Weight change and fracture risk in patients with diabetic kidney disease: A nationwide population-based study

Abstract

Background: The increased risk of fracture has been associated with weight loss in patients with diabetes or chronic kidney disease. However, the relationship between weight changes over time and fracture risk in patients with diabetic kidney disease is still unknown.

Methods: A total number of 78,922 patients with diabetic kidney disease, aged ≥ 40 years, were selected using the Korean National Health Insurance Service database, between 2009 and 2012. They were followed up until the end of 2018. Weight change was defined as the difference in body weight from the index year to 2 years later. Weight changes were then divided into five categories, ranging from weight loss of ≥10% to weight gain of ≥10%.

Results: Fractures were identified in 9,847 patients with diabetic kidney disease, over a median follow-up of 5.2 years. The risk of composite fracture of the vertebral, hip, or other sites increased as the weight change increased. Specifically, patients with ≥10% weight loss (hazard ratio [HR], 1.286; 95% confidence interval [CI], 1.184-1.398) and ≥10% weight gain (HR, 1.198; 95% CI, 1.080-1.330) showed a higher HR compared to those with ≤ 5% weight change after adjusting for several confounding factors. Higher HR of vertebral and hip fractures was also seen with increased weight loss or gain. In particular, patients with ≥10% weight loss showed the highest HR for hip fractures (HR, 1.738; 95% CI, 1.489-2.028).

Conclusions: Both weight loss and weight gain increase the risk of fracture in patients with diabetic kidney disease. Therefore, patients with diabetic kidney disease who experience weight changes should be made aware of the risk of fracture.

Keywords: diabetes; fracture; kidney disease; national health programs; weight.

Figure 1

Flowchart of participant enrolment.

Figure 2

Subgroup analysis for adjusted hazard ratios (HRs) of incidence any fracture. Horizontal lines represent the range for 95% confidence intervals. Models were adjusted for age, sex, smoking, alcohol consumption, regular exercise, income status, use of insulin, more than three classes of oral hypoglycemic agents, diabetes duration, previous histories of hypertension, dyslipidemia and end-stage renal disease, and previous body weight. OHA, oral hypoglycemic agents; BMI, body mass index; DM, diabetes mellitus.