Authors’ reply to ‘Is there a role for C-reactive protein during and after labour?’

Dear Editor,

We wish to acknowledge the importance of the question posed by Dockree et al. in this edition of the Annals of Clinical Biochemistry. While these authors agree that C-reactive protein (CRP) is elevated during pregnancy, they raise concerns regarding the maximum CRP reference value of 34.8 mg/L observed in healthy women at term pregnancy in our prospective study ¹ compared with the upper reference limit (URL) of 19 mg/L established in their study. ² Two main reasons for this discordance are suggested: the rigour in defining the health of the pregnant women recruited and the method employed to establish the URL.

In our study, reference values were established in a well-characterized cohort of healthy pregnant women with an uncomplicated singleton pregnancy and no evidence of clinical infection (i.e. urinary tract infection, upper respiratory tract infection, gastroenteritis or influenza) or any of the following: diabetes, gestational diabetes, hypertensive disease of pregnancy, cardiac disease, autoimmune disease, chronic asthma, renal disease or abnormal liver chemistry tests. ¹

Dockree et al. ² measured CRP by immunoturbidimetry on the Abbott Architect establishing the URL for CRP of 19 mg/L in trimester 3 (gestation age 31–38⁺⁶ weeks) using the International Federation of Clinical Chemistry (IFCC) recommended method. ³ The 90% confidence interval of the URL was wide, ranging from 16.10 to 29.97 mg/L. ² In contrast, we measured CRP using the Beckman Coulter latex-immune-turbidimetric assay on the Olympus AU640 analyser and reported the observed minimum and maximum CRP values in our reference population of pregnant women at 37–40 weeks’ gestation. Both CRP assays are traceable to the IFCC standard CRM 470; however, review of external quality assessment data demonstrates differences in assay bias which may be a contributing factor.

We determined that CRP had poor clinical utility during and after labour, a finding supported by two systematic reviews that reported on the use of CRP at an action-limit of ≥20 mg/L, almost identical to that reported by Dockree et al.^4,5 Trochez-Martinez et al., recognizing an association between CRP concentrations and chorioamnionitis, found no evidence to support its use as an early biomarker of sepsis and cautioned against the use of this potentially misleading threshold. ⁴ Etyang et al. reported poor sensitivity and specificity, 59% (95% CI: 48–69) and 83% (95% CI: 74–89) using the ≥20 mg/L cut-off for chorioamnionitis in preterm-premature-rupture-of-membranes. ⁵ Arguably, the respective sensitivity (73%) and specificity (86%) reported by Dockree et al. ² are only moderately better than those reported by Etyang et al. ⁵ Of note, this is broadly in agreement with our study which determined that implementation of the ≥20 mg/L action-limit for CRP would lead to inappropriate investigations and interventions in 14% of healthy pregnancies.

Hence, we agree with Dockree et al. that the significant variability of CRP in the peripartum precludes its use to diagnose pregnancies complicated by sepsis and that there is a pressing need for a suitably powered diagnostic accuracy study to examine the potential clinical utility of procalcitonin and its application in obstetrics.