The safety of vasectomy: recent concerns.

The safety of vasectomy: recent concerns* T.M.M. Farley,1 0. Meirik,2 S. Mehta,2 & G.M.H. Waites3 Vasectomy has been accepted for family planning by approximately 42 million couples worldwide, the majority of whom live in developing countries. It is a highly reliable and safe contraceptive method, which has been extensively studied. Recently, however, renewed concerns have been raised about a possible effect between vasectomy and cancer of the prostate many years after the procedure has been performed. These concerns are based on research conducted in the USA, where there is a high and rising incidence of prostate cancer. This review discusses the evidence for this association and its potential impact in developing countries. The factors influencing the development and growth of pros- tate cancer are poorly understood and complicate any research into risk factors for the disease. Over- all incidences of prostate cancer in some developed countries, such as the USA, are fifty times higher than in some developing countries, such as China. The majority of epidemiological studies on the rela- tionship between vasectomy and prostate cancer have been based in the USA, but the findings are in- consistent and the reported associations weak. On the basis of currently available data, no changes in family planning policies with regard to vasectomy are warranted, but the concerns raised by these studies require that research into any possible association be undertaken in developing countries where vasectomy is widely practised. Introduction Vasectomy: surgical and other methods Occlusion of the vas deferens (sperm ducts) by sur- gical means is an important family planning method and is popularly referred to as "vasectomy" or male sterilization. In recent years simpler and less inva- sive methods for vasectomy have been introduced (1); for example, the "no scalpel" method, in which the vas deferens is exposed through a small punc- ture hole, is acquiring widespread acceptability, is quicker to perform, and has a lower complication rate (2). Percutaneous injection into the vas deferens of sclerosing or polymerizing agents, such as poly- urethane or silicone, to form plugs is also practised, and has been widely used in China, where it was originally developed. Surgical vasectomy is regarded as a permanent method of family planning, though successful reversal (vasovasostomy) has been per- formed (1). It is claimed that silicone plug vas occlu- sion is more easily reversible, though this has yet to be independently established. The surgical vasecto- my procedure is safe and only rarely are complica- tions reported (3). * From: Special Programme of Research, Development and Research Training in Human Reproduction, World Health Orga- nization, 1211 Geneva 27, Switzerland. I Statistician. 2 Medical Officer. Requests for reprints should be sent to Dr Meirik. 3 Scientist. Reprint No. 5400 Worldwide patterns of vasectomy Vasectomy is a highly effective method of contra- ception and, together with use of condoms, is one of the few contraceptive options available to men. It is estimated that a total of 41.5 million couples rely on vasectomy as their method of family planning (Table 1) (4). Use of vasectomy varies considerably between and within regions and countries. Its popu- larity lags behind female sterilization (estimated world total, 140 million couples) intrauterine devices (100 million), oral contraceptives (65 million) or condoms (45 million). The method is used by a great- er proportion of married couples in developed coun- tries, although the total and annual numbers of vasectomies are greater in developing countries. As yet, vasectomy is only rarely performed in Africa. Long-term safety of vasectomy General morbidity and survival The relationship between vasectomy and long-term morbidity and mortality has been extensively exam- ined (5). A cohort study in the USA involving over 10 000 vasectomized men and a similar number of controls revealed that both groups had similar rates of 98 diseases, including various cancers, autoim- mune diseases, and heart disease. The vasectomized men had a lower death rate than the non-vasectomized men, including half the death rate from cancer (6). A more recent study in the USA of the long-term safety of vasectomy also observed that there was no in- crease in overall cancer mortality; however, there was a significant increase in the risk of lung cancer Bulletin of the World Health Organization, 71 (3/4): 413-419 (1993) © World Health Organization 1993 413 T.M.M. Farley