Subtyping Hyperchloremia among Hospitalized Patients by Machine Learning Consensus Clustering.

Background and Objectives: Despite the association between hyperchloremia and adverse outcomes, mortality risks among patients with hyperchloremia have not consistently been observed among all studies with different patient populations with hyperchloremia. The objective of this study was to characterize hyperchloremic patients at hospital admission into clusters using an unsupervised machine learning approach and to evaluate the mortality risk among these distinct clusters. Materials and Methods: We performed consensus cluster analysis based on demographic information, principal diagnoses, comorbidities, and laboratory data among 11,394 hospitalized adult patients with admission serum chloride of >108 mEq/L. We calculated the standardized mean difference of each variable to identify each cluster's key features. We assessed the association of each hyperchloremia cluster with hospital and one-year mortality. Results: There were three distinct clusters of patients with admission hyperchloremia: 3237 (28%), 4059 (36%), and 4098 (36%) patients in clusters 1 through 3, respectively. Cluster 1 was characterized by higher serum chloride but lower serum sodium, bicarbonate, hemoglobin, and albumin. Cluster 2 was characterized by younger age, lower comorbidity score, lower serum chloride, and higher estimated glomerular filtration (eGFR), hemoglobin, and albumin. Cluster 3 was characterized by older age, higher comorbidity score, higher serum sodium, potassium, and lower eGFR. Compared with cluster 2, odds ratios for hospital mortality were 3.60 (95% CI 2.33-5.56) for cluster 1, and 4.83 (95% CI 3.21-7.28) for cluster 3, whereas hazard ratios for one-year mortality were 4.49 (95% CI 3.53-5.70) for cluster 1 and 6.96 (95% CI 5.56-8.72) for cluster 3. Conclusions: Our cluster analysis identified three clinically distinct phenotypes with differing mortality risks in hospitalized patients with admission hyperchloremia.

Safety and compatibility of smart device heart rhythm monitoring in patients with cardiovascular implantable electronic devices.

INTRODUCTION: Emerging medical technology has allowed for monitoring of heart rhythm abnormalities using smartphone compatible devices. The safety and utility of such devices have not been established in patients with cardiac implantable electronic devices (CIEDs). We sought to assess the safety and compatibility of the Food and Drug Administration-approved AliveCor Kardia device in patients with CIEDs. METHODS AND RESULTS: We prospectively recruited patients with CIED for a Kardia recording during their routine device interrogation. A recording was obtained in paced and nonpaced states. Adverse clinical events were noted at the time of recording. Electrograms (EGMs) from the cardiac device were obtained at the time of recording to assess for any electromagnetic interference (EMI) introduced by Kardia. Recordings were analyzed for quality and given a score of 3 (interpretable rhythm, no noise), 2 (interpretable rhythm, significant noise) or 1 (uninterpretable). A total of 251 patients were recruited (59% with a pacemaker and 41% with ICD). There were no adverse clinical events noted at the time of recording and no changes to CIED settings. Review of all EGMs revealed no EMI introduced by Kardia. Recordings were correctly interpreted in 90% of paced recordings (183 had a score of 3, 43 of 2, and 25 of 1) and 94.7% of nonpaced recordings (147 of 3, 15 of 2, and 9 of 1). CONCLUSION: The AliveCor Kardia device has an excellent safety profile when used in conjunction with most CIEDs. The quality of recordings was preserved in this population. The device, therefore, can be considered for heart rhythm monitoring in patients with CIEDs.

Noninvasive blood potassium measurement using signal-processed, single-lead ecg acquired from a handheld smartphone.

OBJECTIVE: We have previously used a 12-lead, signal-processed ECG to calculate blood potassium levels. We now assess the feasibility of doing so with a smartphone-enabled single lead, to permit remote monitoring. PATIENTS AND METHODS: Twenty-one hemodialysis patients held a smartphone equipped with inexpensive FDA-approved electrodes for three 2min intervals during hemodialysis. Individualized potassium estimation models were generated for each patient. ECG-calculated potassium values were compared to blood potassium results at subsequent visits to evaluate the accuracy of the potassium estimation models. RESULTS: The mean absolute error between the estimated potassium and blood potassium 0.38±0.32 mEq/L (9% of average potassium level) decreasing to 0.6 mEq/L using predictors of poor signal. CONCLUSIONS: A single-lead ECG acquired using electrodes attached to a smartphone device can be processed to calculate the serum potassium with an error of 9% in patients undergoing hemodialysis. SUMMARY: A single-lead ECG acquired using electrodes attached to a smartphone can be processed to calculate the serum potassium in patients undergoing hemodialysis remotely.

