Nephrology literature watch

Citation: Liao C-T, Chen Y-M, Shiao C-C, et al. Rate of decline of residual renal function is associated with all-cause mortality and technique failure in patients on long-term peritoneal dialysis. Nephrol Dial Transplant. 2009;24:2909–2914. Analysis: Numerous investigators have evaluated the impact of treatment choices on the preservation of residual renal function and whether patients with higher levels of residual renal function (RRF) demonstrate improved patient-centered outcomes. Baseline RRF at the initiation of peritoneal dialysis (PD) is associated with improved overall patient survival and technique survival.1, 2 In the current study, Liao and colleagues provided evidence that rate of decline of RRF from initiation of PD rather than simply the level of RRF at baseline may be a stronger predictor of patient and technique survival. In this retrospective cohort, Liao and coworkers analyzed data on 270 PD patients followed for an average of 45 months, evaluating their rate of RRF decline and determining those variables likely contributing to the more accelerated rate of decline seen in some patients. They demonstrated that patients who had diabetes mellitus, had a history of congestive heart failure (CHF), who were obese, used diuretics, or experienced either peritonitis episodes or hypotensive events demonstrated a faster rate of decline of RRF. Survival analysis correcting for comorbid conditions demonstrated that those patients with the greatest decline in RRF had poorer overall life and technique survivals. Validity and threats to validity: Cohort studies have the benefit of evaluating a diverse group of patients over a long period of time and examining multiple outcomes that result from similar potentially harmful exposures. Furthermore, cohort studies can be used to examine questions regarding harm where randomized controlled trials would typically be either unethical or impractical. These types of studies, however, can be prone to both selection bias (systematic errors in group selection, causing groups to differ with respect to prognosis independent of the exposure or intervention of interest) and confounding (measured or unmeasured factors or interventions that affect the outcome of interest but that are not part of the causal pathway), and careful appraisal of data must be employed to insure accurate interpretations3 and to identify true effect modifiers. This type of study design may be the most practical way to evaluate the question posed by Liao and coworkers, that is, whether the rate of RRF decline from any cause predicts increased risk of early mortality and technique failure in patients with end-stage renal disease (ESRD) treated with peritoneal dialysis. This study evaluates a large population of patients with strict inclusion criteria (thus a well defined cohort), and describes the method the investigators employed to eliminate selection bias through restriction. By excluding patients who transferred out of peritoneal dialysis within 6 months, who had a prior transplant, or who were anuric at enrollment, the authors helped reduce prognostic differences that might have affected outcomes. Although this reduces sample size and power of the study and limits applicability to certain types of patients, it also limits the differences in related baseline characteristics3-5 that might affect outcomes separately from the characteristic or exposure under study. Furthermore, the restrictions meet the test of being biologically sound, since, for instance, it is unlikely that one can evaluate rates of decline in RRF in any patient who does not have residual function at baseline (anuric patients). The study also employs regression analysis and stratification of data, two techniques used to minimize or to expose potential confounders.5 Patients were divided into tertiles based on their rate of renal function decline, analyzed for group differences, and then evaluated using a regression analysis to determine statistically significant factors associated with patient survival and technique survival. These robust statistical analyses are described in detail. The results obtained are plausible and consistent with other epidemiologic and experimental studies. Diabetes, CHF, and older age are well-known risk factors for overall mortality in PD patients, and peritonitis, with its systemic inflammation and potential exposure to nephrotoxic antibiotic therapy, likely contributes to decline in residual renal function. Obesity in patients with low residual function may contribute to problems with middle molecule clearance and eventual modality failure, and hypotensive episodes may cause ischemic injury to the kidneys. The authors addressed issues related to study design limitations and the potential limitations in applicability of a single center cohort. Clinical bottom line: This well-designed, methodologically rigorous retrospective cohort study provides evidence supporting the importance of RRF to patient outcomes, and in particular, that the rate of decline of RRF is a stronger predictor of overall mortality and modality failure than the level of baseline RRF itself. Factors such as diabetes, obesity, CHF, diuretic use, peritonitis, and hypotensive episodes are independent prognostic contributors to the rate of RRF decline. Given the new data reported in this study and in conjunction with the findings from previous studies, it appears prudent to intensify effort to preserve residual function through routine measure of care by aggressively managing volume and nutritional status, by preventing hypovolumia and hypotension, through judicious monitoring for optimal glucose control in diabetes mellitus, and by preventing peritonitis. Whether specific novel interventions such as modulation of the PD fluid composition will result in both preservation of RRF and in improved survival awaits confirmation with a randomized controlled trial. The study by Kim and colleagues below begins to examine the question of novel approaches for the preservation of RRF. Citation: Kim S, Oh J, Kim S, et al. Benefits of biocompatible PD fluid for preservation of residual renal function in incident CAPD patients: a 1-year study. Nephrol Dial Transplant. 