Exploring the impact of using measured or estimated values for height and weight on the relationship between BMI and acute hospital mortality

Abstract Purpose Studies have demonstrated an association between height and weight and mortality among patients in the Intensive Care Unit (ICU) and the optimal body mass index (BMI) might be well above the optimal values in the general population. Most of these studies have relied on estimated values, the validity of which is not known. Material and methods Admissions to adult general ICUs from 1 April 2009 to 31 March 2016 in the Case Mix Programme (CMP) Database were described by height and weight assessment methods (measured or estimated). A multilevel logistic regression model was built, which had acute hospital mortality as the outcome and included standard case mix adjustment, BMI, the assessment method and the interactions between BMI and assessment method. Results There were 690,405 eligible admissions and most patients (59.7%) had estimates of height and/or weight recorded. Patients with both height and weight measured had lower severity and mortality. The association between BMI and mortality was reverse J-shaped with the lowest mortality at BMI 34.3 kg/m2. Whether height and weight were measured or estimated did not influence the association between BMI and mortality. Conclusions For epidemiological comparisons of mortality among critically ill adults, estimated values of height and weight appear valid.