[CANCER RESEARCH 51. 629-631. January 15.

1991]

Genital and Urinary Tract Diseases and Bladder Cancer1

Carlo La Vecchia,2 Eva Negri, Barbara D'Avanzo, Renzo Savoldelli, and Silvia Franceschi
Istituto di Ricerche Farmacologiche "Mario Negri, " 20157 Milan, Italy C. L. V., E. N., B. I)., R. S.]; Institute of Social and Preventive Medicine, University of Lausanne, 1005 Lausanne, Switzerland fC. L. I './; and .Jviano Cancer Centre, 33081 Aviano, Pordenone, Italy [S. F.]

ABSTRACT
The relationship between selected urinary tract and genital diseases and the risk of bladder cancer as analyzed using data from a casecontrol study of 364 cases of bladder cancer and 447 controls hospitalized for acute, nonneoplastic, nongenital tract conditions, unrelated to known or suspected risk factors for bladder cancer. Cystitis was reported by 20% of the cases and 8% of the controls, corresponding to a multivariate relative risk (RR) of 3.8 (95% confidence interval, 2.4 to 5.9). No association was observed with urinary tract stones (RR = 1.2). With reference to genital diseases, the RR was elevated for gonorrhea (RR = 2.8,95% confidence interval, 1.0 to 4.5) and condylomata acuminata (RR = 5.9, 95% confidence interval 1.0 to 3.6) but not for syphilis. The risk increased with the number of episodes of cystitis (RR = 5.0 for >4 episodes, x2 for trend = 33.04, P < 0.001), was higher during the last 15 years after the first episode (RR = 5.1 versus 2.3 for over 15 years), and was not heterogeneous across strata of age and sex. The interaction between urinary tract infections and tobacco appeared multiplicative, with RR = 2.4 for ever smoking, 3.2 for cystitis alone, and 10.3 for both exposures. The present study, besides providing further quantitative evidence of a relationship between urinary tract infections (and, possibly, some genital infections, too) and bladder cancer, indicates that the role of infections is probably in one of the latter (promoting) stages of the process of carcinogenesis and suggests a multiplicative interaction with smoking. In terms of prevention and public health, therefore, it is thus important to avoid at least one exposure for subjects with a history of urinary tract infections who smoke tobacco.

of bladder cancer are similar to those of cystitis, and subjects with urinary tract conditions probably tend systematically to recall episodes of cystitis or other urinary tract conditions more accurately. Not surprisingly, therefore, published evidence is largely inconsistent, with relative risks ranging from 1 (9) to 5 (10) for infections and from 1 (9) to 3 (11) for urinary tract stones. Thus, to provide further quantitative information on the issue, we have analyzed data on urinary tract and genital dis eases from a large case-control study conducted in the Greater Milan area. Northern Italy. SUBJECTS AND METHODS
The data were derived from a study of urinary tract neoplasms that has been ongoing since January 1985 in the Greater Milan area. Northern Italy, and the present report is based on data collected before June 1990. The general scheme and method of this investigation has alreadybeen described (6, 12). Briefly, trained interviewers idem liedand questioned patients admitted to a network of hospitals under surveil lance for urological cancers and a wide spectrum of other conditions. All interviews were conducted in the hospital. Less than 3% of eligible subjects (cases and controls) refused to be interviewed. The patients included in the present analysis were 364 histologically confirmed cases of invasive transitional cell bladder cancer (303 males, 61 females; median age, 63 years; range, 28-74) after specific exclusion of papillomas. as well as of squamous cell cancers (which represented less than 2% of bladder cancer cases in this population). The controls were 447 patients (336 males, 111 females; median age, 62 years; range, 25-74), admitted to the same network of hospitals for a wide spectrum of acute, other than urological or genital tract diseases (29% traumas, 17% nontraumatic orthopedic conditions, 13% medical disorders, 7% surgical conditions, 34% other miscellaneous illnesses). The catchment area of cases and controls was comparable: 76% of cases and 78% of the controls came from the same region, Lombardy; 85% of the cases and 87% of the controls were from Northern Italy. The structured questionnaire included information on sociodemographic factors, personal characteristics and habits (including smoking and coffee consumption), selected lists of indicator foods and relevant occupational exposures, and a problem-oriented medical history. Infor mation was specifically elicited on age at first occurrence of five urogenital conditions (cystitis, urinary tract stones, gonorrhea, syphilis, and condylomata acuminata), and on the number of episodes of cystitis. Statistical analysis was based on standard methods for case-control studies, including sex- and age-adjusted RR1 and estimates from mul tiple logistic regression models (13). Included in the regression equa tions were terms for age, sex, education, area of residence, smoking, and exposure to high risk occupations (dyestuff, chemicals, painting, Pharmaceuticals, and coal/tar).

