scholarly journals Machine Learning for Identifying Medication-Associated Acute Kidney Injury

Informatics ◽  
2020 ◽  
Vol 7 (2) ◽  
pp. 18
Author(s):  
Sheikh S. Abdullah ◽  
Neda Rostamzadeh ◽  
Kamran Sedig ◽  
Daniel J. Lizotte ◽  
Amit X. Garg ◽  
...  
Keyword(s):  
Machine Learning ◽  
Acute Kidney Injury ◽  
Clinical Medicine ◽  
Kidney Injury ◽  
Population Based ◽  
Frequent Itemset ◽  
Machine Learning Techniques ◽  
Analysis Techniques ◽  
Stratified Analysis ◽  
The Impact

One of the prominent problems in clinical medicine is medication-induced acute kidney injury (AKI). Avoiding this problem can prevent patient harm and reduce healthcare expenditures. Several researches have been conducted to identify AKI-associated medications using statistical, data mining, and machine learning techniques. However, these studies are limited to assessing the impact of known nephrotoxic medications and do not comprehensively explore the relationship between medication combinations and AKI. In this paper, we present a population-based retrospective cohort study that employs automated data analysis techniques to identify medications and medication combinations that are associated with a higher risk of AKI. By integrating multivariable logistic regression, frequent itemset mining, and stratified analysis, this study is designed to explore the complex relationships between medications and AKI in such a way that has never been attempted before. Through an analysis of prescription records of one million older patients stored in the healthcare administrative dataset at ICES (an independent, non-profit, world-leading research organization that uses population-based health and social data to produce knowledge on a broad range of healthcare issues), we identified 55 AKI-associated medications among 595 distinct medications and 78 AKI-associated medication combinations among 7748 frequent medication combinations. In addition, through a stratified analysis, we identified 37 cases where a particular medication was associated with increasing the risk of AKI when used with another medication. We have shown that our results are consistent with previous studies through consultation with a nephrologist and an electronic literature search. This research demonstrates how automated analysis techniques can be used to accomplish data-driven tasks using massive clinical datasets.

scholarly journals Renin–angiotensin system blocker use and the risk of acute kidney injury after colorectal cancer surgery: a population-based cohort study

BMJ Open ◽  
2019 ◽  
Vol 9 (11) ◽  
pp. e032964
Author(s):  
Charlotte Slagelse ◽  
H Gammelager ◽  
Lene Hjerrild Iversen ◽  
Kathleen D Liu ◽  
Henrik T Toft Sørensen ◽  
...  
Keyword(s):  
Colorectal Cancer ◽  
Acute Kidney Injury ◽  
Cohort Study ◽  
Angiotensin Receptor Blockers ◽  
Kidney Injury ◽  
Population Based ◽  
Adjusted Risk ◽  
Increased Risk ◽  
The Impact ◽  
Postoperative Aki

ObjectivesIt is unknown whether preoperative use of ACE inhibitors (ACE-I) or angiotensin receptor blockers (ARBs) affects the risk of acute kidney injury (AKI) after colorectal cancer (CRC) surgery. We assessed the impact of preoperative ACE-I/ARB use on risk of AKI after CRC surgery.DesignObservational cohort study. Patients were divided into three exposure groups—current, former and non-users—through reimbursed prescriptions within 365 days before the surgery. AKI within 7 days after surgery was defined according to the current Kidney Disease Improving Global Outcome consensus criteria.SettingPopulation-based Danish medical databases.ParticipantsA total of 9932 patients undergoing incident CRC surgery during 2005–2014 in northern Denmark were included through the Danish Colorectal Cancer Group Database.Outcome measureWe computed cumulative incidence proportions (risk) of AKI with 95% CIs for current, former and non-users of ACE-I/ARB, including death as a competing risk. We compared current and former users with non-users by computing adjusted risk ratios (aRRs) using log-binomial regression adjusted for demographics, comorbidities and CRC-related characteristics. We stratified the analyses of ACE-I/ARB users to address any difference in impact within relevant subgroups.ResultsTwenty-one per cent were ACE-I/ARB current users, 6.4% former users and 72.3% non-users. The 7-day postoperative AKI risk for current, former and non-users was 26.4% (95% CI 24.6% to 28.3%), 25.2% (21.9% to 28.6%) and 17.8% (17.0% to 18.7%), respectively. The aRRs of AKI were 1.20 (1.09 to 1.32) and 1.16 (1.01 to 1.34) for current and former users, compared with non-users. The relative risk of AKI in current compared with non-users was consistent in all subgroups, except for higher aRR in patients with a history of hypertension.ConclusionsBeing a current or former user of ACE-I/ARBs is associated with an increased risk of postoperative AKI compared with non-users. Although it may not be a drug effect, users of ACE-I/ARBs should be considered a risk group for postoperative AKI.

