[Apnoea in infants with bronchiolitis: Incidence and risk factors for a prediction model]

Abstract

Introduction: The presence of apnoea in acute bronchiolitis (AB) varies between 1.2% and 28.8%, depending on the series, and is one of its most fearsome complications. The aim of this study is to determine the incidence of apnoea in hospitalised patients diagnosed with AB, and to define their associated risk factors in order to construct a prediction model.

Patients and method: A retrospective observational study of patients admitted to a tertiary hospital in the last 5 years with a diagnosis of AB, according to the classic criteria. Data was collected on the frequency of apnoea and related clinical variables to find risk factors in a binary logistic regression model for the prediction of apnoea. A ROC curve was developed with the model.

Results: Apnoea was recorded during the admission of 53 (4.4%) patients out of a total 1,197 cases found. The risk factors included in the equation were: Female (OR 0.6, 95% CI: 0.27-1.37), Caesarean delivery (OR: 3.44, 95% CI: 1.5-7.7), Postmenstrual age ≤43 weeks (OR: 6.62, 95% CI: 2.38-18.7), Fever (OR: 0.33, 95% CI: 0.09-1.97), Low birth weight (OR: 5.93, 95% CI: 2.23-7.67), Apnoea observed by caregivers before admission (OR: 5.93, 95% CI: 2.64-13.3), and severe bacterial infection (OR: 3.98, 95% CI: 1.68-9.46). The optimal sensitivity and specificity of the model in the ROC curve was 0.842 and 0.846, respectively (P<.001).

Conclusions: The incidence of apnoea during admission was 4.4 per 100 admissions of AB and year. The estimated prediction model equation may be of help to the clinician in order to classify patients with increased risk of apnoea during admission due to AB.

Keywords: Apnea; Apnoea; Bronchiolitis; Bronquiolitis; Caesarean delivery; Cesárea; Infants; Lactantes; Logistic regression model; Modelo de regresión logística.