Department of Biostatistics Seminar/Workshop Series

Analyzing Data on Resistance to Antiretroviral Drugs

Victor De Gruttola, Sc.D.

Chair, Department of Biostatistics, Harvard University, School of Public Health

Wednesday, November 10, 1:30-2:30pm, MRBIII Conference Room 1220

The wide availability of HIV gene sequencing has had a major impact on the understanding and clinical management of antiretroviral drug resistance. Among the issues that have been studied are: 1) Identifying mutations/patterns of mutations that confer phenotypic resistance; 2) Establishing patterns of mutations that are associated with poor clinical response to a combination of drugs; and 3) Determining the genetic pathways to resistance. Regarding 1), we used HIV-1 protease sequences and measures of in-vitro sensitivity to Amprenavir from the Stanford HIV Drug Resistance Database to illustrate statistical approaches for mutations that enhance or decrease drug susceptibility. Analysis of these data shows that while M46I/L mutations are associated with drug resistance, addition of the L88D/S mutation leads to hypersusceptible virus. Further addition of T90M/L mutations results in highly resistant virus. Regarding 2) identifying genetic mutations that cause clinical resistance to antiretroviral drugs requires adjustment for potential confounders, such as the number of active drugs in a HIV-infected patient's regimen other than the one of interest. We investigated resampling-based methods to test equal mean response across multiple groups defined by HIV genotype, after adjustment for covariates. To help preserve Type I error while also improving power in both approaches, we propose resampling approaches based on matching of observations with similar covariate values. Matching reduces the impact of model mis-specification as well as imprecision in estimation. Our focus is on relating HIV genotype to viral susceptibility to abacavir after adjustment for the number of active antiretroviral drugs (excluding abacavir) in the patient's regimen. Regarding 3), methods are described for understanding the genetic pathways that lead to high-level drug resistance under selective drug pressure, as well as for estimating the rates at which viral populations progress along these pathways. These methods can be used to determine whether the presence of certain mutations among drug-sensitive viruses predispose a patient under a particular treatment to develop patterns of mutations that confer high-level drug resistance. We apply our methods to genetic sequences of viruses cloned from the plasma of 170 patients who participated in three phase II clinical studies of efavirenz combination therapy. Multiple viral clones are available from each plasma sample at each time of measurement, allowing for consideration of the effect of minority species on the evolution of the viral populations infecting patients. The sequences can be found in the Stanford HIV RT and Protease Sequence Database.

http://www.hsph.harvard.edu/faculty/victor-de-gruttola/