Background: The epidemiology of neonatal early-onset sepsis (EOS) has changed with time and with changes in prevention strategy. Population-representative contemporary data provide insights on how to further improve EOS prevention and triage strategies. Methods: Neonates born in public hospitals in Hong Kong from January 1, 2006, to December 31, 2017 were included. The epidemiological characteristics of EOS and the use of intrapartum antibiotic prophylaxis (IAP) were compared between the 2 epochs before (January 1, 2006 to December 31, 2011) and after (January 1, 2012 to December 31, 2017) the territory-wide implementation of universal maternal group B Streptococcus (GBS) screening. Results: EOS developed in 1.07‰ of live births (522/490,034). After the implementation of universal GBS screening, the EOS rate decreased in neonates born ≥34 weeks (1.17‰–0.56‰, P < 0.001) and remained similar in those born <34 weeks (7.8‰–10.9‰, P = 0.15), whereas the proportions of IAP coverage increased in both groups [7.6%–23.3% (P < 0.001) and 28.5%–52.0% (P < 0.001), respectively]. The major pathogen for EOS shifted from GBS to Escherichia coli, and for early-onset meningitis from GBS to Streptococcus bovis. IAP was associated with subsequent isolation of pathogens resistant to ampicillin [adjusted odds ratio (aOR) 2.3; 95% confidence interval (CI): 1.3–4.2], and second-generation [aOR 2.0; 95% CI: 1.02–4.3] and third-generation [aOR 2.2; 95% CI: 1.1–5.0] cephalosporins. Conclusions: Pathogen profile of EOS changed with the implementation of universal GBS screening. S. bovis has emerged as a more common pathogen associated with the risk of meningitis. IAP may not be as effective in reducing EOS rate among infants born <34 weeks as compared with those ≥34 weeks, and newer strategies may be needed.

ObjectiveTo determine the fitness of the INTERGROWTH-21st birth weight standards (INTERGROWTH21) for ethnic Chinese babies compared with a local reference (FOK2003).DesignPopulation-based analysis of territory-wide birth data.SettingAll public hospitals in Hong Kong.ParticipantsLive births between 24 and 42 complete weeks’ gestation during 2006–2017.Main outcome measuresBabies’ birth weight Z-scores were calculated using published methods. The two references were compared in three aspects: (1) the proportions of large-for-gestational-age (LGA) or small-for-gestational-age (SGA) infants, (2) the gestation-specific and sex-specific mean birth weight Z-scores and (3) the predictive power for SGA-related complications.Results488 896 infants were included. Using INTERGROWTH21, among neonates born <33 weeks’ gestation, the mean birth weight Z-scores per week were closer to zero (−0.2 to 0.05), while most of them were further from zero (0.06 to 0.34) after excluding infants with a high risk of abnormal intrauterine growth. Compared with FOK2003, INTERGROWTH21 classified smaller proportions of infants as SGA (8.3% vs 9.6%) and LGA (6.6% vs 7.9%), especially SGA among preterm infants (13.1% vs 17.0%). The area under the receiver operating characteristic curve for predicting SGA-related complications was greater with FOK2003 (0.674, 95% CI 0.670 to 0.677) than INTERGROWTH21 (0.658, 95% CI 0.655 to 0.661) (p<0.001).ConclusionsINTERGROWTH21 performed less well than FOK2003, a local reference for ethnic Chinese babies, especially in infants born <33 weeks’ gestation. Although the differences are clinically small, both these references performed poorly for extremely preterm infants, and thus a more robust chart based on a larger sample of appropriately selected infants is needed.

No abstract