Population-based surveillance for congenital zika virus syndrome : a latent class analysis of recorded cases from 2015–2018

Abstract

Abstract Objective: This study aims to describe clinical fndings and determine the medium-term survival of congenital zika syndrome (CZS) suspected cases. Methods: A retrospective cohort study using routine register-based linked data. It included all suspected cases of CZS born in Brazil from January 1, 2015, to December 31, 2018, and followed up from birth until death, 36 months, or December 31, 2018, whichever came frst. Latent class analysis was used to cluster unconfrmed cases into classes with similar combinations of anthropometry at birth, imaging fndings, maternally reported rash, region, and year of birth. Kaplan–Meier curves were plotted, and Cox proportional hazards models were ftted to determine mortality up to 36 months Results: We followed 11,850 suspected cases of CZS, of which 28.3% were confrmed, 9.3% inconclusive and 62.4% unconfrmed. Confrmed cases had almost two times higher mortality when compared with unconfrmed cases. Among unconfrmed cases, we identifed three distinct clusters with diferent mortality trajectories. The highest mor‑ tality risk was observed in those with abnormal imaging fndings compatible with congenital infections (HR=12.6; IC95%8.8–18.0) and other abnormalities (HR=11.6; IC95%8.6–15.6) compared with those with normal imaging fndings. The risk was high in those with severe microcephaly (HR=8.2; IC95%6.4–10.6) and macrocephaly (HR=6.6; IC95%4.5–9.7) compared with normal head size. Conclusion: Abnormal imaging and head circumference appear to be the main drivers of the increased mortality among suspected cases of CZS. We suggest identifying children who are more likely to die and have a greater need to optimise interventions and resource allocation regardless of the fnal diagnoses.