Elicit: EGFR Slope as a Clinical Trial Endpoint in CKD (public)

Clinical trials using EGFR slope as an endpoint in chronic kidney disease

Twenty-two clinical trials used estimated glomerular filtration rate (eGFR) slope as an endpoint in chronic kidney disease studies, with 13 trials using it as a primary outcome and methods for slope calculation varying considerably across studies.

Abstract

Twenty-two trials—21 randomized controlled trials and 1 prospective observational study—address chronic kidney disease using estimated glomerular filtration rate (eGFR) slope endpoints. Thirteen trials defined a true eGFR slope (annualized or total rate of decline) as their primary outcome, while 4 trials included slope measures as secondary endpoints; the remainder employed composite outcomes, time-to-event analyses, or point estimates. In the primary endpoint studies, one trial specified an expected difference of 0.75 ml/min/1.73m² per year with 80% power over a 52‐week period. Methods for slope calculation varied considerably; some studies used linear mixed effects or repeated measurements models, weighted analysis of covariance, simple linear regression, or linear random slope models, whereas several trials provided no detailed description of the analytic method. Populations ranged from chronic kidney disease stages 2 to 5, including diabetic kidney disease, autosomal dominant polycystic kidney disease, and immunoglobulin A nephropathy, with trial durations from 52 weeks to 120 weeks.

Methods

We analyzed 24 sources from an initial pool of 36, using 10 screening criteria. Each paper was reviewed for 6 key aspects that mattered most to the research question.

Papers identified with Elicit search

n = 36

Papers screened using:

Population - CKD Patients
Population - Not AKI Only
Population - Adult Participants
Primary Outcome - eGFR Slope
Outcome Analysis Quality
Study Design Quality
Study Duration
eGFR Measurement Standards
Multiple Time Points
Intervention Type

n = 36

Papers screened out

n = 12

Papers included for extraction

n = 24

Paper search

Using your research question “Clinical trials using EGFR slope as an endpoint in chronic kidney disease”, we searched across all trials from the ClinicalTrials.gov corpus. We retrieved the 36 trials most relevant to the query.

Screening

We screened in sources based on their abstracts that met these criteria:

Population - CKD Patients: Does the study include participants with chronic kidney disease (any stage)?
Population - Not AKI Only: Is the study NOT focused exclusively on acute kidney injury without underlying chronic kidney disease?
Population - Adult Participants: Is the study NOT conducted exclusively in pediatric populations (under 18 years)?
Primary Outcome - eGFR Slope: Does the study report estimated glomerular filtration rate (eGFR) slope as a primary or secondary endpoint?
Outcome Analysis Quality: Is eGFR slope formally analyzed with proper statistical methods (not merely mentioned in passing)?
Study Design Quality: Is the study design a randomized controlled trial, controlled clinical trial, systematic review, or meta-analysis?
Study Duration: Does the study have a follow-up period of at least 12 months?
eGFR Measurement Standards: Does the study use standardized eGFR calculation methods (e.g., CKD-EPI, MDRD)?
Multiple Time Points: Does the study report multiple eGFR measurements over time (not just single time-point measurements)?
Intervention Type: Does the study evaluate therapeutic interventions (pharmacological, dietary, lifestyle, or device-based) with long-term follow-up rather than focusing solely on acute interventions?

We considered all screening questions together and made a holistic judgement about whether to screen in each paper.

Data extraction

We asked a large language model to extract each data column below from each paper.

