Study design, site, aims and sample size
A prospective observational laboratory-based study took place from January 2014 to June 2016 at a tertiary care hospital with more than 2600 annual deliveries. Only one study had evaluated serum PCT concentrations in 53 women with preterm labour compared to 31 healthy pregnant women [
19]. No study has specifically evaluated whether serum PCT levels may predict sPTB within 7 or 14 days in women hospitalized with TPL between 24 and 36 weeks. So, no sample size estimation has been done before the beginning of the study, and an interim analysis had been planned two years after the beginning of the study (with an estimation of 100–120 included women regarding our hospital data).
The aims were to evaluate the association between serum PCT levels and sPTB in women hospitalized with TPL between 24 and 36 weeks, and to evaluate whether serum PCT levels may predict sPTB within 7 or 14 days after testing.
Characteristics of participants and description of methods
The study included pregnant women with a live singleton pregnancy, intact membranes, hospitalized with TPL between 24 and 36 weeks of gestation. Exclusion criteria were multiple gestations, known uterine malformation, cervical cerclage in situ, PPROM, suspected chorioamnionitis at the time of presentation, cervical dilatation of more than three centimetres, previous treatment with tocolysis within 7 days before inclusion, gestational hypertensive diseases, fetal growth restriction (defined as <10th centile for gestational age on Hadlock curves [
20,
21], known congenital anomaly, white blood cell (WBC) count> 15.0 G/L [
22], C-reactive protein (CRP) > 10 mg/L [
23], and contra-indications for tocolysis, such as suspected intra-uterine infections or fetal distress.
TPL was defined as the presence of regular and painful uterine contractions that registered by cardiotocography (at least, 2 uterine contractions/10 min during 20 min), intact membranes before 37 weeks and ultrasound cervical length < 25 mm [
24‐
26], and which may result in preterm birth [
27,
28]. Gestational age at delivery was determined by the crown-rump length at a first-trimester ultrasound examination or by the date of last menstrual period and/or a second- or third-trimester ultrasound if the first-trimester ultrasound was not performed [
29]. Premature births, by definition, occur prior to 37 completed weeks of gestation, and was further categorised into extreme prematurity (less than 28 weeks of gestation), severe prematurity (28–31 weeks), and moderate prematurity (32–36 weeks) [
30].
The technique for ultrasound cervical length measurement has been described in detail in prior reports [
26,
31]. Ultrasound cervical length measurements were subjected to a quality assurance protocol. Each ultrasound examination included 3 transvaginal sonographic cervical length measurements, and the shortest measurement was recorded. Fundal pressure was not used to assess the shortest cervical length. When a cervical funnel was present, the cervical length below the funnel was recorded.
For all patients recruited in to this study, a full examination was conducted by the attending physician. We performed standardized medical management of TPL between 24 and 36 weeks of gestation with intact membranes (sterile speculum examination, digital examination, sonographic cervical length measurement, tocolysis with atosiban, and intramuscular betamethasone 2 × 12 mg/24 h), according to recommendations and international clinical standards [
24]. Although all women with TPL should not be treated with antibiotics, antibiotics in women with TPL were used in cases of group B streptococcal colonization. The clinicians were blinded to the PCT results, and the study did not modify patient management.
Maternal sociodemographic characteristics, clinical characteristics at admission, information regarding pregnancy follow up and standard perinatal outcomes were collected prospectively by the midwife or obstetrician and paediatrician responsible for the delivery and the child. Other data were collected by a research assistant, independent of the local medical team, from a prospectively maintained database of women who were included in the study. In addition, we routinely measured newborns’ umbilical arterial blood gases at birth. A paediatrician examined the newborn in all cases after delivery. Infants in need of close monitoring were transferred to the neonatal intensive care unit (NICU).
Blood samples on admission were collected in the polypropylene tubes were immediately centrifuged and the supernatant was then stored at − 35 °C until analysis. All analysis were done in the Department of Biology by the same biologist (FDR) and all PCT levels were measured in serum within 48 h after blood sampling via the immunoluminometric automatic analyser Cobas 6000 e601 (Roche Diagnostics International Ltd., Rotkreuz, Switzerland). The lower limit of detection of the assay was 0.02 ng/mL and the functional assay sensitivity was 0.06 ng/mL (Cobas 6000 e601; PCT package insert; Roche Diagnostics International Ltd). The assay sensitivity and inter- and intra-assay coefficients of variation (CVs) of the PCT kits, as reported by the manufacturers, are 4–10%. WBC count was determined automatically with XE-2100 Automated Haematology System (Sysmex Corporation, Kobe, Japan). CRP was measured quantitatively by immunoturbidometry with the Cobas 6000 c501 system (Roche Diagnostics International Ltd., Rotkreuz, Switzerland). The lowest detection limit of CRP with this method is 0.3 mg/L. In the study group, the cervicovaginal secretion was cultured for aerobic and anaerobic bacteria.
The endpoint was sPTB before 37 weeks of gestation. Women in the case group were defined as included women who were admitted with TPL between 24 and 36 weeks of gestation but presented ultimately sPTB before 37 weeks of gestation. Women in the control group were defined as included women who delivered at term (≥37 weeks of gestation).
Statistical analysis
Demographic data were evaluated using descriptive statistics. Continuous data were described by their means ± standard deviations and compared by t-tests (or Mann-Whitney tests when appropriate); categorical data were described by percentages and compared by chi-square tests (or Fisher exact tests when appropriate). Serum PCT levels were not normally distributed and therefore are reported as medians and interquartile intervals. Univariate analysis were used to compare serum PCT level at admission for TPL and gestational age at birth, and to compare PCT level at admission and admission-to-delivery interval. Receiver operating characteristic (ROC) curves were generated to identify which cut-off of PCT value produced the best sensitivity to predict sPTB within 7 or 14 days after testing. STROBE guidelines/methodology for a prospective observational study were adhered to. Statistical analyses were performed with Statistical Package for Social Sciences (SPSS) software (version 17.0, SPSS Inc., Chicago, IL, USA). A p-value of < 0.05 was considered statistically significant.