22 First, we excluded patients age <30 and >100 years old. To enroll patients with type 2 diabetes, click here we further excluded those who (1) had a hospital admission

with a discharge diagnosis of insulin dependent diabetes mellitus (ICD-9-CM code 250.x1, 250.x3), or (2) received a catastrophic illness certificate issued by the Department of Health for type 1 diabetes (Fig. 1). Patients were classified as having prevalent or newly diagnosed type 2 diabetes according to the criteria in 1999. Those who had a history of cancer recorded in the National Cancer Registry any time before the cohort entry date, that is, date of diabetes diagnosis for newly diagnosed patients and January 1 2000 for prevalent patients, were also excluded. Patients were followed from January 1 2000 (for prevalent type 2 diabetes patients) or the date of diabetes diagnosis in 2000 (for newly diagnosed type 2 diabetes patients) to the earliest of cancer diagnosis, death, disenrollment from the national health insurance, selleck or December 31 2007. All individuals in the study cohort with the first occurrence of liver, colorectal, lung, and urinary bladder cancer were included as cases. All potential cases were validated by a linkage

through National Cancer Registry. A risk-set sampling (that is, controls sampled from those in the original study cohort who remained free of outcome at the time point when a case occurred) matched by age (within 5 years), sex, and the number of days of follow-up was used to find controls for the

cohort. For newly diagnosed type 2 diabetes patients, cases and controls were also matched on antidiabetic treatment duration (within 30 days) at cancer diagnosis. For newly diagnosed diabetic patients, this scheme that matched follow-up duration would have, by design, also taken diabetes duration into consideration. For prevalent patients with unknown duration, we selected controls with the same follow-up duration to reduce the confounding effect by diabetes duration. Up to four controls were selected for each case. The main exposure of interest was the use of rosiglitazone and pioglitazone, which entered Taiwan’s market in March 2000 and selleck compound June 2001, respectively. We collected information of prescribed drug types (according to the anatomic therapeutic chemical [ATC] classification system, A10BG02 for rosiglitazone and A10BG03 for pioglitazone), dosage, date of prescription, supply days, and total number of pills dispensed from the outpatient pharmacy prescription database. The mean daily dose for each individual was calculated as dividing the cumulative number of pills by the follow-up duration. Subsequently, the defined daily dose (DDD) was then established by an expert panel according to the relative amount compared to the typical maintenance dose for an adult.