Predictors of outpatient kidney function recovery among patients who initiate hemodialysis in the hospital.

BACKGROUND: Recent policy clarifications by the Centers for Medicare & Medicaid Services have changed access to outpatient dialysis care at end-stage renal disease (ESRD) facilities for individuals with acute kidney injury in the United States. Tools to predict "ESRD" and "acute" status in terms of kidney function recovery among patients who previously initiated dialysis therapy in the hospital could help inform patient management decisions. STUDY DESIGN: Historical cohort study. SETTING & PARTICIPANTS: Incident hemodialysis patients in the Mayo Clinic Health System who initiated in-hospital renal replacement therapy (RRT) and continued outpatient dialysis following hospital dismissal (2006 through 2009). PREDICTOR: Baseline estimated glomerular filtration rate (eGFR), acute tubular necrosis from sepsis or surgery, heart failure, intensive care unit, and dialysis access. OUTCOMES: Kidney function recovery defined as sufficient kidney function for outpatient hemodialysis therapy discontinuation. RESULTS: Cohort consisted of 281 patients with a mean age of 64 years, 63% men, 45% with heart failure, and baseline eGFR≥30mL/min/1.73m(2) in 46%. During a median of 8 months, 52 (19%) recovered, most (94%) within 6 months. Higher baseline eGFR (HR per 10-mL/min/1.73m(2) increase eGFR, 1.27; 95% CI, 1.16-1.39; P<0.001), acute tubular necrosis from sepsis or surgery (HR, 3.34; 95% CI, 1.83-6.24; P<0.001), and heart failure (HR, 0.40; 95% CI, 0.19-0.78, P=0.007) were independent predictors of recovery within 6 months, whereas first RRT in the intensive care unit and catheter dialysis access were not. There was a positive interaction between absence of heart failure and eGFR≥30mL/min/1.73m(2) for predicting kidney function recovery (P<0.001). LIMITATIONS: Sample size. CONCLUSIONS: Kidney function recovery in the outpatient hemodialysis unit following in-hospital RRT initiation is not rare. As expected, higher baseline eGFR is an important determinant of recovery. However, patients with heart failure are less likely to recover even with a higher baseline eGFR. Consideration of these factors at hospital discharge informs decisions on ESRD status designation and long-term hemodialysis care.

Incidence of Adverse Events during Continuous Renal Replacement Therapy.

BACKGROUND/AIMS: The incidence of adverse events (AEs) in adults who receive continuous renal replacement therapy (CRRT) is unknown. We report the incidence of mechanical, metabolic, and hemodynamic CRRT AEs. METHODS: This is a retrospective study of all consecutive adult patients (≥18 years) who underwent CRRT from January 1, 2007 to December 31, 2009. RESULTS: Out of 595 patients who underwent CRRT, 366 (62%) were male and 500 (84%) were Caucasian. Regional citrate anticoagulation was used in 98.6% of all patients. The most common clinically significant electrolyte derangements were ionized hypocalcemia (22%), ionized hypercalcemia (23%), and hyperphosphatemia (44%). Almost all (97%) patients had at least one additional AE including new onset hypotension (within the first hour after CRRT initiation) (43%), hypothermia (44%), new onset arrhythmias (29%), new onset anemia (31%) and thrombocytopenia (40%). CONCLUSIONS: ICU patients who require CRRT have a high incidence of AEs. Although the extent to which these complications are attributable to CRRT is not known, clinicians need to be cautious and aware of their high prevalence in this patient population.

Noninvasive potassium determination using a mathematically processed ECG: proof of concept for a novel "blood-less, blood test".

