Diabetes Mellitus, Type 2
Conditions
Brief summary
The aim of the study is to assess the risk of urinary tract malignancies in patients initiating empagliflozin (free or fixed dose combination) compared to patients initiating a dipeptidyl peptidase-4 (DPP-4) inhibitor.
Interventions
DRUGempagliflozin
empagliflozin
DRUGDPP-4 inhibitors
DPP-4 inhibitors
Sponsors
Eli Lilly and Company
Boehringer Ingelheim
Study design
Observational model
COHORT
Time perspective
PROSPECTIVE
Eligibility
Sex/Gender
ALL
Age
18 Years to No maximum
Healthy volunteers
No
Inclusion criteria
* Diagnosis of type 2 diabetes * Age over 18 years at index date * At least 1 year of membership in the medication database prior to index date
Exclusion criteria
* Patients with any cancer (excluding non-melanoma skin cancer) recorded at any time prior to the index date (i.e. during the available look-back time) * Diagnosis of type 1 diabetes or other specific non-type 2 diabetes * Use of any SGLT-2 inhibitor or any DPP-4 inhibitor (including free and fixed-dose combinations) recorded at any time prior to index date (i.e. during the available look-back time). * Use of fixed-dose combinations of SGLT-2 inhibitors with DPP-4 inhibitors * Diagnosis of end stage renal disease or receipt of renal dialysis recorded at any time prior to index date (i.e. during the available look-back time)
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Occurrence of Urinary Tract Cancer | From 181 days after index date until the occurrence of a censoring event or cancer outcome event. Up to 6 (Sweden/Finland) or 7 years (UK). | Occurrence of urinary tract cancer, which include malignant neoplasm and carcinoma in situ of the urinary tract, is reported as incidence rates per 1000 patient years. Incidence rates were estimated using Poisson regression and computed as the number of outcome events divided by the total person-time-at-risk (in years), multiplied by 1,000 for easier comparison. The country-level datasets were analyzed separately. Thereafter, pooled incidence rates were estimated using the aggregated number of outcome events and time-at-risk of all study countries using the Poisson regression. |
| Occurrence of Bladder Cancer | From 181 days after index date until the occurrence of a censoring event or cancer outcome event. Up to 6 (Sweden/Finland) or 7 years (UK). | Occurrence of bladder cancer, malignant and carcinoma in situ, is reported as incidence rates per 1000 patient years. Incidence rates were estimated using Poisson regression and computed as the number of outcome events divided by the total person-time-at-risk (in years), multiplied by 1,000 for easier comparison. The country-level datasets were analyzed separately. Thereafter, pooled incidence rates were estimated using the aggregated number of outcome events and time-at-risk of all study countries using the Poisson regression. |
| Occurrence of Renal Cancer | From 181 days after index date until the occurrence of a censoring event or cancer outcome event. Up to 6 (Sweden/Finland) or 7 years (UK). | Occurrence of malignant renal cancer is reported as incidence rates per 1000 patient years. Incidence rates were estimated using Poisson regression and computed as the number of outcome events divided by the total person-time-at-risk (in years), multiplied by 1,000 for easier comparison. The country-level datasets were analyzed separately. Thereafter, pooled incidence rates were estimated using the aggregated number of outcome events and time-at-risk of all study countries using the Poisson regression. |
Countries
Finland, Sweden, United Kingdom
Participant flow
Recruitment details
This was a non-interventional, comparative, cohort-based Post-authorisation-safety study (PASS) to assess the risk of urinary tract malignancies in patients affected by type 2 diabetes mellitus initiating empagliflozin or dipeptidyl peptidase-4 inhibitor (DPP-4i) between 1-Aug-2014 and 31-Dec-2021 in the United Kingdom (UK), Sweden, and Finland.
Pre-assignment details
Throughout the study, different data extraction timepoints were used, starting from 2016 to 2023, and retrospective data was analyzed. Only subjects that met all inclusion and none of the exclusion criteria were included.
