Background: The adjustment of comorbidity is important in international hemodialysis comparisons. The aim of this study is to verify if it is possible to use the Charlson Comorbidity Index (CCI), in an Italian population of incident hemodialysis patients from the Campania region as outcome predictors. A similar proposal has already been made for an American population of incident patients in peritoneal dialysis.
Methods: The data for this study come from the Uremic Registry of Campania taken in the year 2001. This is an observational study in which demographic, comorbid, laboratory, treatment and insurance data were collected in 111 dialysis units (70%) in Campania. We evaluated 515 hemodialysis incident patients who were hemodialyzed in Campania between January 1 and December 31, 2001. The study was restricted to patients who had already undergone 90 days of hemodialysis. The duration of this study was 15 months. Charlson Comorbidity Index was performed. In 128 patients (24.8%) BIA measurements were performed after dialysis.
Statistics: We used Student's t test for unpaired data and Cox proportional model to analyze predictors of mortality. The variables analyzed were age at start of hemodialysis, sex, CCI, hemoglobin, diabetes, hypertension, albumin, days of hospitalization. The statistically significant variables, analyzed initially by univariate analysis, were chosen for multivariate analysis. We considered p < 0.05 statistically significant.
Results: A total of 515 patients (M = 316, F = 199) (age: 63.62 +/- 15.35 years) presented with the following diseases: NO diagnosed in 93 patients (19%), GN in 64 (13%) IN in 42 (99%), Hereditary in 55 (11%), Vascular in 66 (14%), Diabetes in 135 (28%), others in 30 (6%). Hemoglobin levels were 10.71 +/- 1.51 g/dL and albumin was 3.79 +/- 0.54 g/dL. The days of hospitalization for the population studied were 3364/year. After the study, 75 patients died and the overall mortality rate was 11.65/100 patient/years. Univariate analysis shows that there are significant differences calculated for age (median value 73 and 65 years, respectively for non-survivers and survivers), BMI (median value and 22 and 24 kg/m2, respectively), Hb (median value 9.5 and 11 g/dL, respectively), Albumin (median value 3.5 and 3.8 g/dL, respectively), days of hospitalization (median value 8 and zero days, respectively), CCI (median value 6 and 4 score, respectively, phase angle (median value 3.3 and 4 degree, respectively). The mortality rates (100 dialysis years) by the CCI score: the mortality rate was zero for patients with a CCI of 3; and it increased to approximately 60% of patient years with a CCI score of 6 or greater. The linear correlation between CCI and phase angle in living (y = 18.90 x -3.83; R2 = 0.56) and in the dead (y = 13.01 x -1.87; R2 = 0.29).
Discussion: We found that CCI is a strong predictor of mortality in incident HD patients as has also been indicated in PD patients; CCI correlates with phase angle calculated from Bioelectrical Impedance Analysis and this last factor can be used in the following examinations; several days of hospitalization are a very important determinant in the survival in hemodialysis patients.