Sting-to-Care Interval and Recorded Cardiorespiratory Complications After Scorpion Stings in Brazil, 2008-2022: A Nationwide Cross-Sectional Study of Surveillance Notifications

BackgroundCardiorespiratory complications after scorpion stings can be studied in surveillance data, but their measurement is heterogeneous.We assessed the association between the recorded sting-to-care interval and these complications in Brazil, including an analysis of recorded shock in Pará state. MethodologyWe analyzed scorpion sting notifications from the Brazilian Notifiable Diseases Information System (SINAN), 2008-2022.The exposure comprised six interval categories; the primary outcome was a composite of recorded shock or respiratory complication.Logistic regression was adjusted for age, sex, year, and federative unit, with municipality-clustered robust variance, and estimated standardized risk ratios (RRs). ResultsOf the 1,563,983 notifications, 1,296,606 entered the primary model, with 4,547 events.The interval was globally associated with the composite (Wald p = 3.47 × 10⁻¹⁴); RRs for 1-3, 3-6, 6-12, 12-24, and ≥24 hours versus 0-1 hours were 1.40, 1.70, 2.37, 3.30, and 1.98, respectively.All four secondary outcomes retained global associations after Holm correction.The association persisted without Pará, but a score bootstrap clustered by federative unit did not reach a p-value <0.05 (p = 0.0735).Pará accounted for 2,087 of 3,218 recorded shocks; its adjusted RR for shock fell from 79.46 under the broad definition to 0.97 (95% confidence interval = 0.39-2.38;eight events in Pará) under the strict definition. ConclusionsIn this exploratory analysis of surveillance records, the recorded interval was associated with recorded cardiorespiratory complications, with non-monotonic estimates; evidence varied with the inferential procedure and outcome definition.The heterogeneity of shock records in Pará calls for clinical validation and improved recording; it does not demonstrate a causal effect of delay, a shock mechanism, or individual recording error.