RDW-to-Albumin Ratio Linked to IgA Nephropathy Outcomes

Key Takeaways

• Among patients with biopsy-proven IgA nephropathy at West China Hospital in China, higher baseline RAR was significantly associated with the renal composite after adjustment and matching.
• RAR discriminated the endpoint better than either constituent measure, although its standalone discrimination remained modest.
• An apparent treatment interaction before matching was no longer significant afterward.

Investigators retrospectively analyzed 1169 patients with biopsy-proven IgAN in an established West China Hospital, Sichuan University, cohort study published in the Journal of Clinical Medicine. Enrollment spanned 2009–2018. Exclusions included secondary disease, insufficient data, inadequate biopsy, follow-up shorter than a year, and very low baseline kidney function. Baseline RDW-to-albumin ratio (RAR) was RDW (%) divided by serum albumin (g/L). A receiver operating characteristic (ROC) analysis using the maximum Youden index established a cohort-specific high/low threshold; RAR was also modeled continuously. The composite endpoint was end-stage renal disease, defined as estimated glomerular filtration rate (eGFR) below 15 mL/min/1.73 m² or maintenance renal replacement treatment, and/or more than a 50% decline in eGFR from baseline. Investigators used Kaplan–Meier curves, multivariable Cox models, propensity score matching, and splines, adjusting for demographics, proteinuria, kidney function, hematuria, blood pressure, metabolic and hematologic factors, Oxford biopsy lesions, and treatment.

RAR's area under the ROC curve (AUC) for the composite endpoint was 0.648. This exceeded albumin and RDW individually in significant DeLong comparisons but indicated modest standalone discrimination. The cohort-derived high/low cutoff was 0.359; matching yielded 145 patients in each group. Before matching, the high-RAR group had more adverse clinical and pathological features and more frequent corticosteroid or immunosuppressant use. Most measured characteristics were balanced afterward.

Before matching, high versus low RAR was associated with the composite endpoint in the fully adjusted model: hazard ratio (HR) 1.880 (95% confidence interval [CI] 1.222–2.893; p = 0.004). Endpoint-free survival was poorer with high RAR, and continuous RAR was also significantly associated with the endpoint.

After matching, the fully adjusted high-versus-low estimate was HR 2.128 (95% CI 1.067–4.243; p = 0.032). Endpoint-free survival remained poorer with high RAR, and the continuous-RAR association remained significant. Splines showed significant overall associations before and after matching, without establishing a strictly linear relationship. Associations were generally similar across sex, proteinuria, anemia, kidney-function, and treatment subgroups; the treatment interaction seen before matching was no longer significant afterward.

The authors noted the single-center retrospective design, possible residual confounding, cohort-optimized cutoff, modest discrimination, limited matched renal events with possible overfitting, differing follow-up duration between matched groups, and exploratory subgroup comparisons. Matched models accommodated proportional-hazards violations with time-varying blood-pressure covariates. No comparison with a validated multivariable IgAN risk model established added predictive value, and the cutoff was not validated as a clinical decision threshold. The authors concluded that the association observed in this Chinese cohort needs external validation and larger prospective multicenter studies before RAR's clinical value can be established.

Clinician Questions

Why might red blood cell distribution width-to-albumin ratio track outcomes in IgA nephropathy?

The authors propose that RDW may reflect variation in red-cell size related to inflammation or oxidative stress, whereas serum albumin may be affected by proteinuria and inflammatory activity. Combining them may capture different aspects of disease burden, but the molecular mechanism linking RAR to IgA nephropathy has not been identified.

How were proportional-hazards violations handled in matched IgA nephropathy RAR comparisons?

In the matched IgA nephropathy cohort's RAR analysis, investigators identified global proportional-hazards violations in the more adjusted models, but none for RAR itself. They incorporated time-varying effects for systolic and diastolic blood pressure through interactions with log(time).