Placental α-microglobulin-1 in cervicovaginal fluid and cervical length to predict preterm birth by Thai women with symptoms of labor

The present prospective observational study was approved by the Ethics Committee of the Siriraj Institutional Review Board (Si 625/2016) and was conducted in full compliance with international guidelines for human research protection including the principles outlined in the contemporary revision of the Declaration of Helsinki 1964 (World Medical Association) incorporating the most recent (2013) and earlier amendments, the Belmont Report, Council for International Organizations of Medical Sciences Guidelines, and the International Conference on Harmonization in Good Clinical Practice. We retrospectively registered the study in the Thai Clinical Trials Registry (TCTR20180703002), although we note this was an observational study. We used guidelines in the STARD 2015 list of essential items to ensure appropriate reporting of studies of diagnostic accuracy [33].

A sample of 163 pregnant women was needed based on the expected 85% sensitivity of the tool and an estimated 15% preterm in the study setting [4]. From March 2016 to September 2018, we recruited a total of 180 pregnant women in a consecutive series who attended the labor ward at Siriraj Hospital, a tertiary care, primary teaching hospital of the Faculty of Medicine, Mahidol University, with a capacity of >2,000 beds in Bangkok, Thailand. All participants provided their written informed consent before entering the study.

The included patients were those with a singleton pregnancy who had symptomatic painful uterine contraction every 4 times in 20 min or 8 times in 60 min with a cervical opening <3 cm, effacement <80%, and a gestational age from 20^0/7 weeks to 36^6/7 weeks without ruptured membranes.

Excluded patients were those who had received a digital examination, had bleeding per vagina, had abruption placentae, suspected placenta previa, had been exposed to tocolytic drugs, cervical cerclage, trauma, urinary tract infection, girls <18 years, and those with fetal anomaly.

All patients recruited underwent a full clinical examination by the attending physicians. This included collecting the sample for the PAMG-1 test, sterile speculum examination, transvaginal ultrasound, and digital examination. The sample for the PAMG-1 test was taken before the cervical examination. If cervical opening was ≥3 cm, effacement ≥80%, the patient was excluded from the study. Parameters recorded at presentation included cervical length, cervical dilatation, contraction frequency, cervical effacement, patient history, and PAMG-1 test result. Data regarding the gestational age of delivery, timing of delivery, neonatal Apgar score, and body weight and possible complication after delivery were also recorded.

From the total of 180 pregnant patients recruited, 19 were excluded from the study because of indicated deliveries (10 patients), incomplete data for cervical assessment (5 patients), and loss to follow-up (4 patients). Therefore, we included 161 patients in the study (Figure 1). All 161 patients underwent a PAMG-1 test and cervical length measurement. Their mean age was 28.9 years with an average body weight of 65 kg. Most of the patients were primigravid. Gestational age was calculated from ultrasound in about 43%. Some 145 (91%) patients went into spontaneous vaginal delivery; other deliveries were assisted or by cesarean section (Table 1).

Figure 1

Of the 161 symptomatic pregnant women with cervical length ≤30 mm, 34 (21%) went into spontaneous delivery within 7 days and 11 (7%) went into spontaneous delivery within 14 days (Table 2). Of the 14 symptomatic pregnant women with a cervical length ≤15 mm, 11 (79%) went into spontaneous delivery within 7 days (Table 3).

When PAMG-1 testing was combined with cervical length measurement <30 mm, the NPV was 83% to predict preterm birth within 7 days, and 77% to predict preterm birth within 14 days (Table 4). When PAMG-1 testing and cervical length measurement ≤15 mm was used, the PPV was 83% (Table 5). The optimal cut-off from ROC curve analysis shows that a cervical length <25 mm and a PAMG-1 positive result has a PPV of 80% to predict preterm birth within 7 days and 90% to predict preterm birth within 14 days (Tables 6–8; Figures 2 and 3).

