Cohort Study of Pioglitazone and Bladder Cancer in Patients With Type II Diabetes

• Incident Diagnosis of Bladder Cancer [ Time Frame: January 1, 1997 to December 31, 2010. ] [ Designated as safety issue: No ]
  Incident bladder cancers were identified from the KPNC cancer registry from January 1, 1997 to December 31, 2010. This was supplemented by case identification through surveillance of electronic pathology reports within KPNC for the period from January 1, 2005 to December 31, 2010. Incidence is reported as the number of newly diagnosed cases of bladder cancer per 100,000 person years. The 95% confidence interval was calculated using the Poisson method.
• Number of Participants With Newly Diagnosed Bladder Cancer [ Time Frame: January 1, 1997 to December 31, 2010. ] [ Designated as safety issue: No ]
  Bladder cancers were identified from the KPNC cancer registry from January 1, 1997 to December 31, 2010. This was supplemented by case identification through surveillance of electronic pathology reports within KPNC for the period from January 1, 2005 to December 31, 2010.

• Incident Diagnosis of Bladder Cancer, by Time Since Starting Pioglitazone [ Time Frame: January 1, 1997 to December 31, 2010. ] [ Designated as safety issue: Yes ]
  Incident bladder cancers were identified from the KPNC cancer registry from January 1, 1997 to December 31, 2010. This was supplemented by case identification through surveillance of electronic pathology reports within KPNC for the period from January 1, 2005 to December 31, 2010. Incidence is reported as the number of newly diagnosed cases of bladder cancer per 100,000 person years by time since starting pioglitazone. The 95% confidence interval was calculated using the Poisson method.
• Incident Diagnosis of Bladder Cancer, by Duration of Pioglitazone Therapy [ Time Frame: January 1, 1997 to December 31, 2010. ] [ Designated as safety issue: Yes ]
  Incident bladder cancers were identified from the KPNC cancer registry from January 1, 1997 to December 31, 2010. This was supplemented by case identification through surveillance of electronic pathology reports within KPNC for the period from January 1, 2005 to December 31, 2010. Incidence is reported as the number of newly diagnosed cases of bladder cancer per 100,000 person years and by duration of pioglitazone therapy. The 95% confidence interval was calculated using the Poisson method.
• Incident Diagnosis of Bladder Cancer by Cumulative Dose of Pioglitazone [ Time Frame: January 1, 1997 to December 31, 2010. ] [ Designated as safety issue: Yes ]
  Incident bladder cancers were identified from the KPNC cancer registry from January 1, 1997 to December 31, 2010. This was supplemented by case identification through surveillance of electronic pathology reports within KPNC for the period from January 1, 2005 to December 31, 2010. Incidence is reported as the number of newly diagnosed cases of bladder cancer per 100,000 person years by cumulative dose of pioglitazone. The 95% confidence interval was calculated using the Poisson method.
• Participants With Bladder Cancer by Stage of Cancer [ Time Frame: January 1, 1997 to December 31, 2010. ] [ Designated as safety issue: Yes ]
  Incident bladder cancers were identified from the KPNC cancer registry from January 1, 1997 to December 31, 2010. This was supplemented by case identification through surveillance of electronic pathology reports within KPNC for the period from January 1, 2005 to December 31, 2010. Bladder cancer stages are: PUNLMP (Papillary urethral neoplasm of low malignant potential); In situ (cancer is growing in the inner lining layer of the bladder only); Local (cancer has grown into the thick muscle layer of the bladder wall, but it has not passed completely through the muscle to reach the layer of fatty tissue that surrounds the bladder); Regional (cancer has grown through the bladder wall and into the pelvic or abdominal wall but has not spread to lymph nodes or to distant sites); Distant (cancer has spread to distant sites such as bones, liver, or lungs); Undetermined.
• Comparison of the Results of the Current Analysis and the 5-Year Interim Report for Incidence of Bladder Cancer [ Time Frame: The 5-year analysis includes data from January 1 1997 to 30 April 2008. The 8-year analysis includes data from January 1, 1997 to December 31, 2010. ] [ Designated as safety issue: Yes ]

  Data presented are the hazard ratios for comparison of the group of participants treated with pioglitazone with those who received no pioglitazone therapy. Cox proportional hazards models were used for all calculations of the relative hazard (HR) of bladder cancer with pioglitazone, adjusted for the covariates. The reference group for calculation of the relative hazard associated with use of pioglitazone was no use of pioglitazone.

