European Nephrologists' Attitudes toward the Application of Artificial Intelligence in Clinical Practice: A Comprehensive Survey.

INTRODUCTION: The rapid advancement of artificial intelligence and big data analytics, including descriptive, diagnostic, predictive, and prescriptive analytics, has the potential to revolutionize many areas of medicine, including nephrology and dialysis. Artificial intelligence and big data analytics can be used to analyze large amounts of patient medical records, including laboratory results and imaging studies, to improve the accuracy of diagnosis, enhance early detection, identify patterns and trends, and personalize treatment plans for patients with kidney disease. Additionally, artificial intelligence and big data analytics can be used to identify patients' treatment who are not receiving adequate care, highlighting care inefficiencies in the dialysis provider, optimizing patient outcomes, reducing healthcare costs, and consequently creating values for all the involved stakeholders. OBJECTIVES: We present the results of a comprehensive survey aimed at exploring the attitudes of European physicians from eight countries working within a major hemodialysis network (Fresenius Medical Care NephroCare) toward the application of artificial intelligence in clinical practice. METHODS: An electronic survey on the implementation of artificial intelligence in hemodialysis clinics was distributed to 1,067 physicians. Of the 1,067 individuals invited to participate in the study, 404 (37.9%) professionals agreed to participate in the survey. RESULTS: The survey showed that a substantial proportion of respondents believe that artificial intelligence has the potential to support physicians in reducing medical malpractice or mistakes. CONCLUSION: While artificial intelligence's potential benefits are recognized in reducing medical errors and improving decision-making, concerns about treatment plan consistency, personalization, privacy, and the human aspects of patient care persist. Addressing these concerns will be crucial for successfully integrating artificial intelligence solutions in nephrology practice.

Patient Survival With Extended Home Hemodialysis Compared to In-Center Conventional Hemodialysis.

Introduction: More frequent and/or longer hemodialysis (HD) has been associated with improvements in numerous clinical outcomes in patients on dialysis. Home HD (HHD), which allows more frequent and/or longer dialysis with lower cost and flexibility in treatment planning, is not widely used worldwide. Although, retrospective studies have indicated better survival with HHD, this issue remains controversial. In this multicenter study, we compared thrice-weekly extended HHD with in-center conventional HD (ICHD) in a large patient population with a long-term follow-up. Methods: We matched 349 patients starting HHD between 2010 and 2014 with 1047 concurrent patients on ICHD by using propensity scores. Patients were followed-up with from their respective baseline until September 30, 2018. The primary outcome was overall survival. Secondary outcomes were technique survival; hospitalization; and changes in clinical, laboratory, and medication parameters. Results: The mean duration of dialysis session was 418 ± 54 minutes in HHD and 242 ± 10 minutes in patients on ICHD. All-cause mortality rate was 3.76 and 6.27 per 100 patient-years in the HHD and the ICHD groups, respectively. In the intention-to-treat analysis, HHD was associated with a 40% lower risk for all-cause mortality than ICHD (hazard ratio [HR] = 0.60; 95% confidence interval [CI] 0.45 to 0.80; P < 0.001). In HHD, the 5-year technical survival was 86.5%. HHD treatment provided better phosphate and blood pressure (BP) control, improvements in nutrition and inflammation, and reduction in hospitalization days and medication requirement. Conclusion: These results indicate that extended HHD is associated with higher survival and better outcomes compared to ICHD.

War in Ukraine and dialysis treatment: human suffering and organizational challenges.

