Analysis of data from the RECORD trial of rosiglitazone in patients with type 2 diabetes found an increased risk of heart failure in the rosiglitazone group compared to the control group: the relative risk was roughly doubled.

The thiazolidinedione anti-diabetic drugs (glitazones) are associated with an increased risk of heart failure, however there has been considerable debate over the magnitude of this adverse effect for the two available drugs. The RECORD trial was intended to determine whether the effect of rosiglitazone treatment on overall cardiovascular mortality and morbidity was different to that of an active control. It included heart failure (HF) as a pre-specified outcome, collected prospectively and adjudicated according to accepted criteria: this paper reports the results for the HF outcome.

RECORD compared the combination of rosiglitazone plus either metformin or a sulphonylurea, with metformin plus sulphonylurea in patients with type 2 diabetes inadequately controlled on monotherapy. It was randomised but open label, and the primary outcome was the time to first cardiovascular hospitalization or cardiovascular death including HF. Among the exclusion criteria were pre-existing HF and a recent major cardiovascular event.

There were 4,447 patients randomised to study treatment, 2,220 in the rosiglitazone group and 2,227 in the control group. Mean follow-up was 5.5 years (12,338 person-years in the rosiglitazone group and 12,272 person-years in the control group). For the primary outcome, there was no statistically significant difference between the groups.

There were 90 fatal and non-fatal HF events overall, with a significant excess in the rosiglitazone group (61 vs. 29; HR = 2.10, 95% CI 1.35 to 3.27; P = 0.001). The estimated excess event rate was 2.6 per 1,000 patient-years. As part of the analysis, potential predictive factors were identified. Apart from assignment to rosiglitazone, factors identified with increased risk of HF were age, urinary albumin : creatinine ratio, body mass index, and systolic blood pressure at baseline. A history of previous cardiovascular disease was not predictive.

The authors conclude that their results confirm the increased risk of HF in patients treated with a glitazone. They note that the magnitude of the increase is consistent with that found in a recent meta-analysis, and comment that the mechanism is probably related to fluid retention. They compare their results with those of the PROactive study involving pioglitazone, noting that the absolute incidence of HF was much higher in that trial (probably because all participants had macrovascular disease). Because of multiple differences between the two studies, including patient population and study duration, it was difficult to compare the results directly. Overall, they conclude that their results support current recommendations that these drugs should be avoided in patients with HF and should be discontinued in patients that develop it. If a glitazone is used in a patient with the factors identified as associated with increased risk, they should be closely followed.

[Editor’s comment: although not presented in the paper, it is easy to calculate absolute risks and thus number needed to harm (NNH) from the included data. Absolute risks over the 5.5 year period were 2.75% vs. 1.30%, an absolute risk increase of 1.45% (NNH 69).]