Vascular access for hemodialysis: the impact on morbidity and mortality.

BACKGROUND: In chronic hemodialysis (HD), central venous catheter (CVC) use seems associated with an increased risk of death. This study, using registry data, evaluated the morbidity and mortality risk associated with the use of different permanent vascular access (VA) in a HD patient cohort.
METHOD: We evaluated hospitalization and death rate in prevalent and incident HD patients recorded in the uremic registry of Campania (southern Italy) for 2001. Patients were divided into three groups: CVC, artero-venous graft (AVG) and artero-venous fistula (AVF).
RESULTS: One hundred and eleven dialysis units in the Campania region (69%) provided data. A total of 2201 out of 3387 prevalent HD patients were included: 92 patients (4.2%) were on CVC, 24 patients (1.1%) were on AVG and 2085 patients (94.7%) were on AVF. In comparison with AVF, the CVC group had a greater prevalence of female gender, old age, diabetes, comorbidities, hypoalbuminemia, anemia, erythropoietin (EPO) resistance, and less frequent synthetic membrane use, but had a similar dialysis duration (hr/week). Similar data were collected in the 635 incident patients registered in 2001. During the study, in both prevalent and incident CVC patients, either hospitalization or death rates were enhanced; however, the difference in the relative risk (RR) of death disappeared after correction for age, gender, malnutrition, diabetes, hemoglobin, albumin and comorbidity. Among incident patients, survival analysis was performed in patients remaining on the same VA type throughout the follow-up period; while a similar survival between groups was demonstrated in the 1st year of follow-up, survival was worse in the CVC group during the 2nd year of follow-up; however, this difference also disappeared in the adjusted analysis.
CONCLUSION: This cohort study demonstrates that in chronic dialysis patients CVC choice, with respect to AVF, is mainly associated with female gender, advanced age and worse clinical conditions at baseline, and a worst outcome in both prevalent and incident CVC patients compared to AVF patients. Hospitalization, mortality rate and RR of death increased significantly; however, differences disappeared after correction for comorbidity. Therefore, these data suggest that CVC use per se is not associated with increased mortality risks with respect to AVF.