Clinical Science: Editorial Comment
Combination antiretroviral therapy (ART) is one of the great medical achievements of recent times. Since its introduction in 1996, ART has been key to dramatically reducing HIV mortality and morbidity in the developed world, as well as many countries of the developing world. The rationale for ART is straightforward: although HIV typically develops resistance to regimens using a single antiretroviral, it lacks a similar ability to develop resistance to regimens using appropriately chosen combinations of antiretrovirals. The devil is of course in the details. Over 20 antiretrovirals are FDA approved; in turn, these medications can be divided into six classes based on their probable mechanisms of action . Thus the question: what specific combinations are likely to have the highest level of efficacy, effectiveness, and efficiency? This question is particularly critical when initiating ART in previously untreated adult or adolescent patients.
Carr and Amin's paper gives a superb picture of what we know and don't know about these issues . What we know is what can be learned from randomized trials and prospective cohort studies where the primary end point is undetectable viral load. As valuable as these studies are, they suffer from limitations which affect what we know once we try to move beyond basic issues of efficacy to issues of effectiveness and efficiency.
Some of these limitations result from the practical difficulties associated with this type of clinical research. For example, Carr and Amin found that prior injecting substance use did not affect clinical outcome, which is both counter-intuitive and, as they note, incongruent with the findings of Antiretroviral Cohort Collaboration. The most probable explanation for Carr and Amin's result is not that it represents some underlying clinical reality, but simply that substance users are relatively difficult to recruit and retain and that those who successfully participate in clinical studies are not fully representative of the population of HIV-infected substance users.
Other limitations arise from the deficiencies in the execution of clinical research protocols which can and should be addressed in future studies. Two such deficiencies are appropriately highlighted by Carr and Amin. The first is the incomplete characterization of study participants. For example, there was not a sufficient number of studies reporting HBV and HCV status to make any analysis feasible. Because both of these co-factors could quite plausibly affect treatment outcomes, such lacunae are unfortunate. The second deficiency is the erratic characterization of adverse events across studies. As Carr and Amin note, only about 50% of the studies included reported individual adverse events and even fewer reported all adverse events. Such information is necessary if we are to evaluate the clinical effectiveness and efficiency of particular regimens. The unfortunate effect of this inadequate reporting is illustrated in Carr and Amin's thoughtful discussion of whether didanosine-containing regimens are under-utilized, particularly in resource-limited settings. As they point out, although didanosine has a number of problems (e.g. neurotoxicity), it also has a number of advantages, particularly in resource-limited settings. Unfortunately, because of the inadequacy of adverse event reporting, there is simply no way to perform the necessary cost–benefits analysis to address this question of clinical efficiency.
One can only conclude by endorsing Carr and Amin's call for studies of longer duration and more adequate reporting of adverse events. To which I would add more complete characterization of patients both medically and psycho-socially, as well as recruitment of currently under-represented patient populations, particularly those with current or past substance use issues.
2. Carr A, Amin J. Efficacy and tolerability of initial antiretroviral therapy: a systematic review. AIDS 2009; 23:343–353.