women with ROM. Moreover, all but one case (95%) with a negative placental immunoassay and negative nitrazine test were negative for ROM. Although most (86%) of the cases with a positive placental immunoassay and negative nitrazine test were subsequently shown to have ROM, only one of four cases with a positive nitrazine test and negative placental immunoassay was shown to have ROM. Further studies are required to confirm these observations because no firm conclusions can be reached based on subgroups as small as four patients.
Although preterm PROM is associated with significant perinatal and maternal mortality and morbidity,1–4 an accurate test to confirm the diagnosis has been lacking.7,11 This is due primarily to the fact that the traditional “gold standard” for the diagnosis of ROM is subjective. It relies on the ability of the clinician to document pooling of amniotic fluid in the posterior fornix of the vagina, an alkaline vaginal pH (nitrazine test), and ferning of the cervicovaginal discharge on drying, and each of these clinical tests is subject to interference resulting in high false-positive or false-negative results.11,13 This study demonstrates that the placental alpha-microglobulin-1 immunoassay is an accurate method for diagnosing ROM.
In our cohort, the placental immunoassay alone was an accurate method for confirming the diagnosis of ROM in symptomatic women across a wide gestational age range, with an overall sensitivity of 99%, specificity of 88%, positive predictive value of 98%, and negative predictive value of 91%. Moreover, the placental immunoassay was superior to both conventional combined clinical tests (pooling, nitrazine, ferning) and the nitrazine test alone. Indeed, the placental immunoassay was noted to be positive in the majority of women who were later shown to have ROM but in whom standard clinical tests were negative (Table 2). A false-positive test (defined as a positive placental immunoassay in women who were subsequently determined not to have ROM) was documented in three cases (Table 1, subjects 3–5), all of whom had subsequent uneventful pregnancies and term deliveries. The cause of these false-positive results is not clear, although a small leakage of amniotic fluid that subsequently sealed over cannot be excluded.
Because of the severe limitations with the current “gold standard” for the diagnosis of ROM (namely, clinical assessment of pooling, nitrazine, and ferning), investigators have long been searching for an alternative and more objective test. Such tests are based primarily on the identification in the cervicovaginal discharge of one or more biochemical markers that are present in the setting of ROM but absent in women with intact membranes. Several such markers have been studied—including alpha-fetoprotein,21 fetal fibronectin,22 insulin-like growth factor binding protein-1,13,14,23 and prolactin24,25—but none have yet been found to be reliable. The current test for ROM is based on the presence in the cervicovaginal discharge of high levels of placental alpha-microglobulin-1, a 34 kd fetal glycoprotein that circulates in the amniotic fluid at concentrations that are 1,000- to 10,000-fold higher than that in the cervicovaginal discharge with intact membranes (2,000–25,000 ng/mL versus 0.05–2.0 ng/mL, respectively).18–20 In the current immunoassay, the threshold of detection for placental alpha-microglobulin-1 in the eluted solvent has been set at 5.0 ng/mL.
Despite the importance of an accurate and timely diagnosis of ROM and the theoretic advantages of using placental alpha-microglobulin-1 as a marker of ROM, there is little prospective data on the clinical performance of the placental alpha-microglobulin-1 immunoassay. Cousins et al19 compared the performance of the placental immunoassay with standard clinical assessment for ROM in 203 symptomatic women at 15–42 weeks of gestation. They concluded that the placental immunoassay was highly accurate in diagnosing ROM, with a sensitivity of 98.9%, specificity of 100%, positive predictive value of 100%, and negative predictive value of 99.1%, but their “gold standard” for ROM was based primarily on review in the medical records of the single clinical examination at presentation. Although we applied more stringent criteria for confirming the diagnosis of ROM, our data are largely in agreement with this prior publication. We have shown that the placental alpha-microglobulin-1 immunoassay is superior to both the combined conventional clinical tests and the nitrazine test alone, which is commonly used to identify women with ROM. Moreover, in the majority of cases (87% [20 of 23]) where there was a discrepancy between the placental immunoassay (positive) and conventional clinical tests (negative), the placental immunoassay was found to be accurate. This may be because it is more sensitive in detecting subclinical ROM because of microperforations in the fetal membranes. Further studies are needed to verify this explanation and to determine the clinical significance of such microperforations.
In conclusion, the placental alpha-microglobulin-1 immunoassay is a rapid and accurate method for confirming the diagnosis of ROM. Moreover, its performance appears to be superior to conventional clinical assessment (pooling, nitrazine, ferning) and the nitrazine test alone.
1. Mercer BM, Goldenberg RL, Meis PJ, Moawad AH, Shellhaas C, Das A, et al. The Preterm Prediction Study: prediction of preterm premature rupture of membranes through clinical findings and ancillary testing. The NICHD Maternal-Fetal Medicine Units Network. Am J Obstet Gynecol 2000;183:738–45.
2. Ananth CV, Oyelese Y, Srinivas N, Yeo L, Vintzileos AM. Preterm premature rupture of membranes, intrauterine infection, and oligohydramnios: risk factors for placental abruption. Obstet Gynecol 2004;104:71–7.
3. Shumway JB, Al-Malt A, Amon E, Cohlan B, Amini S, Abboud M, et al. Impact of oligohydramnios on maternal and perinatal outcomes of spontaneous premature rupture of the membranes at 18-28 weeks. J Matern Fetal Med 1999;8:20–3.
