Dr. Xiang, from the University of Southern California Keck School of Medicine, Los Angeles, and colleagues studied 526 Latino women with prior gestational diabetes mellitus who elected to use either DMPA (n = 96) or combination oral contraceptives (n = 430). Median follow-up was about 1 year, with a maximum of 9 years.

The annual incidence rate of diabetes was 19.1% in the DMPA group, the authors report, compared with 11.9% in the combined oral contraceptives group.

Adjustment for baseline body mass index, breast-feeding rates, and family history reduced the increased risk associated with DMPA use from 58% to 18%, the results indicate. Additional adjustment for weight gain during follow-up reduced the increased risk even further, to 7%.

Among women who used DMPA, those with higher than median triglyceride levels had more than double the risk of diabetes, the researchers note, and those who breast-fed also had a similar increase in risk, although this difference was not statistically significant.

"Neither triglycerides nor DMPA use had an important effect to increase diabetes rates alone," the investigators explain, "but together they increased the risk, demonstrating the significant interaction between the two factors."

Dr. Xiang offered this approach to contraceptive choices for Latino women with a history of gestational diabetes: "If breast-feeding, use a non-hormonal approach," and if not breast-feeding "but with relatively high triglycerides, use a non-hormonal method or low-dose combination pills; try to avoid DMPA."

If patients in those categories still elect to use DMPA, "glucose should be carefully monitored, particularly at annual physical exam," she added.