Early start hemodialysis with a catheter may be associated with greater mortality: A propensity score analysis

• Published in: Seminars in dialysis Vol. 36; no. 4; pp. 294 - 302
• Main Authors: Laham, Gustavo, Pujol, Gervasio Soler, Guzman, Jenny, Boccia, Natalia, Abib, Anabel, Diaz, Carlos H.
• Format: Journal Article
• Language: English
• Published: United States 01.07.2023
• Subjects: Catheters; Humans; Kidney Failure, Chronic - therapy; Propensity Score; Renal Dialysis - adverse effects; Retrospective Studies
• ISSN: 0894-0959; 1525-139X
• DOI: 10.1111/sdi.13157

Introduction Deciding when and how to initiate hemodialysis (HD) is still controversial. An early start (ES) seems to show a lack of benefit. “Lead time bias” and comorbidities have been associated with different outcomes in ES groups. On the other hand, it is well accepted that the impact the type of vascular access (VA) has on patient survival. Our aim was to evaluate survival with early start (ES) versus late start (LS) on HD, taking into account the vascular access (VA) used. Methods Between 01/1995 and 06/2018, 503 incidental patients initiated HD at our Dialysis Unit. eGFR was estimated by the CKD‐EPI equation. Diabetes mellitus (DM), coronary disease (CD), and peripheral vascular disease (PVD) were considered comorbid conditions. According to eGFR and VA, patients were divided into four groups: G1: ES (eGFR > 7 mL/min) with catheter (ES + C), G2: ES with fistula or graft (F/G) (ES + F/G), G3: LS (eGFR< 7 mL/min) with catheter (LS + C), and G4: LS with F/G (LS + F/G). The cut‐off value to define ES or LS was based on median eGFR for these 503 patients. We compared patient's survival rates by Kaplan–Meier and log‐rank test. The four groups were compared before and after matching with propensity scores (PS). Cox analysis was performed to determine the impact of predictors of mortality. Results Median eGFR was 7 (5.3–9.5) mL/min/1.73 m2, median follow‐up time was 30.9 (13–50) months, 52.1% had F/G access at entry, and 46.9% died during the observation period. Among the four groups, the ES + C were significantly older, and there were more diabetics and comorbid conditions, while phosphatemia, iPTH, albumin, and hemoglobin were significantly higher in the LS groups. Before propensity score (PS) matching, the ES + C group had a poor survival rate (p < 0.0001), while LS + F/G access had the best survival. After PS, a total of 180 patients were selected in the same four groups and ES + C kept showing a statistically significant poorer survival. Multivariate analysis revealed that ES + C was an independent predictor of mortality. Conclusion In this retrospective study, ES + C on HD was associated with a higher mortality rate than LS. This association persisted after PS matching.