Procalcitonin and C-reactive protein for diagnosing post-operative sepsis in neonates.

Abstract

To determine whether serum procalcitonin (PCT) or C-reactive protein (CRP) can diagnose post-operative sepsis among neonates undergoing major non-cardiac surgery. In this diagnostic study, we included neonates who underwent major non-cardiac surgery and were monitored for post-operative sepsis. We excluded pre-existing septic, inflammatory or life-threatening conditions. Subjects either had 'definite' (culture-positive, n=14), 'probable' (clinical sepsis, culture-negative, n=25) or no sepsis (n=31). We measured serum CRP and PCT at 48 ± 6h, 72 ± 6h and 96 ± 6h post-operatively and compared 'definite or probable sepsis' with 'no sepsis'. Median (Q1, Q3) CRP (mg/L) in 'definite or probable' sepsis group was higher than 'no sepsis' at 72 h (91.48 (57.87, 143.50) vs. 51.32 (33.0, 80.1); P=0.009) and 96 h (87.51 (45.19, 128.22) vs. 31.00 (25.3, 45.2); P < 0.001). Median (Q1, Q3) PCT (ng/mL) in 'definite or probable' sepsis was higher than 'no sepsis' at 72 h (4.22 (2.04, 12.73) vs. 1.78 (0.9, 6.4); P=0.01) and 96 h (3.54 (1.96, 9.65) vs. 0.97 (0.4, 3.0); P < 0.001). Ninety-six-hour CRP and PCT cut-offs (based on Youden's index) were 74.16 mg/L and 1.65 ng/mL, respectively. If both CRP and PCT were positive, specificity was 100% (95% confidence interval: 88.78-100). If either one was positive, sensitivity was 88.89% (95% confidence interval: 73.94-96.89). Septic neonates have significantly higher serum CRP and PCT compared to non-septic neonates at 72 and 96 h post-operatively. If both CRP and PCT are positive at 96 h after surgery, it has 100% specificity, and if either one is positive, 89% sensitivity.