Background: Ethnicity and obesity have been recently established as risk factors for adverse coronavirus disease 19 (COVID-19) outcomes. The prevalence of obesity differs across ethnic groups. Thus, previously reported ethnic differences in COVID-19 outcomes may be due to underlying variations in BMI.
Aims: To investigate the association between ethnicity and in-hospital mortality in COVID-19, and the effect of BMI. The secondary aim was to investigate the effect of BMI on the relationship between ethnicity and incidence of pulmonary embolus (PE), admission to the intensive treatment unit (ITU), and length of hospitalisation.
Methods: Retrospective cohort study of 149 patients with confirmed COVID-19 and thoracic CT scans (for evaluation of PE) from Guys and St Thomas NHS Foundation Trust, London, United Kingdom. Data on self-reported ethnicity and clinical outcomes were extracted from electronic patient records. The 2018 demographic of the local area was obtained from the Office of National Statistics. Multivariate logistic and linear regression analysis was performed to explore ethnicity as a risk factor for the main outcomes of interest, adjusted for BMI.
Results: Mean (± standard deviation) age was 57 ± 16 years; 102 (68%) were male; 59 (40%) were of Black, Asian, and minority ethnic backgrounds; median (interquartile range) of hospitalisation duration was 27 (9 53) days; 31 (21%) died; 97 (65%) were admitted to ITU; and 30 (20%) had a PE. There was a greater percentage of patients of Black ethnicity than the local population (27 vs 20%). BMI was not different across ethnic groups (White vs Black vs Other, 29.3 ± 6.9 vs 29.7 ± 6.8 vs vs 27.2 ± 5.1 kg/m2; P=0.438). After adjustment for BMI, ethnicity was not related to adverse COVID-19 outcomes, including mortality, and presence of PE.
Conclusion: BMI did not modify the effect of ethnicity in predicting adverse COVID-19 outcomes.