scholarly journals Recurrent acute kidney injury: predictors and impact in a large population-based cohort

2019 ◽  
Vol 35 (8) ◽  
pp. 1361-1369 ◽  
Author(s):  
Jennifer Holmes ◽  
John Geen ◽  
John D Williams ◽  
Aled O Phillips
Keyword(s):  
Acute Kidney Injury ◽  
Renal Function ◽  
Patient Outcomes ◽  
Large Population ◽  
Kidney Injury ◽  
Population Based ◽  
First Episode ◽  
Single Episode ◽  
Subsequent Episode ◽  
The Impact

Abstract Background This study examined the impact of recurrent episodes of acute kidney injury (AKI) on patient outcomes. Methods The Welsh National electronic AKI reporting system was used to identify all cases of AKI in patients ≥18 years of age between April 2015 and September 2018. Patients were grouped according to the number of AKI episodes they experienced with each patient’s first episode described as their index episode. We compared the demography and patient outcomes of those patients with a single AKI episode with those patients with multiple AKI episodes. Analysis included 153 776 AKI episodes in 111 528 patients. Results Of those who experienced AKI and survived their index episode, 29.3% experienced a second episode, 9.9% a third episode and 4.0% experienced fourth or more episodes. Thirty-day mortality for those patients with multiple episodes of AKI was significantly higher than for those patients with a single episode (31.3% versus 24.9%, P < 0.001). Following a single episode, recovery to baseline renal function at 30 days was achieved in 83.6% of patients and was significantly higher than for patients who had repeated episodes (77.8%, P < 0.001). For surviving patients, non-recovery of renal function following any AKI episode was significantly associated with a higher probability of a further AKI episode (33.4% versus 41.0%, P < 0.001). Furthermore, with each episode of AKI the likelihood of a subsequent episode also increased (31.0% versus 43.2% versus 51.2% versus 51.7% following a first, second, third and fourth episode, P < 0.001 for all comparisons). Conclusions The results of this study provide an important contribution to the debate regarding the need for risk stratification for recurrent AKI. The data suggest that such a tool would be useful given the poor patient and renal outcomes associated with recurrent AKI episodes as highlighted by this study.

scholarly journals Predicting Acute Kidney Injury: A Machine Learning Approach Using Electronic Health Records

Information ◽  
2020 ◽  
Vol 11 (8) ◽  
pp. 386
Author(s):  
Sheikh S. Abdullah ◽  
Neda Rostamzadeh ◽  
Kamran Sedig ◽  
Amit X. Garg ◽  
Eric McArthur
Keyword(s):  
Machine Learning ◽  
Emergency Department ◽  
Acute Kidney Injury ◽  
Electronic Health Records ◽  
Prediction Models ◽  
Kidney Injury ◽  
Machine Learning Techniques ◽  
Health Records ◽  
Mortality And Morbidity ◽  
Electronic Health

