Women with type 2 diabetes prescribed thiazolidinediones did not see an increased risk or benefit with respect to incident breast cancer, according to a meta-analysis published in Diabetes Metabolism Research and Reviews.
“Previous studies demonstrated that diabetes could contribute to an increased risk of breast cancer-specific mortality,” Jieli Lu, MD, PhD, of the Key Laboratory for Endocrine and Metabolic Diseases at Ruijin Hospital in Shanghai, and colleagues wrote in the study background. “Interestingly, increasing evidence suggests that anti-diabetic therapies, such as metformin, insulin glargine and TZDs, might also affect the risk of breast cancer incidence and mortality.” Results across studies, however, have been inconsistent, the researchers noted.
Lu and colleagues analyzed data from 14 studies examining any association between TZD treatment and risk for breast cancer in women with diabetes published before January 2016, including five randomized controlled trials, seven cohort studies and two case-control studies (n = 1,089,039 with 12,233 breast cancer cases). Exposure to TZDs was defined as TZD monotherapy or ever use of TZDs. Researchers used random-effects models to assess the association between TZD use and risk for incident breast cancer, using RR as the common effect size for cohort studies and randomized controlled trials and ORs for case-control studies.
Researchers found that the strength of any association between TZD use and breast cancer risk depended on the type of study, event number and effect size. In the five randomized controlled trials, women prescribed TZDs did not experience a decrease in breast cancer risk vs. TZD nonusers, which included reference groups assigned placebo, metformin or glyburide, metformin or sulfonylurea, insulin plus sulfonylurea or metformin, or sitagliptin plus metformin (pooled RR = 0.77; 95% CI, 0.39-1.53). Across the seven cohort studies, however, researchers observed that TZD use had a mild, protective effect on breast cancer risk (pooled RR = 0.81; 95% CI, 0.66-0.99). Comparison groups in those studies were described as never pioglitazone users, non-TZD users, other antidiabetic drug users, or metformin or sulfonylurea users, or were assigned to placebo. In the two case-controlled studies, where reference groups were described as other antidiabetic drug users or never TZD users, respectively, the result did not rise to significance (pooled OR = 0.99; 95% CI, 0.76-1.28), according to researchers.
In separate analyses for rosiglitazone and pioglitazone, researchers found no association between the individual therapies and breast cancer risk. Additionally, when removing one cohort study from analysis with the smallest event number and greatest effect size compared with the other included studies, heterogeneity decreased and the association among cohort studies became nonsignificant, according to researchers.
They noted that the randomized controlled trials included in the current study were not designed for breast cancer prevention — breast cancer was considered an adverse event in these trials, and the event numbers were small. In the observational studies, the protective effect was observed primarily in studies with small sample sizes, whereas four of the studies did not adjust for any confounders.
“Our study attests [to] the challenges in epidemiological investigations of this issue,” the researchers wrote. – by Regina Schaffer
Disclosures: The authors report no relevant financial disclosures.