Page 101 - Vitamin D and Cancer

88 E. Giovannucci

was conducted in Linxian, China. The analysis included 545 squamous cell
carcinomas of the esophagus, 353 adenocarcinomas of the gastric cardia, and 81
gastric noncardia adenocarcinomas diagnosed over 5.25 years of follow-up. For
squamous cell carcinomas of the esophagus, when comparing men in the fourth
quartile of serum 25(OH)D concentrations to those in the first, a positive associa-
tion was found (RR = 1.77; 95%CI, 1.16–2.70, P trend = 0.0033). In contrast, no
association was found in women (RR = 1.06 (95% CI = 0.71–1.59), P trend = 0.70),
or for gastric cardia or noncardia adenocarcinoma. The cut-point for the top quartile
was only 48.7 nmol/L.
The other study, from Linxian, China, was a cross-sectional analysis of 720
subjects who underwent endoscopy and biopsy, and were categorized by the pres-
ence or absence of histologic esophageal squamous dysplasia [80]. The mean level
of 25(OH)D in this population was only 35 nmol/L. In this high-risk area, 230 of
720 subjects were diagnosed with squamous dysplasia. In multivariate analyses, the
subjects in the highest compared with the lowest quartile of 25(OH)D were at a
significantly increased risk of squamous dysplasia (RR = 1.86; 95% CI, 1.35–2.62).
This association was observed both in men (RR =1.74; 95% CI, 1.08–2.93) and
women (RR = 1.96; 95% CI, 1.28–3.18).

4.9 Non-Hodgkin Lymphoma

4.9.1 Sun Exposure

The relationship between sun exposure and non-Hodgkin Lymphoma (NHL) is of
special interest because some studies suggest a positive association between NHL
and skin cancer, suggesting that sunlight may increase risk of NHL. Partly based on
this relationship, a number of case–control studies have examined sun exposure and
risk for NHL. The International Lymphoma Epidemiology Consortium (InterLymph)
recently presented results summarizing the association between sun exposure and
NHL risk in a pooled analysis of 10 case–control studies [81]. The studies
comprised 8,243 cases and 9,697 controls of European origin and were conducted
in the USA, Europe, and Australia. Four measures of self-reported personal sun
exposure were assessed at interview; these included time (1) outdoors and not in
the shade in warmer months or summer, (2) in the sun in leisure activities, (3) in
sun light, and (4) sun bathing in summer. The risk of NHL fell significantly with
the composite measure of increasing recreational sun exposure; the multivariate
pooled RR (adjusting for smoking and alcohol) = 0.76 (95% CI 0.63–0.91) for the
highest exposure category, and the trend was significant (p for trend 0.005). For
increasing total sun exposure, a nonsignificant inverse trend was observed with
NHL risk (RR = 0.87; 95% CI 0.71–1.05; P = 0.08). Of note, the inverse association
between recreational sun exposure and NHL risk was statistically significant at
18–40 years of age and in the 10 years before diagnosis, and statistically significant