DIALYSIS. EPIDEMIOLOGY, OUTCOME RESEARCH, HEALTH SERVICES 1

Abstract

Introduction and Aims: Although some guidelines recommend salt restriction, few studies have examined the association between salt restriction and clinical outcomes in hemodialysis (HD) patients. Methods: We conducted a retrospective cohort study of 88,115 adult patients enrolled in the Japanese Society for Dialysis Therapy (JSDT) registry (2008) who had received HD for at least two years and were considered anuric. The primary outcome measure was all-cause mortality at one year, and the secondary outcome was cardiovascular (CV) mortality. Estimated salt intake was the main predictor, and was calculated from interdialytic weight gain and pre-and postdialysis serum sodium levels according to the validated method of Kimura and Ramdeen. Nonlinear logistic regression was used to determine the association of salt intake with mortality, adjusting for age, gender, body mass index, vintage of HD, dialysis time, Kt/V, protein catabolic rate normalized to body weight, comorbid conditions, type of vascular access, serum potassium, phosphate, calcium, CRP level, and endotoxin level in dialysate. Cubic splines were plotted and the reference was median salt intake. Salt consumption was categorized by intake levels of 2 g per day and the association with mortality examined. Results: Median [25th-75th percentile] salt intake at baseline was 6.4 [4.6-8.3] g per day. At one year, all-cause mortality occurred in 1,845 (2.1%) patients, including cardiovascular mortality in 821 (0.9%). We observed an association between low salt intake and clinical outcomes (all-cause and CV mortality) (Fig.1). We observed the highest all-cause mortality in the low salt group (<6g/day) (Fig.2), and no association between all-cause mortality and high salt intake. Further, we observed similar associations between salt intake and CV mortality. Conclusions: Low salt intake is associated with all-cause and CV mortality. These findings do not support current clinical guidelines, which recommend restricting salt intake to less than 6g per day.. We extracted data for only those patients who survived at least 12 months after the start of HD. Ninety-four HD patients with pre-existing cancer were excluded from the analysis. We used available variables to construct regression models to predict cancer development during 2 year follow-up. Candidate predictors included demographic characteristics (age, sex), comorbidities (diabetes, hypertension, cardiovascular disease (CVD), coronary artery disease (CAD), smoking), body mass index (BMI), dialysis parameters (eKtV, urea reduction ratio (URR), vascular access, erythropoietin dosage, intra-dialytic weight gain (IDWG), normalized protein catabolic ratio (nPCR)) and laboratory tests (albumin, hemoglobin (Hgb), sodium, potassium, calcium, phosphorus, ferritin, serum creatinine (sCR), white blood cell count, platelets, and total cholesterol). Continuous variables were converted into categorical ones based on optimal clinical cut off points in a preprocessing step. Categorical variables were pre-screened using Chi-square test (P < 0.1). Stepwise forward method was used for variable selection in the multiple logistic regression. Results: 22024 HD patients were studied (Eastern Europe: 4830, Western Europe: 367, Northern Europe: 1937, Southern Europe: 7189, Western Asia: 2115, Northern America: 5586). The mean (SD) age was 63.2±15.0 years, 58.7% were males. The overall incidence of cancer was 0.84% (185 cases), and 0.3% (57 cases) of HD patients had cancer-related death. The incidence of cancer was highest in Eastern Europe (1.6%, 78 cases) and lowest in Western (0.3%, 1 case) and Northern Europe (0.3%, 5 cases). Men older than 75 years had trend towards higher cancer incidence as compared to older female (1.2% vs 0.7%; P =0.07). The multivariable logistic regression model to predict two-year risk of cancer retained the following variables: age, BMI, ferritin, albumin, Hgb, eKtV, vascular access, CAD, diabetes, and IDWG. The model, significant predictors, and unstandardized β coefficients with 95% CI are presented in Table 1. Conclusions: Our study identifies clinical relevant risk factors to predict cancer in HD patients. This risk assessment model could help clinicians to stratify patients for cancer screening, surveillance, prevention and early therapeutic intervention. Further studies are needed to validate our model in an externally derived cohort to evaluate its generalizability. Introduction and Aims: Assisted peritoneal dialysis (aPD) is now more available as an alternative to hospital haemodialysis (HD) for frail older patients but the lack of outcome data comparing HD with aPD has limited its use. FEPOD part 1 reported no significant difference in the primary outcomes of quality of life and physical functioning, except for higher prevalence of possible depression in the aPD group. This report describes the secondary outcomes for the study group. SECONDARY OUTCOMES: Hospitalisation, falls, symptom burden, cognition and patient satisfaction. Methods: aPD patients and HD patients were recruited from 11 centres. The HD patients were matched to recruited aPD participants by age, sex, diabetes status, time on dialysis, ethnicity and Index of Deprivation. The MiniMental State Examination (MMSE) and the Trail Making Form B were used to assess global cognitive function and executive function respectively. Falls and symptom burden were assessed using a falls questionnaire and the Palliative Outcome Symptom scale (renal) respectively. Patient satisfaction was measured using the Renal Treatment Satisfaction Questionnaire Results: 106 patients (52 HD; 54 aPD) were recruited. 35 % of the study group had at least one hospital admission in the preceding three months. 42% of all admissions were dialysis related. 28% of the study group had at least one fall in the preceding three months.83.3% of them occurred at home, with the HD group sustaining more fractures than the aPD group (26.7% HD vs. 6.7% aPD, p=0.329). Lethargy, pain and poor mobility were predominant in the study group. The median number of reported symptoms were 9 (IQR 7-11) in the HD group and 10 (IQR 7.75 to 13) in the aPD group. 42.3% of HD patients reported no improvement in symptoms since starting dialysis, as against 25.9% of the aPD group. 10.5% of the study group had abnormal MMSE scores (<24). There was no statistical difference in MMSE scores between HD patients and aPD patients [mean MMSE-27 (HD), 28 (aPD), p=0.120]. In contrast, 36.8% had executive dysfunction (trail making B test time > 300 seconds). Executive dysfunction was more prevalent in the aPD group [54.2% aPD vs. 27.7% HD, (p = 0.089)]. Despite the above outcomes, 91.5% of the study group would recommend their therapy to others (mean total renal treatment satisfaction scores-49.6 HD vs. 50.3 aPD, p=0.722). Conclusions: There is a high prevalence of falls, symptom burden, executive dysfunction and hospitalisation in frail elderly dialysis patients, irrespective of dialysis modality. This should be considered during discussions about renal replacement modalities (including non-dialytic care). FEPOD part 2, the longitudinal phase of the study, will provide information on the influence of dialysis modality on the trajectory of these outcomes.