However, it remains possible that KS may impart an independent NVP-BKM120 cell line risk of mortality, as 91% of KS-related mortality occurred in the group with disseminated disease, similar to mortality rates observed in other studies [13-16]. Other studies have noted that improved immunological and virological responses are associated with clinical responses to KS [15]. However, our observation that the median CD4 count at the time of diagnosis for the patients with incident KS was 158 cells/μL compared with 83 cells/μL at baseline implies that

the risk for developing KS continues for some time, even with some degree of immune recovery. Other studies have shown the impact of HAART on KS in HIV-infected patients [17-19]. However, we had the opportunity to examine both risk factors for KS and clinical outcomes among patients predominantly treated with NNRTI-based regimens in a KS endemic country. The fact that regression rates in our study

are similar to those published for industrialized countries, learn more where clinical responses range from 67 to 85% [8, 9], should be somewhat reassuring to patients and physicians who do not have easy access to specific anti-neoplastic therapy or PI-based HAART regimens. Less than half of the patients in this study were able to access any chemotherapy for KS, and only four completed a full course of treatment, which suggests that NNRTI-based HAART may be adequate therapy for most patients who develop PAK6 KS when starting or while receiving HAART. Nevertheless, our study has a number of limitations. Because of the relatively small number of KS patients, we may have lacked sufficient power to detect other risk factors for

KS. This also limited our ability to ascertain differential response rates to different HAART regimens. In many instances we found factors that had ORs or HRs much greater than or less than 1, but with very wide confidence intervals. In particular, we had very few individuals who were switched from NNRTI-based regimens to PI-based regimens, which greatly limited our ability to detect differences in outcomes associated with these regimen changes. Furthermore, subjects were not randomly assigned to switch treatment and the lack of a significant difference in outcomes associated with treatment switching may have been attributable to other confounding factors. However, despite these limitations, these results are somewhat reassuring to patients and clinicians who may not have access to more expensive specific anti-neoplastic KS treatment or PI-based regimens. In conclusion, the use of NNRTI-based HAART regimens appears to induce remission of KS in HIV-infected patients in Uganda, although mortality associated with KS was still very high.