In January 2021, there were 9648 patients in Ukraine on kidney replacement therapy, including 8717 on extracorporeal therapies and 931 on peritoneal dialysis. On 24 February 2022, foreign troops entered the territory of Ukraine. Before the war, the Fresenius Medical Care dialysis network in Ukraine operated three medical centres. These medical centres provided haemodialysis therapy to 349 end-stage kidney disease patients. In addition, Fresenius Medical Care Ukraine delivered medical supplies to almost all regions of Ukraine. Even though Fresenius Medical Care's share of end-stage kidney disease patients on dialysis is small, a brief narrative account of the managerial challenges that Fresenius Medical Care Ukraine and the clinical directors of the Fresenius Medical Care centres had to face, as well as the suffering of the dialysis population, is a useful testimony of the burden imposed by war on these frail, high-risk patients dependent on a complex technology such as dialysis. The war in Ukraine is causing immense suffering for the dialysis population of this country and has called for heroic efforts from dialysis personnel. The experience of a small dialysis network treating a minority of dialysis patients in Ukraine is described. Guaranteeing dialysis treatment has been and remains an enormous challenge in Ukraine and we are confident that the generosity and the courage of Ukrainian dialysis staff and international aid will help to mitigate this tragic suffering.

Real-time prediction of intradialytic hypotension using machine learning and cloud computing infrastructure.

BACKGROUND: In maintenance hemodialysis patients, intradialytic hypotension (IDH) is a frequent complication that has been associated with poor clinical outcomes. Prediction of IDH may facilitate timely interventions and eventually reduce IDH rates. METHODS: We developed a machine learning model to predict IDH in in-center hemodialysis patients 15-75 min in advance. IDH was defined as systolic blood pressure (SBP) <90 mmHg. Demographic, clinical, treatment-related and laboratory data were retrieved from electronic health records and merged with intradialytic machine data that were sent in real-time to the cloud. For model development, dialysis sessions were randomly split into training (80%) and testing (20%) sets. The area under the receiver operating characteristic curve (AUROC) was used as a measure of the model's predictive performance. RESULTS: We utilized data from 693 patients who contributed 42 656 hemodialysis sessions and 355 693 intradialytic SBP measurements. IDH occurred in 16.2% of hemodialysis treatments. Our model predicted IDH 15-75 min in advance with an AUROC of 0.89. Top IDH predictors were the most recent intradialytic SBP and IDH rate, as well as mean nadir SBP of the previous 10 dialysis sessions. CONCLUSIONS: Real-time prediction of IDH during an ongoing hemodialysis session is feasible and has a clinically actionable predictive performance. If and to what degree this predictive information facilitates the timely deployment of preventive interventions and translates into lower IDH rates and improved patient outcomes warrants prospective studies.

Assessing the impact of transitional care units on dialysis patient outcomes: A multicenter, propensity score-matched analysis.

INTRODUCTION: Inadequate predialysis care and education impacts the selection of a dialysis modality and is associated with adverse clinical outcomes. Transitional care units (TCUs) aim to meet the unmet educational needs of incident dialysis patients, but their impact beyond increasing home dialysis utilization has been incompletely characterized. METHODS: This retrospective study included adults initiating in-center hemodialysis at a TCU, matched to controls (1:4) with no TCU history initiating in-center hemodialysis. Patients were followed for up to 14 months. TCUs are dedicated spaces where staff provide personalized education and as-needed adjustments to dialysis prescriptions. For many patients, therapy was initiated with four to five weekly dialysis sessions, with at least some sessions delivered by home dialysis machines. Outcomes included survival, first hospitalization, transplant waiting-list status, post-TCU dialysis modality, and vascular access type. FINDINGS: The study included 724 patients initiating dialysis across 48 TCUs, with 2892 well-matched controls. At the end of 14 months, patients initiating dialysis in a TCU were significantly more likely to be referred and/or wait-listed for a kidney transplant than controls (57% vs. 42%; p < 0.0001). Initiation of dialysis at a TCU was also associated with significantly lower rates of receiving in-center hemodialysis at 14 months (74% vs. 90%; p < 0.0001) and higher rates of arteriovenous access (70% vs. 63%; p = 0.003). Although not statistically significant, TCU patients were more likely to survive and less likely to be hospitalized during follow-up than controls. DISCUSSION: Although TCUs are sometimes viewed as only a means for enhancing utilization of home dialysis, patients attending TCUs exhibited more favorable outcomes across all endpoints. In addition to being 2.5-fold more likely to receive home dialysis, TCU patients were 42% more likely to be referred for transplantation. Our results support expanding utilization of TCUs for patients with inadequate predialysis support.

