A significant gain in weight, which commonly occurs shortly after individuals start their first antiretroviral therapy (ART) regimen, could increase the risk of subsequent cardiovascular disease (CVD) and diabetes in people with HIV, according to an analysis of the Data Collection on Adverse Events of Anti-HIV Drugs (D:A:D) study, presented by Amit Achhra, Ph.D., of the Kirby Institute at the 20th International AIDS Conference (AIDS 2014) in Melbourne, Australia.
In the case of CVD, the effects depended upon the participant's weight before going on ART. A substantial increase in body mass index (BMI) within the first year of going on treatment was associated with an increased risk of CVD in study participants who had a normal BMI before going on ART, but not in those who were underweight, or who were already overweight or obese, when starting treatment. However, regardless of the participant's weight at ART initiation, an increased BMI was associated with an increased risk of diabetes.
The high levels of HIV replication prior to going on treatment can have a significant effect on metabolism (as well as possibly absorption), so weight gains following suppressive ART are common, and have generally been seen as a good prognostic marker. However, gaining too much weight might be cause for concern given the increased metabolic disorders that have also been reported in people on ART. In fact, data from the U.S. Department of Veterans Affairs (VA) suggest that for every 5 lbs of weight gained after starting ART, there is an approximately 10% increased risk of diabetes, though the potential effects of weight gain on CVD have not been as well characterized.
In order to explore the relationship between a short-term change in BMI and the subsequent (after one year) risk of CVD or diabetes mellitus, Achhra and colleagues turned to the D:A:D study, a very large prospective multi-cohort study of people living with HIV that is assessing the incidence of CVD and other comorbidities among study participants and the extent to which these events might be associated with ART. Once those who had had prior CVD or a CVD event within the first year of treatment were excluded, there were 9,321 included in the analysis for whom BMI data were available from before and approximately a year after starting treatment.
CVD in this study was defined as including any of the following: a myocardial infarction (MI), sudden cardiac death, or an invasive procedure (coronary artery bypass graft, carotid endarterectomy or angioplasty) or a confirmed stroke, validated in real time by a defined protocol. Diabetes was verified in D:A:D case report forms or by use of anti-diabetic drugs.
The analyses also allowed for the possibility that pre-ART BMI -- categorized as underweight (BMI <18.5), normal (BMI 18.5-24.9), overweight (BMI >=25) and obese (>=30) -- might affect each of the outcomes, since a gain in weight might mean something different depending upon a person's starting weight.
Overall, the mean change in weight after one year on treatment had been 0.67 BMI units. Notably, however, mean weight gains after initiating ART were quite variable depending on the weight category before going on ART.
- The obese category remained fairly stable or even lost weight.
- Those who were overweight had a stable to slightly increasing rate.
- Those with normal weight had a spike in weight in the first year increasing only slowly thereafter.
- Underweight people gained the most weight (1.82 units at one year), and continued to put on weight, gaining a mean of 2.55 units after five years on treatment.
The proportion of overweight/obese patients increased from 29.6% pre-ART to 35.5% post-ART.
This is a relatively recent cohort -- the median year for starting ART was 2006 -- and 75% of the participants were male. Other than smoking (which was high in the cohort overall, and inversely correlated to weight), this was a low-risk cohort; the mean age was 39 or 40. The median CD4 cell count was lower in the underweight category (170 versus 250-280 in the rest of the cohort). Pre-ART, diabetes was more common in the overweight and obese categories.
CVD events: Overall, there were 97 events -- a rate of 2.21 events/1,000 person-years (95% confidence interval (CI): 1.76-2.68). These included 46 MIs, 33 strokes and 18 invasive procedures. There was a trend toward increased rates based upon increasing pre-ART BMI.
But the pre-ART weight category did seem to have a modifying effect on the subsequent risk of CVD for those who gained weight. Those who were underweight seemed to actually benefit some, although the confidence intervals were wide. After adjustments for demographics and other risk factors, for every unit in BMI gained, there was about an 18% increased risk of CVD in those with “normal” pre-ART BMI -- and this was just statistically significant (P < .041). Perhaps because they were already at increased risk, there was no appreciable increase in risk in those who were overweight or obese.
If pre-ART weight was expressed in quartiles rather than in weight categories, there was a significant increase in CVD risk for every BMI unit gained for the middle two quartiles as well, but again, not in the lowest or highest weight quartile.
Diabetes: Overall, there were 125 events -- a rate of 2.89/1,000 person-years. Rates increased sharply with increasing pre-ART BMI, from 2.04/1,000 person-years in the underweight category up to 9.97/1,000 person-years in the obese.
However, pre-ART BMI did not significantly modify whether someone was at increased risk for diabetes post ART -- for the entire cohort, there was a significant increase in the adjusted risk of incident diabetes per unit BMI. Nevertheless, the confidence intervals at either extreme (underweight or obese) were wide.
The findings were robust for both CVD and diabetes -- holding up in several sensitivity analyses. In one analysis, when people who inject drugs (who tend to have lower BMI, but are at a higher risk for CVD) were excluded, it actually increased the statistical significance of the findings.
One strength of the study was that the cohort was quite heterogeneous. In addition, the outcomes were well documented, with many potential confounders identified. However, it was also a selected sample, and data on lifestyle factors were unavailable. Also, BMI may not really be the best measure of weight -- as it does not account for muscle mass, or necessarily accurately reflect central obesity.
These findings suggest clinicians may want to consider weight management in patients initiating ART. However, some of the findings -- such as that the increase in risk of CVD was largely observed in those with normal-mid-level pre-ART BMI -- may need to be verified in other studies. It is likely that there may be some threshold at which the weight gained becomes riskier -- and those people who put on the most weight could have been the ones having most of the events.
Theo Smart is an HIV activist and medical writer who joined ACT UP New York in 1988 and moved to Cape Town, South Africa, in 2000. As a writer and editor, he has more than 20 years of experience writing and editing about HIV treatment for organizations including ACT UP's Treatment & Data Committee, TAG, the PWA Health Group, GMHC, the Physician's Research Network in New York, amfAR, HIV and AIDS Treatment in Practice (HATIP) and NAM/AIDSmap. He covers human rights, the scale up and delivery of HIV and TB care and treatment services, maternal and child health and revitalizing primary health care and health systems strengthening in resource-limited settings. He also follows the development of targeted health services for people who inject drugs, sexual minorities and commercial sex workers in concentrated epidemics.