Certain protease inhibitors and abacavir linked to heart attacks in two large cohort studies

The latest follow-up data from two large cohort studies, presented on Monday at the Sixteenth Conference on Retroviruses and Opportunistic Infections in Montreal, adds further evidence that specific protease inhibitors (PIs) and nucleos(t)ide reverse transcriptase inhibitors (N(t)RTIs) are associated with a higher risk of cardiovascular problems in people with HIV.

While several past observational studies have seen an increased risk, controlled trials have produced conflicting results. Furthermore, the mechanisms explaining heightened cardiovascular risk in HIV-positive people - whether on or off antiretroviral therapy - are not fully understood.

The D:A:D study (Data Collection on Adverse Events of Anti-HIV Drugs) has collected information about the side-effects of antiretroviral treatment for the past decade. This multinational collaboration includes eleven prospective cohorts from the US, Australia and Europe.

Several years ago D:A:D researchers began reporting a link between PI use and cardiovascular events. At last year’s CROI, they reported the controversial finding that recent use (within six months) of the NRTI abacavir (Ziagen, also in the Kivexa and Trizivir coformulations) appeared to increase the risk of myocardial infarction (MI), or heart attack, by 90%, whilst ddI (didanosine, Videx) increased this risk by 49%.

At last summer's International AIDS Conference in Mexico City, investigators with the large SMART treatment-interruption trial also reported a link between cardiovascular events and recent use of abacavir (though not ddI), but a pooled analysis of more than 50 clinical trials conducted by abacavir manufacturer GlaxoSmithKline did not find any increase in risk.

The D:A:D analysis presented this week included longer follow-up data. A total of 33,308 patients were followed from the time of enrolment through February 2008, reflecting 178,835 total person-years of follow-up. Over this period, 580 participants experienced a heart attack.

The impact of specific antiretroviral drugs on heart attack risk was assessed using Poisson regression analysis. The study looked at seven N(t)RTIs; the non-nucleoside reverse transcriptase inhibitors (NNRTIs) efavirenz (Sustiva, also in the Atripla co-formulation) and nevirapine (Viramune); and the PIs indinavir (Crixivan), saquinavir (Invirase), nelfinavir (Viracept) and lopinavir/ritonavir (Kaletra).

D:A:D has a policy of only reporting findings for drugs that have been used by a significant number of people for a considerable length of time (more than 30,000 total person-years of follow-up). The N(t)RTI tenofovir (Viread, also in Truvada and Atripla) was left out last year due to insufficient data, but included in this year's analysis. There is not yet enough data for the newer PIs, the newly approved NNRTI etravirine (Intelence) or the fusion inhibitor, integrase inhibitor and CCR5-antagonist drug classes.

Since cardiovascular risk is influenced by many variables, risk ratios were adjusted for demographic characteristics, blood lipid levels and other metabolic measurements and other cardiovascular risk factors (such as age, smoking and family history). People with a moderate Framingham cardiovascular risk accounted for about 15% of total follow-up time, while those with a high risk accounted for about 5%. The researchers also adjusted for HIV disease status, including CD4 cell count and viral load.

The investigators found no association between heart attack risk and recent or cumulative exposure to either of the studied NNRTIs, two of the PIs (nelfinavir and saquinavir), or most of the N(t)RTIs.

However, they did see a significant association between higher MI risk and cumulative - but not recent - exposure to indinavir and lopinavir/ritonavir. After adjusting for blood lipid levels, these increased risks were slightly reduced. The pattern did not change after adjusting for ritonavir boosting.

The researchers continued to see the previously reported association between increased odds of having a heart attack and recent use of abacavir or cumulative use of ddI. Furthermore, the risk due to recent abacavir appeared to increase with longer use. Tenofovir - both recent and cumulative use - was not linked to higher MI risk.

Discussing these findings, Dr Lundgren said it was hard to determine whether the observed differences in MI risk are attributable to a "class effect", and emphasised the importance of analysing specific drugs rather than broad drug classes.

Addressing the issue of "channelling bias" - that is, doctors being more likely to prescribe abacavir to patients with a pre-existing risk of cardiovascular disease - he said the study has been "under severe scrutiny" since the 2008 report and the analysis has carefully adjusted for known cardiovascular risk factors and history of heart disease. Furthermore, he added, such a bias should have also directed higher-risk patients toward tenofovir, which was not linked to higher MI risk.

Interestingly, CD4 cell count and HIV viral load did not predict MI risk. This leaves the mechanism(s) of drug-related risk elevation open to question, suggesting that factors other than immune suppression and ongoing HIV replication play a role. Dr Lundgren suggested that the effect of recent abacavir use might be attributable to an inflammatory reaction that damages blood vessel walls.

At the same session, researchers from the French national AIDS research agency (ANRS) presented data on the effect of specific NRTIs and PIs on heart-attack risk. The analysis - which was requested after last year's D:A:D report - was based on a nested case-control study of the more than 11,500 patients enrolled in the French Hospital Database.

A total of 289 HIV-positive individuals in the database experienced a confirmed first heart attack between January 2000 and December 2006. Each case patient was matched with up to five control subjects of the same sex and similar age seen at the same centre who had never had a heart attack (884 total controls). The researchers also controlled for known confounding factors including cardiovascular risk and HIV disease status.

The reported analysis looked at associations between MI risk and recent (again defined as within six months) or cumulative use of seven N(t)RTIs and five PIs - the same four included in the D:A:D study plus amprenavir (Agenerase) and fosamprenavir (Telzir), considered as a single drug. (Other drugs were analysed, but odds ratios were only reported for those taken by at least 100 patients.)

With the exception of saquinavir, cumulative PI exposure was found to be associated with an increased risk of heart attack, although this was only statistically significant for lopinavir/ritonavir and amprenavir or fosamprenavir.

There was a trend toward increased risk with AZT (zidovudine, Retrovir) and d4T (stavudine, Zerit), but these did not reach statistical significance. No association was seen the other N(t)RTIs including ddI and tenofovir.

With regard to abacavir, the investigators reported a "slightly different signal" than the one seen in D:A:D. Patients who recently used abacavir had about twice the odds of having a heart attack. Those who had used abacavir in the past but stopped had a slight, but non-significant, increase in risk, while the risk was unchanged for those who stayed on the drug. Furthermore, there was no observed interaction between abacavir use and other cardiovascular risk factors.