Predictors of shorter- and longer-term mortality after COVID-19 presentation among dialysis patients: parallel use of machine learning models in Latin and North American countries.

BACKGROUND: We developed machine learning models to understand the predictors of shorter-, intermediate-, and longer-term mortality among hemodialysis (HD) patients affected by COVID-19 in four countries in the Americas. METHODS: We used data from adult HD patients treated at regional institutions of a global provider in Latin America (LatAm) and North America who contracted COVID-19 in 2020 before SARS-CoV-2 vaccines were available. Using 93 commonly captured variables, we developed machine learning models that predicted the likelihood of death overall, as well as during 0-14, 15-30, > 30 days after COVID-19 presentation and identified the importance of predictors. XGBoost models were built in parallel using the same programming with a 60%:20%:20% random split for training, validation, & testing data for the datasets from LatAm (Argentina, Columbia, Ecuador) and North America (United States) countries. RESULTS: Among HD patients with COVID-19, 28.8% (1,001/3,473) died in LatAm and 20.5% (4,426/21,624) died in North America. Mortality occurred earlier in LatAm versus North America; 15.0% and 7.3% of patients died within 0-14 days, 7.9% and 4.6% of patients died within 15-30 days, and 5.9% and 8.6% of patients died > 30 days after COVID-19 presentation, respectively. Area under curve ranged from 0.73 to 0.83 across prediction models in both regions. Top predictors of death after COVID-19 consistently included older age, longer vintage, markers of poor nutrition and more inflammation in both regions at all timepoints. Unique patient attributes (higher BMI, male sex) were top predictors of mortality during 0-14 and 15-30 days after COVID-19, yet not mortality > 30 days after presentation. CONCLUSIONS: Findings showed distinct profiles of mortality in COVID-19 in LatAm and North America throughout 2020. Mortality rate was higher within 0-14 and 15-30 days after COVID-19 in LatAm, while mortality rate was higher in North America > 30 days after presentation. Nonetheless, a remarkable proportion of HD patients died > 30 days after COVID-19 presentation in both regions. We were able to develop a series of suitable prognostic prediction models and establish the top predictors of death in COVID-19 during shorter-, intermediate-, and longer-term follow up periods.

Fatigue in incident peritoneal dialysis and mortality: A real-world side-by-side study in Brazil and the United States.

BACKGROUND: We tested if fatigue in incident Peritoneal Dialysis associated with an increased risk for mortality, independently from main confounders. METHODS: We conducted a side-by-side study from two of incident PD patients in Brazil and the United States. We used the same code to independently analyze data in both countries during 2004 to 2011. We included data from adults who completed KDQOL-SF vitality subscale within 90 days after starting PD. Vitality score was categorized in four groups: >50 (high vitality), ≥40 to ≤50 (moderate vitality), >35 to <40 (moderate fatigue), ≤35 (high fatigue; reference group). In each country's cohort, we built four distinct models to estimate the associations between vitality (exposure) and all-cause mortality (outcome): (i) Cox regression model; (ii) competing risk model accounting for technique failure events; (iii) multilevel survival model of clinic-level clusters; (iv) multivariate regression model with smoothing splines treating vitality as a continuous measure. Analyses were adjusted for age, comorbidities, PD modality, hemoglobin, and albumin. A mixed-effects meta-analysis was used to pool hazard ratios (HRs) from both cohorts to model mortality risk for each 10-unit increase in vitality. RESULTS: We used data from 4,285 PD patients (Brazil n = 1,388 and United States n = 2,897). Model estimates showed lower vitality levels within 90 days of starting PD were associated with a higher risk of mortality, which was consistent in Brazil and the United States cohorts. In the multivariate survival model, each 10-unit increase in vitality score was associated with lower risk of all-cause mortality in both cohorts (Brazil HR = 0.79 [95%CI 0.70 to 0.90] and United States HR = 0.90 [95%CI 0.88 to 0.93], pooled HR = 0.86 [95%CI 0.75 to 0.98]). Results for all models provided consistent effect estimates. CONCLUSIONS: Among patients in Brazil and the United States, lower vitality score in the initial months of PD was independently associated with all-cause mortality.