INTRODUCTION Cigarette smoking and occupational exposure to aromatic amines and other chemicals are the best recognized risk factors for bladder cancer in developed countries (1). However, these factors alone cannot totally account for the geographical distri bution of incidence and mortality from the disease, both on an international and national scale (2, 3). Within Italy, for in stance, this is the single tobacco-related site showing only moderate differences between the North (where tobacco spread and the industrialization process occurred earlier) and the South of the country (4, 5). Diet plays probably some role (1, 6), but further potential factors in bladder cancer epidemiology are infections and other diseases of the urinary tract, which may cause chronic irritation and hence favor the action of specific carcinogens. A clear relationship has been established in Egypt and other African regions between infection with Schistosoma haematobium and bladder cancer risk (7, 8), but the role of other infectious agents or aspecific cystitis on bladder carcinogenesis is more difficult to study epidemiologically, particularly because early symptoms
Received 7/27/90; accepted 10/29/90. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1This work was conducted within the framework of the National Research Council (CNR). Applied Projects "Oncology" (Contract 87.01544.44). and "Risk Factors for Disease," and with the contribution of the Italian Association for Cancer Research and the Italian League against Tumours. Milan. 2To whom requests for reprints should be addressed, at Istituto di Ricerche Farmacologiche "Mario Negri." Via Eritrea. 62. 20157 Milan. Italy.

RESULTS Cases were somewhat older than the comparison group, less educated (RR for <7 versus >7 years of education = 1.4; 95% CI 1.0 to 2.1), and more frequently smokers (RR = 3.3 for current and 3.9 for heavy smokers (Table 1).
3The abbreviations used are: RR. relative risk; CI. confidence interval.

629

GENITOURINARY

DISEASES AND BLADDER CANCER

Table

Distribution of 364 cases of bladder cancer and 447 controls according to sex, age, education and smoking: Milan, Italy, 1985-1990 FemalesControlsNf\

Table 3 Relative risk of bladder cancer according to number of episodes of cystitis and time since first episode: Milan, Italy, 1985-1990 estimates\yj risk
\*A)_,\_ /o
PIVll^K.No.

7INO. /o21

/a53

7O.

ontrolsMW-nMI

Age (yr)<5050-5960-69>70Education group 6.973 24.1151 49.858 19.1177 (yr)<77-1112Smoking 58.477 25.449 16.242

cystitis0123>4UnknownXi2 of episodes of

15.8107 31.8146 43.530 8.9176 52.387 25.971 21.196

habitsNever smokersExsmokersSmokers,<15 13.995 28.6107 31.442 31.840 (yr)<55-14not

(trend)28511694764078551751'2.4(0.9-6.3)2.0(0.5-7.0)2.9(0.9-9.4)4.7(2.5-8.7)31.62(P< 0.001)Time 0.001)1*2.4(0.9-6.6)2.1(0.6-7.8)3.2(1.0-10.9)5 since first episode

