Two D:A:D updates: cardiovascular risk from HAART, and predictors of hypertension and changes in blood pressure
Simon Collins, HIV i-Base
The D:A:D (Data Collection on Adverse Events of Anti-HIV Drugs) study is the largest cohort study initiated to investigate whether HAART was linked to increased risk of cardiovascular disease (CVD).
At the 2003 Retrovirus conference an analysis from the first 123 incidents of myocardial infarction (MI) predicted a cumulative relative rate of 1.26 with every year of HAART. While the absolute risk remained low, and was still clearly outweighed by the benefits of HAART, it was clear that health advice given to the general population was equally if not more important for HIV-positive patients.
Two posters were presented at this year’s meeting that included new information from the same study.
In poster 737, an analysis based on other predictors of CVD (rather than just MI as in last year’s analysis) confirmed previous cumulative risk associated with HAART and suggested that the causal mechanism was similar for this composite endpoint. 
The observed rates of MI seen in D:A:D were also compared to the rates that would have been predicted from the Framingham equation. Best predictions of MI rates assumed a five-fold increased risk of MI with previous CVD with upper limits assuming that features of metabolic syndrome (lipodystrophy, BMI >30kg/m2 and triglycerides >2.3 mmol/L) carried a similar increased risk as pre-existing diabetes in the general population; and lower limits assuming that these metabolic changes contribute no clinical risk over five-10 years.
Patients not receiving HAART had fewer events than predicted (three observed vs 7.6 predicted). Patients on HAART had slightly higher numbers of MI than best predicted rates but generally lower than the upper limit predictions. The trend for higher events related to duration of HAART was similar to those that would have been expected based on known risk factors. The study concluded that the observed increase in risk of MI could largely be explained by HAART-induced changes in conventional CVD risk factors. On these data there is no cause to change last year’s recommendations concerning diet, exercise and smoking cessation.
A second study analyses longitudinal changes in blood pressure (BP) from just over 16,000 patients on the D:A:D database. More than 43,000 BP measurements were included (median three per patient over a median of 1.5 years [IQR 0.8-1.7]. 
Risk factors significantly associated with a higher predicted increase in systolic BP (>/= 5mmHg) were older age (+12.8 and 14.5 mmHg at baseline and month 24, respectively, for 60 years old versus 30 years old), male (+7.0 and +6.6 at baseline and month 24, respectively, male versus female), higher BMI (+16.4 and +15.6 at baseline and month 24, respectively, for those with BMI>30kg/m_ versus <18) and blood pressure-lowering drugs (+8.3 mmHg at baseline and month 24 for those treated with an antihypertensive treatment).
In 8,341 patients with normal BP at baseline, 487 developed hypertension providing an incidence of 35.8/1000 patient years. Factors associated with hypertension were similar to those in the general population: older age, male gender and higher BMI. Cumulative exposure to any individual drug class or type of treatment at baseline was not associated with change in BP or risk of hypertension.
Because of the fortunately small number of events even such a large cohort as D:A:D is not powered to allow a subanalysis of the association of cardiovascular events and antiretroviral classes or even specific drugs. Evidence-based recommendations cannot therefore be made for specific antiretroviral drugs.
In addition, due to the huge variability as a result of the low number of events and the limited data base in some aspects D:A:D is not able to exclude additional less pronounced risk factors, which may explain the rapid rise in cardiovascular events after only one year of exposure to antiretrovirals.
- Law M, D’Armino Monforte A, Friis Moller N et al. Cardio- and cerebrovascular events (CVE) and predicted rates of myocardial infarction (MI) in the D:A:D study. 11th CROI 2004, Abstract 737.
- Thiebaart R, El-Sadr W, Chenue G et al. Predictors of hypertension and changes in blood pressure in HIV-infected patients in the D:A:D study. 11th CROI 2004, Abstract 75.