External validation of prediction models for mortality in an incident dialysis population
.

AIMS: Different prediction models have been established to estimate mortality in the dialysis population. This study aims to externally validate the different available mortality prediction models in an incident dialysis population. MATERIALS: This was a retrospective cohort study of incident hemodialysis and peritoneal dialysis patients at two academic tertiary care centers. METHODS: Three previously published prediction models were used: the Liu index, the Urea5 score, and a predictive model estimating the survival probability by Hemke et al. [6]. Models were compared using the C-statistic, net reclassification index, and integrated discrimination improvement. Only the subgroup of 193 patients with enough data to be included in all models was used. RESULTS: 377 patients were started on dialysis in both institutions between 2006 and 2011. Median follow-up was 787 days. 104 patients (27.6%) died during follow-up and 181 were admitted to the hospital (48.0%). All three models were predictive of mortality and hospital admissions. The survival probability model by Hemke et al. [6] performed better than the other two models for mortality (C-statistic 0.72). The Liu index had the highest performance for hospital admissions (C-statistic 0.65). Using reclassification statistics (reference = Urea5), the only model to improve discriminatory ability was the Liu index for the outcome of hospital admission. CONCLUSION: The survival probability model by Hemke et al. [6] may be preferred for mortality prediction in incident dialysis patients. The Liu index could be used to predict hospital admissions in the same population. Available models demonstrated only modest performance in predicting either outcome. Therefore, alternative models need to be developed.
.

Serial versus single troponin measurements for the prediction of cardiovascular events and mortality in stable chronic haemodialysis patients.

AIM: This study aims to describe the variability of pre-dialysis troponin values in stable haemodialysis patients and compare the performance of single versus fluctuating or persistently elevated troponins in predicting a composite of mortality and cardiac arrest, myocardial infarction or stroke. METHODS: A total of 128 stable ambulatory chronic haemodialysis patients were enrolled. Pre-dialysis troponin I was measured for three consecutive months. The patients were followed for 1 year. A troponin elevation (>0.06 µg/L) was considered high risk, and patients were classified into three risk groups: (i) patients who had normal troponin levels on all three measurements; (ii) patients with at least one elevated and one normal troponin value; and (iii) patients with elevated troponin values on all measurements. RESULTS: A total of 81 patients had all three troponin values in the normal range; 29 had fluctuating values; 18 had all three values elevated. Twenty-seven deaths or composite events were observed: eight in the first risk group, 10 in the second and nine in the third. Persistently elevated and fluctuating troponin values were associated with higher mortality and cardiovascular event rate. Serial troponin measurement had a higher sensitivity for the composite outcome than single troponin measurement when either fluctuating or persistently elevated values were considered to confer high risk. CONCLUSION: Most haemodialysis patients do not have elevated troponin levels at baseline. Troponin levels that remain elevated or fluctuate are associated with worse outcomes. A serial troponin measurement strategy is associated with better sensitivity and higher negative predictive value compared with single troponin measurement.

High Ultrafiltration Rates Increase Troponin Levels in Stable Hemodialysis Patients.

BACKGROUND: An elevated troponin level is commonly found in asymptomatic patients on hemodialysis (HD) and is associated with higher risk of mortality and major adverse cardiovascular events. The underlying mechanism for the association between adverse outcomes and elevated troponin levels has not been elucidated. METHODS: Two hundred thirty-six stable chronic HD patients from 2 tertiary care centers were enrolled in this study. We measured pre-dialysis troponin I levels with routine monthly bloods for 3 consecutive months. Troponin I was considered to be elevated if it exceeded the laboratory reference range of 0.06 µg/l. RESULTS: The study population had a mean age of 67.5, 56% were male, 47% had diabetes and 28% had pre-existing coronary artery disease. Eighty-eight positive troponin values were recorded (13% of the available values) in 52 patients. In a repeated measures linear random effects model (univariate analysis), high ultrafiltration (UF), high inter-dialytic weight gain, and duration of the dialysis session, but not intra-dialytic hypotension, were associated with troponin I elevation. In the multivariate model, only high UF explained troponin I elevation (p = 0.04). The intraclass correlation coefficient was found to be 5.8%, suggesting that observed variability is within and not between subjects, with session-related parameters being more important than inter-individual differences. CONCLUSIONS: A high UF rate during HD is associated with a biochemical evidence of myocardial injury. If confirmed, efforts to avoid rapid UF, protect residual kidney function or minimize weight gain between sessions may impact cardiovascular outcomes in this high-risk population.

