Abstract Body

There will be greater need for second-line (2L) antiretroviral treatment (ART) as HIV treatment programs mature. Historically, most 2L regimens in resource-limited settings contained a ritonavir-boosted protease inhibitor (PI), either lopinavir (LPV/r) or atazanavir (ATV/r). The best-in-class PI, ritonavir-boosted darunavir (DRV/r), is now available as a generic with a lower cost than LPV/r. This, together with use of dolutegravir (DTG) in 2L, offer opportunities to optimize 2L regimens.

We used the Applied Cost and Outcomes Research Analysis (ACORA) Model, coded in R, to compare the cost-effectiveness of 2L ART strategies. Markov state transitions were used to simulate an open patient cohort on ART through health states of ART on treatment status, viral load, CD4 cell count, and opportunistic infections (OI). We conducted a 10-year forecast using quarterly time-steps. Disability weights were applied to health states and adverse events (AEs) to estimate quality adjusted life years (QALYs). We included the costs of drugs, lab monitoring, human resources, and OI management. Disease progression was parameterized and drug profiles estimated using published literature and programmatic sources, standardized to account for multiple studies. Transitions were driven by drug-specific profiles of failure rates and discontinuations due to AEs, which were estimated from published clinical trials and standardized to account for multiple studies. We defined 4 scenarios: 1) Standard of Care (SOC): uses ATV/r and LPV/r in 2L; 2) C1: switches LPV/r to DRV/r; 3) C2: switches both ATV/r and LPV/r to DRV/r; and 4) C3: switches LPV/r and ATV/r to DTG, reserving DRV/r for those with 1L DTG failure. The analysis considered a hypothetical sub-Saharan African country with 1,000,000 existing ART patients and 100,000 new patients initiating ART each year; the SOC was either 80% LPV/r and 20% ATV/r, or the opposite. Costs and QALYs are restricted to patients who migrate to 2L. Incremental cost effectiveness ratios (ICER) are presented.

Strategy C3 was cost-saving and maximized QALYs in both LPV/r- and ATV/r-dominant scenarios (Table). Strategies C1 and C2 resulted in higher costs and QALYs than SOC, with ICERs ranging from $34 to $1688.

In most settings It is highly cost-effective to replace LPV/r and ATV/r with DRV/r with a less than 1x GDP threshold. The most optimal strategy is to maximize use of DTG in 2L, which was the most effective and cheapest scenario across all comparisons.