Table 2.

Incidence rate and risk of incident diabetes associated with frequency of fresh fruit consumption

	< 1 time/week	1–3 times/week	4–7 times/week	> 7 times/week	P trend
Incident diabetes, n (%)*	630 (8.00)	1842 (7.51)	2909 (7.28)	505 (6.68)
Model 1	Ref. (1.00)	1.00 (0.91–1.09)	0.96 (0.88–1.05)	0.87 (0.77–0.98)	0.0407
	p	0.9910	0.3650	0.0199
Model 2	Ref. (1.00)	0.99 (0.90–1.08)	0.94 (0.86–1.03)	0.84 (0.75–0.95)	0.0095
	p	0. 7945	0.1796	0.0044
Model 3	Ref. (1.00)	0.98 (0.90–1.08)	0.92 (0.84–1.01)	0.84 (0.74–0.94)	0.0042
	p	0.7195	0.0632	0.0035
Model 4	Ref. (1.00)	0.97 (0.88–1.08)	0.92 (0.83–1.02)	0.84 (0.73–0.97)	0.0527
	p	0.6100	0.1307	0.0106

Data are hazard ratio (HR), 95% confidence interval (CI). P value for trend was for the Wald statistic, the “type III” results, for the additive model from the multivariable Cox regression models. P values were from the maximum likelihood estimate analysis. Model 1, adjusted for age, sex, and study areas; Model 2, adjusted for age, sex, BMI, waist circumference; Model 3, adjusted for age, sex, BMI, waist circumference, physical activity, sedentary time, smoking and drinking status, education level, family history of diabetes, triglycerides, LDL-C and HDL-C; Model 4, further adjusted for other major dietary consumption components based on Model 3. Total n = 79,922. * as P for Mantel–Haenszel chi-square test to compare incident rate of diabetes among the four groups was 0.0013. p value was for the risk of diabetes in relation to each specific group of fruit intake, as compared to the < 1time/week