Rationale & objective: Case-mix adjusted hemodialysis mortality has decreased since 1998. Many factors that influence mortality may have contributed to this trend, and these associations may differ by continental region. We studied changes in hemodialysis facility practices over time and their potential role in mediating changes in patient survival. Study design: Observational prospective cohort study. Setting & participants: Adult hemodialysis patients treated in 500 hemodialysis facilities participating in the Dialysis Outcomes and Practice Patterns Study (DOPPS) between 1999 and 2015 in the United States, Japan, and 4 European countries: Germany, Italy, Spain, and the United Kingdom. Predictors: Four practice measures at each facility: the percentages of patients with Kt/V >= 1.2, interdialytic weight gain [IDWG]<5.7%, phosphorus<6mg/dL, and using arteriovenous fistulas (AVFs). Outcome: Patient survival. Analytical approach: Mediation analyses, adjusted for case mix, were conducted using 3-year study phase as the exposure and facility practice measures as potential mediators. Results: In Europe, we observed a 13% improvement in overall case-mix adjusted survival per decade. Trends in facility practice measures, especially Kt/V and phosphorus, explained 10% improvement in case-mix survival per decade, representing 77% (10% explained of 13% improvement) of the observed improvement. In Japan, 73% of the observed 12%/decade improvement in case-mix adjusted survival could be attributed to facility practices, especially Kt/V and IDWG. In the United States, 56% of the observed 47%/decade improvement in case-mix adjusted survival could be attributed to facility practices, especially AVF use and phosphorus control. Limitations: Unmeasured changes in the characteristics of the patient population over this period may confound the observed associations. Conclusions: The improvements in adjusted hemodialysis patient survival in Europe, Japan, and the United States from 1999 to 2015 can be largely explained by improvements in specific facility practices. Future changes in patient survival may be responsive to further evolution in the implementation of common clinical practices. Plain-language summary: Case-mix adjusted survival of patients treated with hemodialysis has improved over the last 2 decades in the United States, Japan, and Europe. Some of this improvement can be explained by region-specific changes in 4 dialysis practices, namely increases in the proportions of patients achieving (1) Kt/V >= 1.2, (2) serum phosphorus levels<6mg/dL, (3) interdialytic weight gain<5.7% of body weight, and/or (4) use of arteriovenous fistulas as vascular access, with the magnitude varying according to region-specific trends in these practices. These findings suggest that further improvement in these practice measures may be attended by further reductions in mortality among patients treated with maintenance hemodialysis.

