Time-updated hematuria and Kidney Disease Progression in IgA nephropathy: a cohort study

Hematuria, whether macroscopic or asymptomatic microscopic, was widely acknowledged as a typical clinical manifestation of IgAN nephropathy (IgAN). To date, prior studies have demonstrated that both the severity and persistence of hematuria are independent predictors of poor renal outcomes in IgAN, with remission status highlighting therapeutic benefits and better renal survival.Despite its prognostic significance, hematuria remained underutilized in clinical practice, which was partly due to the absence of standardization and daily variability in hematuria. More importantly, throughout the longitudinal follow-up, time-dependent confounders could lead to biased estimates, since hematuria tends to increase as kidney function deteriorates, and, in turn, elevated levels of hematuria may influence future kidney function.To fill these essential knowledge gaps, causal inference analysis methods were used to conduct an in-depth analysis of the association between hematuria and kidney outcomes in IgAN, based on longitudinal hematuria measurements from an observational cohort.

This prospective observational cohort study was conducted at Peking University First Hospital (2003 – June 2024). We enrolled adult patients with biopsy-proven IgAN who had ≥12 months of follow-up and ≥2 clinical visits. Exclusion criteria were: (1) secondary IgAN; (2) CKD stage 5; (3) missing key baseline or longitudinal data.This resulted in 1792 participants being included.The primary exposure of interest was hematuria, which was treated as a time-varying covariate. Data on hematuria status were obtained every 6 months during the study period, measured using a fully automated urine particle analyzer. Both baseline and time-updated hematuria were analyzed as a 3-level categorical variable: (1) Minimal: <20 RBC/µL; (2) Mild: 20 to <100 RBC/µL; (3) Moderate/Severe: ≥100 RBC/µL.The primary outcome was progression of IgAN, defined as a composite of ≥40% decline in eGFR from baseline or the first occurrence of ESKD (eGFR <15 mL/min/1.73 m², dialysis initiation, or kidney transplantation). Participants without a follow-up visit for more than one year were considered lost to follow-up and censored at their last visit date. Death prior to reaching the primary outcome was treated as a competing risk. Baseline covariates, including age, sex, body mass index, albumin, proteinuria, eGFR, blood pressure, diabetes history and MEST-C scores. Time-dependent variables were extracted from the database at 6-month intervals and included age, albumin, proteinuria, eGFR, BP, BP, the usage of RAS inhibitors, other antihypertensive medications (α-blocker, β-blocker, calcium channel blocker, diuretic), hydroxychloroquine, SGLT-2 inhibitors, corticosteroids and other immunosuppressants. When there were multiple measurements within a 6-month period, the closest measurement was used for each time point.To investigate the association of baseline and time-updated hematuria with kidney disease progression in IgAN, we used multivariable Cox proportional hazards regression model and marginal structural model (MSM), respectively. Considering the time-dependent confounding and potential causal effects (Figure 1), stabilized inverse probability of treatment and censoring weights were calculated.

The trends of  patients with different time-updated hematuria were shown in Figure 2. During 10,015 person-years of follow-up (median 4.1 years [IQR, 2.3–7.7]), the primary composite kidney outcome was observed in 489 participants (27.3%). Compared to participants with hematuria < 20 RBC/µL, the incidence rates of the composite kidney outcome were comparable across higher baseline hematuria categories(Table 1). 

We assessed the association between baseline hematuria levels and the risk of the composite kidney outcome using Cox proportional hazards models adjusted for baseline covariates. In the fully adjusted model, no graded association was observed between baseline hematuria and the outcome. Compared with the reference group (hematuria <20 RBC/µL), the HRs (95% CI) for the hematuria categories of 20 to <100 RBC/µL and ≥100 RBC/µL were 0.82 (0.62–1.08) and 0.92 (0.71–1.21), respectively. We then evaluated the association between time-updated hematuria and the composite kidney outcome using marginal structural Cox models, accounting for time-dependent confounders and hematuria fluctuations. Notably, a significantly increased risk was observed at time-updated hematuria levels of ≥100 RBC/µL (HR = 1.56; 95% CI, 1.13–2.15). Conversely, the category 20 to <100 RBC/µL showed no significant risk increase versus the reference group (HR = 1.22; 95% CI, 0.94–1.56) (Table 2).

Finally, we tested interactions between time-updated hematuria and pre-specified subgroups for the primary outcome. MSMs demonstrated a significant graded associations specifically in participants with eGFR <60 mL/min/1.73m² or proteinuria ≥0.5 g/day (P for interaction <0.001). In these high-risk subgroups, even moderate hematuria (20 to <100 RBC/µL) conferred elevated risk(Figure 3).

Persistent hematuria ≥100 RBC/µL was independently associated with kidney disease progression in IgAN. This threshold required external validation and should be considered as a criterion for high-risk patients in future clinical trials.