In STARTMRK, treatment-naïve patients received raltegravir 400 mg bid or efavirenz 600 mg at bedtime (in a 1:1 ratio), both in combination with tenofovir/emtricitabine [11,12]. In BENCHMRK-1 and -2, highly treatment-experienced patients with multi-drug resistant virus and virological failure received raltegravir 400 mg bid or placebo (in a 2:1 ratio), both in combination with

optimized Roscovitine in vitro background therapy (OBT) [13,14]. Patients with chronic HBV and/or HCV coinfection were purposely permitted to enrol if their baseline levels of aspartate aminotransferase (AST), alanine aminotransferase (ALT) and alkaline phosphatase did not exceed five times the upper limit of normal; treatment-experienced patients were also required to have baseline total bilirubin less than twice the upper limit of normal. HBV infection was defined as HBV surface antigen positivity for all studies; HCV infection was defined as HCV RNA

positivity for patients in STARTMRK and as HCV antibody positivity for patients in BENCHMRK. All treated patients were included in the safety and efficacy analyses. For the safety analyses, overall categories of clinical adverse events and selected laboratory abnormalities were tabulated. Adverse events were reported as drug-related if they were judged by the investigator as definitely, probably, or possibly related to any of the study drugs. The severity of laboratory AG-014699 cell line Sulfite dehydrogenase abnormalities was graded according to the 1992 Division of AIDS toxicity guidelines for adults (http://rcc.tech-res-intl.com/tox_tables.htm).

The percentage of patients with a particular laboratory abnormality was calculated as: (number of patients whose highest on-treatment value was a worsened grade from baseline)/(number of patients with a baseline value and at least one on-treatment value). For the BENCHMRK studies, adverse events and laboratory abnormalities are presented in two ways: by frequency and by crude adjustment for duration of follow-up, as the median duration of therapy was substantially greater in the raltegravir group as a result of lower rates of virological failure. A logistic regression model was used to compare virological response rates between treatment groups after adjusting for covariates that might affect the likelihood of achieving HIV-1 RNA suppression. An observed failure approach was used for the exploratory efficacy analyses because it predominantly reflects the antiretroviral effect of treatment; only patients discontinuing the studies because of a lack of efficacy were counted as failures at subsequent time-points. These exploratory subgroup analyses were not specified in the original protocols; formal statistical comparisons between groups were not performed. A total of 743 patients received raltegravir and 519 received comparator across the three studies (Table 1).