Machine Learning Consensus Clustering Approach for Hospitalized Patients with Dysmagnesemia

Background: The objectives of this study were to classify patients with serum magnesium derangement on hospital admission into clusters using unsupervised machine learning approach and to evaluate the mortality risks among these distinct clusters. Methods: Consensus cluster analysis was performed based on demographic information, principal diagnoses, comorbidities, and laboratory data in hypomagnesemia (serum magnesium ≤ 1.6 mg/dL) and hypermagnesemia cohorts (serum magnesium ≥ 2.4 mg/dL). Each cluster’s key features were determined using the standardized mean difference. The associations of the clusters with hospital mortality and one-year mortality were assessed. Results: In hypomagnesemia cohort (n = 13,320), consensus cluster analysis identified three clusters. Cluster 1 patients had the highest comorbidity burden and lowest serum magnesium. Cluster 2 patients had the youngest age, lowest comorbidity burden, and highest kidney function. Cluster 3 patients had the oldest age and lowest kidney function. Cluster 1 and cluster 3 were associated with higher hospital and one-year mortality compared to cluster 2. In hypermagnesemia cohort (n = 4671), the analysis identified two clusters. Compared to cluster 1, the key features of cluster 2 included older age, higher comorbidity burden, more hospital admissions primarily due to kidney disease, more acute kidney injury, and lower kidney function. Compared to cluster 1, cluster 2 was associated with higher hospital mortality and one-year mortality. Conclusion: Our cluster analysis identified clinically distinct phenotypes with differing mortality risks in hospitalized patients with dysmagnesemia. Future studies are required to assess the application of this ML consensus clustering approach to care for hospitalized patients with dysmagnesemia.

Download Full-text

Machine Learning Consensus Clustering Approach for Hospitalized Patients with Phosphate Derangements

Journal of Clinical Medicine ◽

10.3390/jcm10194441 ◽

2021 ◽

Vol 10 (19) ◽

pp. 4441

Author(s):

Charat Thongprayoon ◽

Carissa Y. Dumancas ◽

Voravech Nissaisorakarn ◽

Mira T. Keddis ◽

Andrea G. Kattah ◽

...

Keyword(s):

Machine Learning ◽

Cluster Analysis ◽

Kidney Disease ◽

Serum Phosphate ◽

Older Age ◽

Consensus Clustering ◽

Clustering Approach ◽

Key Features ◽

One Year ◽

Cluster 2

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.

Download Full-text

Subtyping Hyperchloremia among Hospitalized Patients by Machine Learning Consensus Clustering

Medicina ◽

10.3390/medicina57090903 ◽

2021 ◽

Vol 57 (9) ◽

pp. 903 ◽

Cited By ~ 2

Author(s):

Charat Thongprayoon ◽

Voravech Nissaisorakarn ◽

Pattharawin Pattharanitima ◽

Michael A. Mao ◽

Andrea G. Kattah ◽

...

Keyword(s):

Machine Learning ◽

Cluster Analysis ◽

Serum Sodium ◽

Hospitalized Patients ◽

Lower Serum ◽

Comorbidity Score ◽

Mortality Risks ◽

Serum Chloride ◽

One Year ◽

Cluster 2

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.

Download Full-text

Machine Learning Consensus Clustering of Hospitalized Patients with Admission Hyponatremia

Diseases ◽

10.3390/diseases9030054 ◽

2021 ◽

Vol 9 (3) ◽

pp. 54

Author(s):

Charat Thongprayoon ◽

Panupong Hansrivijit ◽

Michael A. Mao ◽

Pradeep K. Vaitla ◽

Andrea G. Kattah ◽

...

Keyword(s):

