Validation of prediction model for successful discontinuation of continuous renal replacement therapy: a multicenter cohort study.

Background: Continuous renal replacement therapy (CRRT) has become the standard modality of renal replacement therapy (RRT) in critically ill patients. However, consensus is lacking regarding the criteria for discontinuing CRRT. Here we validated the usefulness of the prediction model for successful discontinuation of CRRT in a multicenter retrospective cohort. Methods: One temporal cohort and four external cohorts included 1,517 patients with acute kidney injury who underwent CRRT for >2 days in 2018 to 2020. The model was composed of four variables: urine output, blood urea nitrogen, serum potassium, and mean arterial pressure. Successful discontinuation of CRRT was defined as the absence of an RRT requirement for 7 days thereafter. Results: The area under the receiver operating characteristic curve (AUROC) was 0.74 (95% confidence interval, 0.71-0.76). The probabilities of successful discontinuation were approximately 17%, 35%, and 70% in the low-score, intermediate-score, and high-score groups, respectively. The model performance was good in four cohorts (AUROC, 0.73-0.75) but poor in one cohort (AUROC, 0.56). In one cohort with poor performance, attending physicians primarily controlled CRRT prescription and discontinuation, while in the other four cohorts, nephrologists determined all important steps in CRRT operation, including screening for CRRT discontinuation. Conclusion: The overall performance of our prediction model using four simple variables for successful discontinuation of CRRT was good, except for one cohort where nephrologists did not actively engage in CRRT operation. These results suggest the need for active engagement of nephrologists and protocolized management for CRRT discontinuation.

Erratum: Management for Electrolytes Disturbances during Continuous Renal Replacement Therapy.

[This corrects the article on p. 64 in vol. 20, PMID: 36688209.].

Impact of sarcopenia and phase angle on mortality of the very elderly.

BACKGROUND: Sarcopenia is a major component of geriatric syndrome and associated with poor clinical outcomes and mortality. However, diagnosing sarcopenia in the very elderly is difficult, and data on its epidemiology and devastating effects in this group are scarce. Phase angle (PA) is measured using bioimpedance spectroscopy and known to reflect cellular integrity and health. This study aimed to clarify the impact of sarcopenia and PA on mortality risk in very elderly people living in long-term care facilities. METHODS: This prospective cohort study enrolled elderly residents living in nine long-term care facilities. We collected the participants' data, such as body mass index (BMI), comorbidities and laboratory data, from September to October 2017 and mortality data until October 2019. Nutritional status was evaluated using the Mini Nutritional Assessment (MNA) score, and multifrequency bioimpedance spectroscopy was used to assess body composition including PA. Appendicular skeletal muscle mass was calculated using the body composition monitor-derived equation of Taiwan's researchers. Sarcopenia was diagnosed using the European Working Group on Sarcopenia in Older People 2 (EWGSOP2) definition (sarcopenia vs. normal group). We divided the participants into two groups according to the median PA value of 3.65° (high vs. low group) and performed multivariate regression analyses to verify the association with mortality risk according to sarcopenia diagnosis or PA group. RESULTS: A total of 279 elderly participants were enrolled; of them, 238 (85.3%) were diagnosed with sarcopenia according to EWGSOP2 guidelines. The median patient age was 83 years, 211 (75.6%) were female and the median BMI was 20.4 kg/m2 . The sarcopenia group was older than the normal group (84 vs. 81 years; P = 0.002), had a lower mean BMI (19.8 vs. 26.6 kg/m2 , P < 0.001) and had a lower MNA score (9 vs. 12 points, P < 0.001). Sarcopenia was associated with a higher mortality risk after the adjustment for age, sex and diabetes mellitus (hazard ratio [HR], 3.744; 95% confidence interval [CI], 1.155-12.134; P = 0.028). A low PA was associated with sarcopenia, older age, female sex, low MNA score and overhydration volume; it was also a significant predictor of mortality after the adjustment for age, sex, diabetes mellitus and MNA score (HR, 0.593; 95% CI, 0.420-0.837; P = 0.003). CONCLUSIONS: Sarcopenia is prevalent among the very elderly patients in long-term care facilities. Sarcopenia and low PA are significantly associated with higher mortality risk.

Impact of protocolized fluid management on electrolyte stability in patients undergoing continuous renal replacement therapy.

