Abstract
Background and Aims
Chronic kidney disease (CKD) represents a global public health problem, with significant morbidity and mortality due to cardiovascular disease during CKD progression and due to kidney failure. Although non-diabetic CKD accounts for up to 70% of the global CKD burden, its clinical consequences are poorly understood, and data are needed to help identify individuals at high risk of adverse outcomes. This analysis uses real-world evidence to provide insights into clinical characteristics, care and outcomes in individuals with non-diabetic CKD in routine clinical practice.
Method
Individual-level data from the US administrative claims database, Optum Clinformatics Data Mart, from January 1, 2008 to December 31, 2018 were analysed. Adults with non-diabetic CKD stage 3 or 4 and ≥365 days continuous insurance coverage were included and followed until insurance disenrollment, end of data availability or death. Individuals with diabetes mellitus, CKD stage 5 or end-stage kidney disease (ESKD) prior to the index date, or who experienced kidney failure (acute or unspecified), kidney transplant or dialysis in the baseline period, were excluded from the analysis. Study outcomes, captured in the database, were defined using common clinical coding systems. Primary outcomes were hospitalisation for heart failure (HHF), a kidney composite of ESKD/kidney failure/need for dialysis, and worsening of CKD stage from baseline. Individual CKD stage was assigned based on estimated glomerular filtration rate (eGFR) values (priority) or the respective International Classification of Diseases code at index and during follow-up. Further prespecified kidney outcomes included individual components of the kidney composite, acute kidney injury, and absolute and relative change in eGFR from baseline. Event-based outcomes were assessed by time-to-first-event analysis. Summary statistics for time-course analysis of metric outcomes were generated on a quarterly basis.
Results
In total, 504,924 of 64 million individuals in the Optum Clinformatics Data Mart satisfied the selection criteria. Over a median follow-up of 744 (interquartile range 328–1432) days, the incidence rates of primary outcomes of HHF, the kidney composite and worsening of CKD stage from baseline were 3.95, 10.33 and 4.38 events/100 patient-years (PY), respectively.
The incidence rates of the components of the kidney composite outcome, namely ESKD/need for dialysis, kidney failure (acute and unspecified) and need for dialysis were 1.78, 9.53 and 0.49 events/100 PY, respectively. Kidney failure events were driven mainly by acute kidney injury, with an incidence of 8.61 events/100 PY.
In individuals with at least one available eGFR value at baseline and one value during follow-up (n=295,174), the incidence rates of relative decreases in eGFR of ≥30%, ≥40% and ≥57% from baseline were 1.98, 0.97 and 0.30 events/100 PY, respectively; in this cohort, more rapid eGFR decline was associated with increased risk of HHF and the kidney composite outcome. In individuals with a baseline eGFR value and at least one follow-up eGFR value and an available urine albumin-to-creatinine ratio (n=25,824), time-course analysis of eGFR showed that eGFR decline mostly occurred in individuals with moderately-to-severely increased albuminuria (≥30 mg/g).
Conclusion
This analysis generates real-world evidence on clinical outcomes in a cohort of individuals with non-diabetic CKD treated in routine clinical practice in the US. Despite known limitations of claims databases (e.g. low availability of some laboratory data, limited individual follow-up time and tactical coding), individuals with moderate-to-severe non-diabetic CKD are shown to be at high risk of serious clinical outcomes. This highlights the high unmet medical need, and urgency for new treatments and targeted interventions for patients with non-diabetic CKD.
Change in GFR over time in the Oxford Renal Cohort Study
