Combination of cervical interleukin-6 and -8, phosphorylated insulin-like growth factor-binding protein-1 and transvaginal cervical ultrasonography in assessment of the risk of preterm birth

Abstract

Objective To determine the value of combinations of cervical interleukin-6 (IL-6), cervical interleukin-8 (IL-8), the phosphorylated isoform of insulin-like growth-factor binding protein-1 (IGFBP-1), and cervical ultrasonography in the prediction of preterm birth. Design Prospective follow up. Setting Oulu University Hospital maternity clinic from February 1997 to July 1998. Population Women with singleton pregnancies ( n = 77), referred from outpatient clinics at 22–32 weeks of gestation with symptoms (uterine contractions) or signs (cervical change) of threatened preterm birth. Symptomless women ( n = 78) matched for gestational age, parity and maternal age at recruitment were studied as a reference group. Methods A urine sample for bacterial culture was collected, and cervical swab samples for assays of interleukin-6 and -8 and phoshorylated IGFBP-1 were taken before digital cervical examination. A Pap smear for analysis of bacterial vaginosis and samples for analysis of chlamydia and streptococci were also obtained. Cervical measurements were made by transvaginal ultrasonography. The same sampling and cervical measurement were repeated twice at two-week intervals. The cutoff values of the markers were determined by receiver-operating characteristic curve analysis. Main outcome measure Preterm birth (<37 weeks). Results The preterm birth (<37 weeks) rate for women in the study group was 16% (12/77). The cervical interleukin-6 cutoff value (61 ng/L) at first visit had a sensitivity of 73% and a specificity of 61% in predicting preterm birth, with a positive likelihood ratio (LR+) of 1.9 (95% CI 1.2–3.0). An ultrasonographically measured cervical index value of > 0.36 at recruitment predicted preterm birth in 25% (5/20) of the study group compared with 9% (5/54); LR+ 2.2 (95% CI 1.03–4.7). Cervical phosphorylated IGFBP-1 > 6.4μg/L [LR+ 1.8 (95% CI 0.7–2.9)], interleukin-8 > 3739 ng/L [LR+ 1.4 (95% CI 0.9–2.4)], and ultrasonograpic cervical length < 29.3 mm [LR+ 2.7 (95% CI 0.8–9.7)] increased the risk of preterm birth. According to the logistic regression model, a combination of IL-6, and IL-8 and cervical index increased the specificity to 97%, but the sensitivity fell to 30% in detecting preterm birth. There was a significantly increased incidence of puerperal infections if phosphorylated IGFBP-1 concentrations were elevated (> 21.0 μg/L), 36% (4/11) compared with 4.6% (3/65), LR+ 6.7 (95% CI 2.7–17), the sensitivity being 67% (4/6) and the specificity 90% (63/70). Elevated phosphorylated IGFBP-1 concentrations (> 21.6μg/L) were also associated with an increased risk of neonatal infections; LR+ 8.0 (95% CI 3.5–18). Conclusions An increase in cervical IL-6 concentration and the ultrasonographically measured cervical index appear to be associated with preterm birth. A combination of these markers with measurement of cervical IL-8 appears to be the best predictor of preterm birth. Neither the sensitivity nor specificity of the tests used in this study are good enough to predict preterm birth for clinical decision making. Cervical phosphorylated IGFBP-1 seems to be a marker of puerperal and neonatal infectious morbidity in cases of threatened preterm delivery, suggesting early tissue degradation at the choriodecidual interface.