et al. Table 1: Prevalence of vasectomy, 1980s8 %of n couple Region and area vase( All developing countries Asia and Pacific region China India Other Asian and Pacific territories Bangladesh Hong Kong Korea Nepal Sri Lanka Thailand Near East and North Africa Sub-Saharan Africa Latin America Brazil Mexico Guatemala All developed countries Australia and New Zealand Australia New Zealand Europe Great Britain Netherlands Norway North America Other developed countries World total by region, in the late antibodies. Although a WHO meeting convened in 1981 (14) to review the sequelae of vasectomy narried Total relying concluded that there was no clinical or epidemiologi- bsusing on vasectomy cal evidence of an increase in cardiovascular, endo- ctomy (x106 couples) crine, or autoimmune diseases following the pro- 5 32.3 cedure, the concerns raised by the animal stud- ies prompted research on this question among men 8 18.1 in both developed and developing countries (6, 8, 7 13.0 12, 15-17). The results obtained did not substantiate the concerns, and showed that there were no adverse 1 0.8 long-term effects on the cardiovascular system; the 2 consistency of the results from countries with very 1 different patterns of cardiovascular disease risk fac- 11 tors and incidence provided additional evidence that 6 the original animal observations were not relevant to 5 vasectomized men. Also, subsequent research on 6 monkeys could not reproduce the earlier results (18). <0.1 <0.1 nA <0.5 <0.5 0.7 5 13 10 23 3 12 11 4 13 2 5 Testicular cancer Around 1990, epidemiological studies carried out in the USA (19) and Scotland (20) suggested that there was an increased risk of testicular cancer following vasectomy. Methodological concerns about the stud- 9.2 ies were raised during a WHO meeting convened in 0.4 1991 (21) to review these and other results related to prostate cancer; consideration of possible biological mechanisms found no plausible explanation for such an association. The majority of testicular tumours are 3.0 germ cell in origin and the risk factors for testicular cancer appear to be already present at birth. The concerns that vasectomy might cause new, or accel- erate the development of existing, testicular tumours 5.2 have not been substantiated by later research (12). 0.6 41.5 a Adapted from ref. 4. among men who had undergone vasectomy more than 20 years previously (7). A historical cohort study conducted in Sichuan Province, China, show- ed a lower overall mortality among vasectomized men, though cause of death could not be ascertain- ed (8). These observations are consistent with the results from other investigations that have shown no increase in morbidity, as reflected in hospital admission rates, among vasectomized men (9-12). Cardiovascular diseases Concerns about the safety of vasectomy with respect to cardiovascular diseases were first raised in the 1970s following studies on rhesus monkeys that sug- gested an increased risk of atherosclerosis (13), pos- sibly attributable to an elevated level of anti-sperm Prostate cancer Prostate cancer is a disease whose incidence rises steeply with age, and is much commoner among older men. Vasectomy, since it is a permanent contraceptive method, is usually performed on men aged 30-50 years, and became more widely used for family planning in the 1960s and 1970s. Thus it is only recently that the first groups of vasectomized men have reached an age at which there is an appre- ciable risk of prostate cancer; studies of any relation- ship between the two are, therefore, only now fea- sible. Before examining the possible association between vasectomy and prostate cancer, the world- wide and temporal patterns of prostate cancer inci- dence will be briefly reviewed. Incidence. It is important to distinguish between incidental, clinical, and latent prostate cancer. The latent cancer consists of localized, usually well-dif- ferentiated lesions, most often in the peripheral part of the prostate gland. These lesions do not manifest WHO Bulletin OMS. Vol 71 1993 1 414 The safety of vasectomy: recent concerns themselves clinically and are diagnosed at autopsy. Incidental cancers are subclinical lesions with varying potential for malignancy and are usually detected following prostate surgery, which is most frequently performed for the treatment of benign prostatic hyperplasia (BPH), a common condition among elderly men. The prevalence of latent prostate lesions increases with age, reaching almost 50% among men aged >80 years (22), and appears to occur with a similar frequency in different areas