The base cohort for this analysis included all patients who met the eligibility criteria (Table 1) on any day (possible index date) during the study period in the two databases that contained UACR records (Clinformatics Data Mart and RWD). To increase the number of patients in this effectiveness analysis, patients were also included if they had a urinary protein-to-creatinine ratio (UPCR) measurement that corresponded to the UACR thresholds used as inclusion and exclusion criteria for the study. UPCR values were converted using the crude model PCR equation for predicted UACR, as described by Sumida et al. [21 (link)]. To allow a pooled analysis of both countries and thereby increase the study sample size, this analysis applied the more conservative dapagliflozin 10 mg indication for CKD treatment from the USA. Therefore, all included patients were eligible for dapagliflozin 10 mg according to the approved CKD indication. All eligible patients were considered potential comparators until they initiated dapagliflozin 10 mg. Patients who subsequently initiated dapagliflozin 10 mg and met the indication for the treatment of CKD were then categorized as ‘dapagliflozin initiators’. Dapagliflozin initiators could therefore be a matched comparator in the period before initiation.

Effectiveness analysis: eligibility criteria to identify patients with a possible index date

Inclusion criteria
Age≥ 18 years on index date
Meets CKD definition on or within 2 years before index date

UACR ≥ 30 mg/ga

UPCR ≥ 150 mg/g

CKD diagnosis code

Two eGFR measurements ≥ 90 days apart, both < 60 mL/min/1.73 m2

Exclusion criteria
Continuous enrolment before index date< 730 days
eGFR below threshold (on or within 1 year before index date)25 mL/min/1.73 m2
UACRa above threshold (on or within 1 year before index date) or UPCR equivalent200 mg/g
History of comorbidities on or before index date

Type 1 diabetes

Gestational diabetes mellitus

Diagnosis indicating dialysis

Polycystic kidney disease

End-stage kidney disease

History of procedures on or before index date

Dialysis

Procedure indicating end-stage kidney disease

History of prescriptions

Dapagliflozin 10 mg any time before index date

Any SGLT2i on or 1 year before index date

Hydroxychloroquine on index date

Immunotherapy on or 6 months before index date

CKD chronic kidney disease, eGFR estimated glomerular filtration rate, SGLT2i sodium–glucose cotransporter 2 inhibitor, UACR urinary albumin-to-creatinine ratio, UPCR urinary protein-to-creatinine ratio

aAlso included quantitative UPCR converted to UACR

The index date of dapagliflozin initiators was defined as the date of the first dapagliflozin 10 mg prescription. To avoid immortal time bias and selection bias when setting an index date for the comparators [22 (link)] and to ensure a manageable size for each country-specific analysis cohort, up to five potential comparators were randomly sampled for each dapagliflozin initiator in chronological order of their index dates. These comparators were matched on the basis of age, sex, heart failure diagnosis, type 2 diabetes diagnosis and RASi prescription (defined as having a prescription for angiotensin-converting enzyme inhibitors or angiotensin II receptor antagonists). Matching was performed without replacement: once comparators had been matched a dapagliflozin initiator, these comparators were no longer able to be chosen for matching with other dapagliflozin initiators. Index dates for the randomly sampled comparators were set to the index dates of their matched dapagliflozin initiators; the datasets were then merged. To maximize retention of dapagliflozin initiators and minimize potential unmeasured bias, only patients with sufficient post-index data for eGFR slope estimation were retained. Each patient could only contribute once as an initiator and/or once as a comparator.