Preterm delivery predicted by soluble CD163 and CRP in women with symptoms of preterm delivery

OBJECTIVE: To evaluate whether soluble CD163 (sCD163) and C-reactive protein (CRP) can predict spontaneous preterm delivery in women with symptoms of preterm delivery.

DESIGN: Prospective cohort study. Setting Labour ward at a tertiary university hospital.

POPULATION: Ninety-three women with symptoms of preterm delivery before 34 weeks of gestation.

METHODS: sCD163 and CRP were individually examined as predictors of preterm delivery. A model for prediction of preterm delivery was established using exact logistic regression for risk factors individually associated with preterm delivery.

MAIN OUTCOME MEASURES: Gestational age at delivery.

RESULTS: In women with symptoms of preterm delivery, median sCD163 and CRP levels were significantly higher statistically in women delivering preterm (3.4 mg/L, and 62 nmol/L) compared with the women delivering at term (2.7 mg/L, and <48 nmol/L, Mann-Whitney U test, P < 0.01 and P < 0.001) for sCD163 and CRP, respectively. sCD163 above 5 mg/L was associated with an increased risk of preterm delivery (crude OR = 10, [95% CI 1.3-466], adjusted OR = 27, [0.7-infinity]). CRP above 47 mg/L was associated with an increased risk of preterm delivery (crude OR = 5 [1.8-14], adjusted OR = 5 [1.04-31]). Likelihood ratio of a positive test was 8.6 [2.8-14] and 2.8 [0-6.2] for sCD163 and CRP, respectively. The logistic regression model was able to predict 85% of preterm deliveries with 13% false positive, giving a likelihood ratio of 8 [2.2-13.5].

CONCLUSION: High levels of sCD163 or CRP are associated with an increased risk of preterm delivery in women with symptoms of delivery. Good prediction of preterm delivery before 34 weeks of gestation was obtained by a combination of preterm prelabour rupture of membranes (PPROM), overweight, relaxin, CRP and sCD163.