  The fully adjusted includes all potential confounders in the statistical model from the 5 year interim report: age, sex, race/ethnicity, other diabetes medications, smoking, other bladder conditions, median household income, congestive heart failure, cancer other than bladder cancer, renal insufficiency, HbA1c and the interaction with new diagnosis of diabetes, and duration of diabetes. In the 8 year analysis, the fully adjusted model also includes year of cohort entry.

• Comparison of the Results of the Current Analysis and the 5-Year Interim Report for Incidence of Bladder Cancer, by Time Since Starting Pioglitazone [ Time Frame: The 5-year analysis includes data from January 1 1997 to 30 April 2008. The 8-year analysis includes data from January 1, 1997 to December 31, 2010. ] [ Designated as safety issue: Yes ]

  Data presented are the hazard ratios for comparison of the participants treated with pioglitazone, by time since starting pioglitazone with those who received no pioglitazone. Cox proportional hazards models were used for all calculations of the relative hazard (HR) of bladder cancer with pioglitazone, adjusted for the covariates.

  The fully adjusted model includes all potential confounders in the statistical model from the 5 year interim report: age, sex, race/ethnicity, other diabetes medications, smoking, other bladder conditions, median household income, congestive heart failure, cancer other than bladder cancer, renal insufficiency, HbA1c and the interaction with new diagnosis of diabetes, and duration of diabetes. In the 8 year analysis, the fully adjusted model also includes year of cohort entry.

  Row headings state the equivalent number of years since starting pioglitazone therapy first for the 5 year analysis and then for the 8-year analysis.

• Comparison of the Results of the Current Analysis and the 5-Year Interim Report for Incidence of Bladder Cancer, by Duration of Therapy [ Time Frame: The 5-year analysis includes data from January 1 1997 to 30 April 2008. The 8-year analysis includes data from January 1, 1997 to December 31, 2010. ] [ Designated as safety issue: Yes ]

  Data presented are the hazard ratios for comparison of the participants treated with pioglitazone, by duration of pioglitazone therapy with those who received no pioglitazone. Cox proportional hazards models were used for all calculations of the relative hazard (HR) of bladder cancer with pioglitazone, adjusted for the covariates.

  The fully adjusted model includes all potential confounders in the statistical model from the 5 year interim report: age, sex, race/ethnicity, other diabetes medications, smoking, other bladder conditions, median household income, congestive heart failure, cancer other than bladder cancer, renal insufficiency, HbA1c and the interaction with new diagnosis of diabetes, and duration of diabetes. In the 8 year analysis, the fully adjusted model also includes year of cohort entry.

  Row headings state the number of years of duration of therapy first for the 5 year analysis and then the equivalent duration of therapy for the 8-year analysis.

• Comparison of the Results of the Current Analysis and the 5-Year Interim Report for Incidence of Bladder Cancer, by Cumulative Dose [ Time Frame: The 5-year analysis includes data from January 1 1997 to 30 April 2008. The 8-year analysis includes data from January 1, 1997 to December 31, 2010. ] [ Designated as safety issue: Yes ]

  Data presented are the hazard ratios for comparison of the participants treated with pioglitazone, by cumulative dose of pioglitazone with those who received no pioglitazone. Cox proportional hazards models were used for all calculations of the relative hazard (HR) of bladder cancer with pioglitazone, adjusted for the covariates.

  The fully adjusted model includes all potential confounders in the statistical model from the 5 year interim report: age, sex, race/ethnicity, other diabetes medications, smoking, other bladder conditions, median household income, congestive heart failure, cancer other than bladder cancer, renal insufficiency, HbA1c and the interaction with new diagnosis of diabetes, and duration of diabetes. In the 8 year analysis, the fully adjusted model also includes year of cohort entry.