OBJECTIVE: To determine if ECG repolarization measures can be used to detect small changes in serum potassium levels in hemodialysis patients. PATIENTS AND METHODS: Signal-averaged ECGs were obtained from standard ECG leads in 12 patients before, during, and after dialysis. Based on physiological considerations, five repolarization-related ECG measures were chosen and automatically extracted for analysis: the slope of the T wave downstroke (T right slope), the amplitude of the T wave (T amplitude), the center of gravity (COG) of the T wave (T COG), the ratio of the amplitude of the T wave to amplitude of the R wave (T/R amplitude), and the center of gravity of the last 25% of the area under the T wave curve (T4 COG) (Fig. 1). RESULTS: The correlations with potassium were statistically significant for T right slope (P<0.0001), T COG (P=0.007), T amplitude (P=0.0006) and T/R amplitude (P=0.03), but not T4 COG (P=0.13). Potassium changes as small as 0.2mmol/L were detectable. CONCLUSION: Small changes in blood potassium concentrations, within the normal range, resulted in quantifiable changes in the processed, signal-averaged ECG. This indicates that non-invasive, ECG-based potassium measurement is feasible and suggests that continuous or remote monitoring systems could be developed to detect early potassium deviations among high-risk patients, such as those with cardiovascular and renal diseases. The results of this feasibility study will need to be further confirmed in a larger cohort of patients.

Does off-pump coronary artery bypass confer any advantage in patients with end-stage renal failure? A systematic review and meta-analysis.

OBJECTIVES: Patients with end-stage renal disease (ESRD) are often excluded from trials comparing off and on-pump coronary artery bypass grafting (CABG). Thus data in this cohort is limited to small retrospective studies. Hence we compared the adverse clinical events and outcome in patients with ESRD undergoing off (OPCABG) and on-pump surgery (ONCABG). METHODS: Pubmed, Scopus and Web of Science were searched (inception - June 2013) to identify studies comparing clinical results of OPCABG and ONCABG in dialysis dependent patients. A random effect inverse variance weighted meta-analysis was conducted. Results are presented as risk ratios (RR) with 95% confidence intervals; p<0.05 is significant. RESULT: Ten retrospective studies (2762 OPCABG and 11310 ONCABG) fulfilled criteria and were pooled. Patients undergoing off-pump surgery were less than 100 in most of the articles. Early mortality [OPCABG (8.4%); ONCABG (10.4%)] was comparable [RR 0.80(0.51-1.17); p=0.35; I(2)=30%]. Re-exploration for bleeding [RR 0.81(0.47-1.39); p=0.44] and blood transfusion [RR 0.79(0.57-1.08); p=0.14] were also comparable. While patients undergoing off-pump surgery were extubated earlier (p<0.01), other post-operative events like stroke (p=0.34) and atrial fibrillation (p=0.10) were similar. Mid-term survival (three to five years) was also comparable. CONCLUSION: Patients with end-stage renal disease undergoing coronary artery bypass grafting demonstrate comparable results irrespective of method. While available data is limited to retrospective studies, we failed to demonstrate any significant advantage for performing OPCABG in this group of patients.

Use of bortezomib in heavy-chain deposition disease: a report of 3 cases.

Heavy-chain deposition disease (HCDD) is a rare complication of plasma cell dyscrasia in which monoclonal heavy chains deposit in glomerular and tubular basement membranes of the kidney. Clinical and pathologic features of HCDD have been well described in case reports and series, but evidence supporting specific therapies is sparse. Historically, the disease has had a poor prognosis, intensifying the need to clarify optimal treatments. We describe 3 cases of HCDD with biopsy-proven glomerular involvement, severe nephrotic syndrome, and decline in kidney function that were treated successfully with bortezomib, a proteasome inhibitor. None of these patients had multiple myeloma. In all cases, bortezomib-based therapy resulted in sustained resolution of nephrotic syndrome and improvement in kidney function. All 3 patients developed peripheral neuropathy; otherwise, treatment was well tolerated. To our knowledge, this is the first description of the clinical effectiveness of bortezomib against HCDD.

Coronary artery bypass grafting versus drug-eluting stents in patients with end-stage renal disease.

