TZD Fracture Risk Quickly Reverses With Discontinuation

BALTIMORE — A new analysis of the Action to Control Cardiovascular Risk in Diabetes (ACCORD) trial confirms that women taking thiazolidinediones (TZDs) are at a greater risk of sustaining a fracture than men taking the diabetes drugs and shows for the first time that the increased risk diminishes after discontinuation of the drugs.

"Our findings show that there is about a doubling of fracture risk for women who used a TZD," said lead author Ann Schwartz, PhD, an associate professor with the University of California, San Francisco School of Medicine's department of epidemiology and biostatistics, who presented the new findings at the American Society for Bone and Mineral Research (ASBMR) 2013 Annual Meeting.

"This is the first study to show, however, that the fracture risk appears to gradually return to previous levels after discontinuation," she added.

Beata Lecka-Czernik, PhD, a professor with the department of orthopedic surgery and Center for Diabetes and Endocrine Research at the University of Toledo College of Medicine, Ohio, said the findings represent more unexpected insights from the ACCORD trial.

"The ACCORD studies were not designed to look for bone fracture, and it was only by accident in the final analysis [that it] did show a risk, specifically in women, and this is a very interesting new study, showing a significant decrease in the fracture risk with the discontinuation," she told Medscape Medical News.

Type 2 Diabetes Already Increases Fracture Risk

Dr. Schwartz explained that type 2 diabetes, itself, is associated with an increased fracture risk, making the risk associated with the 2 currently available TZD diabetes drugs, rosiglitazone and pioglitazone, all the more concerning. "In the US population of older adults, more than 25% have type 2 diabetes, so the increased risk of fracture is particularly important, as are the effects of diabetes medications on the skeleton," she observed.

Research linking fractures to TZD use has not been entirely consistent, with some trials suggesting different effects for the 2 commercially available TZDs and others pointing to a different degree of risk according to gender.

For this analysis, Dr. Schwartz and colleagues evaluated data on both men and women enrolled in the ACCORD trial, which was conducted to evaluate the effect of standard vs intensive glycemic control on cardiovascular disease outcomes.

The 6865 participants in the trial, with a mean age of 62.4 years, were originally treated with rosiglitazone in the standard- as well as intensive-glycemic-control groups; pioglitazone was added in the fall of 2007. The intensive intervention was stopped early because of higher mortality in that group, and all of the participants remained on the standard intervention until the end of the trial, which concluded in June 2009.

Over a mean follow-up of 4.8 years, 549 of the participants experienced at least 1 confirmed nonspine fracture, including 262 men and 287 women.

Fracture Risk Declines Quickly After Discontinuation

In proportional-hazard models that were adjusted for age and glycemic randomization, women who were on TZDs for 0 to 1 years had a hazard ratio (HR) for nonspine fracture of 1.86, compared with non-TZD users; the HR dropped to 1.27 within a year of discontinuation, however. And at 1 to 2 years after discontinuation, it dropped to 1.05, with a further decrease to 0.79 more than 2 years after discontinuation.

Among women who had taken TZDs for 1 to 2 years, the HR for nonspine fractures was 2.34 during the study, but the risk dropped to 1.59 within a year of discontinuation; it further declined to 1.32 at 1 to 2 years and to 0.99 more than 2 years after discontinuation.

And among women on TZDs for 2 years or more, the HR among current users was 2.04, dropping to 1.39 in the first year of discontinuation, followed by 1.16 after 1 to 2 years and to 0.87 after 2 years.

In men, the association between TZD use and nonspine fracture was not statistically significant, nor was the change in fracture risk following discontinuation of therapy.

The observational nature of the study, with its attendant potential for residual confounding, and the fact that TZD use was not randomly assigned, were limitations, the researchers noted.

In addition, since all nonspine fractures were combined, there is the possibility of differences in effects across skeletal locations, Dr. Schwartz said.

What is Mechanism Behind TZD Increase in Fracture?

Regarding the mechanisms behind the increased fracture risk in women — and the relatively fast reversal, Dr. Schwartz said questions remain.

"Obviously, part of what is going on underlying this is bone loss," she said.

"There have been trials showing bone loss with rosiglitazone and pioglitazone, but one would think it would take longer for bone to…accumulate and for the fracture risk to go back down, so there must be some other aspect of bone quality that we are not really capturing with a [dual-energy X-ray absorptiometry] DXA scan."

Dr. Lecka-Czernik added that a lack of analysis of serum bone turnover markers is another limitation that makes the findings difficult to interpret, she noted. "Mechanistically, we don't know how the bone remodeling occurs.

"But the point is, this shows that when you discontinue the drugs, the fracture risk decreases, which importantly tells us that whatever TZDs are doing to the bone, it is not permanent; it can be reversed and the bone can regenerate."

Dr. Schwartz and Dr. Lecka-Czernick have reported no relevant financial relationships.

American Society for Bone and Mineral Research 2013 Annual Meeting. Abstract 1027, presented October 5, 2013.

http://www.medscape.com/

Graham