of the world (23, 24). Clinical prostate cancer, however, occurs only in a small proportion of men and at widely disparate frequencies in different populations. In developed countries clinical prostate cancer is one of the commonest cancers among men, while in developing countries it has a much lower incidence. Globally, there is a fiftyfold variation in the overall incidence of the disease. For example, in parts of the USA the (age-standardized) annual incidence is 91.2 per 100 000 men, while it is as low as 1.8 per 100 000 in Shanghai. The incidence among men in Bombay is 8.2 per 100 000 (25). Age-specific rates of prostate cancer obtained from cancer registries in various countries are shown in Fig. 1. In the USA the rates among black men are higher than those among white men (data for Detroit are used for purposes of illustration), with no apparent explanation for this dif- ference. In other developed countries, represented in Fig. 1 by Denmark and New Zealand, in both of which vasectomy is a widely used method of family planning, the incidences are lower than in the USA. In contrast, in Bombay and Shanghai, the only devel- oping country populations for which adequate data are available, substantially lower incidences are observed, and the disease is not common. However, the incidence in Bombay and Shanghai rises with age at a similar rate to that in countries with a higher over- all incidence of prostate cancer, though this is not apparent from the plots in Fig. 1 because of the scale. Many studies on the incidence of prostate cancer have not distinguished between the incidental and clinical types. The recent increases in the incidence of prostate cancer in many developed countries may in part be due to an increase in incidental prostate cancer following surgery for BPH. Improved surgical techniques have been introduced, resulting in earlier and more widespread interventions for BPH. Also, new diagnostic techniques, such as the use of prosta- tic-specific antigen levels, ultrasound examinations, and systematic screening for prostate cancer, have led to detection of more prostate cancers, particularly in countries where the incidence is high. In the USA, for example, the age-adjusted incidence of reported prostate cancer increased by 30.3% from 1973 to 1985, while over the same period the age-adjusted mortality increased by only 8.0% (23). Fig. 1. Age-specific incidence of prostate cancer In the 1980s In selected developed and developing countries (from ref. 25). 1500 r §1000 0 SSoo .8 500 .5 0 - Detroit, USA (black) Detroit, USA (white) New Zealand(non-Maori) Donmark India China 505-4 555-9 60-64 65-69 70-74 Age group (yrous) 75-79 Risk factors. Investigation of risk factors for prostate cancer, particularly environmental and behavioural factors, including vasectomy, is difficult because lit- tle is known about the underlying pathophysiological causes of the disease (26, 27). The prostate gland depends on androgens for its development and main- tenance of its function and morphology (28). Some observations implicate dietary factors in the etiology of clinical cancer. For example, increased rates of clinical prostate cancer among relatives of Japanese migrants to Hawaii (29) and the positive correlation with other tumours that have also shown an associa- tion with diet in intemational comparisons (30), sug- gest that diet, possibly its fat content, may play an important role in causing clinical prostate cancer (24, 31); however there is no consensus on this issue. Similarly, it has been suggested that variations in cir- culating hormone levels, in particular testosterone and its metabolites, in different ethnic groups account for differences in prostate cancer incidence (32), but no clear differences have been found in the hormonal levels of prostate cancer cases and con- trols within the same ethnic group (33). Presumably clinical prostate cancer develops from lesions similar to the latent cancers that are highly prevalent among older men. The occurrence of similar prevalences of latent prostate cancer, but wide variations in the incidences of the clinical can- cer among men from markedly different groups, sug- gests that the risk factors for latent and clinical pros- tate cancer, although pporly understood, may be different. Vasectomy and prostate cancer In 1990 two epidemiological studies in the USA sug- gested an increase in the risk of prostate cancer fol- WHO Bulletin OMS. Vol 71 1993 415 T.M.M. Farley et al. lowing vasectomy. The first of these studies was based on results of hospital-based, case-control sur- veillance (34) and the conclusions were considered by the authors to be hypothesis-generating, rather than definitive. The other, also a hospital-based case-control study (35), found a non-significant, moderate increased risk of prostate cancer .13 years after vasectomy had been performed. Prior to 1990, one cohort study (36) and two case-control studies (37, 38) reported no association between vasectomy and prostate cancer, while a further case-control study (39) reported elevated risks .20 years after the procedure. The two 1990 studies resulted in a WHO meeting in 1991 to review available biological and epidemiological evidence on the safety of vasectomy with regard to prostate cancer. At the meeting it was concluded that there was no known biological mechanism to account for any association, and that any causal relationship between vasectomy and can- cer of the prostate was unlikely (21). Epidemiologi- cal studies under way at that time were expected to provide further information on such relationships. Two additional studies on the risk of prostate cancer following vasectomy, conducted in the USA, have recently been published (40, 41). These studies used a prospective design in which groups of men were followed over time; one was a historical and the other a concurrent cohort study. Both studies have been carefully analysed and their prospective nature avoids many of the problems of bias in pre- vious case-control approaches. The results indicate that there is an increased risk of prostate cancer (about 1.6 times) following vasectomy, and a trend of increasing risk with time since the procedure, up to a relative risk of about 1.9 among men vasectom- ized at least 20 years previously. The two studies were conducted among men who were above average socioeconomic status and educational level - one group consisted of health professionals and the other of husbands of registered nurses. However, the results of these new studies are not supported by a re-analy- sis (42) of Sidney's cohort study (36), with more cases and person-years of observation, which con- firmed the earlier findings of the lack of association between vasectomy and subsequent prostate cancer. These latest studies, together with other avail- able information, were reviewed at a recent meeting convened by the National Institutes of Health in Bethesda, MD, on 1-2 March 1993, at which WHO was represented. The meeting endorsed the conclu- sions of the 1991 WHO meeting (21) that no known biological mechanism existed to explain any possible association between vasectomy and prostate cancer and that any causal relationship between the two was unlikely. The meeting also reviewed interim, unpub- lished results from other ongoing studies of vasecto- my and prostate cancer and noted that there was no consistency in the findings and that the reported associations are weak. It was therefore recommended that there was insufficient basis to change clinical and public health practices regarding vasectomy and the procedure should continue to be offered as part of family planning programmes (43). In countries where the incidence of prostate cancer is low (for example, China or India; see Fig. 1) any association of the magnitude observed in the USA, even if sub- stantiated, would be of little significance to indivi- dual or public health. Implications for family planning programmes Recent findings have raised questions about the long-term safety of vasectomy and are of potential public health significance in countries where both prostate cancer and vasectomy are common. In coun- tries where the disease is rare, the potential public health impact of any association is much smaller. However, in countries where vasectomy is a widely used method of contraception, the impact of the recent reports from the USA on family planning pro- grammes and on acceptance of vasectomy could be large. As we have discussed above, the risk factors for prostate cancer are so poorly understood and the incidences so varied between countries that it is not justified to extrapolate the results from the USA to other countries, particularly those with low inci- dences of prostate cancer. Consequently, there are no grounds, based on the currently available epidemio- logical data, for changing the recommendation that vasectomy should be offered as part of national family planning programmes. Health care providers should continue to perform the procedure. Couples seeking family planning should be informed of the benefits and risks of different contra- ceptive methods, and also of the risks and conse- quences of not