Table 3.

Threshold effect analysis of testosterone on the prevalence of kidney stones using piece-wise linear regression.

	Crude ORb (95% CI)	p-value		Adjusted ORc (95% CI)	p-value
All, Testosteronea
Model I one-line	0.904 (0.873, 0.937)	<0.001		0.972 (0.935, 1.011)	0.159
Model II turning point: Q5 (386–443)			Model II turning point: Q5 (386–443)
<Q5	0.942 (0.884, 1.002)	0.059	<Q5	1.033 (0.966, 1.105)	0.336
≥Q5	0.843 (0.764, 0.930)	<0.001	≥Q5	0.875 (0.731, 0.981)	0.010
Age (20–40) Testosteronea
Model I one-line	0.953 (0.880, 1.031)	0.2271		0.995 (0.906, 1.092)	0.9118
Model II turning point: Q3 (313–363)			Model II turning point: Q3 (313–363)
<Q3	0.807 (0.603, 1.080)	0.1499	<Q3	0.809 (0.598, 1.096)	0.1721
≥Q3	1.006 (0.891, 1.136)	0.9257	≥Q3	1.063 (0.932, 1.213)	0.3645
Age 41–60 Testosteronea
Model I one-line	0.892 (0.839, 0.948)	<0.001		0.921 (0.860, 0.987)	0.0193
Model II turning point: Q5 (370–427)			Model II turning point: Q6 (427–492)
<Q5	0.941 (0.844, 1.049)	0.2736	<Q6	0.995 (0.906, 1.092)	0.9084
≥Q5	0.810 (0.680, 0.964)	0.0179	≥Q6	0.657 (0.487, 0.886)	0.006
Age 61–80 Testosteronea
Model I one-line	0.930 (0.883, 0.979)	0.005		0.979 (0.924, 1.038)	0.483
Model II turning point: Q3 (268–318)			Model II turning point: Q5 (360–422)
<Q3	1.141 (0.938, 1.389)	0.187	<Q5	1.084 (0.979, 1.200)	0.120
≥Q3	0.872 (0.805, 0.943)	<0.001	≥Q5	0.831 (0.717, 0.963)	0.014

CI, confidence interval; HR, hazard ratio.

^aPresented in eight quantile (ng/dl) and was conducted the with eight quantiles of the total serum testosterone as a continuous variable.

^bAdjusted for: none.

^cAdjusted for: age, race, BMI, hypertension, diabetes, gout, coronary heart, disease, arthritis, angina, heart attack, stroke, smoking, serum total cholesterol, triglycerides, calcium, and uric acid.