Trajectories of clinical and laboratory characteristics associated with COVID-19 in hemodialysis patients by survival.

INTRODUCTION: The clinical impact of COVID-19 has not been established in the dialysis population. We evaluated the trajectories of clinical and laboratory parameters in hemodialysis (HD) patients. METHODS: We used data from adult HD patients treated at an integrated kidney disease company who received a reverse transcription polymerase chain reaction (RT-PCR) test to investigate suspicion of a severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) infection between May 1 and September 1, 2020. Nonparametric smoothing splines were used to fit data for individual trajectories and estimate the mean change over time in patients testing positive or negative for SARS-CoV-2 and those who survived or died within 30 days of first suspicion or positive test date. For each clinical parameter of interest, the difference in average daily changes between COVID-19 positive versus negative group and COVID-19 survivor versus nonsurvivor group was estimated by fitting a linear mixed effects model based on measurements in the 14 days before (i.e., Day -14 to Day 0) Day 0. RESULTS: There were 12,836 HD patients with a suspicion of COVID-19 who received RT-PCR testing (8895 SARS-CoV-2 positive). We observed significantly different trends (p < 0.05) in pre-HD systolic blood pressure (SBP), pre-HD pulse rate, body temperature, ferritin, neutrophils, lymphocytes, albumin, and interdialytic weight gain (IDWG) between COVID-19 positive and negative patients. For COVID-19 positive group, we observed significantly different clinical trends (p < 0.05) in pre-HD pulse rate, lymphocytes, neutrophils, and albumin between survivors and nonsurvivors. We also observed that, in the group of survivors, most clinical parameters returned to pre-COVID-19 levels within 60-90 days. CONCLUSION: We observed unique temporal trends in various clinical and laboratory parameters among HD patients who tested positive versus negative for SARS-CoV-2 infection and those who survived the infection versus those who died. These trends can help to define the physiological disturbances that characterize the onset and course of COVID-19 in HD patients.

Transmission of SARS-CoV-2 considering shared chairs in outpatient dialysis: a real-world case-control study.

BACKGROUND: SARS-CoV-2 can remain transiently viable on surfaces. We examined if use of shared chairs in outpatient hemodialysis associates with a risk for indirect patient-to-patient transmission of SARS-CoV-2. METHODS: We used data from adults treated at 2,600 hemodialysis facilities in United States between February 1st and June 8th, 2020. We performed a retrospective case-control study matching each SARS-CoV-2 positive patient (case) to a non-SARS-CoV-2 patient (control) treated in the same dialysis shift. Cases and controls were matched on age, sex, race, facility, shift date, and treatment count. For each case-control pair, we traced backward 14 days to assess possible prior exposure from a 'shedding' SARS-CoV-2 positive patient who sat in the same chair immediately before the case or control. Conditional logistic regression models tested whether chair exposure after a shedding SARS-CoV-2 positive patient conferred a higher risk of SARS-CoV-2 infection to the immediate subsequent patient. RESULTS: Among 170,234 hemodialysis patients, 4,782 (2.8 %) tested positive for SARS-CoV-2 (mean age 64 years, 44 % female). Most facilities (68.5 %) had 0 to 1 positive SARS-CoV-2 patient. We matched 2,379 SARS-CoV-2 positive cases to 2,379 non-SARS-CoV-2 controls; 1.30 % (95 %CI 0.90 %, 1.87 %) of cases and 1.39 % (95 %CI 0.97 %, 1.97 %) of controls were exposed to a chair previously sat in by a shedding SARS-CoV-2 patient. Transmission risk among cases was not significantly different from controls (OR = 0.94; 95 %CI 0.57 to 1.54; p = 0.80). Results remained consistent in adjusted and sensitivity analyses. CONCLUSIONS: The risk of indirect patient-to-patient transmission of SARS-CoV-2 infection from dialysis chairs appears to be low.

