Objectives:We evaluated the data requirement for modern AI tools to outperform simpler models in predicting short-term mortality in over 500 000 patients with hemodialysis-dependent kidney failure. Materials and Methods:We compared logistic regression, boosting, and transformers using increasingly complex feature sets (from last-visit data to full trajectories). Performance was measured using the area under the ROC curve (AUC-ROC) and the Precision-Recall curve (AUC-PR) across training data sizes ranging from 500 to 490 197 samples. Results:Using features with temporal information is beneficial across all models. On the full dataset, Transformers (AUC-ROC = 0.8568) and boosting (AUC-ROC = 0.8598) perform similarly. Discussion:Transformers require large datasets to outperform simpler models like boosting, limiting their usefulness in smaller datasets, even on datasets as big as 500K. Conclusion:Modern AI tools require substantial data to justify their computational cost over simpler approaches. However, a more complex feature set seems to be beneficial across all models.

RATIONALE & OBJECTIVE:Pericarditis and pericardial effusions may complicate the clinical course of people with kidney failure receiving dialysis. However, few recent data exist on their incidence or associated outcomes and costs. This study characterized hospitalizations among a nationally representative set of US patients receiving maintenance dialysis who experienced pericarditis or pericardial effusion. STUDY DESIGN:Retrospective cohort study. SETTING & PARTICIPANTS:Adults with dialysis-dependent kidney failure, represented in the National Inpatient Sample (NIS) (2016-2021), a database of ∼20% of nonfederal acute care hospitalizations in the United States, with pericarditis or pericardial effusions identified using administrative codes. EXPOSURE:Calendar year, pericardial drainage procedures, and cardiac tamponade or cardiogenic shock. OUTCOME:Trends in in-hospital mortality, costs, and repeat pericardial drainage procedures. ANALYTICAL APPROACH:Survey methods with weighted least squares regression to characterize trends. Logistic regression for the risk of repeat pericardial drainage. In-hospital mortality was characterized as a proportion of hospital stays. Cost estimation was based on hospitalization charges and cost-to-charge ratios. RESULTS:We identified 18,535 hospitalizations in the NIS sample, representing a total of 92,675 (95% CI, 90,674-94,676) nonfederal acute care hospital stays. The mean age was 57 years, and 49% were female. Pericardial tamponade or shock was diagnosed in 11.1% (95% CI, 10.6-11.6) of hospitalizations, pericardial drainage procedures were performed in 13.1% (95% CI, 12.6-13.6%), and in-hospital mortality was 7.8% (95% CI, 7.4-8.2%). The counts of hospitalizations increased by an average of 1,370 (95% CI, 1,108-1,631; P < 0.001) per year and percutaneous drainage procedures by an average of 131 (95% CI, 15-246; P = 0.04) per year. The rate of hospitalization increased from 2.3 (95% CI, 2.2-2.4) per 100 person-years in 2016 to 3.5 (95% CI, 3.4-3.7) in 2021 (P for trend = 0.001). Hospital costs increased from $347 million (95% CI, $316-378 million) in 2016 to $591 million (95% CI, $545-637 million) in 2021 (P = 0.001). LIMITATIONS:Administrative data. CONCLUSIONS:Hospitalizations for pericarditis or pericardial effusion among patients on dialysis have increased in recent years, with a corresponding increase in pericardial drainage procedures. More than 1 in 10 hospitalizations were complicated by tamponade or shock. Identifying actionable drivers of the increase in this morbidity over time warrants further investigation. PLAIN-LANGUAGE SUMMARY:Inflammation of the connective tissue lining the heart (pericarditis) and accumulation of fluid around the heart (pericardial effusion) can occur in people who receive dialysis for kidney failure. However, little is known about how often this occurs in recent years or the related outcomes and costs. This study estimated how often hospitalizations with pericarditis or pericardial effusion in people with kidney failure on dialysis occur in the United States. We found that the number of hospitalizations with these problems and procedures to treat them have increased from 2016 to 2021. Additionally, the total cost of these hospitalizations increased from 2016 to 2021. The reasons for the observed increases are uncertain and warrant further study.

