December 18, 2001
HIV drug resistance has emerged as a big problem in the U.S. It is the result of many factors, including prior sub-optimal therapy, inadequate adherence, and transmission of drug-resistant strains in newly infected patients. Most of the HIV drug resistance data described in the U.S. comes from clinical trials investigating new drugs or rescue strategies in heavily drug-experienced patients. There have been a few small studies looking at resistance in cases of acute infection or in certain risk groups. The study presented here represents the first relatively large analysis of the prevalence of HIV drug resistance within the U.S. This study was undertaken to determine how big a problem HIV drug resistance was in the U.S. during the beginning of the HAART era.
The study utilized patients being followed in the HIV Cost and Service Utilization Study (HCSUS). The pool of patients represented 208,724 HIV-infected American adults followed from 1996-1999, who were still alive in 1999.
One thousand nine hundred and six patients had plasma samples available. Of those, 1,209 (63 percent) had a viral load >500/ml and were evaluated for phenotypic HIV drug resistance using the PhenoSense assay with previously defined biologic cutoffs for resistance for 15 approved HIV drugs.
A number of factors were considered that might contribute to why drug resistance was likely to be present. The number of patients with any evidence of drug resistance (one drug or greater) was 78 percent in those with detectable viral loads. The estimate for all patients was 50 percent. When examined by class of HIV drugs, 70 percent had resistance to NRTIs, 42 percent to protease inhibitors, and 31 percent to NNRTIs. Fifty-one percent had resistance to more than two classes of drugs and 14 percent had resistance to all three classes of drugs. Resistance to 3TC was the most common phenomena reported. Variables found to correlate with the presence of resistance were: lowest CD4 count recorded (or AIDS diagnosis), being white, men having sex with men, being male, having health insurance and higher education levels. Age and race were not predictive of resistance. Interestingly, living in the Midwest and having a doctor with a large HIV practice size (more than 100 patients) was associated with a reduced likelihood of having resistance.
This retrospective study included a significant number of patients who had been on single- or dual-drug therapy, who may have subsequently gone on HAART and had detectable viral loads from which drug resistance could be measured. The study did not include patients who started HAART after 1999.
Nevertheless, this study indicates what many of us in practice already know: A significant amount of patients have some degree of drug resistance. The study did not address the HIV drug history of these patients or the impact on clinical outcome as a result of drug resistance. It is also possible that newer treatment strategies with more potent agents, less toxicity and better adherence may decrease the amount of newly developing HIV drug resistance in the future. Additional interesting findings concern the types of patients who are more or less likely to have resistance. It appears that those patients who have greater access to care may have been exposed to more drugs and combinations resulting in higher rates of resistance. It is comforting to know that those HIV practitioners who see the most patients are less likely to have patients with HIV drug resistance. This may be the result of increased practitioner knowledge and experience regarding the combination and position of HIV drugs.