Abstract Body

Body mass index (BMI) (weight (kg)/ height(m2)) is a potentially modifiable risk factor for several serious non-AIDS events (SNAEs). However the relationship between BMI and SNAEs in HIV-positive individuals in not well understood.

We followed D:A:D study participants on antiretroviral therapy from their first BMI measurement to the first occurrence of a SNAE or 1/2/2014. The SNAEs of interest, all well-adjudicated, were cardiovascular disease (CVD- composite of myocardial infarction/stroke/invasive cardiovascular procedures); diabetes; non-AIDS-defining malignancies (NADM); BMI-associated cancers (composite of malignancies known to be associated with BMI in general population, including esophagus, pancreas, colon, rectum, breast, endometrium, kidney, thyroid and gallbladder); and all-cause mortality. BMI was time-updated and lagged by 1 year (i.e. there was at least 1 year time-gap between last BMI measurement and a SNAE, so as to minimize bias from reverse causation) and categorised at clinical cut-offs: 18.5, 23, 25, 27.5 and 30 kg/m2. Poisson regression models adjusted for key confounders for each SNAE were used.

During 295,147 person-years of follow-up (PYFU) in 41,149 included individuals, incidence/1000 PYFU of outcomes were: CVD (n=1398): 4.8; diabetes (n=3025): 10.2; NADM (n=1143): 3.9; BMI-cancers (n=184): 0.6 and all-cause mortality (n=3025): 10.2. Participants were largely male (73%) with baseline mean age of 40 years and baseline median BMI of 23.3 (IQR: 21.2- 25.7). A majority of follow-up was in BMI categories of 18.5-23 (41%), 23-25 (22%) and 25-27.5 (17%). Overall, BMI showed a statistically significant J-shaped relationship with the risk of all outcomes except diabetes (Table). There was a higher risk of CVD, NADM, and all-cause mortality at BMI levels <18.5 and at 18.5-23 (especially for NADM and all-cause mortality), compared to the BMI at 23-25. High BMI (>30), compared to that at 23-25, was associated with the higher relative risk of CVD, diabetes, BMI-cancers and possibly all-cause mortality. For diabetes, there was a linear increase in risk with increasing BMI. Results were not sensitive to lagging latest BMI by 2 years (data not shown).

We found that the low BMI, even at the levels of 18-23 and after being lagged by 1 year, was associated with higher short-term risk of CVD, NADM and all-cause mortality in this population. High BMI (>30) was a risk factor for CVD, diabetes, NADM and BMI-cancers.