using any family planning method at all. Furthermore, it is important that couples receive information that is relevant to the health situation in their own setting. It is therefore essential that studies on prostate cancer and vasectomy are carried out particularly in developing countries where vasecto- my is widely used and the availability of other effec- tive and affordable contraceptive methods is limited. Moreover, individuals who have volunteered for vasectomy have a right to be reassured about the long-term safety of their chosen contraceptive method. Such studies, supported by WHO, are under way-their feasibility has been assessed, and pilot studies will be conducted in 1993 in four developing countries. The main epidemiological study, using WHO Bulletin OMS. Vol 71 1993416 The safety of vasectomy: recent concerns case-control methodology, is expected to start in 1994. In the meantine, WHO will continue to moni- tor and disseminate any new information that emerges. Conclusions Vasectomy is one of the most widely studied surgi- cal procedures and there is extensive evidence about its short- and long-term safety. Earlier concems about possible adverse effects on the cardiovascular sys- tem, or an increase in autoimmune disease, have not been substantiated by careful research conducted in a wide range of countries. The recently report- ed results from the USA of a possible increased risk of prostate cancer many years after vasectomy has been performed have yet to be confirmed and there is no known biological mechanism or hypothesis to account for such an association. The relevance of these results to men in the USA who have undergone vasectomy has been carefully scrutinized and current recommendations are that vasectomies should con- tinue to be performed and no change in clinical prac- tice is warranted. There is no justification to extra- polate the U.S. results to other countries, particularly those where prostate cancer is not a common disease. Since these results may have an effect on family planning programmes and also raise concems among men who have been vasectomized, it is essent- ial that the long-term safety of the method be estab- lished in countries where vasectomy is a widely used method of contraception. Also, there are no grounds for suggesting any change in family planning policies with regard to vasectomy, and the method should continue to be provided as part of national family planning programmes. It is a simple, safe and highly effective contraceptive method. Resume S6curit6 de la vasectomie: nouvelles inqui6tudes Environ 42 millions de couples, dont les trois quarts vivent dans des pays en d6veloppement, ont recours a la vasectomie comme moyen de planification familiale. C'est une methode de contraception fiable et sOre, dont la securite a court et a long terme a ete largement etudiee. On a commence a s'inquieter des eventuels effets indesirables de la vasectomie dans les annees 70, des observations sur le singe rhesus ayant 6voque une association avec un risque accru d'atheroscl6rose. Toutefois, des recherches chez l'homme, r6alisees dans les pays developpes et dans les pays en developpement, n'ont mis en 6vidence aucune 6levation du risque de maladie cardio-vasculaire chez les hommes ayant subi une vasectomie. D'autre part, d'aprbs des etudes publiees entre 1988 et 1990, il pourrait y avoir un risque accru du cancer du testicule et de la pros- tate des annees apres que la vasectomie ait ete effectuee. Au vu de ces r6sultats, I'OMS a organi- se, en 1991, une r6union chargee d'examiner les preuves 6pidemiologiques et biologiques d'une telle association. II a ete conclu qu'il n'existait aucun mecanisme biologique connu susceptible d'expliquer une association entre la vasectomie et le cancer du testicule ou de la prostate, et que les travaux actuellement en cours pourraient apporter un complement d'information. En ce qui concerne I'association avec le cancer du testicule, elle n'a ete confirmee par aucune 6tude publiee depuis la reunion. Deux etudes, publiees en fWvrier 1993, ont de nouveau souleve le probleme d'une 6ventuelle association entre la vasectomie et la survenue d'un cancer de la prostate des annees apres l'intervention. Ces deux 6tudes etaient prospec- tives, faisaient appel a des cohortes, et montraient un risque accru d'environ 1,6 fois pour 1'ensemble des hommes vasectomises, et atteignant 1,9 fois chez les hommes ayant subi une vasectomie plus de vingt ans auparavant. Ces observations s'appuient sur des recherches effectuees aux Etats-Unis d'Amerique, ou l'incidence du cancer de la