Real-time prediction of intradialytic relative blood volume: a proof-of-concept for integrated cloud computing infrastructure.

BACKGROUND: Inadequate refilling from extravascular compartments during hemodialysis can lead to intradialytic symptoms, such as hypotension, nausea, vomiting, and cramping/myalgia. Relative blood volume (RBV) plays an important role in adapting the ultrafiltration rate which in turn has a positive effect on intradialytic symptoms. It has been clinically challenging to identify changes RBV in real time to proactively intervene and reduce potential negative consequences of volume depletion. Leveraging advanced technologies to process large volumes of dialysis and machine data in real time and developing prediction models using machine learning (ML) is critical in identifying these signals. METHOD: We conducted a proof-of-concept analysis to retrospectively assess near real-time dialysis treatment data from in-center patients in six clinics using Optical Sensing Device (OSD), during December 2018 to August 2019. The goal of this analysis was to use real-time OSD data to predict if a patient's relative blood volume (RBV) decreases at a rate of at least - 6.5 % per hour within the next 15 min during a dialysis treatment, based on 10-second windows of data in the previous 15 min. A dashboard application was constructed to demonstrate how reporting structures may be developed to alert clinicians in real time of at-risk cases. Data was derived from three sources: (1) OSDs, (2) hemodialysis machines, and (3) patient electronic health records. RESULTS: Treatment data from 616 in-center dialysis patients in the six clinics was curated into a big data store and fed into a Machine Learning (ML) model developed and deployed within the cloud. The threshold for classifying observations as positive or negative was set at 0.08. Precision for the model at this threshold was 0.33 and recall was 0.94. The area under the receiver operating curve (AUROC) for the ML model was 0.89 using test data. CONCLUSIONS: The findings from our proof-of concept analysis demonstrate the design of a cloud-based framework that can be used for making real-time predictions of events during dialysis treatments. Making real-time predictions has the potential to assist clinicians at the point of care during hemodialysis.

Machine learning directed interventions associate with decreased hospitalization rates in hemodialysis patients.

BACKGROUND: An integrated kidney disease company uses machine learning (ML) models that predict the 12-month risk of an outpatient hemodialysis (HD) patient having multiple hospitalizations to assist with directing personalized interdisciplinary interventions in a Dialysis Hospitalization Reduction Program (DHRP). We investigated the impact of risk directed interventions in the DHRP on clinic-wide hospitalization rates. METHODS: We compared the hospital admission and day rates per-patient-year (ppy) from all hemodialysis patients in 54 DHRP and 54 control clinics identified by propensity score matching at baseline in 2015 and at the end of the pilot in 2018. We also used paired T test to compare the between group difference of annual hospitalization rate and hospitalization days rates at baseline and end of the pilot. RESULTS: The between group difference in annual hospital admission and day rates was similar at baseline (2015) with a mean difference between DHRP versus control clinics of -0.008 ± 0.09 ppy and -0.05 ± 0.96 ppy respectively. The between group difference in hospital admission and day rates became more distinct at the end of follow up (2018) favoring DHRP clinics with the mean difference being -0.155 ± 0.38 ppy and -0.97 ± 2.78 ppy respectively. A paired t-test showed the change in the between group difference in hospital admission and day rates from baseline to the end of the follow up was statistically significant (t-value = 2.73, p-value < 0.01) and (t-value = 2.29, p-value = 0.02) respectively. CONCLUSIONS: These findings suggest ML model-based risk-directed interdisciplinary team interventions associate with lower hospitalization rates and hospital day rate in HD patients, compared to controls.

Transmission of SARS-CoV-2 Considering Shared Chairs in Outpatient Dialysis: A Real-World Case-Control Study.

