This was a retrospective study of 245 patients in a Houston HIV clinic who were on HAART. Nearly all of the patients were men (98%), mean age = 46 years and 50% were White, 38% were Black and the remainder were Hispanic. This was a population with advanced HIV disease with nearly two-thirds having been diagnosed with AIDS. About half of the patients had a VL <400 copies and three-quarters had a greater than 30 cell increase in response to HAART.

Of the 245 patients evaluated, 101 (41%) had a defined clinical event during the median follow-up of 14 months. The most common clinical events and incidence/100 person years was for a new laboratory abnormality (41 and 13.5, respectively), a non-infectious event (39 and 12.9, respectively) and infectious events were least likely (19 and 6.3, respectively). The most common laboratory abnormalities were hyperlipidemia (N=32) and hyperglycemia (N=14). The only risk for both hyperlipidemia and hyperglycemia was having been on HAART > 15 months. For the 26 (11%) who were diagnosed with lipodystrophy (N.B. no standard definition is available and may influence the proportion reported in this observational database when compared to other studies), having been on HAART and having had a >30 cell CD4 increase were associated with this event. Eighty percent of patients with lipidystrophy also had an elevated tryglyceride. The use of the combination of ritonavir + saquinavir was associated with the development of lipodystrophy, hyperlipidemia and hyperglycemia, compared to the use of any other single protease inhibitor. Of note, for the 19 patients who had an infectious event, six patients had their event within 3 months after initiating HAART, suggesting the role of recognition of a reactivate infection with immune recovery.

This study, along with others, reaffirms the presence of abnormal lipid control with associated physical features. In that only 80% of patients with lipodystrophy had an elevated triglyceride level, the use of hypertriglyceridemia is perhaps more useful (though not perfect) for screening rather than for diagnosis. The long-term consequences of these conditions are currently unknown.