Risk of cancer in patients with polycystic kidney disease: a propensity-score matched analysis of a nationwide, population-based cohort study.

BACKGROUND: Data for the risk of any solid cancer in patients with polycystic kidney disease are scarce. Therefore, we did a nationwide cohort study in Taiwan to establish the risk of cancer in patients with polycystic kidney disease without either chronic kidney disease or end-stage renal disease. METHODS: From inpatient claims of the Taiwan National Health Insurance Research Database, we included patients aged 20 years and older and diagnosed with polycystic kidney disease between January, 1998 and December, 2010, in the polycystic kidney disease cohort. Patients with a history of cancer, a history of chronic kidney disease or of end-stage renal disease (recorded from the Registry of Catastrophic Illness Patient Database) were excluded. For each patient with polycystic kidney disease, one patient aged older than 20 years with no history of polycystic kidney disease or cancer was randomly selected from the National Health Insurance Research Database, matched 1:1 on the basis of the propensity score calculated by logistic regression, and was included in the control non-polycystic kidney disease cohort. The follow-up period for each patient was estimated from the index date to the date of diagnosis of cancer, or the patient was censored due to withdrawal from the insurance programme (eg, death, immigration, or imprisonment) or on Dec 31, 2011. The primary outcome of interest was a diagnosis of cancer during a 14-year follow-up period. The risk of cancer was represented as a hazard ratio (HR) calculated in Cox proportional hazard regression models. FINDINGS: 4346 patients with polycystic kidney disease and 4346 without were enrolled in the study. The median follow-up period in the polycystic kidney disease cohort was 3·72 years (IQR 1·25-7·31) and in the non-polycystic kidney disease cohort was 4·96 years (2·29-8·38). The overall incidence of cancer was higher in the polycystic kidney disease cohort than in the control cohort (20·1 [95% CI 18·3-21·9] per 1000 person-years vs 10·9 [10·1-11·8] per 1000 person-years; crude hazard ratio (HR) 1·77 [95% CI 1·52-2·07]; HR adjusted for age, sex, frequency of medical visits, and comorbidities was 1·83 [1·57-2·15]). The specific risks (adjusted subhazard ratios) were significantly higher in the polycystic kidney disease cohort than that in the non-polycystic kidney disease cohort for liver cancer (1·49 [95% CI 1·04-2·13]; p=0·030), colon cancer (1·63 [1·15-2·30]; p=0·006), and kidney cancer (2·45 [1·29-4·65]; p=0·006). INTERPRETATION: To our knowledge, this is the first report of the association of polycystic kidney disease without end-stage renal disease with the risk of liver, colon, and kidney cancer. Health-care professionals should be aware of this risk, when treating patients with polycystic kidney disease. FUNDING: Taiwan Ministry of Health and Welfare Clinical Trial and Research Center of Excellence, Academia Sinica Taiwan Biobank, Stroke Biosignature Project, NRPB Stroke Clinical Trial Consortium, Tseng-Lien Lin Foundation, Taiwan Brain Disease Foundation, Katsuzo and Kiyo Aoshima Memorial Funds, China Medical University Hospital, and Taiwan Ministry of Education.