Martins, Rand RandallLeopoldino, Ramon Weyler Duarte2023-09-232023-09-232023-07-12LEOPOLDINO, Ramon Weyler Duarte. Desenvolvimento e validação de instrumento para predição de reações adversas a medicamentos em neonatos internados em unidade de terapia intensiva. Orientador: Rand Randall Martins. 2023. 88f. Tese (Doutorado em Ciências Farmacêuticas) - Centro de Ciências da Saúde, Universidade Federal do Rio Grande do Norte, Natal, 2023.https://repositorio.ufrn.br/handle/123456789/54887Introduction: The Adverse Drug Reactions (ADR) are common in hospitalized neonates, but their detection is made more difficult in these individuals by the severe health status and rapid physiological changes related to organ maturation. The ADR predictive tools are the best care strategies. However, these instruments are not yet available to the neonatal population. Objectives: To develop and validate an instrument for predicting ADR in neonates admitted to a Neonatal Intensive Care Unit (NICU). Methods: This is a nested case-control study in an open cohort conducted from January 2019 to January 2022 with neonates admitted to the NICU of a public maternity hospital in Natal/Brazil. All neonates admitted to the NICU for more than 24 hours were included in the study. ADR were detected by actively searching electronic medical records and analyzing spontaneous notification. For the cohort, ADR were characterized for incidence, prevalence, and risk factors via a multivariate model. Subsequently, patients with ADR were randomly matched with two controls in a 1:2 ratio, allowing the development of a predictive model of ADR by multivariate logistic regression. The discriminative ability of the model was assessed by the area under the ROC curve by bootstrap adjusted cstatistic. From a sample of ADR cases, the performance of the Du and Naranjo algorithms was analyzed by three independent assessors. For these tools, inter-rater, inter-rater and inter-tool agreement (kappa coefficient) were calculated. Results: With a cohort of 600 neonates, the prevalence of ADR was 19.7% (95% CI 16.7 - 23.0%) and incidence of 310 cases per 1,000 neonates (95% CI 267.8 - 357.0). The most frequent reactions were tachycardia (30.6%), polyuria (9.1%) and hypokalemia (8.6%). The long hospital stay (OR 1.025 95%CI 1.011 - 1.038; p < 0.01) and polypharmacy (OR 1.121 95%CI 1.057 - 1.189; p < 0.01) were factors associated with ADR. For the development of the predictive instrument, 450 neonates (150 cases and 300 controls) were included in the analysis. The instrument entered four conditions from the neonate (mechanical ventilation, heart rate ≥ 178 bpm, intravenous medications, 5 or more medications) and one from the mother (gestational hypertension) for ADR prediction and obtained an area under the ROC curve of 0.74 (95% CI 0.68 - 0.80). In the performance of the Du and Naranjo algorithms, the former showed good ability to identify defined ADR (≈60%), but had low reproducibility (overall k = -0.031; 95%CI -0.049 - 0.065). The results of the algorithms were not equivalent (overall k = -0.031; 95% CI -0.049 - 0.065). Conclusion: The ADR frequently observed in the NICU were tachycardia, hyperthermia, polyuria, and hypokalemia. Based on five variables commonly monitored in the NICU, a simple, well-calibrated predictive instrument with adequate discriminative capacity was constructed. Regarding the performance of the causality algorithms, Du's algorithm showed promise for identifying ADR in NICU due to its good ability to classify reactions as defined.Acesso AbertoTerapia intensiva neonatalRecém-nascidoReações adversas a medicamentos - neonatosSegurança do pacienteFarmacovigilânciadoctoralThesisCNPQ::CIENCIAS DA SAUDE::FARMACIA