2009;24:2899–2908. Analysis: Kim and colleagues performed a randomized controlled clinical trial to determine if the use of a peritoneal dialysis fluid of neutral pH and low (relative to standard PD fluid) in glucose degradation products would result in a better preservation of RRF in incident ESRD patients treated with peritoneal dialysis. This study was performed because previous studies examining this question were inconclusive because of limited power6 or by technique/population heterogeneity.7 The investigators reasoned that improved biocompatibility of the PD fluid would result in decreased inflammation, and this in turn would result in improved survival of RRF, improved technique survival and reduced mortality. The hypothesis tested by Kim and colleagues in this study arises from the following reasoning. In 1995, Maiorca and colleagues described an independent relationship between the presence of RRF and survival of dialysis patients. Their multivariate survival analysis on 102 dialysis patients (both hemodialysis [HD] and PD) demonstrated that for each one mL/min increase in glomerular filtration rate (GFR) in RRF, mortality rates decreased by 40% in the entire group, and 50% in PD patients.8 Similar results were shown in the landmark CANUSA (Canadian/USA Peritoneal Dialysis Study Group) study with 680 incident PD patients.9 Other factors that might impact RRF in incident dialysis patients include dialysis modality,10, 11 strict blood-pressure control,12 renin-angiotensin-aldosterone blockade,5 and avoidance of nephrotoxins.13, 14 Animal studies have shown that exposure to traditional PD fluids, which have a “high” concentration of glucose degradation products (GDP), results in increased serum levels of advanced glycation end products and progressive renal injury due to glomerulosclerosis.15 In vitro biocompatibility studies have shown that a PD fluid that is both acidic and hyperosmolar inhibits the leukocyte phagocytosis and reduces bactericidal activity.16 Data on the effects of PD fluid composition on patient-centered outcomes is incomplete. The current study by Kim and colleagues attempts to correct this deficiency. Their study is a multicenter, open-labeled, randomized, prospective trial comparing outcomes in incident patients on PD treated with low-GDP solutions versus standard PD solutions. Patients were followed for a total of 12 months and GFRs were measured by the mean of renal urea and creatinine clearances. The study protocol was rigorously designed to avoid major common threats to validity including selection bias, differential lost to follow-up, and ascertainment bias. Since the study is unmasked, the probability of the latter was reduced by protocolized determination of hard outcomes including measured GFR, survival, and PD membrane transport properties. After masked allocation, as expected there were no significant differences in age, gender, weight, or etiologies of ESRD between the two treatment arms of the study. The study was powered to see a difference in the primary outcome of changes in or preservation of RRF. Data was analyzed according to intention to treat, and these results were compared with additional post hoc analyses of results “per-protocol” and on the subgroup of study patients with baseline residual function of >2 mL/min/1.73 m2. Patients in both arms of the study were treated identically in all aspects of care other than the PD fluid choice. For the primary outcome of the study, that is, preservation of RRF, there was no statistical significant benefit from use of a low GDP solution in the intention-to-treat analysis. However, in a post hoc analysis, when only patients with a baseline GFR of >2 mL/min were included, use of a low-GDP PD solution resulted in a statistically significant improvement in RRF preservation. The authors argue that such an analysis is indicated since one would expect that only those patients with some baseline residual function would likely benefit from an intervention intended to improve the preservation of RRF. Study patients treated in the low-GDP PD fluid arm of the trial maintained higher serum bicarbonate levels. No statistically significant differences in patient, technique, and peritonitis-free survivals were found between the two groups, but the event rates for these important patient-centered outcomes were low and the study was therefore far from sufficiently powered to address these important issues. A “trend” towards an increase in the rates of peritonitis in the low-GDP PD solution group compared to the standard solution group was not statistically significant. Applicability to U.S. population: The population chosen for this study was heterogeneous in nature and included a diverse adult ESRD population in terms of age, gender, comorbidities, and etiology of ESRD. All patients, however, were culturally and ethnically similar and residents of Taiwan. Furthermore, the study excluded patients on automated PD, a common modality in the United States. Hence the results should be applied to the patient population of North America with some caution as the impact of ethnicity and culture, as well as, the organization structure of healthcare on PD outcomes cannot be determined. Clinical bottom line: The trial by Kim and coworkers provides some support for the hypothesis that low-GDP PD fluid may result in better preservation of RRF in incident PD patients on continuous ambulatory PD during their first year on dialysis. The study was a well-designed, randomized controlled trial with masked allocation but unblinded follow-up. The strengths of the conclusions arising from this study are attenuated by: (1) insufficient power to examine any of the secondary end points, including importantly, risk for mortality, peritonitis, and technique failure; (2) inclusion of patients without RRF in a trial where the primary outcome of interest was preservation of RRF thus, necessitating the as-treated and subgroup analyses; and (3) a relatively brief period of observation. Thus, further randomized trials with a larger number of patients should be conducted to validate these results, with sufficient power to address the patient-centered morbidity and mortality outcomes. Pending further RCT evidence, a systematic review with a meta-analysis of current RCTs, assuming the degree of study differences or heterogeneity is not too large, might provide preliminary evidence of potential benefit or harm with regards to the patient-centered outcomes of mortality, peritonitis rates, and technique survival. Citation: Chan KE, Lazarus JM, Wingard RL, Hakim RM. Association between repeat hospitalization and early intervention in dialysis patients following hospital discharge. Kidney Intl. 2009;76:331–341. Analysis: Chan and collaborators have exploited a clinical database of a large dialysis provider (LDO) to ask the