Atrial fibrillation (AF) is the most common sustained arrhythmia, with a prevalence and incidence significantly higher in adults with CKD compared with the general population. This risk increases with reduced kidney function, affecting up to 25% of all patients with CKD and 30% of those on receiving chronic dialysis. AF is associated with increased morbidity and mortality, including higher risks of stroke, heart failure, myocardial dysfunction, and progression to kidney failure. The bidirectional relationship between AF and CKD is driven by a convergence of traditional risk factors—such as hypertension and diabetes—and CKD-specific abnormalities. Key pathophysiologic mechanisms include systemic inflammation, oxidative stress, autonomic dysfunction, and disordered mineral metabolism. These factors promote structural remodeling, atrial fibrosis, and electrical instability, creating a highly arrhythmogenic substrate. Treatment options for AF include anticoagulation, rate and rhythm control medications, other cardiovascular therapies ( e.g ., sodium glucose cotransporter 2 inhibitors), and procedures aimed at normalizing rhythm and reducing stroke risk. Patients with CKD are often undertreated with AF medications and procedures, in part because of lack of robust randomized clinical trial data on efficacy and safety, particularly with advanced CKD. This narrative review summarizes key literature on the epidemiology, risk factors, mechanisms and treatment of AF in adults with CKD, and highlights critical areas for future research.

BACKGROUND:In March 2020, responding to the COVID-19 pandemic, US public health emergency waivers enabled kidney care providers (nephrologists and advanced practice providers) to substitute face-to-face in-center hemodialysis visits with telemedicine encounters. METHODS:We merged an electronic health records database from dialysis patients at clinics owned by DaVita Inc. with Medicare claims-based records. We identified dialysis patients in the first six months of the pandemic. Kidney care providers were divided into those who used telemedicine frequently (above median) versus less frequently (below median) during the first three months of the pandemic. In multivariable regression models controlling for patient, kidney care provider, dialysis facility, and geographic characteristics, and the value of each outcome prior to the pandemic, we examined associations among telemedicine use by kidney care providers and hemodialysis-related surrogate health outcomes. RESULTS:We identified 782 kidney care providers who saw patients at DaVita facilities from 4/2020-9/2020. Higher intensity (vis-à-vis telemedicine) kidney care providers documented the use of telemedicine in 86% (Standard Deviation (SD) 16%) of patients in the first three months of the pandemic, while lower intensity providers documented its use in 15% (SD 13%) of patients. Higher telemedicine use was associated with a 0.7% (95% confidence interval (CI), 0.1%-1%) reduction in the absolute probability of having a serum albumin <3.5 g/dl and a 1% (CI 0.2%-3%) reduction in the absolute probability of having a serum phosphate >5.5 mg/dl. Telemedicine use was not significantly associated with outcomes related to anemia management targets, blood pressure, dialysis efficiency, interdialytic weight gain, and dialysis prescription changes. CONCLUSIONS:Patients of providers who used telemedicine more frequently had slightly better surrogate outcomes in nutritional parameters, but not in other dialysis-related health outcomes. These findings suggest that telemedicine can be employed for in-center hemodialysis visits without compromising conventional metrics of the quality of hemodialysis care.

