Table 2.

Hazard ratio (95%CI) for incident fracture by quartiles of selenium intake.

Quartiles of selenium intake	Q1	Q2	Q3	Q4	P for trend
A. Cumulative average selenium intake
Cases	226	219	253	278
Person-years	38,732	36,754	36,123	36,161
Incident rate (per 1000)	5.8	6.0	7.0	7.7
Model 1	0.99 (0.82–1.20)	1.00	1.18 (0.99–1.42)	1.36 (1.13–1.63)	0
Model 2	1.09 (0.88–1.36)	1.00	1.22 (1.00–1.50)	1.25 (1.01–1.54)	0.092
Model 3	1.07 (0.86–1.33)	1.00	1.25 (1.02–1.53)	1.33 (1.07–1.65)	0.021
Model 4	1.02 (0.83–1.24)	1.00	1.24 (1.03–1.49)	1.30 (1.07–1.58)	0.005
B. Baseline selenium intake
Cases	217	257	258	244
Person-years	40,567	38,197	35,992	33,014
Incident rate (per 1000)	5.3	6.7	7.2	7.4
Model 1	0.76 (0.64–0.91)	1.00	1.08 (0.91–1.29)	1.13 (0.95–1.35)	0
Model 2	0.84 (0.68–1.03)	1.00	1.09 (0.89–1.32)	1.06 (0.87–1.30)	0.023
Model 3	0.82 (0.67–1.01)	1.00	1.11 (0.91–1.34)	1.10 (0.89–1.35)	0.006
Model 4	0.79 (0.65–0.95)	1.00	1.05 (0.88–1.26)	1.10 (0.92–1.32)	0.001
C. Most recent selenium intake
Cases	235	221	252	268
Person-years	38,194	37,183	36,908	35,486
Incident rate (per 1000)	6.2	5.9	6.8	7.6
Model 1	0.99 (0.82–1.20)	1.00	1.17 (0.98–1.41)	1.29 (1.07–1.55)	0.003
Model 2	1.14 (0.92–1.42)	1.00	1.18 (0.96–1.46)	1.21 (0.98–1.50)	0.335
Model 3	1.11 (0.90–1.38)	1.00	1.21 (0.98–1.49)	1.29 (1.03–1.62)	0.112
Model 4	1.12 (0.92–1.36)	1.00	1.19 (0.99–1.44)	1.30 (1.06–1.60)	0.074

Model 1 adjusted age, gender, and energy intake.

Model 2 further adjusted intake of fat, smoking, alcohol drinking, income, urban, education, and physical activities.

Model 3 further adjusted dietary patterns ²⁰(traditional south pattern represented high intake of rice, pork, and vegetables, and low intake of wheat; modern dietary pattern represented high intake of fruit, soy milk, egg, milk, and deep-fried food).

Model 4 further adjusted for self-reported diabetes and hypertension.

All the adjusted variables were treated as time-varying covariates.