  Row headings state the cumulative pioglitazone dose first for the 5-Year analysis and then the equivalent cumulative dose for the 8-Year analysis.

To assess the potential association between pioglitazone and bladder cancer compared with non-pioglitazone users among patients with type 2 diabetes mellitus.

Following guidance from the United States Food and Drug Administration (FDA), the University of Pennsylvania and Kaiser Permanente Northern California (KPNC) designed and is conducting this study from their own database to assess the potential association between pioglitazone and bladder cancer among patients with type 2 diabetes mellitus.

The study is being conducted over the course of 10 years, with a series of interim analyses provided to the sponsor (Takeda) and the appropriate regulatory agencies.

In 2011, the planned 5-year interim analysis of this study was published in Diabetes Care. That report included data from 1 January 1997 to 30 April 2008. Following reporting of these data, there was a request from the FDA for an additional fourth interim analysis at 8 years including data from 1 January 1997 to 31 December 2010.

In August 2011 the FDA requested a sensitivity analysis to assess change of cohort entry criteria to minimize left censoring of exposure. Included in the FDA request was a duration analysis for other antidiabetic medications.

The study is being conducted within Kaiser Permanente Northern California (KPNC), which provides comprehensive healthcare services to approximately 3.2 million members. The source population was identified from the KPNC diabetes registry, which was first constructed in 1993 and has been updated annually since then. The registry identifies patients primarily from four data sources: primary hospital discharge diagnoses of diabetes mellitus (since 1971); two or more outpatient visit diagnoses of diabetes (since 1995); any prescription for a diabetes-related medication (since 1994); or any record of an abnormal hemoglobin A1c (HbA1c) test (>6.7%) (since 1991).

The diabetes registry gathers data from a variety of KPNC electronic medical records (EMR) to build and follow the registry cohort across time. These data include cancer registries, pharmacy records, laboratory records, and inpatient and outpatient medical diagnoses.

• Diabetes
• Bladder Cancer

• Pioglitazone treated
  Participants treated with pioglitazone, defined as having filled at least two prescriptions for the drug within a 6-month period according to the Kaiser Permanente Northern California (KPNC) pharmacy database.
  Intervention: Drug: Pioglitazone
• No pioglitazone treatment
  Participants not treated with pioglitazone, defined as not having filled at least two prescriptions for the drug within a 6-month period according to the KPNC pharmacy database.

Inclusion Criteria:

• In the KPNC diabetes registry and who are age 40 or older as of January 1, 1997 or additional registry members who reach age 40 at any point before December 31 2002; patients age 40 or older who enroll in KPNC between January 1, 1997 and December 21 2002 and who are identified as having diabetes; and all KP members age 40 or older who develop diabetes during this time period.

Inclusion Criteria for the Sensitivity Analysis - Modified Cohort:

• All patients who are in the KPNC Diabetes Registry and who are age 40 or older as of January 1, 1997;
• All additional KPNC Diabetes Registry members who reach age 40 at any point before December 31, 2009 and who have been enrolled in KPNC since January 1, 1997 without a gap in membership;
• All KPNC members age 40 or older who are newly diagnosed with diabetes between January 1, 1997 and December 31, 2009. The definition of newly diagnosed diabetes requires that the patient be a KPNC member for more than 2 years before being identified as having diabetes according the registry criteria.

Exclusion Criteria:

• Diagnosis of bladder cancer recorded in the KPNC cancer registry prior to initiation of observation or within 6 months of entry into KPNC.

Inclusion Criteria for the Sensitivity Analysis - Modified Cohort:

• all patients who are in the KPNC Diabetes Registry and who are age 40 or older as of January 1, 1997;
• all additional KPNC Diabetes Registry members who reach age 40 at any point before December 31, 2009 and who have been enrolled in KPNC since January 1, 1997 without a gap in membership;
• all KPNC members age 40 or older who are newly diagnosed with diabetes between January 1, 1997 and December 31, 2009. The definition of newly diagnosed diabetes requires that the patient be a KPNC member for more than 2 years before being identified as having diabetes according the registry criteria.