Table 3. Age-adjusted and multivariate hazard ratios of death from all causes according to coffee and green tea consumption.

Consumption  category  (cups/day)	Analysis for all respondents				Analysis after excluding respondents with less than 5 years of follow-up
	No. of cases	Person- years	Age-adjusted hazard ratios (95%CI)	Multivariate hazard ratios* (95%CI)	No. of cases	Person- years	Multivariate hazard ratios* (95%CI)
Coffee
Men
<0.5	132	4685	1	1	89	4539	1
0.5-1	71	4162	0.71 (0.53-0.94)	0.70 (0.52-0.94)	49	4082	0.67 (0.47-0.96)
>=2	43	4858	0.50 (0.35-0.71)	0.43 (0.30-0.63)	25	4769	0.34 (0.21-0.54)
Trend P			0.0001	0.0001			0.0001
Women
<0.5	61	4367	1	1	35	4252	1
0.5-1	31	4981	0.63 (0.41-0.98)	0.70 (0.45-1.09)	22	4911	0.88 (0.51-1.52)
>=2	23	5303	0.72 (0.44-1.20)	0.76 (0.45-1.27)	17	5258	1.00 (0.53-1.86)
Trend P			0.10	0.18			0.92
Green tea
Men
<0.5	85	5007	1	1	60	4916	1
0.5-3	84	5181	0.94 (0.69-1.27)	0.97 (0.71-1.31)	51	5063	0.83 (0.57-1.21)
>=4	77	3515	0.94 (0.69-1.28)	0.82 (0.60-1.12)	52	3410	0.78 (0.53-1.13)
Trend P			0.72	0.19			0.22
Women
<0.5	50	6195	1	1	34	6130	1
0.5-3	41	5680	0.78 (0.52-1.18)	0.79 (0.52-1.20)	26	5572	0.70 (0.42-1.17)
>=4	24	2776	0.72 (0.44-1.17)	0.74 (0.45-1.21)	14	2719	0.62 (0.33-1.17)
Trend P			0.18	0.21			0.12

* Coffee and green tea consumption, age (in years), history of selected diseases, physical activity level (active or inactive), educational status (age at final graduation <16, <18, or >=18 years old), and additionally only in men, smoking status (never smoked, smoked in the past, currently smoking =<20, or >20 cigarettes per day), habitual alcohol consumption (never drank, drank in the past, currently drinking =<46g, or >46g of ethanol) were included in the Cox model as covariates.

CI= confidence interval