Study design:
- Extract the specific type of study design from the methods section. Look for key descriptors such as:
  - Interventional/observational
  - Prospective/retrospective
  - Parallel group
  - Blinding type (open-label, double-blind, single-blind)
  - Multicenter or single-site
  - Specific phase of clinical trial (if applicable)
Target population characteristics:
- Extract detailed population characteristics from the methods/participants section:
  - Age range or mean age
  - Gender distribution
  - Specific chronic kidney disease (CKD) stage
  - Inclusion/exclusion criteria related to kidney function
  - Baseline eGFR (with units and range)
  - Presence of comorbidities (e.g., diabetes)
Intervention specifics:
- For intervention studies, extract:
  - Specific drug/treatment name
  - Dosage
  - Frequency of administration
  - Duration of intervention
  - Any dose modifications during study
  - Concurrent treatments or care programs
Comparison/control condition:
- Describe the control group or comparator condition:
  - Specific details of control treatment
  - Whether standard care was maintained
  - Any placebo or alternative intervention used
Primary outcome measure:
- Extract the primary outcome as explicitly stated in the study:
  - Precise outcome definition
  - Measurement method
  - Units of measurement
  - Time point of measurement
Statistical analysis methods:
- Describe key statistical analysis approaches:
  - Analysis sets used (e.g., intention-to-treat)
  - Statistical tests employed
  - Significant p-value threshold
  - Methods for handling missing data
  - Multiplicity or hierarchical testing approaches

Results

Characteristics of Included Studies

Study	Study Design	Population	Intervention	eGFR Slope Definition
NCT05196347	Multicenter, randomized, open-label, blinded-endpoint, phase 3 randomized controlled trial	Chronic kidney disease stage 4–5, estimated glomerular filtration rate 10–30, age ≥20 years	Dapagliflozin 5–10 mg daily plus integrated chronic kidney disease care vs. care alone	Difference in estimated glomerular filtration rate slope (ml/min/1.73m²/year) from randomization to end of study, linear mixed model
NCT02160145	Multicenter, randomized, double-blind, phase 3 randomized controlled trial	Autosomal dominant polycystic kidney disease, late stage 2–early stage 4 chronic kidney disease, age 18–65 years	Tolvaptan vs. placebo	Mean annualized change in estimated glomerular filtration rate (ml/min/1.73m²/year), Chronic Kidney Disease Epidemiology Collaboration formula
NCT02957864	Multicenter, open-label, randomized, phase 4 randomized controlled trial	Human immunodeficiency virus, tenofovir disoproxil fumarate-related nephrotoxicity, estimated glomerular filtration rate ≥30	Switch to tenofovir alafenamide vs. abacavir	Time to first estimated glomerular filtration rate within 75% of baseline (not a slope)
NCT06217302	Multicenter, double-blind, phase 3 randomized controlled trial	Type 1 diabetes, diabetic kidney disease, estimated glomerular filtration rate 20–60, age 18–75 years	Sotagliflozin 200 mg daily vs. placebo	Estimated glomerular filtration rate at end of washout, adjusted by baseline (not a slope)
NCT03789695	Single-center, open-label, phase 4 randomized controlled trial	Atrial fibrillation, type 2 diabetes, chronic kidney disease, age ≥18 years	Dabigatran vs. warfarin	Difference in annual estimated glomerular filtration rate decline (slopes), Chronic Kidney Disease Epidemiology Collaboration formula
NCT03832595	Single-center, pragmatic, cluster randomized controlled trial	High-risk chronic kidney disease, age 18–85 years	Electronic health record-based population health management vs. usual care	≥40% decline in estimated glomerular filtration rate or end-stage renal disease (composite, not slope)
NCT06825676	Multicenter, cluster randomized controlled trial	Chronic kidney disease, estimated glomerular filtration rate 15–59, age ≥18 years	Provider education vs. routine training	≥25% decline in estimated glomerular filtration rate, end-stage renal disease, or death (composite, not slope)
NCT07027774	Multicenter, open-label, phase 4 randomized controlled trial	Chronic kidney disease stage 4, estimated glomerular filtration rate 10–30, age ≥18 years	Henagliflozin 5/10 mg vs. conventional therapy	Estimated glomerular filtration rate total slope change from week 4 to 52
NCT07116928	Multicenter, double-blind, phase 4 randomized controlled trial	Type 2 diabetes, chronic kidney disease, age ≥18 years	Ganagliflozin 50 mg vs. placebo	Average change in estimated glomerular filtration rate slope from baseline to week 120
NCT06291376	Multicenter, double-blind, phase 3 randomized controlled trial	Immunoglobulin A nephropathy, high risk, estimated glomerular filtration rate ≥30 (main), 20–29 (advanced kidney disease)	Ravulizumab vs. placebo	Change in estimated glomerular filtration rate at week 106 (not a slope)