BACKGROUND: The optimal treatment for multivessel coronary artery disease in patients with end-stage renal disease (ESRD) is unresolved. AIM OF STUDY: Compare clinical adverse events after percutaneous intervention with drug-eluting stents (DESs) and coronary artery bypass grafting (CABG) in patients with ESRD. METHODS: MEDLINE, Web of Science, and Scopus were searched for appropriate studies published in the English language (between January 2000 and August 2013). The pooled odds ratio (OR) was estimated by the Peto method with a random effect model. Data are presented with 95% confidence interval; p<0.05 is significant. RESULTS: Five observational studies (12,035 DES patients; 6317 CABG) with a follow-up period of 27.4 ± 6.3 months were included. Early mortality (CABG 8% and DES 2.6%) was less in the DES cohort (OR 0.29 [0.14-0.59]; p=0.0006; I(2)=18%). Repeat intervention (DES 29% and CABG 12%) was more likely in the DES cohort (OR 3.72 [2.24-6.18]: p<0.0001). Late mortality (27.4 ± 7.3 months) was comparable in both cohorts (OR 0.72 [0.40-1.29]; p=0.27). While DES cohort (32%) patients suffered a slightly higher incidence of major adverse cardiac and cerebrovascular events (MACCE) as compared to CABG (25%), this was not significant (1.35 [0.72-2.53]; p=0.35; I(2)=30%). CONCLUSION: Data regarding this topic are limited to small retrospective studies. Early mortality is lower with DESs compared with coronary artery bypass in patients with ESRD. Rate of reintervention is significantly higher in the DES cohort. At a mean pooled follow-up of two years, both mortality and MACCE are comparable in both cohorts.

Validation of Non-invasive Detection of Hyperkalemia by Artificial Intelligence Enhanced Electrocardiography in High Acuity Settings.

BACKGROUND: Artificial intelligence (AI) electrocardiogram (ECG) analysis can enable detection of hyperkalemia. In this validation, we assessed the algorithm's performance in two high acuity settings. METHODS: An emergency department (ED) cohort (February-August 2021) and a mixed intensive care unit (ICU) cohort (August 2017-February 2018) were identified and analyzed separately. For each group, pairs of laboratory-collected potassium and 12 lead ECGs obtained within four hours of each other were identified. The previously developed AI ECG algorithm was subsequently applied to leads I and II of the 12 lead ECGs to screen for hyperkalemia (potassium > 6.0 mEq/L). RESULTS: The ED cohort (N=40,128) had a mean age of 60 years, 48% were male, and 1% (N=351) had hyperkalemia. The area under the curve (AUC) of the AI-ECG to detect hyperkalemia was 0.88, with sensitivity, specificity, positive predictive value, negative predictive value and positive likelihood ratio of, 80%, 80%, 3%, 99.8% and 4.0, respectively, in the ED cohort. Low-eGFR (<30 ml/min) subanalysis yielded AUC, sensitivity, specificity, positive predictive value, negative predictive value and positive likelihood ratio of 0.83, 86%, 60%, 15%, 98% and 2.2, respectively, in the ED cohort. The ICU cohort (N=2,636) had a mean age of 65 years, 60% were male, and 3% (N=87) had hyperkalemia. The AUC for the AI-ECG was 0.88 and yielded sensitivity, specificity, positive predictive value, negative predictive value and positive likelihood ratio of 82%, 82%, 14%, 99% and 4.6, respectively in the ICU cohort. Low-eGFR subanalysis yielded AUC, sensitivity, specificity, positive predictive value, negative predictive value and positive likelihood ratio of 0.85, 88%, 67%, 29%, 97% and 2.7, respectively in the ICU cohort. CONCLUSION: The AI-ECG algorithm demonstrated a high negative predictive value, suggesting that it is useful for ruling out hyperkalemia, but a low positive predictive value, suggesting that it is insufficient for treating hyperkalemia.

Low- and high-molecular-weight urinary proteins as predictors of response to rituximab in patients with membranous nephropathy: a prospective study.

