Being Underweight or Obese Tied to Serious Non-AIDS Illnesses in D:A:D Study
Low body mass index (BMI) -- even below 23 kg/m2 -- predicted serious non-AIDS events and all-cause mortality in a 41,149-person D:A:D study analysis. Obesity, but not being overweight, also predicted mortality and serious non-AIDS events, including cardiovascular disease and cancers.
Overweight and obesity prevalence is rising in HIV-positive people taking antiretroviral therapy (ART). Research in the general population links being overweight and obesity to cardiovascular disease, diabetes, certain cancers and mortality. Low BMI also carries risks. D:A:D: study investigators scrutinized their cohort to gauge the impact of BMI on serious non-AIDS events and death in adults taking ART.
D:A:D is a multi-cohort prospective study in Europe, the United States and Australia. The analysis included D:A:D members with at least one BMI measured when they started ART or afterwards and at least one year of follow-up. The researchers excluded people who already had cardiovascular disease, diabetes or cancer. Follow-up continued until a non-AIDS event developed or until death, six months after the last follow-up visit or February 2014. Centrally adjudicated events were cardiovascular disease (myocardial infarction, stroke or invasive cardiovascular procedures), diabetes, non-AIDS cancers, BMI-related non-AIDS cancers (esophageal, pancreatic, colon, rectal, breast, endometrial, kidney, thyroid, gallbladder) and all-cause mortality.
The investigators considered BMI as a time-updated variable lagged by one or two years (meaning events counted only if they occurred one or two years after the BMI measurement). They divided participants into six BMI groups: under 18.5 kg/m2, 18.5 to 23 kg/m2, 23 to 25 kg/m2, 25 to 27.5 kg/m2, 27.5 to 30 kg/m2 and more than 30 kg/m2. In the general population a BMI below 18.5 kg/m2 is considered underweight, 18.5 to 25 kg/m2 normal, 25 to 30 kg/m2 overweight and above 30 kg/m2 obese. The six BMI ranges in the D:A:D analysis included (in order) 1,929; 17,640; 9,283; 6,998; 2,910 and 2,389 people. Baseline age averaged 40.4 years, 73.5% of participants were men and median baseline CD4 count measured 410 cells/mm3.
Through 295,147 person-years of follow-up, event incidence per 1000 person-years measured 4.8 for cardiovascular disease, 4.2 for diabetes, 3.9 for non-AIDS cancers, 0.6 for BMI-related non-AIDS cancers, 11.2 for all-cause mortality in men and 7.8 for all-cause mortality in women.
Poisson regression models adjusted for key confounders consistently linked BMI below 18.5 or between 18.5 and 23 kg/m2 to higher relative risk of all-cause mortality and of all serious non-AIDS events except diabetes when compared with a BMI of 23 to 25 kg/m2 (the reference value throughout). Regression analysis tied BMI above 30 kg/m2 (but not between 25 and 30 kg/m2) to higher relative risk of cardiovascular disease, BMI-related non-AIDS cancers and all-cause mortality. Diabetes had a linear relationship with BMI, climbing from an incidence rate ratio (IRR) below 1.0 in people below 23 kg/m2 and rising through BMI brackets above 25 kg/m2 to an IRR of about 3.5 in obese people. Results were similar for all outcomes when the researchers lagged time-updated BMI by two years instead of one year.
Except with diabetes, the researchers conclude, BMI had a statistically significant J-shaped relationship to risk of serious non-AIDS events and mortality: In other words, compared with a BMI of 23 to 25 kg/m2, a BMI below 23 kg/m2 or above 30 kg/m2 was associated with higher risk of non-AIDS events or death. Notably, a BMI of 20 to 23 kg/m2 is typically considered normal. The D:A:D: team cautions that the findings are limited by the lower number of cohort members in the upper and lower extremes of BMI. They also note that BMI has limits as a marker of body weight or fat.