Key points from included studies:

Study design:
- 21 randomized controlled trials and 1 prospective observational study were included.
- Of the randomized controlled trials:
  - 20 were multicenter; 3 were single-center (2 studies were both multicenter and single-center, counted in both categories).
  - 2 were cluster randomized controlled trials; 1 was pragmatic; 2 were platform/adaptive designs.
  - 7 were open-label; 2 were single-blind; 7 were double-blind; 5 were quadruple-blind; 1 was blinded-endpoint.
  - 3 were phase 2; 10 were phase 3; 6 were phase 4; 1 was factorial; 2 were single-group; 1 was international.
  - No mention found of phase information for 2 randomized controlled trials.
Estimated glomerular filtration rate slope definition:
- 13 studies used a true estimated glomerular filtration rate slope (annualized, total, or rate of decline) as an outcome.
- 5 studies used a composite outcome including estimated glomerular filtration rate decline (such as ≥25% or ≥40% decline, end-stage renal disease, or death).
- 1 study used time to event (time to first estimated glomerular filtration rate within 75% of baseline).
- 2 studies reported change in estimated glomerular filtration rate at a specific timepoint (not a slope).
- 3 studies used other outcomes not based on estimated glomerular filtration rate slope (kidney fibrosis, drug discontinuation, or chronic kidney disease incidence).
- No mention found of a true estimated glomerular filtration rate slope outcome in 8 of the studies.

Effects

Estimated Glomerular Filtration Rate Slope as Primary Endpoint

Study	Trial Duration	Slope Calculation Method	Expected Effect Size	Statistical Power
NCT05196347	≥52 weeks	Linear mixed effects model, estimated glomerular filtration rate slope (ml/min/1.73m²/year)	0.75 ml/min/1.73m²/year difference	80% (power calculation)
NCT02160145	Up to 61 weeks	Chronic Kidney Disease Epidemiology Collaboration, annualized, weighted analysis of covariance	No mention found	No mention found
NCT03789695	24 months	Mixed model of repeated measurements, Chronic Kidney Disease Epidemiology Collaboration	No mention found	No mention found
NCT06676384	24 months	No mention found	No mention found	No mention found
NCT01194154	24 months	Modification of Diet in Renal Disease-4, simple linear regression	No mention found	No mention found
NCT06058585	108 weeks	No mention found	No mention found	No mention found
NCT05047263	32 months	No mention found	No mention found	No mention found
NCT07116928	120 weeks	No mention found	No mention found	No mention found
NCT07027774	52 weeks	No mention found	No mention found	No mention found
NCT05056727	104 weeks	No mention found	No mention found	No mention found

Summary of primary endpoint studies:

Trial duration:
- 1 study had a duration of 1 year or less (52 weeks).
- 5 studies had durations greater than 1 year and up to 2 years (ranging from 61 weeks to 24 months or 104 weeks).
- 3 studies had durations greater than 2 years (ranging from 108 weeks to 32 months or 120 weeks).
- We found mention of trial duration for all studies.
Slope calculation method:
- 1 study used a linear mixed effects model for estimated glomerular filtration rate slope.
- 1 study used Chronic Kidney Disease Epidemiology Collaboration with annualized, weighted analysis of covariance.
- 1 study used a mixed model of repeated measurements with Chronic Kidney Disease Epidemiology Collaboration.
- 1 study used Modification of Diet in Renal Disease-4 with simple linear regression.
- No mention found of the slope calculation method for 6 studies.
Expected effect size and statistical power:
- 1 study specified an expected effect size (0.75 ml/min/1.73m²/year difference) and statistical power (80%).
- No mention found of expected effect size or statistical power for the other 9 studies.