BACKGROUND: Selective urinary biomarkers have been considered superior to total proteinuria in predicting response to treatment and outcome in patients with membranous nephropathy (MN). METHODS: We prospectively tested whether urinary (U) excretion of retinol-binding protein (RBP), α1-microglobulin (α1M), albumin, immunoglobulin IgG and IgM and/or anti-phospholipase 2 receptor (PLA(2)R) levels could predict response to rituximab (RTX) therapy better than standard measures in MN. We also correlated changes in antibodies to PLA(2)R with these urinary biomarkers. RESULTS: Twenty patients with MN and proteinuria (P) >5 g/24 h received RTX (375 mg/m(2) × 4) and at 12 months, 1 patient was in complete remission (CR), 9 were in partial remission (PR), 5 had a limited response (LR) and 4 were non-responders (NR). At 24 months, CR occurred in 4, PR in 12, LR in 1, NR in 2 and 1 patient relapsed. By simple linear regression analysis, UIgG at baseline (mg/24 h) was a significant predictor of change in proteinuria at 12 months (Δ urinary protein) (P = 0.04). In addition, fractional excretion (FE) of IgG, urinary alpha 1 microglobulin (Uα1M) (mg/24 h) and URBP (µg/24 h) were also predictors of response (P = 0.05, 0.04, and 0.03, respectively). On the other hand, UIgM, FEIgM, albumin and FE albumin did not predict response (P = 0.10, 0.27, 0.22 and 0.20, respectively). However, when results were analyzed in relation to proteinuria at 24 months, none of the U markers that predicted response at 12 m could predict response at 24 m (P = 0.55, 0.42, 0.29 and 0.20). Decline in anti-PLA(2)R levels was associated with and often preceded urinary biomarker response but positivity at baseline was not a predictor of proteinuria response. CONCLUSIONS: The results suggest that in patients with MN, quantification of low-, medium- and high-molecular-weight urinary proteins may be associated with rate of response to RTX, but do not correlate with longer term outcomes.

Risk factors for hospitalization among older, incident haemodialysis patients.

AIMS: The number of elderly persons with end-stage renal disease is increasing with many requiring hospitalizations. This study examines the causes and predictors of hospitalization in older haemodialysis patients. METHODS: We reviewed hospitalizations of older (≥65 years) incident chronic haemodialysis patients initiating therapy between 1 January 2007 and 31 December 2009 under the care of a single Midwestern United States dialysis provider. RESULTS: Of 125 patients, the mean age was 76 ± 7 years and 72% were male. At first dialysis, 68% used a central venous catheter (CVC) and 51% were in the hospital. Mean follow-up was 1.8 ± 1.0 years. At least one hospitalization occurred in 89 (71%) patients and half of all patients were hospitalized once within the first 223 days. Total hospital admission rate was 1.48 per patient year with hospital days totalling 8.54 days per patient year. The three most common reasons for first admission were cardiac (33%), infection (18%) and gastrointestinal (12%). Predictors of future hospitalization included the first dialysis occurring in hospital (hazard ratios (HR) 2.1, 95% CI 1.4-3.3, P = 0.0005) and the use of a CVC at first haemodialysis (HR 2.6, CI 1.6-4.4, P < 0.0001). CONCLUSION: Hospitalizations are common in older incident haemodialysis patients. Access preparation and overall burden of illness leading to the initial hospitalization appear to play a role. Identification of additional factors associated with hospitalization will allow for focused interventions to reduce hospitalization rates and increase the value of care.

Rituximab therapy for Type I membranoproliferative glomerulonephritis.

AIMS: Type I membranoproliferative glomerulonephritis (MPGN) is an immune-complex disease with a relatively poor prognosis. It has no established treatment in adults. Our hypothesis was that this disease would respond to B cell depletion with rituximab, an anti-CD20 monoclonal antibody. METHODS: We conducted an openlabel trial, in Canada and the United States, of rituximab in 6 adult patients with Type I MPGN (4 idiopathic, 2 with cryoglobulinemia). The rituximab dose was 1,000 mg intravenously on Day 1 and on Day 15. The patients were followed for 1 year. The primary outcome was the change in proteinuria. RESULTS: Peripheral blood B cells were suppressed, after rituximab, in all patients. The mean urinary protein excretion was 3.9 ± 2.0 g/d before treatment. Proteinuria fell in all patients, at all-time points, after rituximab administration. The difference was statistically significant (p < 0.05) at 6, 9 and 12 months, but not at 3 months. The minimum mean urinary protein excretion was 1.4 ± 1.4 g/d at 9 months. There were 2 complete and 3 partial remissions among the 6 patients. The creatinine clearance did not change significantly over the course of the study. There were no adverse effects. CONCLUSIONS: Rituximab reduced proteinuria among patients with Type I MPGN. This trial suggests that B cells may play a role in this disease and that additional study of B-cell suppression is warranted.

Subtyping hospitalized patients with hypokalemia by machine learning consensus clustering and associated mortality risks.