Machine Learning ◽

Kidney Function ◽

Serum Sodium ◽

Learning Approach ◽

Serum Bicarbonate ◽

Unsupervised Machine Learning ◽

Disease Cluster ◽

Machine Learning Approach ◽

One Year ◽

Cluster 2

Background: The objective of this study was to characterize patients with hyponatremia at hospital admission into clusters using an unsupervised machine learning approach, and to evaluate the short- and long-term mortality risk among these distinct clusters. Methods: We performed consensus cluster analysis based on demographic information, principal diagnoses, comorbidities, and laboratory data among 11,099 hospitalized adult hyponatremia patients with an admission serum sodium below 135 mEq/L. The standardized mean difference was utilized to identify each cluster’s key features. We assessed the association of each hyponatremia cluster with hospital and one-year mortality using logistic and Cox proportional hazard analysis, respectively. Results: There were three distinct clusters of hyponatremia patients: 2033 (18%) in cluster 1, 3064 (28%) in cluster 2, and 6002 (54%) in cluster 3. Among these three distinct clusters, clusters 3 patients were the youngest, had lowest comorbidity burden, and highest kidney function. Cluster 1 patients were more likely to be admitted for genitourinary disease, and have diabetes and end-stage kidney disease. Cluster 1 patients had the lowest kidney function, serum bicarbonate, and hemoglobin, but highest serum potassium and prevalence of acute kidney injury. In contrast, cluster 2 patients were the oldest and were more likely to be admitted for respiratory disease, have coronary artery disease, congestive heart failure, stroke, and chronic obstructive pulmonary disease. Cluster 2 patients had lowest serum sodium and serum chloride, but highest serum bicarbonate. Cluster 1 patients had the highest hospital mortality and one-year mortality, followed by cluster 2 and cluster 3, respectively. Conclusion: We identified three clinically distinct phenotypes with differing mortality risks in a heterogeneous cohort of hospitalized hyponatremic patients using an unsupervised machine learning approach.

Download Full-text

Echocardiographic phenotypes of patients with sickle cell disease. An unsupervised analysis based on etendard cohort

European Heart Journal - Cardiovascular Imaging ◽

10.1093/ehjci/jeaa356.023 ◽

2021 ◽

Vol 22 (Supplement_1) ◽

Author(s):

T D"humieres ◽

J Inamo ◽

S Deswarte ◽

T Damy ◽

G Loko ◽

...

Keyword(s):

Cluster Analysis ◽

Cardiac Output ◽

Sickle Cell Disease ◽

Diastolic Function ◽

Sickle Cell ◽

Left Ventricular ◽

Integrative Approach ◽

Cell Disease ◽

Clinical Profiles ◽

Cluster 2

Abstract Funding Acknowledgements Type of funding sources: Public grant(s) – National budget only. Main funding source(s): PHRC Backgroung Echocardiography is the cornerstone in the diagnosis of cardiopulmonary involvement in sickle cell disease (SCD). However, given the unique pathophysiology of SCD associating high cardiac output, and various degrees of peripheral vasculopathy, differentiate the pathological from the physiological using echocardiography can be particularly challenging. Purpose This study sought to link cardiac phenotypes in homozygous SCD patients with clinical profiles and outcomes using cluster analysis. Methods We analyzed data of 379 patients with a sufficient echographic dataset included in the French Etendard Cohort, a prospective cohort initially designed to assess the prevalence of pulmonary hypertension. A cluster analysis was performed on echocardiographic variables, and the association between clusters and clinical profiles and outcomes was assessed. Results Three clusters were identified. Cluster 1 (N = 122) patients had the lowest cardiac output, only mild left cavities remodeling, diastolic dysfunction, and high tricuspid regurgitation velocity (TRV). They were predominantly female, as old as cluster 2, and displayed the most severe functional limitation. Cluster 2 (N = 103) patients had the highest cardiac output, left ventricular mass and a severely dilated left atrium. Diastolic function and TRV were similar to cluster 1. These patients had a higher blood pressure and a severe hemolytic anemia. Cluster 3 (N = 154) patients had mild left cavities remodeling, the best diastolic function and the lowest TRV. They were younger patients with the highest hemoglobin and lowest hemolytic markers. Right heart catheterization was performed in 94 patients. Cluster 1 gathered the majority of precapillary PH while cluster 2 gathered postcapillary PH and no PH was found in cluster 3. After a follow-up of 9.9 years (IQR: 9.3 to 10.5 years) death occurred in 38 patients (10%). Clusters 2 had the worst prognosis with 18% mortality rate vs. 12% in cluster 2 and 5% in cluster 1 (P log-rank = 0,02). Results are summarized in the central illustration. Conclusions Cluster analysis of echocardiographic variables identified 3 phenotypes among SCD patients, each associated with different clinical features and outcome. These findings underlines the necessity to rethink echocardiographic evaluation of SCD patients, with an integrative approach based on simultaneous evaluation of TRV along with left cavities remodeling and diastolic parameters. Abstract Figure.