We welcome the letter from Hecking et al. titled "Dialysate Sodium and Mortality: 140 mmol/L at Hemodialysis Initiation by Default?,"1 published in this issue of JASN, regarding our publication "Effect of Dialysate and Plasma Sodium on Mortality in a Global Historical Hemodialysis Cohort"2 and suspect we are largely in agreement with the writers. We agree that despite careful efforts to adjust for potential confounders, residual confounding remains a concern in observational analyses such as these. As such, we feel the results of the analysis are hypothesis-generating but should not be used to change practice. We feel the results raise questions about dialysate sodium policy that can only be addressed through randomized comparisons. Hence, our strongest conclusions are the imperative for randomized evidence. We can offer the following comments in response to the particular points: Potential confounding by dialysate sodium concentrations by country: We acknowledge that completely controlling for confounding by country is impossible due to the limited variation in dialysate sodium concentrations within most participating countries. However, on the basis of our data, we do not find evidence for confounding by country. Our analysis stratified the data by country, adjusting for heterogeneity of mortality risk across countries using a country-specific baseline hazard function. The overall dialysate sodium versus high dialysate sodium effect reported in our article is a weighted average of these country-specific estimates. Only countries with both variability in dialysate sodium during follow-up and sufficient event counts contributed to the estimation of the country-specific effects. Table 1 presents these estimates: Three of five countries showed positive risk associations (hazard ratio >1), with two countries (Argentina and Portugal) demonstrating very strong and statistically significant effects. Potential confounding by fluid status: The core model used to estimate hazard ratios did not adjust for fluid overload. However, sensitivity analysis showed that including relative fluid status predialysis in the model did not substantially change the hazard ratios for low dialysate sodium ≤138 mmol/L, suggesting minimal confounding by fluid status. Potential confounding by baseline dialysate sodium prescription and limited change during follow-up: We have analyzed the data as documented in the European Clinical Database5. Dialysate sodium prescription is typically determined by country or clinic, rather than being individualized for patients. This limited variation in dialysate sodium prescription during the study could introduce confounding. To mitigate this, we exclude patients with predefined ultrafiltration and dialysate sodium prescription profiles from the analysis. Most of the dialysis clinics (78.6%) within our study were practicing a default dialysate sodium policy. It is a matter of debate which policy is feasible and subject to future randomized studies, what the best next step for outcome improvement could be, i.e., automated hemodialysis machine–based personalization of dialysate sodium on the basis of the dialysate–plasma sodium concentration gradient?3 Table 1 - Country-specific estimates of effect of low dialysate sodium versus high dialysate sodium on all-cause mortality Country % Person Months with DNa ≤138 mmol/L Patients Events HR 95% CI Argentina 7.6 4065 2346 2.48 1.91 to 3.19 Chile 6.7 2313 837 0.80 0.61 to 1.04 Poland 0.3 8717 3102 1.10 0.61 to 1.99 Portugal 1.1 3144 1529 4.38 1.64 to 11.73 Russia 0.1 5865 960 0.68 0.09 to 4.86 CI, confidence interval; DNa, dialysate sodium; HR, hazard ratio.

Drinking water contaminated by pathogenic micro-organisms increases the risk of infectious gastrointestinal disease which could potentially lead to acute kidney injury and even death, particularly amongst the young and the elderly. Earlier studies have shown a substantial reduction in the incidence of diarrheal disease over a period of one year using a polysulfone membrane water gravity-powered water filtration device. The current report is a continuation of these studies to assess the long-term effects of the innovative method on diarrheal incidence rates over a 4-year follow-up period. This follow-up study monitored the trend of self-reported diarrheal events in all households in the previously studied villages for 5 months, in the last half of each study year, using the same questionnaire utilized in the earlier study. Three villages that had no device yet installed served as controls. We computed monthly diarrheal incidence rates for all study years (standardized to per 100 person-months) and compared these to the pre-device incidence rate in 2018 and in the control group, using the Wilcoxon rank sum exact test. The average diarrheal incidence rates of 1.5 p100pm in 2019, 2.19 p100pm in 2021, and 0.54p100pm in 2022 were significantly different from an earlier study that reported 17.8 p100pm rates before the devices were installed in 2018, (all p-values < 0.05). Concomitantly, self-reported diarrheal infections were substantially higher in the “control villages” not yet having the filtration device installed (80.9, 77.6, and 21.5 per 100 pm). The consistent and large reduction in diarrhea incidence documents the long-term efficacy of the use of the membrane filtration device. This simple water purification method using gravity flow improves public health in remote regions with limited resources.

Every hemodialysis session starts with the question of how much fluid should be removed, which can currently not be answered precisely. Herein, we first revisit the "probing-dry-weight" concept, using the historical example of Tassin/France (practicing also "long, slow dialysis"): Mortality outcomes were, in the 1980s, better than registry data, but are nowadays similar to European average. In view of the negative primary end point in a recent trial on dry weight assessment, based on lung ultrasound-guided evaluation of fluid excess in the lungs, and a meta-analysis of prospective studies failing to show that bioimpedance-based interventions for correction of volume overload had a direct effect on all-cause mortality, we ask how to ever move forward. Clinical reasoning demands that as much information as possible should be gathered on the fluid status of patients undergoing dialysis. Besides body weight and blood pressure, measurements of bioimpedance and dialysate bolus-derived absolute blood volume can in principle be automatized, whereas lung ultrasound can be obtained routinely. In the era of machine learning, fluid management could consist of flexible target weight prescriptions, adjusted on a daily basis and accounting even for fluctuations in fluid-free body mass. In view of all the negative prospective results surrounding fluid management in hemodialysis, we propose this as a "never-give-up" approach.

