This article is more than 5 years old. Information may no longer be current.
RECORD: No overall increase in CV risk with rosiglitazone
Adverse findings include increased risk for fracture in women, HF
Click Here to Manage Email Alerts
Rosiglitazone was not associated with an increased risk for overall cardiovascular morbidity or mortality compared with standard glucose-lowering therapies, according to the final analysis data of the RECORD study.
At 5.5 years, the primary outcome – first occurrence of CV death or hospitalization – was equivalent (14.5%) in patients assigned to rosiglitazone (Avandia, GlaxoSmithKline) or metformin and sulfonylurea, according to Philip D. Home, DM, DPhil, chairman of the RECORD Steering Committee and professor of diabetes medicine at Newcastle University, UK. The researchers reported 321 events among patients assigned rosiglitazone as add-on therapy and 323 events among patients assigned metformin or sulfonylurea only (HR=0.99; 95% CI, 0.85-1.16).
“Overall, in CV terms, the drug is safe. Would rosiglitazone meet current FDA criteria as a safe drug in diabetes? The answer to that is yes,” Home said during a press conference on Friday.
The RECORD study was designed to evaluate the long-term effect of rosiglitazone on CV outcomes and blood glucose control compared with metformin and sulfonylureas. The open-label study was conducted at 338 centers in 23 countries in Europe, Australia and New Zealand. The researchers randomly assigned 4,447 patients with type 2 diabetes who were already on metformin or sulfonylurea monotherapy to either add-on rosiglitazone (n=2,220) or a combination of metformin and sulfonylurea (n=2,227). All doses were progressively increased toward achieving and maintaining a target HbA1c of <7%. If HbA1c rose to 8.5% or more, either a third oral glucose-lowering drug was added (for rosiglitazone group) or insulin was started (for active control group).
Adverse findings
The primary endpoint result is in contrast to other study results, including a 2007 meta-analysis that reported a 43% increased risk for myocardial infarction with rosiglitazone, which prompted the RECORD Steering Committee to release an interim analysis at 3.75 years. The final results add about 50% more patient-years of follow-up.
However, data remain “inconclusive” on risk for MI with rosiglitazone. The researchers reported 64 fatal or nonfatal MI events among patients assigned rosiglitazone compared with 56 events among patients assigned metformin and sulfonylurea only (HR=1.14; 95% CI, 0.80-1.63).
“We do know that [rosiglitazone] is not resulting in increased CV death because both total CV death and deaths from MI were decreased in the rosiglitazone group,” Home said.
Despite the positive news concerning CV deaths, one adverse finding was an increased risk for heart failure, which is known with thiazolidinediones. The risk approximately doubled with rosiglitazone compared with active controls, a finding that is in line with results of the interim analysis, according to Home. The rosiglitazone group experienced 61 fatal or nonfatal HF events compared with 29 in the active control group (HR=2.10; 95% CI, 1.35-3.27). The excess rate of HF events was 2.6 per 1,000 patient-years.
Rosiglitazone was also associated with an increased risk for fractures compared with active controls (185 vs. 118), and fractures were significantly more common among women than men (154 vs. 71).
The RECORD data verify earlier data reported from the ADOPT trial, according to Home. “There is an increase in fracture with rosiglitazone compared with the other drugs, particularly with women and affecting distal fractures not hip or spine,” he said. However, a small effect in men cannot be excluded.
Secondary outcomes
A key secondary outcome is a composite of CV death, stroke and MI; results were slightly but not statistically significantly in favor of rosiglitazone vs. metformin and sulfonylurea (HR=0.93; 95% CI, 0.74-1.15).
None of the differences in secondary endpoints were statistically significant. CV death (HR=0.84; 95% CI, 0.59-1.18) and all-cause death (HR=0.86; 95% CI, 0.68-1.08) were in favor of rosiglitazone compared with active controls. In addition, the rosiglitazone group had a reduced number of fatal or nonfatal strokes (HR=0.72; 95% CI, 0.49-1.06).
“Positive findings in other areas – especially CV death and stroke – almost exactly balanced out the total numbers for the primary outcome, thus meeting the criterion of noninferiority for rosiglitazone,” Home said.
After five years, rosiglitazone was associated with progressively better maintenance of glucose control. Patients assigned to rosiglitazone had HbA1c levels 0.29% lower than patients assigned sulfonylureas and 0.26% lower than metformin.
Rosiglitazone was also associated with higher HDL, LDL and body weight compared with active control. Diabetes eye and foot events were similar between groups. Progression of albumin excretion was less with rosiglitazone.
Interpreting RECORD
“Overall, the CV safety of rosiglitazone has been confirmed in more than 4,000 participants. In people with type 2 diabetes, rosiglitazone can be used without concern that there is increased overall risk of CV morbidity or CV or any other cause mortality,” Home said.
However, he cautioned against use in patients with a history of HF, myocardial damage and women at higher risk for fractures.
Of note, the trial had less statistical power than initially planned because the overall primary event rate was substantially lower than anticipated because with better treatment of CV risk factors the overall CV event rate in the population has fallen, according to the researchers.
Addressing concerns with rosiglitazone based on previous data, the researchers said loss to follow-up for the primary endpoint was 7.2% and excess discontinuation from rosiglitazone was only reported in 1.4% of the study population.
“This is the kind of study that drives guidelines and recommendations,” Home said.
The results of RECORD were also published online in The Lancet. – by Katie Kalvaitis
For more information:
- Home PH. Results of the RECORD clinical trial. Presented at: American Diabetes Association’s 69th Annual Scientific Sessions; June 5-9, 2009; New Orleans.
I chaired the consensus algorithm group that developed consensus recommendations on how to treat type 2 diabetes, what medications to use and in what order. When this whole rosiglitazone flap started with the meta-analysis, we just cautiously recommended against using rosiglitazone. With our most recent recommendation, we took it off the list of drugs that we are using. We recommended pioglitazone (Actos, Takeda) as a second-tier agent but we specifically said that because of this concern with the safety of rosiglitazone that we no longer recommend the use of rosiglitazone. I am a clinical trialist myself and usually clinical trials in general trump observational, meta-analytic data, and we will take this back to our committee and discuss it. There are issues about fractures, HF and expense.
David M. Nathan, MD
Director of the Diabetes Center, Massachusetts General Hospital, Boston
The good thing about RECORD is that it was absolutely neutral. The bad thing about RECORD is that it was absolutely neutral. There was one really interesting finding about reduced number of strokes with rosiglitazone. In PROactive, the patients who had prior stroke had a 50% reduction in subsequent stroke on pioglitazone. In RECORD, the most substantial evidence of benefit was in stroke. So, because this was also observed in PROactive, I think it may actually be something real. We are slightly more focused on cardiac events than strokes for some reason and we have so many more interventions and do much more testing for cardiac events than we do for cerebrovascular events. The fact is that cerebrovascular events are a huge problem. This should not be thought of as a minor thing; looking at stroke may be something exciting.
- Zachary T. Bloomgarden, MD
Clinical Professor, Department of Medicine, Mount Sinai School of Medicine