**Background:** Hemodialysis (HD) is the most common kidney replacement therapy for end-stage kidney disease, yet mortality remains high—especially from cardiovascular disease (CVD), which accounts for nearly 50% of deaths. Chronic inflammation, immune dysfunction, and malnutrition are known risk factors for adverse outcomes in HD patients, but existing single-marker prognostic indicators have limited predictive value. The hemoglobin, albumin, lymphocyte, and platelet (HALP) composite score has emerged as a novel index integrating nutritional and inflammatory status and has shown prognostic value in cancers, heart failure, and stroke, but its role in HD patients was previously unknown.
**Methods:** This retrospective multicenter study analyzed data from 4 large HD centers in China, sourced from the Chinese National Renal Data System (www.cnrds.net). Patients aged >18 years on maintenance HD for >3 months between January 2008 and October 2022 were included. Exclusion criteria were recent invasive operations (except dialysis access), acute infection, autoimmune disease, malignancy, hemopathy, use of anti-inflammatory/antiplatelet/immunosuppressive drugs, missing baseline data, or no outcome data. The final cohort comprised 4,796 patients. HALP was calculated as: hemoglobin (g/L) × albumin (g/L) × lymphocytes (/L) / platelets (/L). Patients were divided into tertiles: Tertile 1 (<20.3, n=1,599), Tertile 2 (≥20.3 to ≤31.6, n=1,599), and Tertile 3 (>31.6, n=1,598). Primary endpoints were all-cause death and CVD death (including cardiac failure, coronary heart disease, pulmonary edema, stroke, arrhythmia, and other fatal CVD). Statistical methods included Kaplan-Meier curves with log-rank tests, Cox proportional hazards regression, Fine-Gray competing risk models, restricted cubic splines, time-dependent ROC curves, and C-index analyses.
**Key Results:** Mean follow-up was 65 months (IQR: 36–96). At study end, 1,907 patients had died, of whom 1,024 (53.6%) died from CVD. Baseline characteristics differed significantly across HALP tertiles: patients in Tertile 1 were older (57.6 vs. 53.5 years, P<.001), had more diabetes (37.9% vs. 27.1%, P<.001), higher prior CVD history (34.3% vs. 21.3%, P<.001), lower hemoglobin (80.7 vs. 111.9 g/L, P<.001), lower albumin (33.4 vs. 39.0 g/L, P<.001), lower lymphocyte count (1.0 vs. 1.7 ×10⁹/L, P<.001), higher platelet count (209.9 vs. 177.0 ×10⁹/L, P<.001), higher CRP (3.8 vs. 2.3 mg/L, P<.001), higher phosphate (1.79 vs. 1.56 mmol/L, P<.001), and lower calcium (2.06 vs. 2.22 mmol/L, P<.001). Kaplan-Meier analysis showed significantly better overall survival in Tertile 3 (log-rank χ²=62.4, P<.001). In fully adjusted Cox models, continuous HALP was associated with lower all-cause mortality (HR=0.96, 95% CI: 0.95–0.97, P<.001). Compared to Tertile 1, Tertile 3 had a 34% lower risk of all-cause death (HR=0.66, 95% CI: 0.49–0.86, P=.007). The Fine-Gray competing risk model showed Tertile 3 had a 49% lower risk of CVD death (sub-distribution HR=0.51, 95% CI: 0.34–0.80, P=.005). Restricted cubic splines confirmed a linear negative relationship between HALP and all-cause mortality (P for non-linearity=0.436). Adding HALP to a traditional risk factor model improved predictive performance: the traditional factors + HALP model had an AUC of 0.687 at 1 year, 0.752 at 3 years, and 0.763 at 5 years (vs. 0.641, 0.689, and 0.693 for the traditional model alone; all P<.001). The C-index improved from 0.71 (95% CI: 0.66–0.75) to 0.76 (95% CI: 0.72–0.79), with net reclassification improvement of 0.09 (P=.001) and integrated discrimination improvement of 0.33 (P<.001).
**Clinical Implications:** This large, multicenter study provides the first evidence that lower baseline HALP is independently and linearly associated with increased all-cause and CVD mortality in hemodialysis patients. HALP integrates multiple pathophysiological domains—anemia, malnutrition, inflammation, and platelet/immune function—into a single, inexpensive, and routinely available score. Adding HALP to traditional risk models significantly improved mortality prediction, suggesting it could be a practical tool for risk stratification in clinical practice. The findings support HALP as a novel prognostic indicator that may help identify high-risk HD patients who could benefit from intensified nutritional support, anti-inflammatory interventions, or closer cardiovascular monitoring. Limitations include the retrospective design, potential selection bias, inability to account for dynamic changes in medications or dialysis adequacy, and lack of an established optimal HALP cutoff. Prospective validation studies are needed.