Download Full-text

Comment on: Comorbidity burden in axial spondyloarthritis: a cluster analysis: reply

Rheumatology ◽

10.1093/rheumatology/kez573 ◽

2019 ◽

Vol 59 (3) ◽

pp. 692-693 ◽

Cited By ~ 2

Author(s):

Sizheng Steven Zhao ◽

Daniel H Solomon ◽

Nicola J Goodson

Keyword(s):

Cluster Analysis ◽

Axial Spondyloarthritis ◽

Comorbidity Burden

Download Full-text

Review of Neurologic Comorbidities in Hospitalized Patients with Opioid Abuse

Neurology Clinical Practice ◽

10.1212/cpj.0000000000001080 ◽

2021 ◽

pp. 10.1212/CPJ.0000000000001080

Author(s):

Kevin R. Nelson ◽

Katelyn Dolbec ◽

William Watson ◽

Hanwen Yuan ◽

Mam Ibraheem

Keyword(s):

Disease Burden ◽

Neuromuscular Disorders ◽

Hospitalized Patients ◽

Opioid Abuse ◽

Central Nervous System Infections ◽

Hospital Discharges ◽

The Public ◽

Urban Rural ◽

One Year ◽

Spine Disorders

AbstractPurpose of review:Determine the prevalence and burden of neurological comorbidities in hospitalized patients with opioid abuse.Recent findings:From one year of hospital discharges 2,182 opioid abuse patients were identified (prevalence 6.3%), with abuse greater among younger patients (p<0.0001), women (p<0.0001), whites (p<0.0001), and urban population (p=0.028). Matching for age, sex, race, and urban-rural residence, 347 patients were reviewed and 179 (52%) had a neurological comorbidity. The comorbidities frequently overlapped and included: encephalopathy (130), neuromuscular disorders (42), seizures (23), spine disorders (23), strokes (20), central nervous system infections (3), and movement disorders (2). Abuse patients with neurologic comorbidities experienced substantially greater number of hospital and intensive care unit days and mortality, independent of overdose.Summary:Neurological comorbidities are a frequent and heretofore underappreciated contributor to the disease burden of hospitalized patients with opioid abuse. The importance of neurological comorbidities should be included in the public health discussions surrounding the opioid epidemic.

Download Full-text

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

Journal of Nephrology ◽

10.1007/s40620-021-01163-2 ◽

2021 ◽

Author(s):

Charat Thongprayoon ◽

Michael A. Mao ◽

Mira T. Keddis ◽

Andrea G. Kattah ◽

Grace Y. Chong ◽

...

Keyword(s):

Machine Learning ◽

Patient Survival ◽

Hospitalized Patients ◽

Consensus Clustering

Download Full-text

Evaluation of a gait scoring system for cattle by using cluster analysis and Krippendorff’s α reliability

Veterinary Record ◽

10.1136/vr.105059 ◽

2018 ◽

Vol 184 (7) ◽

pp. 220-220 ◽

Cited By ~ 4

Author(s):

Nina Volkmann ◽

Jenny Stracke ◽

Nicole Kemper

Keyword(s):

Cluster Analysis ◽

Dairy Cows ◽

Scoring System ◽

Partial Agreement ◽

Claw Lesions ◽

Cluster 2 ◽

Locomotion Score

The aim of the presented study was to validate a three-point locomotion score (LS) classifying lameness in dairy cows. Therefore, locomotion of 144 cows was scored and data on claw lesions were collected during hoof trimming. Based on latter data a cluster analysis was performed to objectively classify cows into three groups (Cluster 1–3). Finally, the congruence between scoring system and clustering was tested using Krippendorff’s α reliability. In total, 63 cows (43.7 per cent) were classified as non-lame (LS1), 38 (26.4 per cent) were rated as LS2 with an uneven gait and 43 (29.9 per cent) cows were ranked as clearly lame (LS3). In comparison, hoof-trimming data revealed 64 cows (44.4 per cent) to show no diagnosis, 37 (25.7 per cent) one diagnosis, 33 animals (22.9 per cent) two diagnoses and 10 (7.0 per cent) more than two. Comparing the respective categorisation received by either the cluster analysis or LS in between groups, a high correspondence (79.4 per cent and 83.7 per cent) could be found for LS1 and cluster 1 as well as for LS3 and cluster 3. Only LS2 had partial agreement (21.1 per cent) to cluster 2. However, Krippendorff’s α was 0.75 (95 per cent CI 0.68 to 0.81), indicating a good degree of reliability. Therefore, the results of this study suggested that the presented LS is suitable for classifying the cows’ state of lameness representing their claw diseases.

