Table 3.

Relationship between RFM and incident diabetes in different models.

	Variable	Model 1 (HR,95%CI, P)	Model 2 (HR,95% CI, P)	Model 3 (HR,95% CI, P)
All	RFM	1.24 (1.21, 1.27) < 0.0001	1.13 (1.08, 1.19) < 0.0001	1.08 (1.02, 1.14) 0.0055
	RFM (quartile)
	Q1	ref	ref	ref
	Q2	2.01 (1.46, 2.76) < 0.0001	0.98 (0.69, 1.38) 0.9003	0.72 (0.51, 1.03) 0.0719
	Q3	5.19 (3.79, 7.11) < 0.0001	1.32 (0.87, 2.00) 0.1972	0.96 (0.63, 1.47) 0.8539
	Q4	19.03 (12.27, 29.53) < 0.0001	1.65 (0.84, 3.25) 0.1484	1.18 (0.58, 2.39) 0.6543
	P for trend	< 0.0001	0.1033	0.6038
Female	RFM	1.28 (1.22, 1.33) < 0.0001	1.19 (1.09, 1.30) < 0.0001	1.13 (1.04, 1.24) 0.0061
	RFM (quartile)
	Q1	ref	ref	ref
	Q2	2.11 (0.76, 5.81) 0.1502	1.40 (0.50, 3.90) 0.5190	1.34 (0.48, 3.75) 0.5814
	Q3	3.59 (1.39, 9.29) 0.0084	1.61 (0.60, 4.29) 0.3442	1.28 (0.47, 3.49) 0.6335
	Q4	14.42 (6.17, 33.68) < 0.0001	2.86 (1.04, 7.88) 0.0419	1.63 (0.56, 4.72) 0.3711
	P for trend	< 0.0001	0.0240	0.3942
Male	RFM	1.22 (1.18, 1.26) < 0.0001	1.11 (1.04, 1.18) 0.0023	1.05 (0.98, 1.13) 0.1511
	RFM (quartile)
	Q1	ref	ref	ref
	Q2	1.40 (0.86, 2.30) 0.1791	0.96 (0.58, 1.60) 0.8867	0.60 (0.36, 1.01) 0.0547
	Q3	1.99 (1.26, 3.15) 0.0034	0.97 (0.59, 1.59) 0.9002	0.50 (0.30, 0.85) 0.0107
	Q4	6.09 (4.05, 9.15) < 0.0001	1.76 (1.04, 3.00) 0.0367	0.82 (0.47, 1.46) 0.5031
	P for trend	< 0.0001	0.0085	0.7088

Model 1: we did not adjust for any covariants.

Model 2: we adjusted for gender, age, BMI, alcoholic intake, smoking status, exercise habits, SBP, and DBP.

Model 3: we adjusted for gender, age, BMI, alcoholic intake, smoking status, exercise habits, SBP, DBP, ALT, AST, GGT, HDL-C, TC, TG, HbA1c, and FPG.

Note: The models were not adjusted for gender variables in both male and female models.

HR: hazard ratio; CI: confidence interval; Ref: Reference; RFM: relative fat mass.