HIV and ID Observations

The FDA has issued one of its new “early communications” indicating that it has opened an investigation into the safety of abacavir and didanosine based on analyses showing higher rates of myocardial infarction with these drugs than with other NRTIs in the Data Collection on Adverse Events of Anti-HIV Drugs (D:A:D) study.

The pace of our field is sometimes remarkably fast: Immersed in HIV care or research or policy, we think of this as old news — after all, it was presented over a month ago at the Retrovirus Conference. These D:A:D results linking abacavir and ddI to increased MI risk have already been the subject of extensive discussions in clinics, conference rooms, and meetings. But I have to remind myself that no, it’s not old news at all — the full paper has not yet been published, although it will be soon — and that we still need some time to process the data, and to consider the questions raised by the findings. Is this a causal relationship, or just an association? Why has it not been seen in other studies? What is the mechanism? Would it still be the case with HLA-B*5701 screening? Why is the risk not cumulative? When will we be seeing data on tenofovir? (I suspect soon, given when tenofovir was approved.) What do abacavir and ddI have in common that would cause this? What should we be doing with our patients on abacavir who are doing well? I confess my answer to all of these questions is the same: I just don’t know.

D:A:D is an extremely important study that has already provided enormous insight into HIV treatment. But as the D:A:D investigators no doubt would agree, there are limitations to ascribing toxicities to treatment based on observational data. So I guess I’m proposing that as of today (March 31, 2008) we view these results as suggestive, and hypothesis-generating, rather than defining standard of care right now.

UPDATE

The D:A:D data have now been published in the Lancet. There is also an accompanying editorial here (subscription required).