Participants by arm
| Arm | Count |
|---|---|
| Empagliflozin Initiators - UK Participants with Type 2 diabetes mellitus (T2D) initiating (first purchase/prescription) empagliflozin (study drug), in free form or in fixed-dose combination with metformin, between 1-Aug-2014 and 31-Dec-2021 (index date), and with more than 6 months of possible follow-up (index date less than 6 months prior to end of study date). The study population was identified from Clinical Practice Research Datalink (CPRD) (General practitioner Online Data \[GOLD\], and Aurum) in the United Kingdom (UK). | 27,410 |
| DPP4-i Initiators - UK Participants with T2D initiating (first purchase/prescription) a dipeptidyl peptidase-4 inhibitor (DPP-4i) (comparator), in free form or in fixed-dose combination with metformin, between 1-Aug-2014 and 31-Dec-2021 (index date), and with more than 6 months of possible follow-up (index date less than 6 months prior to end of study date). The study population was identified from CPRD (General practitioner Online Data \[GOLD\], and Aurum) in the UK. | 123,334 |
| Empagliflozin Initiators - Sweden Participants with T2D initiating (first purchase/prescription) empagliflozin (study drug), in free form or in fixed-dose combination with metformin, between 1-Aug-2014 and 31-Dec-2020 (index date), and with more than 6 months of possible follow-up (index date less than 6 months prior to end of study date). The study population was identified from the nationwide registers in Sweden. | 35,954 |
| DPP4-i Initiators - Sweden Participants with T2D initiating (first purchase/prescription) a dipeptidyl peptidase-4 inhibitor (DPP-4i) (comparator), in free form or in fixed-dose combination with metformin, between 1-Aug-2014 and 31-Dec-2020 (index date), and with more than 6 months of possible follow-up (index date less than 6 months prior to end of study date). The study population was identified from the nationwide registers in Sweden. | 73,623 |
| Empagliflozin Initiators - Finland Participants with T2D initiating (first purchase/prescription) empagliflozin (study drug), in free form or in fixed-dose combination with metformin, between 1-Aug-2014 and 31-Dec-2020 (index date), and with more than 6 months of possible follow-up (index date less than 6 months prior to end of study date). The study population was identified from the nationwide registers in Finland. | 21,942 |
| DPP4-i Initiators - Finland Participants with T2D initiating (first purchase/prescription) a dipeptidyl peptidase-4 inhibitor (DPP-4i) (comparator), in free form or in fixed-dose combination with metformin, between 1-Aug-2014 and 31-Dec-2020 (index date), and with more than 6 months of possible follow-up (index date less than 6 months prior to end of study date). The study population was identified from the nationwide registers in Finland. | 62,732 |
| Total | 344,995 |
Baseline characteristics
| Characteristic | Empagliflozin Initiators - UK | DPP4-i Initiators - UK | Empagliflozin Initiators - Sweden | DPP4-i Initiators - Sweden | Empagliflozin Initiators - Finland | DPP4-i Initiators - Finland | Total |
|---|---|---|---|---|---|---|---|
| Age, Continuous | 56.51 Years STANDARD_DEVIATION 10.96 | 63.50 Years STANDARD_DEVIATION 13.24 | 62.75 Years STANDARD_DEVIATION 11.18 | 66.65 Years STANDARD_DEVIATION 12.66 | 60.11 Years STANDARD_DEVIATION 11.84 | 66.04 Years STANDARD_DEVIATION 13.42 | 63.78 Years STANDARD_DEVIATION 12.93 |
| Race and Ethnicity Not Collected | — | — | — | — | — | — | 0 Participants |
| Sex: Female, Male Female | 16922 Participants | 72124 Participants | 24825 Participants | 44352 Participants | 13048 Participants | 34233 Participants | 205504 Participants |
| Sex: Female, Male Male | 10488 Participants | 51210 Participants | 11129 Participants | 29271 Participants | 8894 Participants | 28499 Participants | 139491 Participants |
Adverse events
| Event type | EG000 affected / at risk | EG001 affected / at risk | EG002 affected / at risk | EG003 affected / at risk | EG004 affected / at risk | EG005 affected / at risk |
|---|---|---|---|---|---|---|
| deaths Total, all-cause mortality | 0 / 0 | 0 / 0 | 0 / 0 | 0 / 0 | 0 / 0 | 0 / 0 |
| other Total, other adverse events | 0 / 0 | 0 / 0 | 0 / 0 | 0 / 0 | 0 / 0 | 0 / 0 |
| serious Total, serious adverse events | 0 / 0 | 0 / 0 | 0 / 0 | 0 / 0 | 0 / 0 | 0 / 0 |
Outcome results
Primary
Occurrence of Bladder Cancer
Occurrence of bladder cancer, malignant and carcinoma in situ, is reported as incidence rates per 1000 patient years. Incidence rates were estimated using Poisson regression and computed as the number of outcome events divided by the total person-time-at-risk (in years), multiplied by 1,000 for easier comparison. The country-level datasets were analyzed separately. Thereafter, pooled incidence rates were estimated using the aggregated number of outcome events and time-at-risk of all study countries using the Poisson regression.
Time frame: From 181 days after index date until the occurrence of a censoring event or cancer outcome event. Up to 6 (Sweden/Finland) or 7 years (UK).