BACKGROUND: Hospitalized patients with hypokalemia are heterogeneous and cluster analysis, an unsupervised machine learning methodology, may discover more precise and specific homogeneous groups within this population of interest. Our study aimed to cluster patients with hypokalemia at hospital admission using an unsupervised machine learning approach and assess the mortality risk among these distinct clusters. METHODS: We performed consensus clustering analysis based on demographic information, principal diagnoses, comorbidities and laboratory data among 4763 hospitalized adult patients with admission serum potassium ≤3.5 mEq/L. We calculated the standardized mean difference of each variable and used the cutoff of ±0.3 to identify each cluster's key features. We assessed the association of the hypokalemia cluster with hospital and 1-year mortality. RESULTS: Consensus cluster analysis identified three distinct clusters that best represented patients' baseline characteristics. Cluster 1 had 1150 (32%) patients, cluster 2 had 1344 (28%) patients and cluster 3 had 1909 (40%) patients. Based on the standardized difference, patients in cluster 1 were younger, had less comorbidity burden but higher estimated glomerular filtration rate (eGFR) and higher hemoglobin; patients in cluster 2 were older, more likely to be admitted for cardiovascular disease and had higher serum sodium and chloride levels but lower eGFR, serum bicarbonate, strong ion difference (SID) and hemoglobin, while patients in cluster 3 were older, had a greater comorbidity burden, higher serum bicarbonate and SID but lower serum sodium, chloride and eGFR. Compared with cluster 1, cluster 2 had both higher hospital and 1-year mortality, whereas cluster 3 had higher 1-year mortality but comparable hospital mortality. CONCLUSION: Our study demonstrated the use of consensus clustering analysis in the heterogeneous cohort of hospitalized hypokalemic patients to characterize their patterns of baseline clinical and laboratory data into three clinically distinct clusters with different mortality risks.

Hypernatremia subgroups among hospitalized patients by machine learning consensus clustering with different patient survival.

BACKGROUND: The objective of this study was to characterize hypernatremia patients at hospital admission into clusters using an unsupervised machine learning approach and to evaluate the mortality risk among these distinct clusters. METHODS: We performed consensus cluster analysis based on demographic information, principal diagnoses, comorbidities, and laboratory data among 922 hospitalized adult patients with admission serum sodium of > 145 mEq/L. We calculated the standardized difference of each variable to identify each cluster's key features. We assessed the association of each hypernatremia cluster with hospital and 1-year mortality. RESULTS: There were three distinct clusters of patients with hypernatremia on admission: 318 (34%) patients in cluster 1, 339 (37%) patients in cluster 2, and 265 (29%) patients in cluster 3. Cluster 1 consisted of more critically ill patients with more severe hypernatremia and hypokalemic hyperchloremic metabolic acidosis. Cluster 2 consisted of older patients with more comorbidity burden, body mass index, and metabolic alkalosis. Cluster 3 consisted of younger patients with less comorbidity burden, higher baseline eGFR, hemoglobin, and serum albumin. Compared to cluster 3, odds ratios for hospital mortality were 15.74 (95% CI 3.75-66.18) for cluster 1, and 6.51 (95% CI 1.48-28.59) for cluster 2, whereas hazard ratios for 1-year mortality were 6.25 (95% CI 3.69-11.46) for cluster 1 and 4.66 (95% CI 2.73-8.59) for cluster 2. CONCLUSION: Our cluster analysis identified three clinically distinct phenotypes with differing mortality risk in patients hospitalized with hypernatremia.

Explainable Preoperative Automated Machine Learning Prediction Model for Cardiac Surgery-Associated Acute Kidney Injury.

BACKGROUND: We aimed to develop and validate an automated machine learning (autoML) prediction model for cardiac surgery-associated acute kidney injury (CSA-AKI). METHODS: Using 69 preoperative variables, we developed several models to predict post-operative AKI in adult patients undergoing cardiac surgery. Models included autoML and non-autoML types, including decision tree (DT), random forest (RF), extreme gradient boosting (XGBoost), and artificial neural network (ANN), as well as a logistic regression prediction model. We then compared model performance using area under the receiver operating characteristic curve (AUROC) and assessed model calibration using Brier score on the independent testing dataset. RESULTS: The incidence of CSA-AKI was 36%. Stacked ensemble autoML had the highest predictive performance among autoML models, and was chosen for comparison with other non-autoML and multivariable logistic regression models. The autoML had the highest AUROC (0.79), followed by RF (0.78), XGBoost (0.77), multivariable logistic regression (0.77), ANN (0.75), and DT (0.64). The autoML had comparable AUROC with RF and outperformed the other models. The autoML was well-calibrated. The Brier score for autoML, RF, DT, XGBoost, ANN, and multivariable logistic regression was 0.18, 0.18, 0.21, 0.19, 0.19, and 0.18, respectively. We applied SHAP and LIME algorithms to our autoML prediction model to extract an explanation of the variables that drive patient-specific predictions of CSA-AKI. CONCLUSION: We were able to present a preoperative autoML prediction model for CSA-AKI that provided high predictive performance that was comparable to RF and superior to other ML and multivariable logistic regression models. The novel approaches of the proposed explainable preoperative autoML prediction model for CSA-AKI may guide clinicians in advancing individualized medicine plans for patients under cardiac surgery.

