Background Early onset neonatal sepsis is an important cause of morbidity and mortality but remains a diagnostic challenge. C-reactive protein (CRP) is widely used to assist diagnosis of infection and monitor treatment response. CRP is synthesised by the liver in response to interleukin 6 within 6–8 h of tissue injury. Despite widespread use in infants with suspected sepsis, little is known about CRP responses after birth in the absence of infection.

Objective To describe postnatal CRP responses in asymptomatic term and preterm infants with no evidence of infection

Methods Data was collected from infants admitted to a tertiary NICU during two discrete time periods (Sept–Oct 2012; Apr–May 2013) who had blood taken for culture. CRP values and time from birth were recorded. Infants with symptoms of infection or a positive blood culture were excluded from analysis, as were those with congenital anomalies. Remaining infants were subdivided by gestational age. Normalised CRP curves were generated by linear interpolation between measured values, assuming CRP half-life of 18h. Individual subjects were ranked by CRP response and centile curves derived.

Results 219 babies were screened in the study period. After exclusions, 85 infants (70 term, 15 preterm) remained. The median (range) birthweight and gestation were 3560 g (2080–4820 g) and 40⁺⁴ weeks (37–43) respectively in the term group and 1814 g (975–3020 g) and 34⁺² weeks (27⁺⁵–36⁺⁴) in the preterm group. In asymptomatic term neonates, CRP peaked at 24 h of age, with the 75^th centile 9.3 g/dL, 90^th centile 21 g/dL, 95^th centile 24 g/dL. In preterm babies, CRP peaked at 48 h, 90^th centile=5 g/dL, 95^th centile=19 g/dL. Excluding infants with risk factors for infection, 90^th centile values remained raised in both groups.

Conclusions We describe CRP centiles in infants without symptoms of infection in the first days after birth. A substantial proportion of uninfected term and preterm infants had raised CRP values, with more than 10% being sufficiently high (>10g/dL) to prompt further invasive investigations (current UK guidelines). We demonstrate raised CRP values, unrelated to infection, in a significant number of infants. Further characterisation of CRP responses in non-infected infants is needed to optimise use of this common test.