Abstract BACKGROUND AND AIMS Loop diuretic use during maintenance haemodialysis (HD), including use of very high doses in Europe, is thought to maintain diuresis and reduce interdialytic weight gain, but is based on little evidence, with impact on clinical outcomes unclear. To help better inform international therapeutic strategies, we investigate regional variation in loop diuretic use and dosing strategies. METHOD This study included 71 756 HD patients from the international Dialysis Outcomes and Practice Patterns Study (DOPPS) phases 2–5 (2002–15), including 27 759 patients with dialysis vintage (time since HD initiation) <1 year. We report the proportion of patients in each country prescribed a diuretic (loop, thiazide or other) by dialysis vintage. Among patients with vintage < 1 year at enrollment and prescribed a loop diuretic, we report the distribution of loop diuretic dose (mg/day) by country. Doses of torsemide (4:1) and bumetanide (80:1) were converted to oral furosemide-equivalent dose. RESULTS Diuretic use varied widely by country, ranging at vintage < 3 months from >80% in Germany and Sweden to < 35% in the US, Spain, and Gulf Cooperation Council (GCC) countries (Figure 1). The proportion of patients prescribed a diuretic decreased with dialysis vintage in all countries. At vintage >5 years, diuretic use was 30%–35% in Germany and Sweden and < 20% in other countries. In all countries, >90% of diuretic prescriptions were for loop diuretics. Loop diuretic dose varied widely across countries. Among patients with vintage < 1 year, median dose ranged from 400–500 mg/day in Belgium, France, Germany and Sweden to < 100 mg/day in the GCC, Japan, Spain and the US (Figure 2). CONCLUSION We observed substantial international differences in diuretic prescription patterns to HD patients, with usage and doses much higher in some European countries than the US, where high-dose formulations are not available. The potential impact of diuretic prescription patterns on outcomes, including residual urine volume, volume-related complications and others should be investigated.

Introduction: More men than women start kidney replacement therapy (KRT) although the prevalence of chronic kidney disease (CKD) is higher in women than men. We therefore aimed at analyzing sex-specific differences in clinical outcomes among 8237 individuals with CKD in stages 3 to 5 from Brazil, France, Germany, and the United States participating in the Chronic Kidney Disease Outcomes and Practice Patterns Study (CKDopps). Methods: Fine and Gray models, evaluating the effect of sex on time to events, were adjusted for age, Black race (model A); plus diabetes, cardiovascular disease, albuminuria (model B); plus estimated glomerular filtration rate (eGFR) slope during the first 12 months after enrollment and first eGFR after enrollment (model C). Results: There were more men than women at baseline (58% vs. 42%), men were younger than women, and men had higher eGFR (28.9 +/- 11.5 vs. 27.0 +/- 10.8 ml/min per 1.73 m(2)). Over a median follow-up of 2.7 and 2.5 years for men and women, respectively, the crude dialysis initiation and pre-emptive transplantation rates were higher in men whereas that of pre-KRT death was more similar. The adjusted subdistribution hazard ratios (SHRs) between men versus women for dialysis were 1.51 (1.27-1.80) (model A), 1.32 (1.10-1.59) (model B), and 1.50 (1.25-1.80) (model C); for pre-KRT death, were 1.25 (1.02-1.54) (model A), 1.14 (0.92-1.40) (model B), and 1.15 (0.93-1.42) (model C); for transplantation, were 1.31 (0.73-2.36) (model A), 1.44 (0.76-2.74) (model B), and 1.53 (0.79-2.94) (model C). Conclusion: Men had a higher probability of commencing dialysis before death, unexplained by CKD progression alone. Although the causal mechanisms are uncertain, this finding helps interpret the preponderance of men in the dialysis population.

