The relationship between glycemic control and falls in older adults

OBJECTIVES: To determine whether glycemic control contributes to fall risk in frail and nonfrail elderly adults with diabetes mellitus. DESIGN: Retrospective, case-controlled design. SETTING: Health maintenance organization in the Denver, Colorado, metropolitan area. PARTICIPANTS: One hundred eleven community-dwelling adults aged 75 and older who receive care through Kaiser Permanente of Colorado. All subjects had been diagnosed with diabetes mellitus, had at least one hemoglobin A1C (HbA1c) measurement in the previous 12 months, and were using oral hypoglycemic medication or insulin to control their diabetes mellitus. MEASUREMENTS: Measurements of risk factors (Vulnerable Elders Survey (VES-13) with a cutpoint of 3 to determine frailty status, self-reported number of falls over the prior 12-month period, HbA1c, fasting low-density lipoprotein cholesterol, average blood pressure, and other factors related to fall risk) were obtained through telephone interview and medical chart review. The outcome measure was falls. RESULTS: Bivariate analyses to assess correlations between falls and risk factors determined that only HbA1c, frailty, and peripheral neuropathy were significantly associated with falls. A stepwise logistic regression determined that fall risk markedly increased when HbA1c was 7 or below, regardless of frailty status. CONCLUSION: In this retrospective study of a convenience sample of frail older adults with diabetes mellitus, tighter glycemic control was associated with greater risk of falling. Prospective studies that further evaluate the risks and benefits of relaxed glucose control in high-risk older adults are needed to confirm this finding.