An individual patient meta-analysis assessed the associations between change in albuminuria and acute changes in the glomerular filtration rate in randomized controlled trials of chronic kidney disease progression.

INTRODUCTION

Early changes in albuminuria and glomerular filtration rate (GFR) have been proposed as markers of treatment efficacy for individual patients. Here, we evaluated the relationship between changes in GFR and in albuminuria in the immediate period following initiation of medication across trials and within individual patients.

METHODS

In 49 randomized controlled trials encompassing 66,449 eligible participants, we estimated early changes in GFR from baseline to three months and in albuminuria from baseline to six months. We first used Bayesian mixed-effects meta-regression to relate treatment effects on early GFR change (acute effects) with treatment effects on albuminuria across trials. Second, we examined associations between an individual's change in GFR and albuminuria using adjusted linear regression in both the treatment and control arms.

RESULTS

There was no evidence of association between mean treatment effects on GFR and albuminuria across studies [meta-regression slope 0.00 (95% Bayesian credible intervals, BCI -0.08, 0.09), R2 0.03 (95% BCI 0.03, 0.27)]. For individual patients, similar magnitude of associations between changes in albuminuria and in GFR were observed in both arms: for each standard deviation greater decline in GFR, the geometric mean albumin/creatinine ratio declined by 8.0% (7.1%, 8.9%) in participants in the treatment arm and by 6.6% (5.6%, 7.6%) in those in the control arm. The effect was seen in each of the intervention classes except for the immunosuppressive studies, where this association was not observed within either the intervention or control arms.

CONCLUSIONS

The similar associations between changes in GFR and albuminuria for individual patients in both treatment and control arms and no association between the treatment effects on GFR and on albuminuria at the trial level. This suggests that early changes in GFR and albuminuria may be more likely related to natural history of the disease or other factors, rather than the effect of treatments.