For personal use. Only reproduce with permission from The Lancet Publishing Group. RESEARCH LETTERS Our results underline the importance of cultural factors in health promotion, but also point to the improvements in health that can be brought about by simple behavioural changes. 1 Anon. Elimination of neonatal tetanus will save 1 million newborns annually. Soz Praventivmed 1996; 41: 262–63. 2 Bennett J, Schooley M, Traverso H, Agha SB, Boring J. Bundling, a newly identified risk factor for neonatal tetanus: implications for global control. Int J Epidemiol 1996; 25: 879–84. 3 Mustafa BE, Omer MI, Aziz MI, Karrar ZE. Neonatal tetanus in rural and displaced communities in the East Nile Province. J Trop Pediatr 1996; 42: 110–12. 4 Anita-Obong OE, Young MU, Effiong CE. Neonatal tetanus: prevalence before and subsequent to implementation of the Expanded Programme on Immunization. Ann Trop Pediatr 1993; 13: 7–11. 5 Mull DS, Anderson JW, Mull JD. Cow dung, rock salt, and medical innovation in the Hindu Kush of Pakistan: the cultural transformation of neonatal tetanus and iodine deficiency. Soc Sci Med 1990; 30: 675–91. ICROSS Kenya, PO Box 507, Ngong Hills, Kenya (M E Meegan PhD, S Ole Lengeny); Royal College of Surgeons in Ireland, Mercer Building, Dublin 2, Ireland (R M Conroy MusB); Faculty of Medicine, McMaster University, Hamilton, Ontario L8S 4L8, Canada (K Renhault); and Ministry of Health, Kajiado District, Kenya (J Nyangole MD) Correspondence to: Dr Ronán M Conroy (e-mail: rconroy@rcsi.ie) mortality are poorly defined, although environmental factors, including latitude, are implicated. This association with latitude has been interpreted to indicate that ultraviolet radiation (UVR) protects against prostate cancer. 1 We have examined this hypothesis by studying 210 patients with sporadic prostate cancer and 155 patients with benign prostatic hypertrophy. We obtained ethics committee approval and written informed consent from patients. We recruited white northern-European people who were not related to one another from the North Staffordshire Hospital, UK between October, 1999 and May, 2000. We attempted to recruit all patients with histological evidence of prostatic adenocarcinoma or, if histological results were unavailable, with a malignant prostate gland on digital rectal examination, positive bone-scan results, and prostatic specific antigen concentration more than 30 ng/ml (age- related reference range 0–6·5 ng/ml). During the study, 217 patients with prostate cancer were seen in urology clinics, and seven did not fulfil our inclusion criteria. Out of the remaining 210 patients, histological evidence of prostate cancer was available for 190. Patients with benign prostatic hypertrophy were chosen as controls. We attempted to recruit all patients with histological evidence of this condition or, if histology was unavailable, with prostatic specific antigen concentration in serum in the appropriate reference range and benign results on digital rectal examination. We identified 167 patients with benign prostatic hypertrophy, but 12 were excluded because their prostatic specific antigen concentration rose above the reference range in follow-up. Histological evidence was available for 123 patients. More than 85% of cases and controls who attended hospital during the study were asked to participate. Two patients with benign prostatic hypertrophy refused consent. We used logistic regression analysis to obtain odds ratios of having prostate cancer with various UVR exposure levels. Because mean age at diagnosis in cases (70·6 years, SD 7·3) was greater (p0·0001) than in controls (67·0 years, 7·9), we included age at diagnosis in all models. We used a validated questionnaire to study UVR exposure (table) and attempted to eliminate potential bias by allowing self-administration. 3,4 About 25% of participants needed help to complete the questionnaire. Accordingly, the interviewer (CJL) used a neutral script to guide the questions. Questionnaires with a closed-answer design similar to ours are commonly validated for blinded and non- blinded investigators and for self-administered use, especially when a set script is followed by the interviewer. The table shows cumulative lifetime exposure to UVR in cases (355 weeks) and controls (393 weeks). Sun exposure was estimated from weekday and weekend activity, and, thus, measures occupational and recreational exposure. UVR exposure had a significantly protective effect on cancer risk (table). The table also shows the proportions of cases and controls in each quartile of exposure. Comparison of the THE LANCET • Vol 358 • August 25, 2001 641 Exposure to ultraviolet radiation: association with susceptibility and age at presentation with prostate cancer Christopher J Luscombe, Anthony A Fryer, Michael E French, Samson Liu, Mark F Saxby, Peter W Jones, Richard C Strange A positive association between latitude and prostate cancer mortality has been interpreted to indicate that ultraviolet radiation (UVR) protects against development of this cancer. We aimed to examine this hypothesis. We compared exposure between 210 cases and 155 controls. Childhood sunburn (odds ratio 0·18, 95% CI 0·08–0·38), regular foreign holidays (0·41, 0·25–0·68), sunbathing score (0·83, 0·76–0·89), and low exposure to UVR (3·03, 1·59–5·78) were associated with development of prostate cancer. Furthermore, cases with low UVR exposure developed cancer at a younger median age (67·7 years, IQR 61·5–74·6) than cases with higher exposure (72·1 years, 67·5–76·4); p0·006. These findings are compatible with UVR having a protective role against prostate cancer. Lancet 2001; 358: 641–42 Sporadic prostate cancer is a leading cause of cancer death. 1,2 Factors that affect prostate cancer susceptibility and BPH Cancer Odds ratio (95% CI) p Chronic exposure Mean weeks cumulative exposure (SD) 393 (201) 355 (194) 0·998 (0·997–0·999)* 0·006 Lowest exposure quartile (%) 18·7 29·0 3·03 (1·59–5·78) 0·008 25–50% exposure quartile (%) 24·5 25·2 1·51 (0·83–2·76) 0·182 50–75% exposure quartile (%) 27·1 22·9 1·18 (0·65–2·16) 0·588 Highest exposure quartile (%) 29·7 22·9 1·00 Living abroad in sunny country for >6 months (%) 32·9 31·0 0·71 (0·45–1·14) 0·161 Acute exposure Positive childhood sunburn (%) 21·9 4·3 0·18 (0·08–0·38) 0·0001 Mean sunbathing score (SD) 7·7 (2·8) 6·2 (2·8) 0·83 (0·76–0.89)† 0·0001 History of regular foreign holidays (%) 34·8 17·1 0·41 (0·25–0·68) 0·005 Mean weeks foreign holiday/year (SD) 1·05 (1·7) 0·56 (1·8) 0·85 (0·74–0·98)* 0·030 BPH=benign prostatic hypertrophy. Odds ratio derived from logistic regression analysis including age at diagnosis. *Per week. †Per unit score. Exposure to UV light derived from questionnaire responses