The nephrology community grieves for Dr. med. Jürgen Bommer, who passed away in Heidelberg, Germany on February 7, 2022, shortly after his 80th birthday. His patients, peers and colleagues will miss a passionate phy-sician and outstanding pioneer of both dialysis technology and clinical research with a worldwide recognition in the realm of clinical nephrology. His commitment to application science in dialysis care and his dedication as a teacher has motivated and stimulated students, nurses, and academic colleagues during more than the past 40 years. His list of scientific publications shows all aspects of dialysis treatment and can be taken as a perfect exam-ple of the concept of dialysis as a systems approach . and engage-ment for medical therapies in to improve and pro-mote patient outcomes.

The provision of clean water to remote communities is a major goal of both the World Health Organization and the United Nations. We report on the long-term sustainability of filter-sterilizing polluted water in remote villages in Ghana that lack electricity. Contaminated water pumped several times a week via a gasoline pump into a 1000 L elevated tank is filtered through polysulfone hemodialyzers on demand. The 3 nm fiber pore size rejects all bacteria, parasites, and viruses. Villagers flush organic matter from the dialyzers thrice daily to maintain a flow of up to 250 L/h. Having previously reported a 73% reduction in diarrheal episodes, we now address system sustainability. After passing through the hemodialyzer filters, a fecally polluted water source remains consistently free of pathogens even after the system has been in place for >1 year in most villages. Filters are easily replaced when needed. Daily cost for unlimited clean water is less than USD 2.22 per village over five years. Villagers have continued to independently fill the tank and flush the system, because they appreciate the clean water and health benefits. We demonstrate that over 2–6 years this system providing pathogen-free drinking water can be maintained independently by villagers for long-term sustainability. It does not require electricity nor disinfectants to be added to the product water and is ready for far broader application in similarly remote settings.

Low sodium dialysate was commonly used in the early year of hemodialysis to enhance diffusive sodium removal beyond its convective removal by ultrafiltration. However, disequilibrium syndrome was common, particularly when dialysis sessions were reduced to 4 h. The recent trend of lowering the DNa from the most common level of 140 mEq/L has been associated with intradialytic hypotension and increased risk of hospitalization and mortality. Higher DNa also has disadvantages, such as higher blood pressure and greater interdialytic weight gain, likely due to increased thirst. My assessment of the evidence leads me to choose DNa at the 140 level for most patients and to avoid DNa below 138. Patients with intradialytic symptoms may benefit from DNa 142 mEq/L, if they can avoid excessive fluid weight gains.

Background:Metabolic acidosis is a common threat for patients on hemodialysis, managed by alkaline dialysate. The main base is bicarbonate, to which small amounts of acetic, citric, or hydrochloric acid are added. The first two are metabolized to bicarbonate, mostly by the liver. Citric acid-containing dialysate might improve dialysis efficiency, anticoagulation, calcification propensity score, and intradialytic hemodynamic stability. However, a recent report from the French dialysis registry suggested this dialysate increases mortality risk. This prompted us to assess whether citric acid-containing bicarbonate-based dialysate was associated with mortality in the international Dialysis Outcomes and Practice Patterns Study (DOPPS).Methods:Detailed patient-based information on dialysate composition was collected in DOPPS phases 5 and 6 (2012-2017). Cox regression was used to model the association between baseline bicarbonate dialysate containing citric acid versus not containing citric acid and mortality among DOPPS countries and phases where citric acid-containing dialysate was used.Results:Citric acid-containing dialysate was most commonly used in Japan, Italy, and Belgium (25%, 25%, 21% and of patients who were DOPPS phase 6, respectively) and used in <10% of patients in other countries. Among 11,306 patients in DOPPS country and phases with at least 15 patients using citric acid-containing dialysate, patient demographics, comorbidities, and laboratories were similar among patients using (14%) versus not using (86%) citric acid-containing dialysate. After accounting for case mix, we did not observe a directional association between citric acid-containing dialysate use (any versus none) and mortality (HR, 1.14; 95% CI, 0.97 to 1.34), nor did we find evidence of a dose-dependent relationship when parameterizing the citric acid concentration in the dialysate as 1, 2, and 3+ mEq/L.Conclusions:The use of citric acid-containing dialysate was not associated with greater risk of all-cause mortality in patients on hemodialysis participating in DOPPS. Clinical indications for the use of citric acid-containing dialysate deserve further investigation.