BACKGROUND: SARS-CoV-2 is primarily transmitted through aerosolized droplets; however, the virus can remain transiently viable on surfaces. OBJECTIVE: We examined transmission within hemodialysis facilities, with a specific focus on the possibility of indirect patient-to-patient transmission through shared dialysis chairs. DESIGN: We used real-world data from hemodialysis patients treated between February 1 st and June 8 th , 2020 to perform a case-control study matching each SARS-CoV-2 positive patient (case) to a non-SARS-CoV-2 patient (control) in the same dialysis shift and traced back 14 days to capture possible exposure from chairs sat in by SARS-CoV-2 patients. Cases and controls were matched on age, sex, race, facility, shift date, and treatment count. SETTING: 2,600 hemodialysis facilities in the United States. PATIENTS: Adult (age ≥18 years) hemodialysis patients. MEASUREMENTS: Conditional logistic regression models tested whether chair exposure after a positive patient conferred a higher risk of SARS-CoV-2 infection to the immediate subsequent patient. RESULTS: Among 170,234 hemodialysis patients, 4,782 (2.8%) tested positive for SARS-CoV-2 (mean age 64 years, 44% female). Most facilities (68.5%) had 0 to 1 positive SARS-CoV-2 patient. We matched 2,379 SARS-CoV-2 positive cases to 2,379 non-SARS-CoV-2 controls; 1.30% (95%CI 0.90%, 1.87%) of cases and 1.39% (95%CI 0.97%, 1.97%) of controls were exposed to a chair previously sat in by a shedding SARS-CoV-2 patient. Transmission risk among cases was not significantly different from controls (OR=0.94; 95%CI 0.57 to 1.54; p=0.80). Results remained consistent in adjusted and sensitivity analyses. LIMITATION: Analysis used real-world data that could contain errors and only considered vertical transmission associated with shared use of dialysis chairs by symptomatic patients. CONCLUSIONS: The risk of indirect patient-to-patient transmission of SARS-CoV-2 infection from dialysis chairs appears to be low. PRIMARY FUNDING SOURCE: Fresenius Medical Care North America; National Institute of Diabetes and Digestive and Kidney Diseases (R01DK130067).

Relationship between serum phosphate levels and survival in chronic hemodialysis patients: interactions with age, malnutrition and inflammation.

BACKGROUND: Evidence indicates that the inverse relationships between phosphate levels and mortality maybe modified by age. Furthermore, malnutrition and inflammation could strengthen the risk associated with phosphate abnormalities. This study aimed to assess the associations between phosphate levels and mortality while accounting for the interactions with age and parameters associated with malnutrition and inflammation in hemodialysis (HD) patients. METHODS: Adult HD patients (n = 245 853) treated in Fresenius Medical Care North America clinics from January 2010 to October 2018 were enrolled. Baseline was defined as Months 4-6 on dialysis, with the subsequent 12 months as the follow-up period. Univariate and multivariate Cox proportional hazard models with spline terms were applied to study the nonlinear relationships between serum phosphate levels and mortality. The interactions of phosphate levels with albumin, creatinine, normalized protein catabolic rate (nPCR) and neutrophil-lymphocyte ratio (NLR) were assessed with smoothing spline analysis of variance Cox proportional hazard models. RESULTS: Older patients tended to have lower levels of serum phosphate, albumin, creatinine and nPCR. Additionally, both low (<4.0 mg/dL) and high (>5.5 mg/dL) phosphate levels were associated with higher risk of mortality across all age strata. The U-shaped relationships between phosphate levels and outcome persisted even for patients with low or high levels of serum albumin, creatinine, nPCR and NLR, respectively. CONCLUSION: The consistent U-shaped relationships between serum phosphate and mortality across age strata and levels of inflammatory and nutritional status should prompt the search for underlying causes and potentially nutritional intervention in clinical practice.

Hepatitis B Vaccination Response in Hemodialysis Patients: The Impact of Dialysis Shift.