Machine Learning Consensus Clustering Approach for Hospitalized Patients with Phosphate Derangements.

BACKGROUND: The goal of this study was to categorize patients with abnormal serum phosphate upon hospital admission into distinct clusters utilizing an unsupervised machine learning approach, and to assess the mortality risk associated with these clusters. METHODS: We utilized the consensus clustering approach on demographic information, comorbidities, principal diagnoses, and laboratory data of hypophosphatemia (serum phosphate ≤ 2.4 mg/dL) and hyperphosphatemia cohorts (serum phosphate ≥ 4.6 mg/dL). The standardized mean difference was applied to determine each cluster's key features. We assessed the association of the clusters with mortality. RESULTS: In the hypophosphatemia cohort (n = 3113), the consensus cluster analysis identified two clusters. The key features of patients in Cluster 2, compared with Cluster 1, included: older age; a higher comorbidity burden, particularly hypertension; diabetes mellitus; coronary artery disease; lower eGFR; and more acute kidney injury (AKI) at admission. Cluster 2 had a comparable hospital mortality (3.7% vs. 2.9%; p = 0.17), but a higher one-year mortality (26.8% vs. 14.0%; p < 0.001), and five-year mortality (20.2% vs. 44.3%; p < 0.001), compared to Cluster 1. In the hyperphosphatemia cohort (n = 7252), the analysis identified two clusters. The key features of patients in Cluster 2, compared with Cluster 1, included: older age; more primary admission for kidney disease; more history of hypertension; more end-stage kidney disease; more AKI at admission; and higher admission potassium, magnesium, and phosphate. Cluster 2 had a higher hospital (8.9% vs. 2.4%; p < 0.001) one-year mortality (32.9% vs. 14.8%; p < 0.001), and five-year mortality (24.5% vs. 51.1%; p < 0.001), compared with Cluster 1. CONCLUSION: Our cluster analysis classified clinically distinct phenotypes with different mortality risks among hospitalized patients with serum phosphate derangements. Age, comorbidities, and kidney function were the key features that differentiated the phenotypes.

Machine Learning Consensus Clustering Approach for Patients with Lactic Acidosis in Intensive Care Units.

BACKGROUND: Lactic acidosis is a heterogeneous condition with multiple underlying causes and associated outcomes. The use of multi-dimensional patient data to subtype lactic acidosis can personalize patient care. Machine learning consensus clustering may identify lactic acidosis subgroups with unique clinical profiles and outcomes. METHODS: We used the Medical Information Mart for Intensive Care III database to abstract electronic medical record data from patients admitted to intensive care units (ICU) in a tertiary care hospital in the United States. We included patients who developed lactic acidosis (defined as serum lactate ≥ 4 mmol/L) within 48 h of ICU admission. We performed consensus clustering analysis based on patient characteristics, comorbidities, vital signs, organ supports, and laboratory data to identify clinically distinct lactic acidosis subgroups. We calculated standardized mean differences to show key subgroup features. We compared outcomes among subgroups. RESULTS: We identified 1919 patients with lactic acidosis. The algorithm revealed three best unique lactic acidosis subgroups based on patient variables. Cluster 1 (n = 554) was characterized by old age, elective admission to cardiac surgery ICU, vasopressor use, mechanical ventilation use, and higher pH and serum bicarbonate. Cluster 2 (n = 815) was characterized by young age, admission to trauma/surgical ICU with higher blood pressure, lower comorbidity burden, lower severity index, and less vasopressor use. Cluster 3 (n = 550) was characterized by admission to medical ICU, history of liver disease and coagulopathy, acute kidney injury, lower blood pressure, higher comorbidity burden, higher severity index, higher serum lactate, and lower pH and serum bicarbonate. Cluster 3 had the worst outcomes, while cluster 1 had the most favorable outcomes in terms of persistent lactic acidosis and mortality. CONCLUSIONS: Consensus clustering analysis synthesized the pattern of clinical and laboratory data to reveal clinically distinct lactic acidosis subgroups with different outcomes.