cigarettes/day1 13.988 11.962 cigarettes/day^25 5-24 29.036 18.531 9.2CasesMrJNO.4172713342433928111_6.627.944.321.355.739.34.963.93.313.118.01.6ControlsMrINO.8275323791814866883m/O7.22 cigarettes/dayfl 11.9les The sum of strata doesMeCasesMr 15Table add up to the total because of missing values. 2 Distribution of 364 cases of bladder cancer and 447 controls accordingUnknown47111561431585.1(2.7-9.4)5.5(1.5-20.2)1.9(0.9-4.0)5.1(2.6-9.8)5.4(1.4-20.2)2.3(1.0-5.2) to history of selected urinary diseases: Milan, Italy, 1985-1990 M-H, Mantel-Haenszel estimates adjusted for age and sex; MLR estimates No. (%) of subjects with the disease diseaseCystitisUrinary Type of cancer73(20.1)17(4.7)21 (7.8)20 stonesGonorrheaSyphilisCondylomata tract (5.8)3 (0.4)2 (0.8)9(2.5)Controls35 acuminataBladder (4.5)12(2.7)2 Table 4 Relationship of bladder cancer to history of cystitis in separate strata of age and sex: Milan, Italy, 1985-1990 No. (%) of subjects with history of cystitis Relative risk estimates (95% CI) from multiple logistic equations including terms for age, sex, education, area of residence, smoking, and occupation (chemical industry, dyestuff, painting, pharmaceuticals, and coal/tar). * Reference category.

Relative risk estimates Cl)4.1(2.3-7.1)2.4(1.2-5.0)3.2(1.7-6 cancer54(17.8)19(31.1)28 (0.4)M-H"3.4(2.2-5.3)1.0(0.5-2.0)2.0(1.0-4.1)2.9(0.3-12.4)5.8(1.4-24.6)MLR3.8(2.4-5.9)1.2(0.6-2.6)2.1(1.0-4.5)1.1(0.2-7.0)5.9(1.1-30.6) SexMalesFemalesAge

M-H, Mantel-Haenszel estimates adjusted for age and sex; MLR, estimates from multiple logistic regression equations including terms for age, sex, education, area of residence, smoking, and occupation (chemical industry, dyestuff, painting. Pharmaceuticals, and coal/tar).

History of selected urinary and genital diseases is considered (24.3)45 in Table 2. Cystitis was reported by 20% of the cases and 8% (18.1)Controls17(5.1)18(16.2)19(9.7)16(6.3)(95% yrBladder of the controls corresponding to a multivariate relative risk of 3.8 (95% CI, 2.4-5.9). No association was observed with kidney " Mantel-Haenszel estimates adjusted for age and sex (when appropriate). or bladder stones (RR = 1.2, 95% CI 0.6-2.6). With reference to genital diseases, the relative risk was elevated for gonorrhea Table 5 Interaction between history of cystitis and cigarette smoking and the risk (RR = 2.1, 95% CI 1.0-4.5), and for condylomata acuminata of bladder cancer: Milan, Italy, 1985-1990 (reported by 9 cases and 2 controls, for a point estimate of 5.9, Relative risk (95% CI) for his tory of cystitis" of borderline significance), but not for syphilis (reported by 3 cases and 2 controls only). None of these RR estimates was materially modified by allowance for major identified cosmokersEver Never variates, using multiple logistic regression. (1.6-6.3) 59:162C2.4 22:2010.3 The relationship with cystitis was further examined in Table 3 in terms of number of episodes and time since first episode. smokersNo1" The RR was approximately 2-fold elevated for one or two (5.3-20.1) (1.6-3.6)232:250Yes3.2 51:15 episodes and rose to 3.2 for three and to 5.1 for four or more ' Mantel-Haenszel estimates adjusted for age and sex. episodes. About two-thirds (46 of 73) of cases with history of ^ Reference category. cystitis reported four or more episodes. With reference to time r Number of cases:number of controls. since first episode, the RR was over 5-fold elevated up to 15 years but declined to 2.3 at 15 years or more after the first presented in Table 5. Compared with individuals who reported episode. not to have ever smoked or had cystitis, the RR was 2.4 for Separate strata of sex and age group are considered in Table 4. The risk estimates were somewhat higher in males and in ever smoking and 3.2 for individuals with history of cystitis older age groups but not heterogeneous across various sex and alone and rose to 10.3 for exposure to both risk factors. Thus, the interaction between these two risk factors seemed to be age groups. Likewise, no significant heterogeneity across strata multiplicative (or overmultiplicative, but not additive) on the of sex and age was observed for gonorrhea or genital warts. The interaction between history of cystitis and smoking is relative risk. 630