Objective: Continuous renal replacement therapy (CRRT) is the standard treatment for critically ill patients with acute kidney injury (AKI). Electrolyte disturbance such as hypokalemia or hypophosphatemia occurs paradoxically in patients undergoing CRRT due to high clearance. We developed a fluid management protocol for dialysate and replacement fluid that depends on serum electrolytes and focuses on potassium and phosphate levels to prevent electrolyte disturbance during CRRT. The impact of our new fluid protocol on electrolyte stability was evaluated. Methods: Adult patients who received CRRT between 2013 and 2017 were included. Patients treated 2 years before (2013-2014; pre-protocol group) and 2 years following development of the fluid protocol (2016-2017; protocol group) were compared. The primary outcomes were individual coefficient of variation (CV) and abnormal event rates of serum phosphate and potassium. Secondary outcomes were frequency of electrolyte replacement and incidence of cardiac arrhythmias. Individual CV and abnormal event rates for each electrolyte were analyzed using the Wilcoxon rank-sum test and Chi-square test with Yates' continuity correction. Results: A total of 1,448 patients was included. Both serum phosphate and potassium were higher in the protocol group. The CVs of serum phosphate (pre-protocol vs. protocol, 0.275 [0.207-0.358] vs. 0.229 [0.169-0.304], p < 0.01) and potassium (0.104 [0.081-0.135] vs. 0.085 [0.064-0.110], p < 0.01) were significantly lower in the protocol group. The abnormal event rates of serum phosphate (rate [95% CI], 0.410 [0.400-0.415] vs. 0.280 [0.273-0.286], p < 0.01) and potassium (0.205 [0.199-0.211] vs. 0.083 [0.079-0.087], p < 0.01) were also significantly lower in the protocol group. Conclusion: The protocolized management of fluid in CRRT effectively prevented hypophosphatemia and hypokalemia by inducing excellent stability of serum phosphate and potassium levels.

Deep Learning Model for Predicting Intradialytic Hypotension Without Privacy Infringement: A Retrospective Two-Center Study.

Objective: Previously developed Intradialytic hypotension (IDH) prediction models utilize clinical variables with potential privacy protection issues. We developed an IDH prediction model using minimal variables, without the risk of privacy infringement. Methods: Unidentifiable data from 63,640 hemodialysis sessions (26,746 of 79 patients for internal validation, 36,894 of 255 patients for external validation) from two Korean hospital hemodialysis databases were finally analyzed, using three IDH definitions: (1) systolic blood pressure (SBP) nadir <90 mmHg (Nadir90); (2) SBP decrease ≥20 mmHg from baseline (Fall20); and (3) SBP decrease ≥20 mmHg and/or mean arterial pressure decrease ≥10 mmHg (Fall20/MAP10). The developed models use 30 min information to predict an IDH event in the following 10 min window. Area under the receiver operating characteristic curves (AUROCs) and precision-recall curves were used to compare machine learning and deep learning models by logistic regression, XGBoost, and convolutional neural networks. Results: Among 344,714 segments, 9,154 (2.7%), 134,988 (39.2%), and 149,674 (43.4%) IDH events occurred according to three different IDH definitions (Nadir90, Fall20, and Fall20/MAP10, respectively). Compared with models including logistic regression, random forest, and XGBoost, the deep learning model achieved the best performance in predicting IDH (AUROCs: Nadir90, 0.905; Fall20, 0.864; Fall20/MAP10, 0.863) only using measurements from hemodialysis machine during dialysis session. Conclusions: The deep learning model performed well only using monitoring measurement of hemodialysis machine in predicting IDH without any personal information that could risk privacy infringement.

Management for Electrolytes Disturbances during Continuous Renal Replacement Therapy.

