A greater number of selected clinical abnormalities is associated with higher erythropoiesis resistance index in maintenance hemodialysis: a repeated patient-month analysis

Background Erythropoiesis resistance in hemodialysis patients is associated with multiple clinical factors, including nutritional impairment, inflammation, and intercurrent illness. We examined whether the number of selected clinical abnormalities present within the same patient-month was associated with concurrent erythropoiesis resistance index (ERI). Methods We conducted a single-center retrospective study using repeated patient-month data over 12 months. Patients receiving hypoxia-inducible factor prolyl hydroxylase (HIF-PH) inhibitors were excluded. For each patient-month, we calculated an unweighted count of four prespecified clinical abnormalities: low body mass index (BMI; < 20.0 kg/m 2 ), hypoalbuminemia (albumin < 3.5 g/dL), elevated C-reactive protein (CRP ≥ 0.3 mg/dL), and an acute clinical event. Linear mixed-effects models with patient-specific random intercepts were used to evaluate the contemporaneous association between the abnormality count and log(ERI + 1), with adjustment for TSAT and ferritin. Results A total of 56 patients contributed 672 patient-month observations. ERI increased across categories of 0, 1, and ≥ 2 selected clinical abnormalities. In the mixed-effects model, each additional abnormality was associated with higher log(ERI + 1) ( β = 0.082, 95% CI 0.049–0.115, p < 0.001), corresponding to approximately 8.6% higher ERI + 1. Findings were consistent in sensitivity analyses using alternative component thresholds, categorical modeling of the abnormality count, weekly ESA dose as the outcome, and exclusion of patient-months with absolute iron deficiency. Conclusions In maintenance hemodialysis patients, a greater number of selected clinical abnormalities was associated with higher concurrent ERI across repeated patient-month observations. This contemporaneous association does not establish temporal precedence or validate the abnormality count as a clinical prediction score.