Population: Main analysis population: participants that were matched using propensity scores (PS) conditional on variables such as index date, disease progression, comorbidities, treatment history, age, and sex; and who did not have observations of any censoring events or outcome cancer events within the latency period (spanning 180 days after the index date).
| Arm | Measure | Value (NUMBER) |
|---|---|---|
| Empagliflozin Initiators - UK, PS-matched | Occurrence of Bladder Cancer | 0.45 Events per 1000 patient years |
| DPP4-i Initiators - UK, PS-matched | Occurrence of Bladder Cancer | 0.67 Events per 1000 patient years |
| Empagliflozin Initiators - Sweden, PS-matched | Occurrence of Bladder Cancer | 0.92 Events per 1000 patient years |
| DPP4-i Initiators - Sweden, PS-matched | Occurrence of Bladder Cancer | 1.03 Events per 1000 patient years |
| Empagliflozin Initiators - Finland, PS-matched | Occurrence of Bladder Cancer | 0.51 Events per 1000 patient years |
| DPP4-i Initiators - Finland, PS-matched | Occurrence of Bladder Cancer | 0.64 Events per 1000 patient years |
| Empagliflozin Initiators - All Countries, PS-matched | Occurrence of Bladder Cancer | 0.63 Events per 1000 patient years |
| DPP4-i Initiators - All Countries, PS-matched | Occurrence of Bladder Cancer | 0.74 Events per 1000 patient years |
Comparison: Crude HRs were calculated using a Cox proportional hazards model with as-treated exposure (continuous exposure to the study drugs, defined as having consecutive treatment episodes separated by a grace period of 30 days, starting from index date) as the only included variable. Country-level HRs were analyzed separately. Thereafter, random-effects meta-analyses were conducted to pool the HRs from all countries, accounting for potential heterogeneity between the country-level effect sizes.p-value: 0.2295% CI: [0.55, 1.15]Regression, Cox
Comparison: Base HRs were calculated using a Cox proportional hazards model by adjusting for variables unbalanced after PS matching. To be included in the Cox model, each covariate is required to have a minimum of 5 events per category level to avoid converge failure.~Country-level HRs were analyzed separately. Thereafter, random-effects meta-analyses were conducted to pool the HRs from all countries, accounting for potential heterogeneity between the country-level effect sizes.p-value: 0.62895% CI: [0.63, 1.32]Regression, Cox
Comparison: Adjusted HRs were calculated using a Cox proportional hazards model by adjusting for variables unbalanced after PS matching, and pre-specified time-varying covariates. To be included in the Cox model, each covariate is required to have a minimum of 5 events per category level. Country-level HRs were analyzed separately. Thereafter, random-effects meta-analyses were conducted to pool the HRs from all countries, accounting for potential heterogeneity between the country-level effect sizes.p-value: 0.63295% CI: [0.63, 1.33]Regression, Cox
Primary
Occurrence of Renal Cancer
Occurrence of malignant renal cancer is reported as incidence rates per 1000 patient years. Incidence rates were estimated using Poisson regression and computed as the number of outcome events divided by the total person-time-at-risk (in years), multiplied by 1,000 for easier comparison. The country-level datasets were analyzed separately. Thereafter, pooled incidence rates were estimated using the aggregated number of outcome events and time-at-risk of all study countries using the Poisson regression.
Time frame: From 181 days after index date until the occurrence of a censoring event or cancer outcome event. Up to 6 (Sweden/Finland) or 7 years (UK).
Population: Main analysis population: participants that were matched using propensity scores (PS) conditional on variables such as index date, disease progression, comorbidities, treatment history, age, and sex; and who did not have observations of any censoring events or outcome cancer events within the latency period (spanning 180 days after the index date).