Machine Learning Prediction Models for Mortality in Intensive Care Unit Patients with Lactic Acidosis.

BACKGROUND: Lactic acidosis is the most common cause of anion gap metabolic acidosis in the intensive care unit (ICU), associated with poor outcomes including mortality. We sought to compare machine learning (ML) approaches versus logistic regression analysis for prediction of mortality in lactic acidosis patients admitted to the ICU. METHODS: We used the Medical Information Mart for Intensive Care (MIMIC-III) database to identify ICU adult patients with lactic acidosis (serum lactate ≥4 mmol/L). The outcome of interest was hospital mortality. We developed prediction models using four ML approaches consisting of random forest (RF), decision tree (DT), extreme gradient boosting (XGBoost), artificial neural network (ANN), and statistical modeling with forward stepwise logistic regression using the testing dataset. We then assessed model performance using area under the receiver operating characteristic curve (AUROC), accuracy, precision, error rate, Matthews correlation coefficient (MCC), F1 score, and assessed model calibration using the Brier score, in the independent testing dataset. RESULTS: Of 1919 lactic acidosis ICU patients, 1535 and 384 were included in the training and testing dataset, respectively. Hospital mortality was 30%. RF had the highest AUROC at 0.83, followed by logistic regression 0.81, XGBoost 0.81, ANN 0.79, and DT 0.71. In addition, RF also had the highest accuracy (0.79), MCC (0.45), F1 score (0.56), and lowest error rate (21.4%). The RF model was the most well-calibrated. The Brier score for RF, DT, XGBoost, ANN, and multivariable logistic regression was 0.15, 0.19, 0.18, 0.19, and 0.16, respectively. The RF model outperformed multivariable logistic regression model, SOFA score (AUROC 0.74), SAP II score (AUROC 0.77), and Charlson score (AUROC 0.69). CONCLUSION: The ML prediction model using RF algorithm provided the highest predictive performance for hospital mortality among ICU patient with lactic acidosis.

Clinically Distinct Subtypes of Acute Kidney Injury on Hospital Admission Identified by Machine Learning Consensus Clustering.

BACKGROUND: We aimed to cluster patients with acute kidney injury at hospital admission into clinically distinct subtypes using an unsupervised machine learning approach and assess the mortality risk among the distinct clusters. METHODS: We performed consensus clustering analysis based on demographic information, principal diagnoses, comorbidities, and laboratory data among 4289 hospitalized adult patients with acute kidney injury at admission. The standardized difference of each variable was calculated to identify each cluster's key features. We assessed the association of each acute kidney injury cluster with hospital and one-year mortality. RESULTS: Consensus clustering analysis identified four distinct clusters. There were 1201 (28%) patients in cluster 1, 1396 (33%) patients in cluster 2, 1191 (28%) patients in cluster 3, and 501 (12%) patients in cluster 4. Cluster 1 patients were the youngest and had the least comorbidities. Cluster 2 and cluster 3 patients were older and had lower baseline kidney function. Cluster 2 patients had lower serum bicarbonate, strong ion difference, and hemoglobin, but higher serum chloride, whereas cluster 3 patients had lower serum chloride but higher serum bicarbonate and strong ion difference. Cluster 4 patients were younger and more likely to be admitted for genitourinary disease and infectious disease but less likely to be admitted for cardiovascular disease. Cluster 4 patients also had more severe acute kidney injury, lower serum sodium, serum chloride, and serum bicarbonate, but higher serum potassium and anion gap. Cluster 2, 3, and 4 patients had significantly higher hospital and one-year mortality than cluster 1 patients (p < 0.001). CONCLUSION: Our study demonstrated using machine learning consensus clustering analysis to characterize a heterogeneous cohort of patients with acute kidney injury on hospital admission into four clinically distinct clusters with different associated mortality risks.