Estimated Glomerular Filtration Rate Slope as Secondary Endpoint

Study	Trial Duration	Slope Calculation Method	Expected Effect Size	Statistical Power
NCT02160145	Up to 61 weeks	Linear mixed model (secondary)	No mention found	No mention found
NCT06291376	106 weeks	No mention found	No mention found	No mention found
NCT00285298	1 year	No mention found	No mention found	No mention found
NCT03253172	2 years	No mention found	No mention found	No mention found

Summary of secondary endpoint studies:

Trial duration was more than 1 year in 3 studies and exactly 1 year in 1 study; no studies had a duration less than 1 year.
The slope calculation method was specified in 1 study (using a linear mixed model) and no mention found in 3 studies.
No mention found of expected effect size or statistical power in any of the studies.

Methodological Variations in Slope Calculation

Study	Slope Calculation Method	Analysis Set	Statistical Test	Multiplicity/Hierarchical Testing
NCT05196347	Linear mixed effects model	Intention-to-treat, per-protocol	Linear mixed model, Cox proportional hazards for secondary	Closed testing, hierarchical
NCT02160145	Weighted analysis of covariance, linear mixed model	No mention found	Analysis of covariance, linear mixed model	No mention found
NCT03789695	Mixed model of repeated measurements	No mention found	Mixed model	No mention found
NCT01194154	Simple linear regression (per patient)	Full analysis set, safety	Wilcoxon (Mann-Whitney)	No mention found
NCT04881448	Linear random slope model	Full analysis set	Shift analysis	No mention found

Key methodological points:

Slope calculation methods varied:
- Linear mixed effects models or mixed models were used in 2 studies.
- Weighted analysis of covariance was used in 1 study.
- Mixed model of repeated measurements was used in 1 study.
- Simple linear regression (per patient) was used in 1 study.
- Linear random slope model was used in 1 study.
Statistical tests included:
- Linear mixed models (2 studies)
- Analysis of covariance (1 study)
- Cox proportional hazards (for secondary analysis, 1 study)
- Mixed model of repeated measurements (1 study)
- Wilcoxon (Mann-Whitney) (1 study)
- Shift analysis (1 study)
Multiplicity or hierarchical testing:
- Closed testing with hierarchical procedures was described in 1 study.
- No mention found of multiplicity or hierarchical testing in the other 4 studies.

Summary

Use of estimated glomerular filtration rate slope as an endpoint:
- Among the included studies, 13 used estimated glomerular filtration rate slope as a primary outcome and 4 as a secondary outcome.
- The remaining studies used composite outcomes, time-to-event, or other non-slope outcomes.
Heterogeneity in methodology:
- There was substantial variation in how estimated glomerular filtration rate slope was defined, calculated, and reported.
- Slope calculation methods and statistical analyses were inconsistently described, with many studies not specifying key details.
Study populations and interventions:
- The studies included diverse populations (e.g., chronic kidney disease of various etiologies, diabetic kidney disease, kidney transplant recipients) and a wide range of interventions (pharmacologic, care models, education, diagnostics).
Reporting of effect size and power:
- Only one study reported both an expected effect size and statistical power for the estimated glomerular filtration rate slope endpoint.
- Most studies did not mention these parameters.
Generalizability:
- The diversity of study designs, populations, and analytic approaches limits direct comparability and synthesis of findings across studies.

Overall, these studies demonstrate frequent use of estimated glomerular filtration rate slope as a clinical trial endpoint in chronic kidney disease, but highlight the need for more consistent definitions and analytic methods to improve interpretability across trials.