group<60yrS60

GENITOURINARY

DISEASES AND BLADDER CANCER

DISCUSSION The present study provides further quantitative evidence of a relationship between urinary tract infections (and, possibly, some genital infections as well) and bladder cancer. This is therefore consistent with most (10, 11, 14-16) [though not all (9)] previous investigations, in terms of both existence and size of the association, since a relative risk of 3-4 appears to be a reasonable estimate of the bladder cancer/urinary infection association from published studies [Refs. 10, 11, 14-16, and, for a summary tabulation, Ref. 9]. In terms of population attributable risk (17), urinary tract infections could account for about 10% of the cases of bladder cancer in this population. A further, more innovative indication that emerged is the associ ation with gonorrhea and condylomata acuminata, which may suggest similar etiopathogenic mechanism(s) for genital and lower urinary tract neoplasms. A major problem of the present and most previous casecontrol investigations is recall bias, since cases of bladder cancer are probably more sensitized towards recalling urinary tract conditions that patients admitted to hospital for other diseases (and population-based controls as well). Thus, although the interviewers were specifically trained to avoid or reduce this potential problem, it is likely that the relative risks are some what overestimated, although to an extent difficult to quantify. However large this bias may be, it is still unlikely by itself to have produced relative risks of the order of those observed in this study, and such strong direct trends in risk. Selection (diagnostic) bias is also possible, since patients with cystitis are probably cystoscoped more than patients without it. However, only cases of histologically confirmed invasive blad der cancer were included, and this should have reduced the scope for bias. Other possible sources of bias are probably less important. Although this is a typical hospital-based case-control study and, as such, has all the widely discussed relative possible weakness (18), it also has the strengths deriving from almost complete participation, a comparable catchment area of cases and con trols, as well as the absence of important confounding by a number of potentially relevant covariates (including socioeconomic indicators, smoking and occupation) and the advantage of a similar recall of medical histories by cases and hospital controls (19). The observation that the relative risk for cystitis tends to decrease with time since first episode indicates that urinary tract infections have a late-stage effect [in terms of the multi stage theory of carcinogenesis (20, 21)] on bladder cancer risk, which is consistent with the promotional effect of wound heal ing and irritants in animal experiments of skin carcinogenesis (22, 23). Within the multistage theory of carcinogenesis, it is also of interest to note that the interaction between tobacco and urinary tract infections was multiplicative, implying that these two factors act with different mechanisms and on different stages of the process (24). A similar analysis of the tobacco/occupation interaction based on a subset of the same study (25) indicated, in contrast, an additive effect between tobacco and occupation, suggesting that the same bladder carcinogens are probably shared by these two exposures. A large, multicentric study from the United States (15), however, found that the interaction between cigarette smoking and urinary tract infections was additive (or slightly beyond additive) rather than multiplicative. Although the issue is still

controversial, nonetheless, examination of the tobacco/urinary tract infection interaction and indicators from theories of car cinogenesis have important preventive and public health impli cations, since the grossly elevated relative risk found in this study among subjects with a history of cystitis and smoking indicates and underlines the importance for intervention on at least one factor in subjects exposed to both. ACKNOWLEDGMENTS
The authors wish to thank J. Baggott. M. P. Bonifacino. and the G. A. Pfeiffer Memorial Library for editorial assistance.