Acute kidney injury (AKI) is a common complication in hospitalized patients and can result in increased hospital stay, health-related costs, mortality and morbidity. A number of recent studies have shown that AKI is predictable and avoidable if early risk factors can be identified by analyzing Electronic Health Records (EHRs). In this study, we employ machine learning techniques to identify older patients who have a risk of readmission with AKI to the hospital or emergency department within 90 days after discharge. One million patients’ records are included in this study who visited the hospital or emergency department in Ontario between 2014 and 2016. The predictor variables include patient demographics, comorbid conditions, medications and diagnosis codes. We developed 31 prediction models based on different combinations of two sampling techniques, three ensemble methods, and eight classifiers. These models were evaluated through 10-fold cross-validation and compared based on the AUROC metric. The performances of these models were consistent, and the AUROC ranged between 0.61 and 0.88 for predicting AKI among 31 prediction models. In general, the performances of ensemble-based methods were higher than the cost-sensitive logistic regression. We also validated features that are most relevant in predicting AKI with a healthcare expert to improve the performance and reliability of the models. This study predicts the risk of AKI for a patient after being discharged, which provides healthcare providers enough time to intervene before the onset of AKI.

scholarly journals Preadmission kidney function and risk of acute kidney injury in patients hospitalized with acute pyelonephritis: A Danish population-based cohort study

PLoS ONE ◽  
2021 ◽  
Vol 16 (3) ◽  
pp. e0247687
Author(s):  
Henriette Vendelbo Graversen ◽  
Mette Nørgaard ◽  
Dorothea Nitsch ◽  
Christian Fynbo Christiansen
Keyword(s):  
Acute Kidney Injury ◽  
Risk Factor ◽  
Cohort Study ◽  
Kidney Function ◽  
Acute Pyelonephritis ◽  
Laboratory Data ◽  
Kidney Injury ◽  
Population Based ◽  
The Impact ◽  
First Time

Background and objectives Only few smaller studies have examined if impaired kidney function increases the risk of acute kidney injury in patients with acute pyelonephritis. Therefore, we estimated 30-day risk of acute kidney injury by preadmission kidney function in patients with acute pyelonephritis. Furthermore, we examined if impaired kidney function was a risk factor for development of acute kidney injury in pyelonephritis patients. Methods This cohort study included patients with a first-time hospitalization with pyelonephritis from 2000 to 2017. Preadmission kidney function (estimated glomerular filtration rate (eGFR) <30, 30–44, 45–59, 60–89, and ≥90 ml/min/1.73 m2) and acute kidney injury within 30 days after admission were assessed using laboratory data on serum creatinine. The absolute 30-days risk of acute kidney injury was assessed treating death as a competing risk. The impact of eGFR on the odds of acute kidney injury was compared by odds ratios (ORs) with 95% confidence intervals estimated using logistic regression adjusted for potential confounding factors. Results Among 8,760 patients with available data on preadmission kidney function, 25.8% had a preadmission eGFR <60. The 30-day risk of acute kidney injury was 16% among patients with preadmission eGFR ≥90 and increased to 22%, 33%, 42%, and 47% for patients with preadmission eGFR of 60–89, 45–59, 30–44, and <30 respectively. Compared with eGFR≥90, the adjusted ORs for the subgroups with eGFR 60–89, 45–59, 30–45, and <30 were 0.95, 1.32, 1.78, and 2.19 respectively. Conclusion Acute kidney injury is a common complication in patients hospitalized with acute pyelonephritis. Preadmission impaired kidney function is a strong risk factor for development of acute kidney injury in pyelonephritis patients and more attention should be raised in prevention of pyelonephritis in patients with a low kidney function.

scholarly journals Unsupervised Machine learning to subtype Sepsis-Associated Acute Kidney Injury

2018 ◽  
Author(s):  
Kumardeep Chaudhary ◽  
Aine Duffy ◽  
Priti Poojary ◽  
Aparna Saha ◽  
Kinsuk Chauhan ◽  
...  
Keyword(s):  
Machine Learning ◽  
Acute Kidney Injury ◽  
Tertiary Care ◽  
Vital Signs ◽  
Kidney Injury ◽  
Machine Learning Techniques ◽  
Care Hospital ◽  
Unsupervised Machine Learning ◽  
Laboratory Variables ◽  
Cluster 2