Significance Statement Sustained (or overt) diabetes mellitus after kidney transplantation is strongly associated with hyperglycemia during the early perioperative period. In a multicenter trial with 263 participants randomized to strict blood glucose monitoring and an early basal insulin intervention versus control (a more liberal approach consisting of sporadic corrections of hyperglycemia and otherwise oral antidiabetics), overt post-transplantation diabetes mellitus (PTDM) was ascertained by antidiabetic treatment and an oral glucose tolerance test (2 hour glucose ≥200 mg/dl). The intervention resulted in modestly reduced PTDM rates at 12 months and 24 months at the cost of higher rates of hypoglycemia. In a per-protocol analysis that excluded protocol violators and accounted for baseline differences in polycystic kidney disease, the reduction in PTDM at 12 months was significant, suggesting the approach merits further study. Background Post-transplantation diabetes mellitus (PTDM) might be preventable. Methods This open-label, multicenter randomized trial compared 133 kidney transplant recipients given intermediate-acting insulin isophane for postoperative afternoon glucose ≥140 mg/dl with 130 patients given short-acting insulin for fasting glucose ≥200 mg/dl (control). The primary end point was PTDM (antidiabetic treatment or oral glucose tolerance test–derived 2 hour glucose ≥200 mg/dl) at month 12 post-transplant. Results In the intention-to-treat population, PTDM rates at 12 months were 12.2% and 14.7% in treatment versus control groups, respectively (odds ratio [OR], 0.82; 95% confidence interval [95% CI], 0.39 to 1.76) and 13.4% versus 17.4%, respectively, at 24 months (OR, 0.71; 95% CI, 0.34 to 1.49). In the per-protocol population, treatment resulted in reduced odds for PTDM at 12 months (OR, 0.40; 95% CI, 0.16 to 1.01) and 24 months (OR, 0.54; 95% CI, 0.24 to 1.20). After adjustment for polycystic kidney disease, per-protocol ORs for PTDM (treatment versus controls) were 0.21 (95% CI, 0.07 to 0.62) at 12 months and 0.35 (95% CI, 0.14 to 0.87) at 24 months. Significantly more hypoglycemic events (mostly asymptomatic or mildly symptomatic) occurred in the treatment group versus the control group. Within the treatment group, nonadherence to the insulin initiation protocol was associated with significantly higher odds for PTDM at months 12 and 24. Conclusions At low overt PTDM incidence, the primary end point in the intention-to-treat population did not differ significantly between treatment and control groups. In the per-protocol analysis, early basal insulin therapy resulted in significantly higher hypoglycemia rates but reduced odds for overt PTDM—a significant reduction after adjustment for baseline differences—suggesting the intervention merits further study. Clinical Trial registration number: NCT03507829

