BACKGROUND: Epidemiological evidence suggests lack of neonatal circumcision as the strongest risk factor for penile cancer, but the role of sexually transmitted diseases in the etiology of penile cancer has remained unclear. PURPOSE: To further clarify risk factors for penile cancer, we examined the role of circumcision, personal characteristics and habits (such as smoking), sexually transmitted diseases, past sexual activity, and medical conditions of the penis. METHODS: A population-based, case-control study was conducted in western Washington state and in the province of British Columbia. We interviewed 110 men with penile cancer diagnosed from January 1979 to July 1990 and 355 control subjects from the general population, frequency matched to case subjects on age and date of diagnosis. Tumor tissue from 67 case subjects was tested for human papillomavirus (HPV) DNA by polymerase chain reaction. Results of blood tests from 69 case subjects and 208 control subjects were available for study. STATISTICALLY SIGNIFICANT RESULTS: Relative to men circumcised at birth, the risk for penile cancer was 3.2 times greater among men who were never circumcised and 3.0 times greater among men who were circumcised after the neonatal period. For current smokers, the risk was 2.8 times that of men who never smoked. The risk among men reporting a history of genital warts was 5.9 times that of men reporting no such history. Of 67 tumors tested for HPV DNA, 49% were positive; the majority of these positive tumors (70%) were type 16, which has been associated with anogenital carcinoma. Relative risks (RRs) associated with a reported history of penile rash or penile tear were 9.4 and 3.9, respectively. Among men not circumcised at birth, RRs associated with presence of smegma and difficulty in retracting the foreskin were 2.1 and 3.5, respectively. Twenty-eight percent of case subjects, compared with only 10% of control subjects, reported 30 or more sexual partners, and men with HPV-positive tumors were more likely to report a greater number of sexual partners. CONCLUSIONS: These results suggest that the absence of neonatal circumcision and potential resulting complications are associated with penile cancer. Additionally, medical conditions of the penis, sexual activity, infection with HPV, and smoking may increase the risk for penile cancer. IMPLICATIONS: A larger study would allow examination of interrelationships of circumcision, infection with HPV, and smoking as risk factors.