Elevated troponin I levels but not low grade chronic inflammation is associated with cardiac-specific mortality in stable hemodialysis patients.

BACKGROUND: Elevated cardiac troponin I (TnI) levels are associated with all-cause mortality in stable hemodialysis patients. Their relationship to cardiac-specific death has been inconsistent, and the reason for their elevation is not well understood. We hypothesized that elevated TnI levels in chronic stable hemodialysis patients more specifically track with cardiac mortality, and this mechanism is independent of other contributors of cardiac mortality, such as inflammation. METHODS: We conducted a single-centre, cohort study of prevalent hemodialysis patients at a tertiary care hospital. Plasma TnI levels were measured with routine monthly blood tests in clinically stable patients for two consecutive months. Plasma TnI was measured by immunoassay and a value above the laboratory reference range (0.06 µg/L) was considered elevated. The primary outcome of death was adjudicated separately for this study, and classified as cardiac, non-cardiac, or unknown. Cox proportional hazard models were used to examine the association of TnI with the all-cause and cardiac-specific mortality, adjusting for potential confounders, including C-reactive protein (CRP) as a marker of inflammation. RESULTS: Of 133 patients followed for a median of 1.7 years, there were 38 deaths (58% non-cardiac, 39% cardiac, 3% unknown). Elevated TnI was associated with adjusted HR for all-cause mortality of 2.57 (95% CI 1.30-5.09) and an adjusted HR for cardiac death of 3.14 (95% CI 1.07-9.2), after accounting for age, time on dialysis, diabetes status, prior coronary artery disease history, and C-reactive protein. Although CRP was also independently associated with all-cause mortality, it did not add prognostic information to TnI for cardiac-specific death. CONCLUSION: Elevated TnI levels are independently associated with cardiac and all-cause mortality in asymptomatic hemodialysis patients. The mechanism for this risk is likely independent of inflammation, but may reflect chronic subclinical myocardial injury or unmask those with subclinical atherosclerotic heart disease. Whether those with elevated TnI levels may benefit from additional investigations or more aggressive therapies to treat cardiovascular disease remains to be determined.

A prospective evaluation of renal replacement therapy modality eligibility.

BACKGROUND: Patient eligibility for renal replacement therapy (RRT) modalities is frequently debated, but little prospective data are available from large patient cohorts. METHODS: We prospectively evaluated medical and psychosocial eligibility for the three RRT modalities in patients with chronic kidney disease (CKD) stages III-V who were enrolled in an ongoing prospective cohort study conducted at seven North American nephrology practices. RESULTS: Ninety-eight percent of patients were considered medically eligible for haemodialysis (HD), 87% of patients were assessed as medically eligible for peritoneal dialysis (PD) and 54% of patients were judged medically eligible for transplant. Age was the leading cause of non-eligibility for both PD and transplant. Anatomical concerns (adhesions, hernias) were the second most frequent concern for PD eligibility followed by weight. Weight was also a concern for transplant eligibility. The proportion of patients medically eligible for RRT did not vary by CKD stage. There was, however, significant inter-centre variation in the proportion of patients medically eligible for PD and transplant. Ninety-five percent of patients were considered psychosocially eligible for HD, 83% of patients were assessed as psychosocially eligible for PD and 71% of patients were judged psychosocially eligible for transplant. The percentage of patients who were assessed as having neither medical nor psychosocial contraindications for RRT was 95% for HD, 78% for PD and 53% for transplant. CONCLUSIONS: Most CKD patients are considered by their medical care providers to be suitable for PD. Enhanced patient education, promotion of home dialysis for suitable patients and empowerment of patient choice are expected to augment growth of home dialysis modalities.