Download Full-text

EFFECT OF UNILATERAL NEPHRECTOMY ON KIDNEY FUNCTION IN FERAL DOMESTIC CATS (Felis catus)

International Journal of Research -GRANTHAALAYAH ◽

10.29121/granthaalayah.v7.i2.2019.1012 ◽

2019 ◽

Vol 7 (2) ◽

pp. 101-107

Author(s):

Dionesio A. Estopa

Keyword(s):

Kidney Function ◽

Block Design ◽

Age Groups ◽

Unilateral Nephrectomy ◽

Domestic Cats ◽

Experimental Animals ◽

Operative Care ◽

Significant Difference ◽

Randomized Complete Block Design ◽

One Year

This experimental study was conducted to determine the effect of unilateral nephrectomy on kidney function through hematological and urological values after surgery using feral domestic cats. Four (4) apparently healthy stray male cats were caught and used as experimental animals in this study. The animals were grouped into two (2) with two cats in each group. Grouping of the animals was based according to age, the 1st group – ages one year and above (>1 year) and the 2nd group – ages one year and below (<1 year). A split plot randomized complete block design was used in the experiment. The result shows that rectal temperature and packed cell volume have been affected on both age group of cats and other urological values like the presence of urobilinogen, protein, leukocytes, erythrocytes and crystals were elevated twenty four hours after surgery. A significant increase on the level of creatinine and blood urea nitrogen has been noted from the 3rd and 15th post – operative day. However, no significant difference was noted between experimental animals and the two age groups, all of which have similar result. The findings of the study revealed that the remaining kidney could not completely compensate the function of the other kidney up to the 15th post – operative day, it may take longer than fifteen days. Hence, proper supportive treatment, post-operative care & management of animals subjected to unilateral nephrectomy is a must and should continue beyond fifteen days.

Download Full-text

NF-κB-Related Metabolic Gene Signature Predicts the Prognosis and Immunotherapy Response in Gastric Cancer

BioMed Research International ◽

10.1155/2022/5092505 ◽

2022 ◽

Vol 2022 ◽

pp. 1-30

Author(s):

Qiuxiang Chen ◽

Xiaojing Du ◽

Sunkuan Hu ◽

Qingke Huang

Keyword(s):

Gastric Cancer ◽

Risk Score ◽

Cancer Metabolism ◽

Gene Signature ◽

Sufficient Evidence ◽

Consensus Clustering ◽

Lasso Regression ◽

Immune Infiltration ◽

Related Risk ◽

Cluster 2

Background. Sufficient evidence indicated the crucial role of NF-κB family played in gastric cancer (GC). The novel discovery that NF-κB could regulate cancer metabolism and immune evasion greatly increased its attraction in cancer research. However, the correlation among NF-κB, metabolism, and cancer immunity in GC still requires further improvement. Methods. TCGA, hTFtarget, and MSigDB databases were employed to identify NF-κB-related metabolic genes (NFMGs). Based on NFMGs, we used consensus clustering to divide GC patients into two subtypes. GSVA was employed to analyze the enriched pathway. ESTIMATE, CIBERSORT, ssGSEA, and MCPcounter algorithms were applied to evaluate immune infiltration in GC. The tumor immune dysfunction and exclusion (TIDE) algorithm was used to predict patients’ response to immunotherapy. We also established a NFMG-related risk score by using the LASSO regression model and assessed its efficacy in TCGA and GSE62254 datasets. Results. We used 27 NFMGs to conduct an unsupervised clustering on GC samples and classified them into two clusters. Cluster 1 was characterized by high active metabolism, tumor mutant burden, and microsatellite instability, while cluster 2 was featured with high immune infiltration. Compared to cluster 2, cluster 1 had a better prognosis and higher response to immunotherapy. In addition, we constructed a 12-NFMG (ADCY3, AHCY, CHDH, GUCY1A2, ITPA, MTHFD2, NRP1, POLA1, POLR1A, POLR3A, POLR3K, and SRM) risk score. Followed analysis indicated that this risk score acted as an effectively prognostic factor in GC. Conclusion. Our data suggested that GC subtypes classified by NFMGs may effectively guide prognosis and immunotherapy. Further study of these NFMGs will deepen our understanding of NF-κB-mediated cancer metabolism and immunity.

Download Full-text