Abstract Background Hemodialysis (HD) patients are commonly prescribed phosphate binders (PBs) to manage serum phosphorus levels, as hyperphosphatemia is strongly associated with poorer survival. Nonadherence with the PB prescription is associated with elevated serum phosphorus levels. We studied associations between patient satisfaction with their PB and serum phosphorus levels and mortality rates. Methods Adult HD patients in Germany, Italy, Spain and the UK in the Dialysis Outcomes and Practice Patterns Study were administered a survey instrument in late 2017. Patients were asked about their satisfaction with their PBs, as measured through three questions (difficulty, inconvenience and dissatisfaction) on a 5-point Likert scale, with each dichotomized into average worst versus good responses. These were used as predictors in linear regression models of continuous serum phosphorus levels and in Cox proportional hazards models of mortality, with adjustments for demographics, comorbidities and laboratory values. Results Patients having greater difficulty, inconvenience and dissatisfaction with their PB had higher serum phosphorus levels in adjusted models {+0.21 mg/dL [95% confidence interval (CI) ±0.23], +0.30 (±0.21) and 0.36 (±0.22), respectively}, and higher odds of having serum phosphorus levels ≥6.0 mg/dL. Measures of dissatisfaction were also associated with an elevated risk of mortality, with adjusted hazard ratios of 2.2 (95% CI 1.3–3.6), 1.6 (1.0–2.6) and 1.7 (1.1–2.7), respectively; this association was not strongly affected by adjustment for baseline serum phosphorous level. Conclusions Self-reported difficulty, inconvenience and dissatisfaction in taking one’s prescribed PBs were associated with elevated serum phosphorus levels and serum phosphorus levels above clinically meaningful thresholds. While the mechanism for the association with mortality is unclear, patient-reported satisfaction should be considered when attempting to manage patient serum phosphorus levels.

BackgroundSystematic analyses about sex differences in wait-listing and kidney transplantation after dialysis initiation are scarce. We aimed at identifying sex-specific disparities along the path of kidney disease treatment, comparing two countries with distinctive health care systems, the US and Austria, over time.MethodsWe analyzed subjects who initiated dialysis from 1979–2018, in observational cohort studies from the US and Austria. We used Cox regression to model male-to-female cause-specific hazard ratios (csHRs, 95% confidence intervals) for transitions along the consecutive states dialysis initiation, wait-listing, kidney transplantation and death, adjusted for age and stratified by country and decade of dialysis initiation.ResultsAmong 3,053,206 US and 36,608 Austrian patients starting dialysis, men had higher chances to enter the wait-list, which however decreased over time [male-to-female csHRs for wait-listing, 1978–1987: US 1.94 (1.71, 2.20), AUT 1.61 (1.20, 2.17); 2008–2018: US 1.35 (1.32, 1.38), AUT 1.11 (0.94, 1.32)]. Once wait-listed, the advantage of the men became smaller, but persisted in the US [male-to-female csHR for transplantation after wait-listing, 2008–2018: 1.08 (1.05, 1.11)]. The greatest disparity between men and women occurred in older age groups in both countries [male-to-female csHR for wait-listing after dialysis, adjusted to 75% age quantile, 2008–2018: US 1.83 (1.74, 1.92), AUT 1.48 (1.02, 2.13)]. Male-to-female csHRs for death were close to one, but higher after transplantation than after dialysis.ConclusionsWe found evidence for sex disparities in both countries. Historically, men in the US and Austria had 90%, respectively, 60% higher chances of being wait-listed for kidney transplantation, although these gaps decreased over time. Efforts should be continued to render kidney transplantation equally accessible for both sexes, especially for older women.

Given the need for treating polluted drinking water, our NGO Easy Water for Everyone has produced pure water in remote villages without power and achieved health benefits. With the goal of reaching more needy populations we report our experience and successful implementation in Ghana. In 20 villages polluted water is pumped every few days to an elevated water tank connected to a filtration device leading to a faucet. Repurposed hemodialyzers with polysulfone membranes, having a filter pore size of 0.003 micrometres, prevent passage of pathogens. Gravity from a 3 m height pushes water through the membrane whenever the faucet is open. Backflushing of the hemodialyzer membrane three times daily removes built-up organic material and maintains flow rates of 250 L/hour for at least two years. Filtered water has been culture-negative. Management of problems and optimization are reported. The five-year cost per village of <1,500 population averaged <2 US$ per day.

