Skip to main content
. 2016 Jun 13;101(9):3297–3305. doi: 10.1210/jc.2016-1366

Table 4.

Multivariable-Adjusted HR Estimates for Incident Fracture by Fasting Plasma TG Levels at Baseline and in Time-Varying Analyses

TG level (mg/dl) Number (%) of Repeat-Measures That Contributed to Fracture vs Nonfracture Full Model HR (95% CI)a
For Fracture Incidence 2 or More Y Later, Based on Baseline Fasting TG Level Time-Varying Analyses: TG Level at Visit i and Fracture Risk 2–5 Y Later
Single TG Level Mean TG Level of 2 Consecutive Measuresb
<150 793 (73)/7776 (77) 1 (reference) 1 (reference) 1 (reference)
150–299 242 (22)/2067 (20) 0.89 (0.48–1.65) 1.13 (0.80–1.59) 1.03 (0.70–1.52)
≥300 49 (5)/298 (3) 2.48 (1.13–5.44) 1.94 (1.14–3.30) 2.21 (1.29–3.79)
a

Full model included age at baseline, race/ethnicity, study site, smoking, drinking, physical activity, menopausal stage (including exogenous hormone use), BMI, diabetes, and lumbar spine BMD.

b

Mean TG level for TG at visit i and the last TG level measured before visit i in relation to fractures between i + 2 and i + 5 y.