Boosted darunavir is associated with higher cardiovascular risk in D:A:D study, but not atazanavir

Simon Collins, HIV i-Base

Compared to HIV positive people using other HIV drugs, cumulative use of the boosted protease inhibitor darunavir/ritonavir was associated with an approximately 60% higher rate of developing heart disease over five years, in an analysis from the D:A:D study. Similar results were not found for atazanavir/ritonavir.

Importantly, the analysis adjusted for other factors linked to cardiovascular risk and were independent of dyslipidaemia. The study was authored by Lene Ryom and colleagues and published online in Lancet HIV. [1]

The D:A:D study is a large prospective cohort study designed to look for serious non-AIDS events including cardiovascular risk of HIV drugs. The current analysis included data from 35,711 HIV positive people with follow-up data from 2009 (approximately 70% of the original D:A:D cohort) to look at the currently used protease inhibitors atazanavir and darunavir. Cardiovascular disease (CVD) was defined as myocardial infarction, stroke, sudden cardiac death, or use of invasive cardiovascular procedures, including coronary bypass, coronary angioplasty, and carotid endarterectomy.

D:A:D is drawn from 11 cohorts mainly in Europe, but also Australia and the USA. Baseline demographics include median age 44 (IQR: 38 to 51), 74% male and 48% white (43% ethnicity unknown). Median CD4 count was 501 cells/mm3(IQR: 360 to 689), more than 85% were ART-experienced and 76% were virologically suppressed. Traditional cardiovascular risk factors were common: 40% had dyslipidaemia, 39% were current smokers (23% previous), 10% had hypertension, 5% had diabetes, and 1% had previous CVD. 

At baseline, exposure to atazanavir/ritonavir and daraunavir/ritonavir was 18% and 4% respectively, increasing to 26% and 22%, respectively, at the time of the last study visit.

During a median 6.96 years of follow-up (IQR 6.28 to 7.08), 1157 people developed CVD,with an overall incidence rate of 5.34 events per 1000 person-years (PY) (95% CI: 5.03 to 5.65).

The most common CVD events were angioplasty (n=459), type 1 myocardial infarction (n=454), stroke (n=379) and bypass (n=93), with multiple events possible for one individual on the day the CVD event was registered.

The CVD impact of darunavir/ritonavir was cumulative with longer duration of use. The event rate (54% CI) was 4.91 (4.59 to 5.23) per 1000 PY in individuals who had not used to ritonavir-boosted darunavir compared to 13.67 (8.51 to 18.82) for people who used darunavir/ritonavir for more than six years. 

This produced an adjusted incidence rate ratio (IRR) of 1.59 (95% CI: 1.33 to 1.91) per five years of exposure to darunavir/ritonavir. This compared to a non-significant IRR of 1.03 per 5 years 0.90 to 1.18) with atazanavir/ritonavir.

This association was supported in several sensitivity analyses.

The analysis also calculated the number needed to treat with darunavir/ritonavir to produce one case of harm (NNTH), stratified by the underlying estimated 5 year CVD risk (using the online D:A:D calculator [2]). This produced an NNTH of 15 (95% CI: 13 to 17) for those at high risk and 533 (95% CI: 314 to 706) for those at low risk (defined at >10% and <1% risk over five years, respectively).

Although D:A:D doesn’t collect  details on drug dosing similar associations with darunavir/ritonavir were seen in both people using first-line ART and those with previous history of viral failure (when the higher dose of darunavir does would be more likely to be used.


As with all observational studies, these results include a caution that they only show association rather than a causative link.

However, the magnitude of the association with darunavir/ritonavir is similar to that previous reported in D:A:D analysis for earlier protease inhibitors indinavir and lopinavir/ritonavir. The lack of association with atazanavir/ritonavir supports this being a drug-specific rather than class effect.

The low number needed to treat to harm for people with high CVD risk (15 people would need to be treated with darunavir/ritonavir for five years to produce one related CVD event), suggest baseline CVD risk is an important factor before considering darunavir/ritonavir.

This further supports the decision by US guidelines to only recommend integrase inhibitor based combinations for first-line ART. [3]

Editorial commentary by Padraig McGettrick and Paddy Mallon in the same issue of Lancet HIV, stressed that higher rates of cardiovascular disease in HIV positive people compared to the general population are generally underestimated and the importance of the association observed in D:A:D not being driven by traditional CVD risk factors.  [4] It also noted that other research groups, including MACS, have associated cumulative use of darunavir/ritonavir (> 4 years) with surrogates for subclinical CVD. [5]


  1. Ryon L et al. Cardiovascular disease and use of contemporary protease inhibitors: the D:A:D international prospective multicohort study. Lancet HIV(2018). DOI: 10.1016/S2352-3018(18)30043-2. (03 May 2018).
  2. D:A:D online risk calculators.
  3. DHHS guidelines for the use of antiretroviral agents in adults and adolescents living with HIV (May 2018).
  4. PMC and Mallon P. HIV and cardiovascular disease: defining the unmeasured risk. Lancet HIV (2018), DOI: 10.1016/S2352-3018(18)30061-4. (03 May 2018).
  5. Thomas GP et al. Associations between antiretroviral use and subclinical coronary atherosclerosis. AIDS. 2016; 30: 2477–2486.

Links to other websites are current at date of posting but not maintained.