1Arbor Research Collaborative for Health, Ann Arbor, MI, USA; 2Departments of Epidemiology and Environmental Health Sciences, School of Public Health, and Department of Urology, Medical School, University of Michigan, Ann Arbor, MI, USA; 3Department of Biostatistics, University of Michigan School of PublicHealth, AnnArbor, MI, USA; 4Vanderbilt University Medical Center, Nashville, TN, USA; 5Division of Nephrology, Department of Medicine, Showa University Fujigaoka Hospital, Yokohama, Japan, and Showa University Research Administration Center (SURAC), Showa University, Tokyo, Japan; 6Cliniques universitaires St Luc, Université Catholique de Louvain, Brussels, Belgium; 7Institute of Public Health, Charité – Universitätsmedizin Berlin, Berlin, Germany; 8Department of Nephrology, University Hospital Ghent, Ghent, Belgium; 9Department of Epidemiology, University of Michigan School of Public Health, and Department of Internal Medicine-Nephrology, University of Michigan Medical School, Ann Arbor, MI, USA Purpose: Mortality among first-year hemodialysis (HD) patients remains unacceptably high.

In rural regions with limited resources, the provision of clean water remains challenging. The resulting high incidence of diarrhea can lead to acute kidney injury and death, particularly in the young and the old. Membrane filtration using recycled hemodialyzers allows water purification. This study quantifies the public health effects. Between 02/2018 and 12/2018, 4 villages in rural Ghana were provided with a high-volume membrane filtration device (NuFiltration). Household surveys were collected monthly with approval from Ghana Health Services. Incidence rates of diarrhea for 5-month periods before and after implementation of the device were collected and compared to corresponding rates in 4 neighboring villages not yet equipped. Data of 1,130 villagers over 10 months from the studied communities were studied. Incidence rates showed a decline following the implementation of the device from 0.18 to 0.05 cases per person-month (ppm) compared to the control villages (0.11 to 0.08 ppm). The rate ratio of 0.27 for the study villages is revised to 0.38 when considering the non-significant rate reduction in the control villages. Provision of a repurposed hemodialyzer membrane filtration device markedly improves health outcomes as measured by diarrhea incidence within rural communities.

BACKGROUND: Decisions about dialysis for advanced kidney disease are often strongly shaped by sociocultural and system-level factors rather than the priorities and values of individual patients. We examined international variation in the uptake of conservative approaches to the care of patients with advanced kidney disease, in particular discontinuation of dialysis. METHODS: We employed an observational cohort study design using data collected from patients maintained on long-term hemodialysis between 1996 and 2015 in facilities across 12 developed countries participating in the Dialysis Outcomes and Practice Patterns Study (DOPPS). The main outcome was discontinuation of dialysis therapy. We analyzed the association between several patient characteristics and time to dialysis discontinuation by country and phase of study entry. RESULTS: A total of 259 343 DOPPS patients contributed data to the study, of whom 48 519 (18.7%) died during the study period. Of the decedents, 5808 (12.0%) discontinued dialysis before death. Rates of discontinuation were higher within the first few months after initiation of dialysis, among older adults, among those with a greater number of comorbidities and among those living in an institution. After adjustment for age, sex, dialysis duration, diabetes and dialysis era, rates of discontinuation were highest in Canada, the United States and Australia/New Zealand (33.8, 31.4 and 21.5 per 1000/yr, respectively) and lowest in Japan and Italy (< 0.1 per 1000/yr). Crude discontinuation rates were highest in dialysis facilities that were more likely to offer comprehensive conservative renal care to older adults. INTERPRETATION: We found persistent international variation in average rates of dialysis discontinuation not explained by differences in patient case-mix. These differences may reflect physician-, facility- and society-level differences in clinical practice. There may be opportunities for international cross-collaboration to improve support for patients with end-stage renal disease who prefer a more conservative approach.