AbstractObjectiveAcute kidney injury (AKI) is highly prevalent in critically ill patients with sepsis. Sepsis-associated AKI is a heterogeneous clinical entity, and, like many complex syndromes, is composed of distinct subtypes. We aimed to agnostically identify AKI subphenotypes using machine learning techniques and routinely collected data in electronic health records (EHRs).DesignCohort study utilizing the MIMIC-III Database.SettingICUs from tertiary care hospital in the U.S.PatientsPatients older than 18 years with sepsis and who developed AKI within 48 hours of ICU admission.InterventionsUnsupervised machine learning utilizing all available vital signs and laboratory measurements.Measurements and Main ResultsWe identified 1,865 patients with sepsis-associated AKI. Ten vital signs and 691 unique laboratory results were identified. After data processing and feature selection, 59 features, of which 28 were measures of intra-patient variability, remained for inclusion into an unsupervised machine-learning algorithm. We utilized k-means clustering with k ranging from 2 – 10; k=2 had the highest silhouette score (0.62). Cluster 1 had 1,358 patients while Cluster 2 had 507 patients. There were no significant differences between clusters on age, race or gender. We found significant differences in comorbidities and small but significant differences in several laboratory variables (hematocrit, bicarbonate, albumin) and vital signs (systolic blood pressure and heart rate). In-hospital mortality was higher in cluster 2 patients, 25% vs. 20%, p=0.008. Features with the largest differences between clusters included variability in basophil and eosinophil counts, alanine aminotransferase levels and creatine kinase values.ConclusionsUtilizing routinely collected laboratory variables and vital signs in the EHR, we were able to identify two distinct subphenotypes of sepsis-associated AKI with different outcomes. Variability in laboratory variables, as opposed to their actual value, was more important for determination of subphenotypes. Our findings show the potential utility of unsupervised machine learning to better subtype AKI.

scholarly journals Machine Learning for the Prediction of Progression in Patients With Acute Kidney Injury in Critical Care

2020 ◽  
Author(s):  
Lifan Zhang ◽  
Canzheng Wei ◽  
Xuepeng Zhang ◽  
Aijia Ma ◽  
Jiangli Cheng ◽  
...  
Keyword(s):  
Machine Learning ◽  
Acute Kidney Injury ◽  
Intensive Care ◽  
Regression Model ◽  
Cox Regression ◽  
Kidney Injury ◽  
Machine Learning Techniques ◽  
Gradient Boosting ◽  
Cox Regression Model ◽  
Extreme Gradient Boosting

Abstract Background Acute kidney injury (AKI) is a serve and harmful syndrome in the intensive care unit. Our previous study has shown that patients who will progress to AKI 3 stage are considered to receive RRT. This study aimed to develop a prediction model that can predict whether progression to AKI stage 3. Methods Patients with AKI stage 1/2, when they were first diagnosed with AKI in the Medical Information Mart for Intensive Care (MIMIC-III), were included. Patients who receive RRT or progress to AKI 3 stage within 72 hours of first AKI diagnosis were excluded. We build two predictive models, respectively using machine learning extreme gradient boosting (XGBoost) and logistic regression, to predict patients who will progress to AKI stage 3. Established models were evaluated by cross-validation and area under receiver operating characteristic curve (AU-ROC). Results Of the 29238 patients included in the analysis, 3237 (11.1%) patients progressed to AKI stage 3. Creatinine, blood urea nitrogen (BUN), sepsis and respiratory failure were the important predictors of AKI progression. The machine learning XGBoost model has a better performance than the Cox regression model on predicting AKI stage 3 progression (AU-ROC, 0.860 vs. 0.728, respectively). Conclusions The XGBoost model was able to identify patients with AKI progression better than the Cox regression model. Machine learning techniques may improve predictive modeling in medical research.

scholarly journals Acute kidney injury and 1-year mortality after colorectal cancer surgery: a population-based cohort study

BMJ Open ◽  
2019 ◽  
Vol 9 (3) ◽  
pp. e024817 ◽  
Author(s):  
Charlotte Slagelse ◽  
Henrik Gammelager ◽  
Lene Hjerrild Iversen ◽  
Henrik Toft Sørensen ◽  
Christian F Christiansen
Keyword(s):  
Colorectal Cancer ◽  
Acute Kidney Injury ◽  
Cohort Study ◽  
Postoperative Complication ◽  
Kidney Injury ◽  
Population Based ◽  
Colorectal Cancer Surgery ◽  
Absolute Increase ◽  
Cancer Group ◽  
The Impact

