Background:  With the use of 4 classes of antiretroviral drugs and the likely availability of 2 new classes and second generation drugs within the existing classes, the opportunity for improving treatment strategies for persons with HIV disease has become more pressing. High priority areas of strategic intervention include simplification, treatment interruption, and early treatment. Current studies of simplification have arisen because of the toxicity of thymidine analogue nucleoside reverse transcriptase inhibitors (NRTI) and the potency and high genetic barrier for resistance of protease inhibitors (PI). Now, thymidine analogues are used less often in first line treatment, and drug classes other than PI are available, so the momentum for this strategy seems to be decreasing. Nevertheless the availability of new classes will allow further examination of options for simplification to preserve drug options but still allow excellent antiretroviral control. Treatment interruption has been studied under a number of paradigms; including CD4-guided interruption and fixed-time or scheduled interruptions. For the CD4-guided interruption strategy, the SMART study has shown that this is not a viable option and has revealed an important new aspect of HIV disease—the vulnerability of HIV-infected persons to serious non-AIDS events, such as cardiovascular disease, stroke, liver disease, renal disease, and death—not directly attributable to manifestations of conventional opportunistic disorders. This will require further intensive study especially to explain the mechanism and further quantify the extent of this phenomenon. For fixed-time interruptions, the data are less robust, but seem to imply that this could be less risky than the CD4-guided strategy. This will require more study, but should not at this stage prevent us from studying it in the context of analytic treatment interruptions to evaluate new therapeutic vaccines or gene therapies. Lastly, there is new momentum to consider the “when to start” question. Given the substantial serious non-AIDS event rates at higher CD4 counts that have been documented from several cohorts, the availability of new and safer drugs with less likelihood of resistance, and that treatment interruption is unsafe, now may be the time to consider randomized clinical trials to ask whether early treatment at higher CD4 levels, eg, >500, should be recommended. Moreover in the developing world, the high early death rate from starting patients too late in the course of the disease and the increasing burden of co-infections (tuberculosis, bacterial sepsis, and malaria) make it important to consider a trial to investigate starting at higher CD4 counts than are currently recommended, namely >350. Clearly cost-benefit will need to be taken into account in the development and execution of any such trials.

Conclusions:  The most pressing need for a strategy trial at this point is the development and execution of “when to start” studies in both the developed and developing world. This question can only be reliably answered in the context of a randomized clinical trial that minimizes the known and unknown factors that bias even well-executed prospective cohort studies.