BACKGROUND:Chronic kidney disease is a common comorbidity of rheumatoid arthritis. Because there has been scarce and conflicting evidence on the use of biological and targeted synthetic therapy in patients with rheumatoid arthritis and chronic kidney disease, we aimed to examine the effectiveness and persistence of biological and targeted synthetic disease-modifying antirheumatic drugs (DMARDs) for patients with rheumatoid arthritis and chronic kidney disease. METHODS:This multicentre, prospective cohort study used data from the CorEvitas Rheumatoid Arthritis registry on patients with rheumatoid arthritis who initiated each type of biological and targeted synthetic DMARD (TNF inhibitors, CTLA4 immunoglobulin, IL-6 inhibitors, B-cell depletion therapy, and JAK inhibitors) recruited from 160 rheumatology practices in the USA. Patients were required to have moderate or high disease activity according to the Clinical Disease Activity Index (CDAI; score >10). Chronic kidney disease was defined as an estimated glomerular filtration rate (eGFR) of less than 60 mL/min per 1·73 m2 at biological or targeted synthetic DMARD initiation. The primary outcome was the attainment of remission according to CDAI score. Propensity scores were used for overlap weighting to improve the balance of characteristics for patients with and without reduced eGFR. Cox regressions estimated the unadjusted and adjusted hazard ratio (HR) of reduced eGFR for CDAI-based remission. People with lived experience of rheumatoid arthritis were not involved in the study design or conduct. FINDINGS:Between Oct 1, 2001, and Dec 31, 2023, 12 123 biological and targeted synthetic DMARD treatment initiations in 9601 patients with rheumatoid arthritis, with 51 931 person-years of follow-up, were identified. Of 12 123 eligible biological and targeted synthetic DMARD treatment initiations, 10 857 (89·6%) were in patients with preserved eGFR and 1266 (10·4%) were in patients with reduced eGFR. Across both groups, median age was 59·0 years (IQR 50·0-67·0), 9720 (80·2%) of 12 123 treatment initiations were in female patients and 2403 (19·8%) were in male patients, and 9908 (81·7%) were in non-Hispanic White patients. 3025 (27·9%) of 10 857 treatment initiations in patients with preserved eGFR resulted in CDAI-based remission versus 246 (19·4%) of 1266 treatment initiations in patients with reduced eGFR. Reduced eGFR was associated with a lower likelihood of attaining CDAI-based remission (unadjusted HR 0·71 [95% CI 0·62-0·82], adjusted HR 0·76 [0·66-0·88]) compared with preserved eGFR. INTERPRETATION:Although biological and targeted synthetic DMARDs are generally well tolerated and effective options, patients with rheumatoid arthritis and reduced eGFR have lower likelihood of attaining remission. Tailored treatment strategies for this high-risk population should be established. FUNDING:National Institute of Arthritis and Musculoskeletal and Skin Diseases.

KEY POINTS:Many major adverse cardiovascular events happen within the first 6 months after hemodialysis initiation. Using nondialyzable β -blockers at hemodialysis initiation was associated with fewer cardiovascular events than dialyzable β -blockers. BACKGROUND:How β -blocker dialyzability affects cardiovascular outcomes in patients receiving hemodialysis remains unclear. Previous observational studies show conflicting results, and conducting clinical trials in this population has proven challenging. This uncertainty has contributed to substantial variation in β -blocker prescribing patterns across regions and clinicians. METHODS:Using data from the United States Renal Data System, we identified patients receiving β -blockers on the day of starting hemodialysis and classified them into nondialyzable and dialyzable exposure groups. Geographical variations in nondialyzable and dialyzable β -blocker prescription rates across neighboring regions in each state were leveraged as an instrumental variable. The treatment effect of nondialyzable versus dialyzable β -blockers on major adverse cardiovascular events (MACE) at 6 months, 1 year, and 3 years was evaluated using instrumental variable g -estimation methods for survival outcomes. MACE comprised myocardial infarction, stroke, and all-cause mortality. RESULTS:The cohort of 40,313 participants from 41 US states experienced 25,720 MACE over the 3-year follow-up period (median years [interquartile range], 1.4 [0.4-3.0]), with more than 42% occurring during the first 6 months after hemodialysis initiation. The instrumental variable met all statistical assumptions. Nondialyzable β -blockers were associated with significantly lower risk of MACE at all three follow-up time points compared with dialyzable β -blockers (hazard ratios [95% confidence intervals], 0.82 [0.74 to 0.90], 0.85 [0.79 to 0.92], and 0.90 [0.85 to 0.96], respectively). Similar patterns were observed for each separate component of MACE. CONCLUSIONS:The use of nondialyzable β -blockers compared with dialyzable β -blockers was associated with a lower risk of MACE during all time points, including the early, high-risk period in the 6 months following hemodialysis initiation.

The Tracking Treatment Pathways in Adult Patients with Hyperkalemia (TRACK) study demonstrated that clinical practice guidelines for hyperkalemia are inconsistently implemented and a conservative approach to management is often preferred. We report additional analyses focusing on routine clinical practice in the USA. TRACK was an observational prospective cohort study (July 2022–December 2024) including adults in the USA with a recent index hyperkalemia episode (serum potassium [K+] > 5.0 mmol/L ≤ 21 days of enrollment). Primary and secondary data were obtained from participants, their respective healthcare professionals (HCPs), and medical records. Clinical characteristics, hyperkalemia management decisions and rationales, and treatment outcomes over 12 months were evaluated. The final cohort included 229 participants, of whom 66.4