| Arm | Measure | Value (NUMBER) |
|---|---|---|
| Empagliflozin Initiators - UK, PS-matched | Occurrence of Renal Cancer | 0.24 Events per 1000 patient years |
| DPP4-i Initiators - UK, PS-matched | Occurrence of Renal Cancer | 0.34 Events per 1000 patient years |
| Empagliflozin Initiators - Sweden, PS-matched | Occurrence of Renal Cancer | 0.42 Events per 1000 patient years |
| DPP4-i Initiators - Sweden, PS-matched | Occurrence of Renal Cancer | 0.50 Events per 1000 patient years |
| Empagliflozin Initiators - Finland, PS-matched | Occurrence of Renal Cancer | 0.76 Events per 1000 patient years |
| DPP4-i Initiators - Finland, PS-matched | Occurrence of Renal Cancer | 0.82 Events per 1000 patient years |
| Empagliflozin Initiators - All Countries, PS-matched | Occurrence of Renal Cancer | 0.45 Events per 1000 patient years |
| DPP4-i Initiators - All Countries, PS-matched | Occurrence of Renal Cancer | 0.51 Events per 1000 patient years |
Comparison: Crude HRs were calculated using a Cox proportional hazards model with as-treated exposure (continuous exposure to the study drugs, defined as having consecutive treatment episodes separated by a grace period of 30 days, starting from index date) as the only included variable. Country-level HRs were analyzed separately. Thereafter, random-effects meta-analyses were conducted to pool the HRs from all countries, accounting for potential heterogeneity between the country-level effect sizes.p-value: 0.4295% CI: [0.54, 1.29]Regression, Cox
Comparison: Base HRs were calculated using a Cox proportional hazards model by adjusting for variables unbalanced after PS matching. To be included in the Cox model, each covariate is required to have a minimum of 5 events per category level to avoid converge failure.~Country-level HRs were analyzed separately. Thereafter, random-effects meta-analyses were conducted to pool the HRs from all countries, accounting for potential heterogeneity between the country-level effect sizes.p-value: 0.60695% CI: [0.57, 1.38]Regression, Cox
Comparison: Adjusted HRs were calculated using a Cox proportional hazards model by adjusting for variables unbalanced after PS matching, and pre-specified time-varying covariates. To be included in the Cox model, each covariate is required to have a minimum of 5 events per category level. Country-level HRs were analyzed separately. Thereafter, random-effects meta-analyses were conducted to pool the HRs from all countries, accounting for potential heterogeneity between the country-level effect sizes.p-value: 0.60995% CI: [0.57, 1.38]Regression, Cox
Primary
Occurrence of Urinary Tract Cancer
Occurrence of urinary tract cancer, which include malignant neoplasm and carcinoma in situ of the urinary tract, is reported as incidence rates per 1000 patient years. Incidence rates were estimated using Poisson regression and computed as the number of outcome events divided by the total person-time-at-risk (in years), multiplied by 1,000 for easier comparison. The country-level datasets were analyzed separately. Thereafter, pooled incidence rates were estimated using the aggregated number of outcome events and time-at-risk of all study countries using the Poisson regression.
Time frame: From 181 days after index date until the occurrence of a censoring event or cancer outcome event. Up to 6 (Sweden/Finland) or 7 years (UK).
Population: Main analysis population: participants that were matched using propensity scores (PS) conditional on variables such as index date, disease progression, comorbidities, treatment history, age, and sex; and who did not have observations of any censoring events or outcome cancer events within the latency period (spanning 180 days after the index date).
| Arm | Measure | Value (NUMBER) |
|---|---|---|
| Empagliflozin Initiators - UK, PS-matched | Occurrence of Urinary Tract Cancer | 0.73 Events per 1000 patient years |
| DPP4-i Initiators - UK, PS-matched | Occurrence of Urinary Tract Cancer | 1.07 Events per 1000 patient years |
| Empagliflozin Initiators - Sweden, PS-matched | Occurrence of Urinary Tract Cancer | 1.34 Events per 1000 patient years |
| DPP4-i Initiators - Sweden, PS-matched | Occurrence of Urinary Tract Cancer | 1.52 Events per 1000 patient years |
| Empagliflozin Initiators - Finland, PS-matched | Occurrence of Urinary Tract Cancer | 1.27 Events per 1000 patient years |
| DPP4-i Initiators - Finland, PS-matched | Occurrence of Urinary Tract Cancer | 1.49 Events per 1000 patient years |
| Empagliflozin Initiators - All Countries, PS-matched | Occurrence of Urinary Tract Cancer | 1.10 Events per 1000 patient years |
| DPP4-i Initiators - All Countries, PS-matched | Occurrence of Urinary Tract Cancer | 1.29 Events per 1000 patient years |
Comparison: Crude HRs were calculated using a Cox proportional hazards model with as-treated exposure (continuous exposure to the study drugs, defined as having consecutive treatment episodes separated by a grace period of 30 days, starting from index date) as the only included variable. Country-level HRs were analyzed separately. Thereafter, random-effects meta-analyses were conducted to pool the HRs from all countries, accounting for potential heterogeneity between the country-level effect sizes.p-value: 0.12295% CI: [0.61, 1.06]Regression, Cox
Comparison: Base HRs were calculated using a Cox proportional hazards model by adjusting for variables unbalanced after PS matching. To be included in the Cox model, each covariate is required to have a minimum of 5 events per category level to avoid converge failure.~Country-level HRs were analyzed separately. Thereafter, random-effects meta-analyses were conducted to pool the HRs from all countries, accounting for potential heterogeneity between the country-level effect sizes.p-value: 0.41795% CI: [0.67, 1.18]Regression, Cox
Comparison: Adjusted HRs were calculated using a Cox proportional hazards model by adjusting for variables unbalanced after PS matching, and pre-specified time-varying covariates. To be included in the Cox model, each covariate is required to have a minimum of 5 events per category level. Country-level HRs were analyzed separately. Thereafter, random-effects meta-analyses were conducted to pool the HRs from all countries, accounting for potential heterogeneity between the country-level effect sizes.p-value: 0.37695% CI: [0.66, 1.17]Regression, Cox