Figure 2

Figure 3

We sought to determine whether testing cervicovaginal fluid for PAMG-1 could be used with cervical length to improve the predictability of preterm delivery within 7 days or 14 days in our hospital setting. The 161 patients in our present study were all treated with the same protocol, which did not affect the time to delivery. Of the 161 pregnant women with labor pain who had a cervical length <30 mm, 45 (28%) went into delivery within 7 days, and of those, 14 women with cervical length <15 mm, 11 (79%) went into delivery in <7 days. A positive PAMG-1 test result had a high NPV of 82% for preterm birth within 7 days and 76% within 14 days. When a cervical length <30 mm was combined with a PAMG-1 positive result, the NPV increased to 83% for preterm birth within 7 days, and 77% within 14 days. When a cervical length <15 mm was combined with a PAMG-1 positive result, the PPV was higher than it was when using cervical length alone. Cervical length <15 mm alone is a good predictor of preterm birth within both 7 days and 14 days and explains this finding [25]. Nikolova et al. [31] found that a positive PAMG-1 test result has a higher PPV than fetal fibronectin, or cervical length <25 mm. When cervical length is used as an initial screen and later combined with PAMG-1 testing, it has the greatest utility to predict delivery within 7 days in patients with cervical length 15–35 mm [31, 32]. Without cervical length screening, a positive PAMG-1 test result is more accurate than fetal fibronectin even when fetal fibronectin is combined with cervical length [31, 32]. The present study found combining a cervical length <15 mm with a positive PAMG-1 test result was the best predictor for delivery within 7 days in women with preterm labor. Combining the cervical length with the PAMG-1 test result increases the accuracy of the test to predict delivery within 7 days, achieving our goal of guiding the rational use of antenatal corticosteroids and tocolytic agents, decreasing unnecessary admissions, and correcting triage of patients that are need in of utero transfer [14, 15, 31, 32]. Therefore, the pregnant women who are not at risk of premature delivery can be advised to stay at home without unnecessary tocolytics or admissions. Patients who are at risk of premature delivery can be referred for tertiary care at specialized perinatal centers. The test allows physicians time to make the best decision to support their patients and manage them optimally. Psychological and physical burdens, the length of hospital stay, and health care costs can be reduced.

The lower cervical length cutoff point is associated with the highest specific prediction of imminent delivery within either 7 days or 14 days. A higher cervical length cutoff point lowers the specificity of the test. We sought to determine whether there is a specific cervical length range in which the PAMG-1 test would be most applicable and found that when a cervical length ≤15 mm was combined with a positive PAMG-1 result, 122/144 patients (85%) delivered within 7 days. When cervical length >30 mm was combined with a positive PAMG-1 result, 9/97 patients (9%) delivered within 7 days. However, patients with cervical length measurements between these two cutoffs had varying rates of delivery within 7 days. In our present study, 14/50 (28%) patients with cervical length 15–30 mm delivered within 7 days of testing.

The present study suggests that a PAMG-1 test result is useful when combined with a cervical length of 15–30 mm. The ROC curve analysis found that an optimal cut off of cervical length <25 mm with a positive PAMG-1 result has a PPV of 80% to predict preterm birth within 7 days, and 90% within 14 days.

A limitation of the study is that patients with any bleeding per vagina, rupture of membranes, or who had undergone a vaginal examination had to be excluded. Therefore, the target population is not fully representative and the test can only be used in a limited population excluding those with the conditions described. Another limitation is that we did not meet our expected sample size target of 163 patients because of an unexpectedly high loss to follow up. Nevertheless, the sample size of 161 patients obtained did not unduly affect the results of the present study.

It is already known that a short cervix or cervical length <15 mm is a good predictor of imminent spontaneous preterm delivery within 7 days. When cervical length measurement 15–30 mm was used as a cutoff, it was the least accurate predictor of imminent spontaneous preterm delivery within 7 days. This study adds that a cervical length 15–30 mm combined with a positive PAMG-1 test result will provide a higher predictor of imminent spontaneous preterm delivery within 7 days than either method alone, and should be considered for clinical practice. Our present findings support those of Nikolova et al. [31] and Bolotskikh and Borisova [32], who reported predictive results for a cervical length 15–35 mm combined with a positive PAMG-1 test result.

Cervical length 15–30 mm combined with testing cervicovaginal fluid for PAMG-1 has a high accuracy to predict imminent spontaneous delivery within 7 days by pregnant women with preterm labor and cervical dilatation <3 cm in our clinical practice setting.