Background: Older adults with type 2 diabetes and those taking insulin have an increased risk of falls. However, previous studies in these populations have been cross-sectional or have not included specific data about glycemic control and diabetic complications. To better evaluate the contributions of diabetes progression, treatment, or complications in the risk of falls, Schwartz and colleagues analyzed longitudinal data of patients with diabetes from the Health, Aging and Body Composition (Health ABC) study.

The Study: Health ABC, a prospective study, included 3,075 men and women 70 to 79 years of age who were recruited at the Universities of Pittsburgh and Tennessee in 1997 or 1998. Patients who reported difficulties with activities of daily living, climbing up 10 steps, or walking one fourth of one mile at enrollment were excluded. Participants were evaluated at annual visits. Diabetes was reported or diagnosed in 719 patients (24 percent) at baseline; 446 of these patients received a peripheral nerve function measurement after three years and were included in the analysis. At baseline, the 273 patients who were excluded had A1C values and diabetes duration similar to the study participants, but had reduced renal function, poorer physical performance, and more frequent falls.

At the annual visits, patients reported how many times they had fallen in the previous 12 months (i.e., zero, one, two or three, four or five, or six or more). Tests to assess diabetes-related complications (e.g., those for visual acuity, contrast, depth perception, and peripheral nerve conduction) were completed at specific follow-up visits. Physical performance, fasting plasma glucose levels, oral glucose tolerance, serum creatinine levels, and A1C levels were also tested regularly throughout the study.

Results: Multivariate analysis included the additional variables of change in body weight, change in diastolic blood pressure, and the effect of insulin therapy on A1C levels. It also accounted for potential confounders that were identified in earlier studies, including depression; use of antidepressants, selective serotonin reuptake inhibitors, loop diuretics, or calcium channel blockers; history of stroke; and mental status.

The average age of participants at baseline was 74 years, and the average A1C level was 7.6 percent; 14 percent of participants were using insulin. In the first year, 23 percent of participants reported falls. In the subsequent years, 22, 26, 30, and 31 percent reported falls. In the small number of patients who used insulin, those who had an A1C level of 6 percent or less had an increased risk of falls compared with those who had A1C levels greater than 8 percent (odds ratio = 4.36). Intensive glycemic control was not associated with falls in those using only oral hypoglycemic medications. Progression of diabetic complications was associated with increased falls, but not necessarily with poor glycemic control.

Conclusion: The authors conclude that intensive glycemic control with insulin, but not with oral hypoglycemic medications, appears to increase the risk of falls in older patients with diabetes. They suggest that reducing diabetes-related complications may reduce falls in this population.