OBJECTIVES: The objective was to compare the cost-effectiveness of antidiabetic therapies in a middle-income country.

METHODS: Electronic databases were searched between August-April 2021. Inclusion criteria were: randomized clinical trials with T2D; articles published in English; double-blind, phase 3 or 4 trials; ≥100 patients per arm; follow-up >24 weeks; and the pre-specified endpoints: death from any cause, major adverse cardiovascular events, and hospitalization for heart failure.

Two prospective cohorts of T2D patients were used to feed the Markov Influence Diagram (MIDs) transition probabilities.

Bayesian Networks and MIDs were used to estimate the incremental cost-effectiveness ratios (ICER) per disease-adjusted life years [DALYs] averted by using OpenMarkov software. The perspective of the Brazilian public health system was considered for costs and values were converted to international dollars (Int$). The main outcomes were the mean change in relative risk of cardiovascular disease with the therapies and their incremental cost-effectiveness ratio.

RESULTS: We selected a total of 157 trials generating a sample size of 684,389 patient-years. Modeling was based on a primary and secondary prevention cohort of 851 and 1,337 patients, respectively. SGLT2i [HR:0.81 (95%CI: 0.69; 0.96)], GLP-1A [HR:0.79 (95%CI: 0.67; 0.94)], and pioglitazone [HR:0.73 (95%CI: 0.59; 0.91)] reduced the relative risk of non-fatal cardiovascular events. In primary prevention, pioglitazone resulted in 0.2339 DALYs averted, with ICER of Int$7,082 (95%CI: 4,521; 10,770) per DALY averted, when compared to standard care. SGLT2i and GLP-1A also increased the effectiveness (0.261 and 0.259 DALYs averted, respectively) but with higher ICERs [Int$12,061 (95%CI: 7,227; 18,121) and Int$29,119 (95% CI:23,811; 35,367) per DALY averted, respectively]. In secondary prevention, the three classes were cost-effective at a maximum willingness to pay threshold of Int$26,700. In both scenarios, however, pioglitazone had the highest probability of being cost-effective.

CONCLUSIONS: In a middle-income country, pioglitazone presented a higher probability of being cost-effective followed by SGLT2i and then GLP-1A.