INTRODUCTION:Apixaban reduces thromboembolic events, bleeding, and mortality compared with warfarin in patients with nonvalvular atrial fibrillation (NVAF). However, evidence in patients with kidney failure (KF) receiving peritoneal dialysis (PD) is limited. We evaluated the safety and effectiveness of apixaban versus warfarin in this population. METHODS:Using the US Renal Data System, we identified Medicare beneficiaries on PD with newly diagnosed NVAF (2014-2019) who had no prior oral anticoagulation (OAC) use. We included patients initiating apixaban or warfarin within 30 days after NVAF diagnosis. After propensity matching, we estimated the hazard ratios for thromboembolic events, bleeding events, and mortality using Cox models accounting for competing risks, where applicable, and using both intent-to-treat and as-treated (AT) approaches. RESULTS:Among 6,571 OAC-naive PD patients with newly diagnosed NVAF, 334 initiated apixaban and 717 initiated warfarin within 30 days. After 1:1 propensity matching, all observed characteristics of 330 apixaban and 330 warfarin users were balanced. The median CHA2DS2-VASc score was 4 (IQR 3-5). In intent-to-treat analyses, hazards of ischemic stroke (hazard ratio [HR] = 0.92; 95% CI: 0.42-2.01) and the composite of other thromboembolic outcomes with cardiovascular death (HR = 0.90; 95% CI: 0.66-1.22) did not differ between groups. The hazards of clinically important bleeding (HR = 0.81; 95% CI: 0.62-1.07), major bleeding, intracranial hemorrhage, and all-cause mortality also did not differ between groups. Findings were consistent in AT analyses. CONCLUSION:Among PD patients with newly diagnosed NVAF, we observed no statistically significant differences between apixaban and warfarin for effectiveness or bleeding outcomes. Given the challenges of warfarin management in KF, apixaban may represent a more practical option when anticoagulation is pursued.

Little is known about the implementation of guideline recommendations and healthcare provider decision-making when managing hyperkalemia. We aimed to address these knowledge gaps and better understand hyperkalemia management in routine clinical care and the baseline characteristics and longitudinal (12-month) clinical variables of participants with hyperkalemia. This was an observational, prospective, longitudinal, cohort study including adults with recent (≤ 21 days of enrollment) hyperkalemia (serum potassium [K+] > 5.0 mmol/L) during standard of care, across Germany, Italy, Spain, the UK, and the USA from July 2022 to December 2024. Primary and secondary data were collected from participants’ medical records and their respective healthcare providers. Healthcare providers’ hyperkalemia management decisions, treatment rationale, and expectations were measured overall, by incident versus recurrent cases, and by hyperkalemia severity. In total, 1330 participants with hyperkalemia were included (458 incident, 870 recurrent). The most common treatment rationale for index hyperkalemia was ease of treatment (41.4

Atrial fibrillation (AF) and chronic kidney disease (CKD) frequently coexist, share risk factors and pathophysiology, and confer a high burden of stroke, bleeding, heart failure, and premature death. Risk scores such as CHA₂DS₂-VASc, which estimates thromboembolic risk, and HAS-BLED, which estimates bleeding risk, were derived in populations including relatively few patients with advanced CKD. Consequently, they may not adequately account for CKD-specific modifiers of thromboembolic and hemorrhagic risk. Although these scores do not formally incorporate competing non-stroke mortality, higher CHA₂DS₂-VASc scores may indirectly identify patients with a greater competing risk of death. Meanwhile, widespread use of renin-angiotensin-aldosterone system blockade, mineralocorticoid receptor antagonists, sodium-glucose cotransporter 2 inhibitors, and glucagon-like peptide‑1 receptor agonists has reshaped cardiorenal prognosis, further challenging historical calibrations of stroke and bleeding scores. In this perspective, we critically appraise conventional risk scores in AF with CKD, highlight their limitations in the context of modern preventive therapies and competing mortality, and outline priorities for more dynamic, kidney‑focused risk stratification.

Factor [F]XI/XIa inhibitors have emerged as a promising class of investigational anticoagulants that aim to preserve hemostasis while attenuating thrombosis. By selectively targeting the intrinsic pathway, FXI/XIa inhibition may provide effective antithrombotic protection with a lower risk of bleeding compared with conventional therapies. Early clinical studies across a range of indications suggest that FXI/XIa inhibition is associated with favorable safety profiles and predictable pharmacokinetics. Ongoing phase 3 trials are still evaluating the role of FXI/XIa inhibitors in multiple settings. If efficacy and safety are confirmed in larger studies, FXI/XIa inhibitors may represent an evolution in anticoagulant therapy, potentially expanding access to effective thromboprophylaxis for patients currently underserved by existing options. Continued investigation will be essential to define optimal dosing strategies, identify patient populations most likely to benefit, and establish how these agents integrate into evolving standards of care.

