Table 4. Relative risks (RRs) of melanoma by daily servings of fruits and vegetables and by frequency of consumption of orange juice.

	Cases	Person-years (thousands)	Multivariateb RR (95% CIc)
Total fruits
<1.0	62	349	1.00
1.0–1.4	58	274	1.16 (0.80–1.66)
1.5–2.4	151	524	1.42 (0.99–2.04)
2.5–3.4	83	303	1.37 (0.95–1.98)
⩾3.5	60	198	1.37 (0.86–2.20)
P for linear trend			0.22
Total vegetables
<2.0	86	448	1.00
2.0–2.9	126	479	1.22 (0.92–1.62)
3.0–3.9	106	356	1.32 (0.97–1.78)
4.0–4.9	53	196	1.11 (0.76–1.62)
⩾5.0	43	169	1.01 (0.68–1.50)
P for linear trend			0.81
Fruits and vegetables high in carotenoids
<0.25	50	263	1.00
0.25–0.49	116	447	1.30 (0.51–3.33)*
0.50–0.74	85	347	1.18 (0.49–2.85)*
0.75–0.99	68	236	1.40 (0.73–2.68)
⩾1.00	95	355	1.24 (0.67–2.29)
P for linear trend			0.78
Fruits and vegetables high in vitamin C
<0.5	57	317	1.00
0.5–0.9	77	376	1.10 (0.77–1.55)
1.0–1.4	96	378	1.33 (0.94–1.87)
1.5–1.9	92	280	1.54 (1.06–2.22)
⩾2.0	92	297	1.43 (0.99–2.07)
P for linear trend			0.01
Orange juiced
Never	39	228	1.00
1–3 month−1	70	364	1.09 (0.73–1.62)
1 week−1	44	197	1.31 (0.83–2.05)
2–6 week−1	203	799	1.44 (1.00–2.08)
⩾1 day−1	147	499	1.61 (0.92–2.84)
P for linear trend			0.008

^*P<0.05 in test for heterogeneity of results from the NHS and NHS II.

^aResults from NHS (1984–1998) and NHS II (1991–1999) were pooled to obtain risk estimates for the combined cohorts.

^bRelative risks were adjusted for the covariates listed in Table 2 footnote c plus total energy intake, multivitamin use, and use of vitamin C, vitamin E, and β-carotene supplements.

^c95% confidence interval.

^dFollow-up for NHS is 1980 – 1998.