REFERENCES
1. Matanoski, G. M.. and Elliott. E. A. Bladder cancer epidemiology. Epidemiol. Rev., J: 203-229. 1981. 2. Muir, C.. Waterhouse, J., Mack, T.. Powell, J., and Whelan. S. (eds.). Cancer Incidence in Five Continents, Vol. 5, No. 88. Lyon, France: IARC Scientific Publications. 1987. 3. Levi, F., Maisonneuve, P.. Filiberti. R., La Vecchia, C., and Boyle, P. Cancer incidence and mortality in Europe. Soz. Praventivmed.. 34 (Suppl. 2): SlS84, 1989. 4. Cislaghi, C., Decarli, A., La Vecchia, C., Laverda, N., Mezzanotte, G., and Smans, M. Dati, Indicatori e Mappe di Mortalit Tumorale. Data, Statistics and Maps on Cancer Mortality. Italia 1975-1977. Bologna, Italy: Pitagora Editrice, 1986. 5. Facchini, U., Camnasio, M., Cantaboni, A., Decarli, A., and La Vecchia, C. Geographical variation of cancer mortality in Italy. Int. J. Epidemiol.. 14: 538-548, 1985. 6. La Vecchia, C., Negri, E., Decarli, A., D'Avanzo, B.. Liberati, C., and Franceschi, S. Dietary factors in the risk of bladder cancer. Nutr. Cancer, 12: 93-101, 1989. Mustacchi, P.. and Shimkin. M. B. Cancer of the bladder and infestation with Schistosoma hemalobium. J. Nati. Cancer Inst.. 20: 825-842, 1958. Gelfand, M., Weinberg, R. W., and Castle, W. M. Relation between carci noma of the bladder and infestation with Schistosoma hemalobium. Lancet, /: 1249-1251. 1967. Kjaer, S. K., Knudsen, J. B., Sorensen, B. L., and Jensen, O. M. The Copenhagen case-control study of bladder cancer. V. Review of the role of urinary-tract infection. Acta Oncol., 28: 631-636. 1989. Wyndcr, E. L.. Onderdonk. J., and Mantel. N. An epidemiological investi gation of cancer of the bladder. Cancer (Phila.), 16: 1388-1407, 1963. Dunham. L. J., Rabson, A. S.. Stewart. H. L., Frank. A. S.. and Young, J. L. Rates, interview, and pathology study of cancer of the urinary bladder in New Orleans, Louisiana. J. Nati. Cancer Inst., 41: 683-709. 1968. La Vecchia. C., Negri, E.. D'Avanzo, B.. and Franceschi, S. Occupation and the risk of bladder cancer. Int. J. Epidemiol., 19: 264-268, 1990. Breslow. N. E., and Day. N. E. Statistical Methods in Cancer Research, Vol. 1. p. 32. Lyon. France: IARC Scientific Publications. 1980. Howe. G. R., Burch. J. D.. Miller, A. B., Cook, G. M., Estve,J., Morrison. B., Gordon, P., Chambers, L. W., Fodor. G., and Winsor, G. M. Tobacco use, occupation, coffee, various nutrients and bladder cancer. J. Nati. Cancer Inst., 64: 701-713, 1980. Kantor. A. F.. Hartge. P.. Hoover. R. N.. Narayana. A. S.. Sullivan, J. W., and Fraumeni. J. F., Jr. Urinary tract infection and risk of bladder cancer. Am. J. Epidemiol., 119.-510-515. 1984. Piper. J. M., Matanoski, G. M., and Tonascia. J. Bladder cancer in young women. Am. J. Epidemiol.. 123: 1033-1042. 1986. Bruzzi. P., Green, S. B., Byar, D. P., Brinton, L. A., and Schairer. C. Estimating the population attributable risk for multiple risk factors using case-control data. Am. J. Epidemiol.. 122: 904-914, 1985. Mantel, N.. and Haenszel, W. Statistical aspects of the analysis of data from retrospective studies of disease. J. Nati. Cancer Inst., 22: 719-748. 1959. Kelly. J. P., Rosenberg. L.. Kaufman. D. W.. and Shapiro. S. Reliability of personal interview data in a hospital-based case-control study. Am. J. Epi demiol., 131:79-90. 1990. Armitage, P., and Doll, R. A two-stage theory of carcinogenesis ion relation to the age distribution of human cancer. Br. J. Cancer. //: 161-169. 1961. Day. N. E., and Brown. C. C. Multistage models and primary prevention of cancer. J. Nati. Cancer Inst.. 64: 977-989. 1980. Friedewald, W. F.. and Rous. P. The initiating and promoting elements in tumor production. J. Exp. Med.. SO: 101-126, 1944. Shubik. P. Progression and promotion. J. Nati. Cancer Inst., 73:1005-1011. 1984. Peto. R. Epidemiology, multistage models, and short-term mutagenesis tests. In: H. H. Hiatt. J. D. Watson, and J. A. Winsten (eds.). Origins of Human Cancer, pp. 1403-1428. Cold Spring Harber, NY: Cold Spring Harbor Laboratories, 1977. D'Avanzo, B., Negri, E., La Vecchia. C.. Gramenzi, A., Bianchi, C., Fran ceschi, S., and Boyle, P. Cigarette smoking and bladder cancer. Eur. J. Cancer. 26: 714-718. 1990.

7. 8. 9. 10. 11. 12. 13. 14.

15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

25.