Background How beta-blocker dialyzability affects cardiovascular outcomes in patients receiving hemodialysis remains unclear. Previous observational studies show conflicting results, and conducting clinical trials in this population has proven challenging. This uncertainty has contributed to substantial variation in beta-blocker prescribing patterns across regions and clinicians.Methods Using data from the United States Renal Data System, we identified patients receiving beta-blockers on the day of starting hemodialysis and classified them into nondialyzable and dialyzable exposure groups. Geographical variations in nondialyzable and dialyzable beta-blocker prescription rates across neighboring regions in each state were leveraged as an instrumental variable. The treatment effect of nondialyzable versus dialyzable beta-blockers on major adverse cardiovascular events (MACE) at 6 months, 1 year, and 3 years was evaluated using instrumental variable g-estimation methods for survival outcomes. MACE comprised myocardial infarction, stroke, and all-cause mortality.Results The cohort of 40,313 participants from 41 US states experienced 25,720 MACE over the 3-year follow-up period (median years [interquartile range], 1.4 [0.4-3.0]), with more than 42% occurring during the first 6 months after hemodialysis initiation. The instrumental variable met all statistical assumptions. Nondialyzable beta-blockers were associated with significantly lower risk of MACE at all three follow-up time points compared with dialyzable beta-blockers (hazard ratios [95% confidence intervals], 0.82 [0.74 to 0.90], 0.85 [0.79 to 0.92], and 0.90 [0.85 to 0.96], respectively). Similar patterns were observed for each separate component of MACE.Conclusions The use of nondialyzable beta-blockers compared with dialyzable beta-blockers was associated with a lower risk of MACE during all time points, including the early, high-risk period in the 6 months following hemodialysis initiation.

Rationale & Objective During the coronavirus disease 2019 (COVID-19) pandemic, the US government expanded originating telemedicine sites to include outpatient dialysis units. For the first time, nephrology care providers (nephrologists and affiliated advanced practice providers) across the United States could use telemedicine in lieu of face-to-face visits to deliver care for patients receiving in-center hemodialysis. In this study, we describe perspectives and experiences of nephrologists using telemedicine to deliver in-center hemodialysis care. Study Design Qualitative research study. Setting & Participants Nephrologists in 3 health systems who used telemedicine for in-center hemodialysis during the COVID-19 pandemic. Analytical Approach We conducted 16 semistructured telephone interviews. Transcripts were thematically analyzed. Results We identified 5 themes and respective subthemes: maintaining safety and quality of care (making up missed appointments, fostering continuity of care, addressing urgent medical issues); maximizing efficiency (reducing nephrologists' travel burden, allowing for flexibility); operational complexities (dependence on facility resources; challenges coordinating with facility staff; modifying visit duration/length); diminished depth of clinical encounters (excess formality, constrained communication, incomplete physical examinations); supporting confidence in telemedicine (complementing in-person care, accounting for patient preferences, requiring reimbursement). Limitations The transferability of the findings outside of an urban academic setting is uncertain. Conclusions Although nephrologists encountered operational (both technical and personal level) challenges such as communication constraints when using telemedicine for in-center hemodialysis care, they reported improvements in aspects of care quality and enhanced efficiency. These findings inform the potential use of a hybrid in-center hemodialysis care delivery model in which telemedicine supplements in-person visits. Plain-Language Summary We describe nephrologists’ perspectives and experiences with delivery of care using telemedicine at dialysis facilities during the COVID-19 pandemic. The nephrologists participating in this study suggested telemedicine is beneficial to maintaining patient safety and quality of care and maximizing physicians’ efficiency. Nephrologists experienced barriers given operational complexities such as technical and personnel support from dialysis facilities. Clinical encounters were also less personal and more formal. Nevertheless, the experiences of participating nephrologists suggest that a hybrid care model with both in-person and telemedicine visits may help maintain high-quality care while increasing the efficiency of delivering clinical care. These findings inform the potential use of a hybrid in-center hemodialysis care delivery model.