question whether specific aspects of the care of dialysis patients in the outpatient dialysis units, in the period immediately after their discharge from hospitalization, might impact whether these patients are re-admitted quickly to the hospital (within 30 days) after discharge. This study represents a retrospective look at data collected prospectively on all individuals in the LDO cohort. The authors demonstrated four principal findings. First that anemia, mineral-bone disease parameters, and dry weights deteriorated in direct relationship to the length of hospital stay. Second, that simply obtaining the laboratory values for calcium, phosphorus, and parathyroid hormone (PTH) did not appear to impact hospital readmission rates but obtaining a hemoglobin value did. The latter is more likely to translate into changes in erythropoietin (EPO) prescription because of the impact of Medicare reimbursement policies and hence, the management of anemia largely by protocol. Third, that a 16% reduction in risk of repeat hospitalization was observed for those patients in whom the EPO dose was modified in conjunction with a hemoglobin order. As noted by the authors, the hazard ratio for repeat hospitalization changed most steeply in the first 7 days (a 28% risk benefit in the first 14 days versus a 15% risk benefit for the full 30 day period) consistent with a relatively immediate effect of the administration of EPO. Fourth, that resumption of vitamin D was associated with improved re-hospitalization risk independent of whether calcium, phosphorus and PTH were checked. Administering vitamin D reduced the risk by 6% and not resuming vitamin D in those patients previously on vitamin D increased risk by 9%. Similar to EPO, the greatest benefit from resumption of vitamin D also occurred early after hospital discharge. Given that this is an observational study, it is likely that should the same question be interrogated with a randomized controlled clinical trial that the magnitude of demonstrated benefit might be somewhat diminished. Validity and threats to validity: There are two major sources of error that might arise from an observational study such as the study by Chen and colleagues: error introduced by biases such as measurement, ascertainment, and spectrum biases and error arising from confounding. A retrospective look at data, even data that might have been collected prospectively for clinical or non-study purposes is particularly susceptible to errors in identification of exposure. In the current study, the exposure can be defined as whether patients were evaluated for anemia and mineral metabolism abnormalities where such evaluations resulted, either in resumption of or in changes in their vitamin D and ESA orders within 7 days post-discharge. Since the latter are billable events, it is unlikely that there were too many opportunities for failure to record occurrences. It is more likely that ascertainment of hospitalizations, either the case-defining initial hospitalization or the primary study outcome (re-hospitalization) might have been missed in an outpatient dialysis unit generated database. It seems plausible that short, less severe hospitalizations would have been under-recorded. If such under-recording occurred, it is more likely that record keeping might have been most complete in those units exhibiting best practices, where early resumption of outpatient anemia and mineral-bone disease management might have been most likely. Therefore, the likely direction of this bias would have been to diminish the observed magnitude of benefit seen in an early resumption of therapies. The authors attempt to account for these potential biases by performing an analysis that is propensity score-based risk adjusted; a maneuver that can identify and adjust for some of the unit and physician specific differences in care. This method cannot however, eliminate entirely the possibility of confounding, that is, the notion that the benefit associated with early therapy is simply a marker for something else, such as better overall care. The possibility that the results are due to confounding is reduced somewhat by the size and scope of the dataset. Furthermore, the authors have explored this possibility, rigorously using a nested case-control methodology where the concurrent controls are drawn from the same dialysis units/nephrologists as the cases. Since a study such as this cannot eliminate entirely the potential for bias and confounding, the results must be interpreted and applied with caution. Clinical bottom line: The study by Chen and colleagues has particular relevance in light of the discussions regarding healthcare reform including the impact of bundling on ESRD care. It has been recognized for some time that many dialysis-dependent patients demonstrate deterioration in hemoglobin, nutrition, and mineral-bone parameters during prolonged hospitalizations. This study confirms this deterioration. Additionally, this observational study provides some of the most compelling evidence to date that aggressive management of hospitalization-related deterioration in metabolic and clinical parameters in the immediate post-hospitalization period might significantly and favorably impact rehospitalization risk. Any cost savings from currently proposed reforms such as bundling could be significantly attenuated if there were an unanticipated negative impact of reform on aggressive pharmacologic management in the dialysis patient post-hospitalization. Furthermore, it is likely that repeat hospitalization will be associated with increased risk of death further impacting quality of life and cost-effectiveness, an issue that could not be addressed by the current study, but one that is relevant to current national quality of care efforts. This robust study supports the view that clinical assessment immediately post-hospitalization might improve patient outcomes and more specifically, that some of the benefits might be related to EPO and vitamin D therapies. This study does not provide any insights into why EPO and vitamin D therapies might have such an important impact on immediate outcomes. The magnitude of the benefits and the optimal type and timing of dose modification awaits further exploration with randomized controlled trials. Pending RCT confirmation, it seems prudent to consider measures to eliminate barriers to thoughtful therapeutic modification of the care of the dialysis patient within the first 7 days after discharge from hospital.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.