Introduction: Underweight is associated with the development of vertebral fractures. However, the impact of time-burden of underweight on the risk of vertebral fractures is unknown. We investigated the effect of the cumulative longitudinal burden of underweight on the development of vertebral fracture.

Methods: We included 561,779 individuals without vertebral fracture who underwent four annual health examinations during 2009–2012 from the database of the Korean national health insurance service. Underweight burdens were evaluated in the following way: cumulative number of underweight (body mass index < 18.5) diagnosed at each health examination (0–3 times). Baseline (index date) demographic data including age, sex, smoking, drinking, regular exercise, low income, height, weight, BMI, waist circumference, and comorbidities (diabetes, hypertension, dyslipidemia, and chronic renal failure) was also evaluated. The risk of vertebral fractures according to the underweight and underweight burden was estimated using Cox proportional-hazards models.

Results: During a mean follow-up of 8.3 years, 7049 new vertebral fracture was occurred. The incidence rate per 1000 person years was 1.5. The proportions of the underweight, normal, overweight, stage 1 obesity, and stage 2 or 3 obesity populations at the index date were 1.8% (n=10,121), 35.8 % (n=201,152), 28.3 % (n=159,052), 31.5 % (n=177,198), and 2.5% (n=14,256), respectively. In the Cox proportional hazards model, the risk of vertebral fractures in the underweight population increased approximately 1.2 fold compared to normal body weight population (p = 0.02). Of all individuals, 97.2%, 1.2%, 0.7%, and 0.9% met the underweight diagnostic criteria 0, 1, 2, and 3 times, respectively. The incidence of vertebral fractures decreased when underweight occurred once compared to reference (no underweight), but the incidence of vertebral fracture increased when underweight was accumulated twice (1.4 times) or three times (1.2 times).

Discussion: Given the positive correlations between the cumulative underweight burdens and the risk of vertebral fracture, maximal effort to detect and correct underweight might be important to prevent vertebral fracture.