Despite the lack of proven superiority in mortality compared to intermittent hemodialysis, continuous renal replacement therapy (CRRT) is the preferred renal replacement therapy modality for critically ill patients with acute kidney injury (AKI) due to better hemodynamic stability and steady correction of electrolytes disturbances and volume overload. Multiple and complex electrolyte disorders in patients with AKI can be managed effectively with CRRT because controlled and predictable correction is feasible. Thus, CRRT has an advantage with safety over conventional hemodialysis, especially in patients with both renal dysfunction and electrolyte disorder that require a sophisticated treatment with avoidance of rapid correction. On the contrary, CRRT can potentially lead to paradoxical disturbance of electrolytes such as hypokalemia or hypophosphatemia, especially in patients under high dose or prolonged duration of CRRT treatment. These electrolytes related complications can be prevented with close monitoring followed by the appropriate use of CRRT fluids. Although there is a lack of solid evidence and standardized guideline for CRRT prescriptions, optimal management of various electrolyte disturbances can be achieved with individualized and tailored dialysate and replacement fluid prescriptions. Several commercially available CRRT solutions with varying compositions provide flexibility to manage electrolyte disorders and maintain the stability of electrolyte. In this review, we discuss various prescription methods to manage common electrolyte imbalances as well as preventative strategies to maintain electrolyte homeostasis during CRRT providing detailed protocols used in our center. This review may contribute to future research that can lead to the development of clinical practice guidelines.

A Scoring Model with Simple Clinical Parameters to Predict Successful Discontinuation of Continuous Renal Replacement Therapy.

BACKGROUND: Continuous renal replacement therapy (CRRT) is the standard treatment for severe acute kidney injury in critically ill patients. However, a practical consensus for discontinuing CRRT is lacking. We aimed to develop a prediction model with simple clinical parameters for successful discontinuation of CRRT. METHODS: Adult patients who received CRRT at Samsung Medical Center from 2007 to 2017 were included. Patients with preexisting ESRD and patients who progressed to ESRD within 1 year or died within 7 days after CRRT were excluded. Successful discontinuation of CRRT was defined as no requirement for renal replacement therapy for 7 days after discontinuing CRRT. Patients were assigned to either a success group or failure group according to whether discontinuation of CRRT was successful or not. RESULTS: A total of 1,158 patients were included in the final analyses. The success group showed greater urine output on the day before CRRT discontinuation (D-1) and the discontinuation day (D0). Multivariable analysis identified that urine output ≥300 mL on D-1, and mean arterial pressure 50∼78 mm Hg, serum potassium <4.1 mmol/L, and BUN <35 mg/dL (12.5 mmol/L) on D0 were predictive factors for successful discontinuation of CRRT. A scoring system using the 4 variables above (area under the receiver operating curve: 0.731) was developed. CONCLUSIONS: Scoring system composed of urine output ≥300 mL/day on D-1, and adequate blood pressure, serum potassium <4.1 mmol/L, and BUN <35 mg/dL (12.5 mmol/L) on D0 was developed to predict successful discontinuation of CRRT.

Association between diuretics and successful discontinuation of continuous renal replacement therapy in critically ill patients with acute kidney injury.

BACKGROUND: Despite aggressive application of continuous renal replacement therapy (CRRT) in critically ill patients with acute kidney injury (AKI), there is no consensus on diuretic therapy when discontinuation of CRRT is attempted. The effect of diuretics on discontinuation of CRRT in critically ill patients was evaluated. METHODS: This retrospective cohort study enrolled 1176 adult patients who survived for more than 3 days after discontinuing CRRT between 2009 and 2014. Patients were categorized depending on the re-initiation of renal replacement therapy within 3 days after discontinuing CRRT or use of diuretics. Changes in urine output (UO) and renal function after discontinuing CRRT were outcomes. Predictive factors for successful discontinuation of CRRT were also analyzed. RESULTS: The CRRT discontinuation group had a shorter duration of CRRT, more frequent use of diuretics after discontinuing CRRT, and greater UO on the day before CRRT discontinuation [day minus 1 (day - 1)]. The diuretics group had greater increases in UO and serum creatinine elevation after discontinuing CRRT. In the CRRT discontinuation group, continuous infusion of furosemide tended to increase UO more effectively. Multivariable regression analysis identified high day - 1 UO and use of diuretics as significant predictors of successful discontinuation of CRRT. Day - 1 UO of 125 mL/day was the cutoff value for predicting successful discontinuation of CRRT in oliguric patients treated with diuretics following CRRT. CONCLUSIONS: Day - 1 UO and aggressive diuretic therapy were associated with successful CRRT discontinuation. Diuretic therapy may be helpful when attempting CRRT discontinuation in critically ill patients with AKI, by inducing a favorable fluid balance, especially in oliguric patients.