prostate est elevee et en augmentation. Leurs resultats s'cartent de ceux des prec6- dentes 6tudes cas-t6moins et de cohorte realis6es aux Etats-Unis d'Am6rique et en Europe qui, a une exception pres, n'ont montre aucune associa- tion syst6matique entre la vasectomie et le risque de cancer de la prostate. Les facteurs qui influent sur I'apparition et le d6veloppement du cancer de la prostate sont mal connus et compliquent les recherches portant sur les facteurs de risque, y compris les risques environnementaux et compor- tementaux, dont la vasectomie. L'incidence globa- le du cancer de la prostate, dans certains pays d6veloppes comme les Etats-Unis, est cinquante fois superieure a celle que l'on observe dans cer- tains pays en d6veloppement, par exemple la Chine. Dans ces derniers pays, ou l'incidence de ce cancer est faible, l'impact de ces associations sur la sant6 individuelle et publique est faible; de plus, il reste a savoir si les facteurs de risque observes aux Etats-Unis sont egalement pr6sents dans ces pays. Compte tenu des donnees epidemiologiques disponibles, il n'y a pas lieu de changer la recom- mandation selon laquelle la vasectomie doit etre WHO Bulletin OMS. Vol 71 1993 417 T.M.M. Farley et al. proposee dans le cadre des programmes natio- naux de planification familiale. Les agents de soins de sante devront continuer a pratiquer ces interventions. Les couples consultant pour la pla- nification familiale devront etre informes des avan- tages et des risques des differentes methodes contraceptives, et egalement des risques et des cons6quences de l'abstention de toute methode de planification familiale. De plus, il importe que les couples regoivent une information adaptee a la situation sanitaire de leur propre environnement. II est par consequent necessaire d'entreprendre des etudes sur le rapport entre vasectomie et cancer de la prostate, en particulier dans les pays en d6veloppement ou la vasectomie est largement utilis6e et ou le nombre de m6thodes contracep- tives efficaces et abordables est limite. De plus, les sujets qui se sont portes volontaires pour une vasectomie sont en droit d'etre inform6s quant a la securit6 a long terme de cette methode contra- ceptive. De telles 6tudes, soutenues par l'OMS, sont en cours; leur faisabilite a et evaluee, et des 6tudes pilotes seront effectuees en 1993 dans quatre pays en d6veloppement. La principa- le 6tude epidemiologique, une etude cas-temoins, devrait d6marrer en 1994. Dans l'intervalle, I'OMS continuera a surveiller et diffuser toutes nouvelles informations paraissant sur ce sujet. References 1. Hargreave, T.B. Towards reversible vasectomy. Inter- national journal of andrology, 15: 455-459 (1992). 2. Nirapathpongporn, A. et al. No-scalpel vasectomy at the King's birthday vasectomy festival. Lancet, 335: 894-895 (1990). 3. Liskin, L. et al. Vasectomy - safe and simple. Bal- timore, MD, Johns Hopkins University, Population Information Program, November-December 1983 (Population Reports, Series D, No. 4). 4. Liskin, L. et al. Vasectomy: new opportunities. Bal- timore, MD, Johns Hopkins University, Population Information Program, March 1992 (Population Reports, Series D, No. 5). 5. Petitti, D.B. Epidemiologic studies of vasectomy. In: Zatuchni, G.L. et al., ed. Male contraception: advances and future prospects. New York, Harper & Row, 1986, pp. 24-33. 6. Massey, E.J. et al. Vasectomy and health: results from a large cohort study. Journal of the American Medical Association, 252: 1023-1029 (1984). 7. Giovanucci, E. et al. A long-term study of mortality in men who have undergone vasectomy. New England journal of medicine, 326: 1392-1398 (1992). 8. Tang Guang-hua et al. Vasectomy and health: car- diovascular and other diseases following vasectomy in Sichuan Province, People's Republic of China. International journal of epidemiology, 17: 608-617 (1988). 9. Goldacre, M.J. et al. Follow-up of vasectomy using medical record linkage. American journal of epide- miology, 108: 176-180 (1978). 10. Walker, A.M. et al. Hospitalization rates in vasecto- mized men. Journal of the American Medical Asso- ciation, 245: 2315-2317 (1981). 11. Petitti, D.B. et al. Vasectomy and the incidence of hospitalized illness. Journal of urology, 129: 760-762 (1983). 12. Nienhuis, H. et al. Incidence of disease after vasectomy: a record linkage retrospective cohort study. British medical journal, 304: 743-746 (1992). 13. Alexander, N.J. & Clarkson, T.B. Vasectomy increases the severity of diet-induced atherosclero- sis in Macaca fascicularis. Science, 201: 538-541 (1978). 