ObjectivesAcute kidney injury (AKI) is a frequent postoperative complication, but the mortality impact within different postoperative time frames and severities of AKI are poorly understood. We examined the occurrence of postoperative AKI among colorectal cancer (CRC) surgery patients and the impact of AKI on mortality during 1 year after surgery.DesignObservational cohort study. We defined the exposure, AKI, as a 50% increase in plasma creatinine or initiation of renal replacement therapy within 7 days after surgery or an absolute increase in creatinine of 26 µmol/L within 48 hours.SettingPopulation-based Danish medical databases.ParticipantsA total of 6580 patients undergoing CRC surgery in Northern Denmark during 2005–2011 were included from the Danish Colorectal Cancer Group database.Outcomes measureOccurrence of AKI and 8–30, 31–90 and 91–365 days mortality in patient with or without AKI.ResultsAKI occurred in 1337 patients (20.3%) of the 6580 patients who underwent CRC surgery. Among patients with AKI, 8–30, 31–90 and 91–365 days mortality rates were 10.1% (95% CI 8.6% to 11.9%), 7.8% (95% CI 6.4% to 9.5%) and 12.0% (95% CI 10.3% to 14.2%), respectively. Compared with patients without AKI, AKI was associated with increased 8–30 days mortality (adjusted HR (aHR)=4.01,95% CI 3.11 to 5.17) and 31–90 days mortality (aHR 2.08,95% CI 1.60 to 2.69), while 91–365 days aHR was 1.12 (95% CI 0.89 to 1.41). We observed no major differences in stratified analyses.ConclusionsAKI after surgery for CRC is a frequent postoperative complication associated with a substantially increased 90-day mortality. AKI should be considered a potential target for reducing 90-day mortality.

scholarly journals Enhancing Military Burn- and Trauma-Related Acute Kidney Injury Prediction Through an Automated Machine Learning Platform and Point-of-Care Testing

2021 ◽  
Vol 145 (3) ◽  
pp. 320-326
Author(s):  
Hooman H. Rashidi ◽  
Amy Makley ◽  
Tina L. Palmieri ◽  
Samer Albahra ◽  
Julia Loegering ◽  
...  
Keyword(s):  
Machine Learning ◽  
Acute Kidney Injury ◽  
Point Of Care ◽  
Total Body Surface Area ◽  
Kidney Injury ◽  
Reference Method ◽  
Trauma Patients ◽  
Learning Platform ◽  
Delayed Recognition ◽  
The Impact

Context.— Delayed recognition of acute kidney injury (AKI) results in poor outcomes in military and civilian burn-trauma care. Poor predictive ability of urine output (UOP) and creatinine contribute to the delayed recognition of AKI. Objective.— To determine the impact of point-of-care (POC) AKI biomarker enhanced by machine learning (ML) algorithms in burn-injured and trauma patients. Design.— We conducted a 2-phased study to develop and validate a novel POC device for measuring neutrophil gelatinase-associated lipocalin (NGAL) and creatinine from blood samples. In phase I, 40 remnant plasma samples were used to evaluate the analytic performance of the POC device. Next, phase II enrolled 125 adults with either burns that were 20% or greater of total body surface area or nonburn trauma with suspicion of AKI for clinical validation. We applied an automated ML approach to develop models predicting AKI, using a combination of NGAL, creatinine, and/or UOP as features. Results.— Point-of-care NGAL (mean [SD] bias: 9.8 [38.5] ng/mL, P = .10) and creatinine results (mean [SD] bias: 0.28 [0.30] mg/dL, P = .18) were comparable to the reference method. NGAL was an independent predictor of AKI (odds ratio, 1.6; 95% CI, 0.08–5.20; P = .01). The optimal ML model achieved an accuracy, sensitivity, and specificity of 96%, 92.3%, and 97.7%, respectively, with NGAL, creatinine, and UOP as features. Area under the receiver operator curve was 0.96. Conclusions.— Point-of-care NGAL testing is feasible and produces results comparable to reference methods. Machine learning enhanced the predictive performance of AKI biomarkers including NGAL and was superior to the current techniques.

Sign in / Sign up

Export Citation Format

Share Document