BACKGROUND/AIMS: Hepatitis B (HB) vaccination in hemodialysis patients is important as they are at a higher risk of contracting HB. However, hemodialysis patients have a lower HB seroconversion rate than their healthy counterparts. As better sleep has been associated with better seroconversion in healthy populations and early hemodialysis start has been linked to significant sleep-wake disturbances in hemodialysis patients, we examined if hemodialysis treatment start time is associated with HB vaccination response. METHODS: Demographics, standard-of-care clinical, laboratory, and treatment parameters, dialysis shift data, HB antigen status, HB vaccination status, and HB titers were collected from hemodialysis patients in Fresenius clinics from January 2010 to December 2015. Patients in our analysis received 90% of dialysis treatments either before or after 8:30 a.m., were negative for HB antigen, and received a complete series of HB vaccination (Engerix B® or Recombivax HB™). Univariate and multivariate regression models examined whether dialysis start time is a predictor of HB vaccination response. RESULTS: Patients were 65 years old, 57% male, and had a HD vintage of 10 months. Patients whose dialysis treatments started before 8:30 a.m. were more likely to be younger, male, and have a greater dialysis vintage. Patients receiving Engerix B® and starting dialysis before 8:30 a.m. had a significantly higher seroconversion rate compared to patients who started dialysis after 8:30 a.m. Early dialysis start was a significant predictor of seroconversion in univariate and multivariate regression including male gender, but not in multivariate regression including age, neutrophil-to-lymphocyte ratio, and vintage. CONCLUSION: While better sleep following vaccination is associated with seroconversion in the general population, this is not the case in hemodialysis patients after multivariate adjustment. In the context of end-stage kidney disease, early dialysis start is not a significant predictor of HB vaccination response. The association between objectively measured postvaccination sleep duration and seroconversion rate should be investigated.

Machine Learning for Prediction of Patients on Hemodialysis with an Undetected SARS-CoV-2 Infection.

Background: We developed a machine learning (ML) model that predicts the risk of a patient on hemodialysis (HD) having an undetected SARS-CoV-2 infection that is identified after the following ≥3 days. Methods: As part of a healthcare operations effort, we used patient data from a national network of dialysis clinics (February-September 2020) to develop an ML model (XGBoost) that uses 81 variables to predict the likelihood of an adult patient on HD having an undetected SARS-CoV-2 infection that is identified in the subsequent ≥3 days. We used a 60%:20%:20% randomized split of COVID-19-positive samples for the training, validation, and testing datasets. Results: We used a select cohort of 40,490 patients on HD to build the ML model (11,166 patients who were COVID-19 positive and 29,324 patients who were unaffected controls). The prevalence of COVID-19 in the cohort (28% COVID-19 positive) was by design higher than the HD population. The prevalence of COVID-19 was set to 10% in the testing dataset to estimate the prevalence observed in the national HD population. The threshold for classifying observations as positive or negative was set at 0.80 to minimize false positives. Precision for the model was 0.52, the recall was 0.07, and the lift was 5.3 in the testing dataset. Area under the receiver operating characteristic curve (AUROC) and area under the precision-recall curve (AUPRC) for the model was 0.68 and 0.24 in the testing dataset, respectively. Top predictors of a patient on HD having a SARS-CoV-2 infection were the change in interdialytic weight gain from the previous month, mean pre-HD body temperature in the prior week, and the change in post-HD heart rate from the previous month. Conclusions: The developed ML model appears suitable for predicting patients on HD at risk of having COVID-19 at least 3 days before there would be a clinical suspicion of the disease.

Artificial intelligence enabled applications in kidney disease.

Artificial intelligence (AI) is considered as the next natural progression of traditional statistical techniques. Advances in analytical methods and infrastructure enable AI to be applied in health care. While AI applications are relatively common in fields like ophthalmology and cardiology, its use is scarcely reported in nephrology. We present the current status of AI in research toward kidney disease and discuss future pathways for AI. The clinical applications of AI in progression to end-stage kidney disease and dialysis can be broadly subdivided into three main topics: (a) predicting events in the future such as mortality and hospitalization; (b) providing treatment and decision aids such as automating drug prescription; and (c) identifying patterns such as phenotypical clusters and arteriovenous fistula aneurysm. At present, the use of prediction models in treating patients with kidney disease is still in its infancy and further evidence is needed to identify its relative value. Policies and regulations need to be addressed before implementing AI solutions at the point of care in clinics. AI is not anticipated to replace the nephrologists' medical decision-making, but instead assist them in providing optimal personalized care for their patients.