14. WHO Special Programme of Research, Develop- ment and Research Training in Human Reproduc- tion. Sequelae of vasectomy - report of a meeting. International journal of andrology, 5: 1-5 (1982). 15. Chi, I.C. et al. Vasectomy and non-fatal acute myo- cardial infarction: a hospital based case-control study in Seoul, Korea. International journal of epi- demiology, 19: 32-41 (1990). 16. Chi, I.C. et al. Vasectomy and cardiovascular deaths in Korean men: a community-based case-control study. International journal of epi- demiology, 19: 1113-1115 (1990). 17. Indian Council for Medical Research. Long-term effects of vasectomy, part I: biochemical param- eters. An ICMR Task Force study on regulation of male fertility (surgical approaches). Contraception, 28: 423-435 (1983). 18. Clarkson, T.B. et al. Atherosclerosis of cynomolgus monkeys hyper- and hypo-responsive to dietary cholesterol: lack of effect of vasectomy. Arterioscle- rosis, 8: 488-498 (1988). 19. Strader, C.H. et al. Vasectomy and the incidence of testicular cancer. American journal of epidemiology, 128: 56-63 (1988). 20. Cale, A.R.J. et al. Does vasectomy accelerate tes- ticular tumour? Importance of testicular examina- tions before and after vasectomy. British medicaljournal, 300: 370 (1990). 21. Vasectomy and cancer. Lancet, 338: 1586 (1991). 22. Yatani, R. et al. Geographic pathology of latent prostatic carcinoma. International journal of cancer, 29: 611-616 (1982). 23. Carter, H.B. & Coffey, D.S. The prostate: an increas- ing medical problem. Prostate, 16: 39-48 (1990). 24. Nomura, A.M.Y. & Kolonel, L.N. Prostate cancer: a current perspective. Epidemiological review, 13: 200-227 (1991). 25. Muir, C. et al. Cancer incidence in five continents, vol. 5. Lyon, International Agency for Research on Cancer, 1987 (IARC Scientific Publications No. 88). 26. Howards, S.S. & Peterson, H.B. Vasectomy and prostate cancer: chance, bias, or a causal relation- ship? Journal of the American Medical Association, 269: 913-914 (1993). 27. Howards, S.S. Possible biological mechanisms for a relationship between vasectomy and prostatic cancer. European journal of cancer, (in press). 28. National Cancer Institute Prostate Cancer Work- 418 WHO Bulletin OMS. Vol 71 1993 The safety of vasectomy: recent concerns ing Group. Prostate cancer: future research direc- tions. Cancer research, 51: 2498-2505 (1991). 29. Kolonel, L.N. et al. Nutrient intakes in relation to cancer incidence in Hawaii. British journal of cancer, 44: 332-339 (1981). 30. Armstrong, B. & Doll, R. Environmental factors and cancer incidence and mortality in different coun- tries, with special reference to dietary practices. International journal of cancer, 15: 617-631 (1975). 31. Berg, J.W. Can nutrition explain the pattern of inter- national epidemiology of hormone-dependent can- cers? Cancer research, 35: 3345-3350 (1975). 32. Gittes, R.F. Carcinoma of the prostate. New England journal of medicine, 324: 236-245 (1991). 33. de Jong, F.H. et al. Peripheral hormone levels in controls and patients with prostatic cancer or benign prostatic hyperplasia: results from the Dutch- Japanese case-control study. Cancer research, 51: 3445-3450 (1991). 34. Rosenberg, L. et al. Vasectomy and the risk of prostate cancer. American journal of epidemiology, 132: 1051-1055 (1990). 35. Mettlin, C. et al. Vasectomy and prostate cancer risk. American journal of epidemiology, 132: 1056- 1061 (1990). 36. Sidney, S. Vasectomy and the risk of prostatic can- cer and benign prostatic hypertrophy. Journal of urology, 138: 795-797 (1987). 37. Ross, R.K. et al. The etiology of prostate cancer. Prostate, 4: 333-344 (1983). 38. Newell, G.R. et al. A case-control study of prostate cancer. American journal of epidemiology, 130: 395-398 (1989). 39. Honda, G.D. et al. Vasectomy, cigarette smoking, and age at first sexual intercourse as risk factors for prostate cancer in middle-aged men. British journal of cancer, 57: 326-331 (1988). 40. Giovannucci, E. et al. A prospective cohort study of vasectomy and prostate cancer in U.S. men. Journal of the American Medical Association, 269: 873-877 (1993). 41. Giovannucci, E. et al. A retrospective cohort study of vasectomy and prostate cancer in U.S. men. Journal of the American Medical Association, 269: 878-882 (1993). 42. Sidney, S. et al. Vasectomy and the risk of prostate cancer in a cohort of multiphasic health-checkup examinees: second report. Cancer causes control, 2: 113-116 (1991). 43. Healy, B. News from NIH: Does vasectomy cause prostate cancer? Journal of the American Medical Association, 269: 2620 (1993). WHO Bulletin OMS. Vol 71 1993 419