 |
 |
 |
 |
 |
 |
| Home > Evidence of the Month |  |
 |
|
Education |
| Evidence of the Month |
 |
Commentaries on both new and classic studies of importance for the treatment of diabetes are posted here monthly.
The RECORD Study of Rosiglitazone and Cardiovascular Risk
Alan Vaag |
Click here to download PowerPoint slides of this commentary. |
Background
The oral antidiabetic drug rosiglitazone has been used for Type 2 diabetes as monotherapy or in combination with metformin or sulfonylureas since 2000. This thiazolidinedione has been associated with durable control of blood glucose, an effect attributed to improved insulin sensitivity via activation of peroxisome proliferator-activated receptors (specifically, PPARγ). But it has also been the focus of concern because of its possible role in increased cardiovascular risk. The predominant concern initially was increased risk of heart failure, stemming from fluid retention, a known effect of thiazolidinediones. In 2007, the evidence on heart failure resulted in black box warnings on the labels of both rosiglitazone and pioglitazone, strengthening caveats already in place about use of these drugs in patients with or at risk of heart failure.
At about the same time, concerns over ischaemic heart disease intensified when evidence tied rosiglitazone therapy to increased myocardial infarction (MI) risk. The most influential data came from clinical trial meta-analyses, performed by the drug maker and independent investigators, that showed a potential 30% to 40% increased relative risk of MI with use of rosiglitazone (GlaxoSmithKline, 2007; Nissen, Wolski, 2007). An interim analysis of the Rosiglitazone Evaluated for Cardiac Outcomes and Regulation of Glycaemia in Diabetes (RECORD) trial, which was designed specifically to address the cardiovascular concerns, showed a non-significant 16% increase in MI with rosiglitazone (Home et al, 2007). Although a black box warning was added to rosiglitazone’s label, many regulators and investigators concluded that overall evidence on MI was not definitive. The final results of RECORD have been widely anticipated, not only for cardiovascular information but for data on fractures, cancers, and other adverse effects possibly associated with rosiglitazone therapy (Kahn et al, 2007; Ramos-Nino et al, 2007).
Methods and Key Results
RECORD was a multicentre, open-label, randomized trial with 4447 patients with Type 2 diabetes. It compared cardiovascular outcomes over 5 to 7 years in people assigned either to rosiglitazone, added to metformin or a sulfonylurea, or to active control therapy with metformin and sulfonylurea combined. The primary end point was hospitalization or death from cardiovascular causes, with a hazard ratio (HR) non-inferiority margin of 1.20 (that is, for the therapy to be considered non-inferior, any excess risk for the end point could not exceed 20%). Heart failure, MI, and stroke rates were compared in secondary analyses, as were prespecified subgroups of patients.
Treatment in both arms targeted an HbA1c of ≤7 %. Rosiglitazone was begun at 4 mg/day and titrated to 8 mg/day if needed after 8 weeks. Starting doses and titration of metformin and sulfonylurea doses varied by local practices.
The primary end point occurred in 14.5% of patients in both RECORD treatment groups over a mean follow-up of 5.5 years. The rosiglitazone versus control HR for the primary end point was 0.99 (95% confidence interval [CI], 0.85 to 1.16), meaning that rosiglitazone met the criterion for non-inferiority. However, heart failure was more than twice as common in the rosiglitazone group (HR, 2.10; CI, 1.35 to 3.27; P=0.0010). There were no significant differences between the groups in the risk of cardiovascular death, MI, or stroke; the excess risk for MI with rosiglitazone was, at 14%, about the same risk seen in the interim analyses. In a prespecified RECORD subgroup with previous ischaemic heart disease, the risk for the primary end point was increased by 26%, but this was not statistically significant (CI, 0.95 to 1.68; P=0.055). No other significant differences were observed based on gender, age, background diabetes therapy, duration of diabetes, body mass index, or use of statins, angiotensin-converting enzyme inhibitors, or nitrates.
Mean HbA1c was lower with rosiglitazone than control therapy in RECORD, as was the risk of pancreatic cancer. However, the overall incidence of bone fractures was significantly higher with rosiglitazone (risk ratio, 1.57; CI, 1.26 to 1.97; P<0.0001). Risk appeared to be greater in women than men and due mainly to upper and distal lower limb fractures.
Clinical Implications
The RECORD investigators concluded that rosiglitazone does not increase overall cardiovascular morbidity or mortality; that it clearly increases the risk of heart failure; and that a small increased risk of MI with use of the drug cannot be ruled out. They said that rosiglitazone is not recommended for people with a history of heart failure or ‘problems that might have led to myocardial dysfunction’. They also said it should be used with caution in women at high risk of fractures. The RECORD investigators acknowledged that due to a lower than expected primary event rate, the trial had less statistical power than was predicted in the trial design. Other experts have expressed skepticism about RECORD’s open-label design and dismay that the trial did not clearly answer the question of ischaemic risk (Psaty, Furberg, 2007).
RECORD produced several other clinically useful findings. The use of statins increased during the trial more in the rosiglitazone group than in the active controls (9.2% difference between groups at 5 years). The difference may have meant greater cardioprotection in the rosiglitazone group, but the authors calculated that even if statins had reduced ischaemic events by 25%, the non-inferiority criterion for rosiglitazone would still have been met. Similarly, loop-diuretic use went up more sharply with rosiglitazone, a difference that might have decreased the rate of heart failure in that group. Body weight and high-density lipoprotein cholesterol increased more with rosiglitazone, and low-density lipoprotein cholesterol was reduced less; these are all important factors in managing diabetic patients.
In late 2007, the American Diabetes Association (ADA) and European Association for the Study of Diabetes (EASD) issued a consensus statement urging clinicians to ‘consider more carefully’ whether to use thiazolidinediones (rather than insulin or sulfonylureas) for second-line antidiabetic therapy. In 2008, the ADA/EASD unanimously advised against using rosiglitazone in the diabetes treatment algorithm; instead the ADA/EASD currently recommend use of pioglitazone (which has demonstrated cardiovascular benefit in several trials) when choosing thiazolidinedione therapy (Nathan et al, 2009). Whether further guideline revisions will follow on the RECORD findings is not yet clear. Rosiglitazone has also been used in numerous recent clinical trials of cardiovascular outcomes, including the Veterans Affairs Diabetes Trial (VADT) (Duckworth et al, 2009) and the Bypass Angioplasty Revascularization Investigation 2 Diabetes (BARI 2D) trial (BARI 2D Study Group, 2009). None of these trials reported safety concerns with the drug, but none was specifically designed to address the connection between rosiglitazone and cardiovascular outcomes.
There are many options for diabetes management. The argument can be made that unless rosiglitazone is shown in trials like RECORD to have uniquely desirable benefits, there seems to be no compelling reason at present to include it in the treatment plan. When it is used, clinicians should remember that caution and monitoring are advised with respect to cardiovascular risks.
|
 |
|
Print this page |
|
||
 |
Add to Favourites |
|
|