Mineral bone disease in autosomal dominant polycystic kidney disease.

Mice with disruption of Pkd1 in osteoblasts demonstrate reduced bone mineral density, trabecular bone volume and cortical thickness. To date, the bone phenotype in adult patients with autosomal dominant polycystic kidney disease (ADPKD) with stage I and II chronic kidney disease has not been investigated. To examine this, we characterized biochemical markers of mineral metabolism, examined bone turnover and biology, and estimated risk of fracture in patients with ADPKD. Markers of mineral metabolism were measured in 944 patients with ADPKD and other causes of kidney disease. Histomorphometry and immunohistochemistry were compared on bone biopsies from 20 patients with ADPKD with a mean eGFR of 97 ml/min/1.73m2 and 17 healthy individuals. Furthermore, adults with end stage kidney disease (ESKD) initiating hemodialysis between 2002-2013 and estimated the risk of bone fracture associated with ADPKD as compared to other etiologies of kidney disease were examined. Intact fibroblast growth factor 23 was higher and total alkaline phosphatase lower in patients with compared to patients without ADPKD with chronic kidney disease. Compared to healthy individuals, patients with ADPKD demonstrated significantly lower osteoid volume/bone volume (0.61 vs. 1.21%) and bone formation rate/bone surface (0.012 vs. 0.026 µm3/µm2/day). ESKD due to ADPKD was not associated with a higher risk of fracture as compared to ESKD due to diabetes (age adjusted incidence rate ratio: 0.53 (95% confidence interval 0.31, 0.74) or compared to other etiologies of kidney disease. Thus, individuals with ADPKD have lower alkaline phosphatase, higher circulating intact fibroblast growth factor 23 and decreased bone formation rate. However, ADPKD is not associated with higher rates of bone fracture in ESKD.

Impacts of dialysis adequacy and intradialytic hypotension on changes in dialysis recovery time.

BACKGROUND: Dialysis recovery time (DRT) surveys capture the perceived time after HD to return to performing regular activities. Prior studies suggest the majority of HD patients report a DRT > 2 h. However, the profiles of and modifiable dialysis practices associated with changes in DRT relative to the start of dialysis are unknown. We hypothesized hemodialysis (HD) dose and rates of intradialytic hypotension (IDH) would associate with changes in DRT in the first years after initiating dialysis. METHODS: We analyzed data from adult HD patients who responded to a DRT survey ≤180 days from first date of dialysis (FDD) during 2014 to 2017. DRT survey was administered with annual KDQOL survey. DRT survey asks: "How long does it take you to be able to return to your normal activities after your dialysis treatment?" Answers are: < 0.5, 0.5-to-1, 1-to-2, 2-to-4, or > 4 h. An adjusted logistic regression model computed odds ratio for a change to a longer DRT (increase above DRT > 2 h) in reference to a change to a shorter DRT (decrease below DRT < 2 h, or from DRT > 4 h). Changes in DRT were calculated from incident (≤180 days FDD) to first prevalent (> 365-to- ≤ 545 days FDD) and second prevalent (> 730-to- ≤ 910 days FDD) years. RESULTS: Among 98,616 incident HD patients (age 62.6 ± 14.4 years, 57.8% male) who responded to DRT survey, a higher spKt/V in the incident period was associated with 13.5% (OR = 0.865; 95%CI 0.801-to-0.935) lower risk of a change to a longer DRT in the first-prevalent year. A higher number of HD treatments with IDH episodes per month in the incident period was associated with a 0.8% (OR = 1.008; 95%CI 1.001-to-1.015) and 1.6% (OR = 1.016; 95%CI 1.006-to-1.027) higher probability of a change to a longer DRT in the first- and second-prevalent years, respectively. Consistently, an increased in incidence of IDH episodes/months was associated to a change to a longer DRT over time. CONCLUSIONS: Incident patients who had higher spKt/V and less sessions with IDH episodes had a lower likelihood of changing to a longer DRT in first year of HD. Dose optimization strategies with cardiac stability in fluid removal should be tested.