Predialysis psychoeducational intervention extends survival in CKD: a 20-year follow-up.

BACKGROUND: Predialysis psychoeducational interventions increase patient knowledge about chronic kidney disease (CKD) and its treatment and extend time to dialysis therapy without compromising physical well-being in the short run. The present research examines long-term survival after predialysis psychoeducational intervention. In addition, we examined whether survival differed because of early (ie, > or = 3 months) versus late referral to nephrology. METHODS: We collected follow-up data for patients with CKD who participated in a multicenter randomized controlled trial of predialysis psychoeducational interventions in the mid-1980s. We gathered 20-year survival data from clinical records and databases. RESULTS: Participants included 335 patients with CKD, including 172 patients randomly assigned to receive predialysis psychoeducational interventions (63.0% men; mean age, 50.8 years) and 163 patients assigned to usual care (62.1% men; mean age, 52.7 years). Two hundred forty-six patients (66.8%) died during the course of the study. Mean duration of follow-up was 8.5 +/- 7.23 (SD) years. Analyses were by intention to treat. Adjusting for age, general nonrenal health at inception, and time between identification and predialysis psychoeducational intervention or usual care, Cox proportional hazards multiple regression analyses indicated that median survival was 2.25 years longer after patients with CKD received predialysis psychoeducational interventions compared with usual care (chi-square-change [1] = 3.75; P = 0.053; hazard ratio, 1.32; 95% confidence interval, 1.0 to 1.74). Predialysis psychoeducational intervention recipients survived a median of 8.0 months longer than usual-care patients after the initiation of dialysis therapy (chi-square-change [1] = 4.39; P = 0.036; hazard ratio, 1.35; 95% confidence interval, 1.02 to 1.775). No significant survival advantage was evident for early referral to nephrology or the combination of early referral plus predialysis psychoeducational interventions. CONCLUSION: Predialysis psychoeducational intervention is a safe and useful intervention that contributes valuably to multidisciplinary predialysis care.

Cobalamin dose regimen for maximum homocysteine reduction in end-stage renal disease.

Plasma total homocysteine (tHcy) concentrations are markedly increased in end-stage renal disease and only partially corrected by folic acid supplementation. We and others have reported that cobalamin, administered parenterally, reduces plasma tHcy substantially below the lowest concentrations attainable with folic acid. We have now carried out a randomized controlled clinical trial to compare the plasma Hcy-lowering effect of 3 intravenous cyanocobalamin dose regimens in maintenance hemodialysis patients: 1 mg postdialysis every 28, 14, and 7 days in addition to routine oral vitamin B supplementation. All patients in the hemodialysis unit where the study was carried out routinely received 1 mg intravenous cyanocobalamin every month, so participants who were randomized to receive the vitamin every 28 days simply continued with their existing treatment program. Serum cobalamin and plasma tHcy concentrations in the control group did not change over the course of the study. As measured after 8 weeks of therapy, intravenous cyanocobalamin every 14 days increased serum cobalamin approximately 2.5-fold and reduced plasma tHcy by 11.5% ( P = .035) below the concentration previously attained with monthly administration, whereas treatment every 7 days increased serum cobalamin concentrations approximately 5-fold and reduced plasma tHcy by 11.0% ( P = .013). These results show that intravenous cyanocobalamin at 7- or 14-day intervals reduces plasma tHcy concentrations of hemodialysis patients below the levels brought about by prior long-term administration every 4 weeks and confirms that plasma tHcy lowering with parenteral cobalamin is a true pharmacological effect and not merely correction of a latent deficiency state.

Predialysis psychoeducational intervention and coping styles influence time to dialysis in chronic kidney disease.