4. Marlowe SE, Greenwald J, Anwar M, Hiatt M, Hegyi T. Prolonged rupture of membranes in the term newborn. Am J Perinatol 1997;14:483–6.
5. Shim SS, Romero R, Hong JS, Park CW, Jun JK, Kim BI, et al. Clinical significance of intra-amniotic inflammation in patients with preterm premature rupture of membranes. Am J Obstet Gynecol 2004;191:1339–45.
6. Park JS, Yoon BH, Romero R, Moon JB, Oh SY, Kim JC, et al. The relationship between oligohydramnios and the onset of preterm labor in preterm premature rupture of membranes. Am J Obstet Gynecol 2001;184:459–62.
7. Garite TJ. Management of premature rupture of membranes. Clin Perinatol 2001;28:837–47.
8. Hannah ME, Hodnett ED, Willan A, Foster GA, Di Cecco R, Helewa M. Prelabor rupture of the membranes at term: expectant management at home or in hospital? The Term PROM Study Group. Obstet Gynecol 2000;96:533–8.
9. Healy AJ, Veille JC, Sciscione A, McNutt LA, Dexter SC. The timing of elective delivery in preterm premature rupture of the membranes: a survey of members of the Society of Maternal-Fetal Medicine. Am J Obstet Gynecol 2004;190:1479–81.
10. Gorodeski IG, Haimovitz L, Bahari CM. Reevaluation of the pH, ferning and nile blue sulphate staining methods in pregnant women with premature rupture of the fetal membranes. J Perinat Med 1982;10:286–92.
11. Friedman ML, McElin TW. Diagnosis of ruptured fetal membranes: clinical study and review of the literature. Am J Obstet Gynecol 1969;104:544–50.
12. de Haan HH, Offermans PM, Smits F, Schouten HJ, Peeters LL. Value of the fern test to confirm or reject the diagnosis of ruptured membranes in modest in nonlaboring women presenting with nonspecific vaginal fluid loss. Am J Perinatol 1994;11:46–50.
13. Gaucherand P, Salle B, Sergeant P, Guibaud S, Brun J, Bizollon CA, et al. Comparative study of three vaginal markers of the premature rupture of membranes: insulin-like growth factor binding protein, 1 diamine-oxidase, and PH. Acta Obstet Gynecol Scand 1997;76:536–40.
14. Erdemoglu E, Mungan T. Significance of detecting insulin-like growth factor binding protein-1 in cervicovaginal secretions: comparison with nitrazine test and amniotic fluid volume assessment. Acta Obstet Gynecol Scand 2004;83:622–6.
15. Reece EA, Chervenak FA, Moya FR, Hobbins JC. Amniotic fluid arborization: effect of blood, meconium, and pH alterations. Obstet Gynecol 1984;64:248–50.
16. Rosemond RL, Lombardi SJ, Boehm FH. Ferning of amniotic fluid contaminated with blood. Obstet Gynecol 1990;75:338–40.
17. Smith RP. A technique for the detection of rupture of the membranes: a review and preliminary report. Obstet Gynecol 1976;48:172–6.
18. Petrunin DD, Griaznova IM, Petrunina IuA, Tatarinov IuS. Immunochemical identification of organ specific human placental alphaglobulin and its concentration in amniotic fluid [in Russian]. Akush Ginekol 1977;1:64–5.
19. Cousins LM, Smok DP, Lovett SM, Poeltler DM. Amnisure placental alpha macroglobulin-1 rapid immunoassay versus standard diagnostic methods for detection of rupture of membranes. Am J Perinatol 2005;22:317–20.
20. Boltovskaia MN, Zaraiskii EI, Fuks BB, Sukhikh GT, Kalafati TI, Starosvetskaia NA, et al. Histochemical and clinical-diagnostic study of placental alpha 1-microglobulin using monoclonal antibodies [in Russian]. Bull Eksp Biol Med 1991;112:397–400.
21. Kishida T, Yamada H, Negishi H, Sagawa T, Makinoda S, Fujimoto S. Diagnosis of premature rupture of the membranes in preterm patients, using an improved AFP kit: comparison with ROM-check and/or nitrazine test. Eur J Obstet Gynecol Reprod Biol 1996;69:77–82.
22. Lockwood CJ, Senyei AE, Dische MR, Casal D, Shah KD, Thung SN, et al. Fetal fibronectin in cervical and vaginal secretions defines a patient population at high risk for preterm delivery. N Engl J Med 1991;325:669–74.
23. Jeurgens-Borst AJ, Bekkers RL, Sporken JM, van den Berg PP. Use of insulin like growth factor binding protein-1 in the diagnosis of ruptured fetal membranes. Eur J Obstet Gynecol Reprod Biol 2002;102:11–4.
24. Buyukbayrak EE, Turan C, Unal O, Dansuk R, Cengizoglu B. Diagnostic power of the vaginal washing-fluid prolactin assay as an alternative method for the diagnosis of premature rupture of membranes. J Matern Fetal Neonatal Med 2004;15:120–5.
25. Koninckx PR, Trappeniers H, Van Assche FA. Prolactin concentration in vaginal fluid: a new method for diagnosing ruptured membranes. Br J Obstet Gynaecol 1981;88:607–10.