Adiposity and risk of prostate cancer death: a prospective analysis in UK Biobank and meta-analysis of published studies

Table 2 Sensitivity analyses. Multivariable-adjusted hazard ratios (95 % CI) for prostate cancer death in relation to adiposity measurements at recruitment in 218,237 men from UK Biobank

	BMI			Body fat percentage			Waist circumference			Waist to hip ratio
	n prostate cancer deaths	Per 5 kg/m² increase	p-trend¹	n cases	Per 5 % increase	p-trend¹	n cases	Per 10 cm increase	p-trend¹	n cases	Per 0.05 unit increase	p-trend¹
Overall	659	1.07 (0.97–1.17)	0.190	645	1.00 (0.94–1.08)	0.892	659	1.06 (0.99–1.14)	0.116	659	1.07 (1.01–1.14)	0.028
Excluding first 5 years of follow-up	568	1.06 (0.96–1.18)	0.265	559	1.01 (0.93–1.09)	0.802	568	1.05 (0.98–1.14)	0.180	568	1.07 (1.00–1.14)	0.060
Excluding men with BMI ≥ 25 kg/m², per 1 SD increment	144	0.93 (0.79–1.10)	0.404	140	0.87 (0.74–1.03)	0.118	144	1.04 (0.87–1.24)	0.667	144	1.13 (0.96–1.33)	0.144
Excluding extreme values: percentiles 1–99	659	1.07 (0.97–1.18)	0.170	645	1.01 (0.94–1.08)	0.826	659	1.06 (0.99–1.14)	0.101	659	1.07 (1.01–1.14)	0.023
Excluding men < 50 years of age	650	1.06 (0.96–1.17)	0.223	636	1.00 (0.93–1.08)	0.950	650	1.06 (0.98–1.14)	0.137	650	1.07 (1.01–1.14)	0.034
Per 1 SD increment	659	1.06 (0.98–1.15)	0.170	645	1.01 (0.93–1.10)	0.826	659	1.07 (0.99–1.16)	0.101	659	1.10 (1.01–1.19)	0.023
Residuals^a							659	1.08 (0.92–1.27)	0.325	659	1.07 (0.99–1.15)	0.072

Abbreviations: BMI body mass index
Cox regression analyses. All models are stratified by region and age at recruitment and adjusted for age (underlying time variable), Townsend deprivation score, ethnicity, lives with a wife or partner, smoking, physical activity, alcohol consumption, height, diabetes, and history of PSA test. Full details for each covariate are provided in the statistical section
¹P-values for trend are obtained by entering the anthropometric variable per increment in the Cox regression model
^aHR (95% CI) is from the multiple adjusted model (above) after accounting for the residuals of waist circumference and waist to hip ratio regressed on BMI for analyses of waist circumference and waist to hip ratio as exposures

ISSN: 1741-7015