BACKGROUND: Consensus endorses predialysis intervention before the onset of end-stage renal disease. In a previous study, predialysis psychoeducational intervention (PPI) extended time to dialysis therapy by a median of 6 months. We undertook to replicate and extend this finding by examining hypothesized mechanisms. METHODS: We used an inception-cohort, prospective, randomized, controlled trial with follow-up to evaluate an intervention that included an interactive 1-on-1 slide-supported educational session, a printed summary (booklet), and supportive telephone calls once every 3 weeks. Participants were sampled from 15 Canadian (tertiary care) nephrology units and included 297 patients with progressive chronic kidney disease (CKD) expected to require renal replacement therapy (RRT) within 6 to 18 months. The main outcome was time to dialysis therapy (censored at 18 months if still awaiting RRT). RESULTS: Time to dialysis therapy was significantly longer (median, 17.0 months) for the PPI group than the usual-care control group (median, 14.2 months; Cox's proportional hazards analysis, controlling for general nonrenal health, P < 0.001). Coping by avoidance of threat-related information (called blunting) was associated with shorter times to dialysis therapy (P < 0.032). A group x blunting interaction (P < 0.069) indicated: (1) time to dialysis therapy was shortened in the usual-care group, especially when patients coped by blunting; but (2) time to dialysis therapy was extended with PPI, even among patients who coped by blunting. Knowledge acquisition predicted time to dialysis therapy (r = 0.14; P < 0.013). Time to dialysis therapy was unrelated to depression or social support. CONCLUSION: PPI extends time to dialysis therapy in patients with progressive CKD. The mechanism may involve the acquisition and implementation of illness-related knowledge. Routine follow-up also may be especially important when patients cope by avoiding threat-related information.

Convergence of health ratings across nephrologists, nurses, and patients with end-stage renal disease.

We evaluated a health rating for renal failure patients that was completed by patients, nurses, and nephrologists. The study was a prospective inception-cohort follow-up design. Measurements were taken before initiating dialysis (n=206) or at the initiation of dialysis (n=200) and at 18 (n=225), 30 (n=181), 42 (n=162), 54 (n=137), and 66 (n=112) months after initiating dialysis. Patients, nurses, and nephrologists independently rated patients' health at each measurement occasion. Objective measures of health status, abstracted from the medical record, included emergency and non-emergency admissions, smoking, diabetes mellitus, pulmonary edema, history and number of myocardial infarctions (MI), basal rales, comorbid illnesses, and uremic symptoms. Simultaneous multiple regression analyses examined the correspondence between objective measures of health status and subjective health ratings separately for each rater and measurement occasion. Health ratings were averaged and submitted to the same analyses. Raters showed good agreement (average Pearson r=.43 overall), although agreement was higher between nephrologists and nurses (average r=.64) than between health professional and patients (average r=.34 and .31, respectively). All three ratings and the combined rating corresponded significantly to objective measures of health status. Uremic symptoms, emergency hospital admissions, diabetes mellitus, and recent MI correlated uniquely and most consistently with subjective health ratings. Despite overall convergence, objective measures of health status related to the groups' ratings in a complementary fashion. The health rating is reliable and relates to the current status of the patient. Performance was superior for the combined score that incorporated ratings by patients, nurses, and nephrologists.

Serial change in echocardiographic parameters and cardiac failure in end-stage renal disease.

Echocardiographic abnormalities are the rule in patients starting dialysis therapy and are associated with the development of cardiac failure and death. It is unknown, however, whether regression of these abnormalities is associated with an improvement in prognosis. As part of a prospective cohort study with mean follow-up of 41 mo, 227 patients had echocardiography at inception and after 1 yr of dialysis therapy. Improvements in left ventricular (LV) mass index, volume index, and fractional shortening were seen in 48, 48, and 46%, respectively. Ninety patients had developed cardiac failure by 1 yr of dialysis therapy. Twenty-six percent of the remaining 137 patients subsequently developed new-onset cardiac failure. The mean changes in LV mass index were 17 g/m(2) in those who subsequently developed cardiac failure compared with 0 g/m(2) among those who did not (P = 0.05). The corresponding values were -8 versus 0% for fractional shortening (P < 0.0001). The associations between serial change in both LV mass index and fractional shortening and subsequent cardiac failure persisted after adjusting for baseline age, diabetes, ischemic heart disease, and the corresponding baseline echocardiographic parameter. Regression of LV abnormalities is associated with an improved cardiac outcome in dialysis patients. Serial echocardiography adds prognostic information to one performed at baseline.