Unsupervised clustering reveals phenotypes of AKI in ICU COVID-19 patients
BackgroundAcute Kidney Injury (AKI) is a very frequent condition, occurring in about one in three patients admitted to an intensive care unit (ICU). AKI is a syndrome defined as a sudden decrease in glomerular filtration rate. However, this unified definition does not reflect the various mechanisms...
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2022-10-01
|
| Series: | Frontiers in Medicine |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fmed.2022.980160/full |
| _version_ | 1811158662393102336 |
|---|---|
| author | David Legouis David Legouis Gilles Criton Benjamin Assouline Christophe Le Terrier Sebastian Sgardello Jérôme Pugin Elisa Marchi Frédéric Sangla |
| author_facet | David Legouis David Legouis Gilles Criton Benjamin Assouline Christophe Le Terrier Sebastian Sgardello Jérôme Pugin Elisa Marchi Frédéric Sangla |
| author_sort | David Legouis |
| collection | DOAJ |
| description | BackgroundAcute Kidney Injury (AKI) is a very frequent condition, occurring in about one in three patients admitted to an intensive care unit (ICU). AKI is a syndrome defined as a sudden decrease in glomerular filtration rate. However, this unified definition does not reflect the various mechanisms involved in AKI pathophysiology, each with its own characteristics and sensitivity to therapy. In this study, we aimed at developing an innovative machine learning based method able to subphenotype AKI according to its pattern of risk factors.MethodsWe adopted a three-step pipeline of analyses. First, we looked for factors associated with AKI using a generalized additive model. Second, we calculated the importance of each identified AKI related factor in the estimated AKI risk to find the main risk factor for AKI, at the single patient level. Lastly, we clusterized AKI patients according to their profile of risk factors and compared the clinical characteristics and outcome of every cluster. We applied this method to a cohort of severe COVID-19 patients hospitalized in the ICU of the Geneva University Hospitals.ResultsAmong the 248 patients analyzed, we found 7 factors associated with AKI development. Using the individual expression of these factors, we identified three groups of AKI patients, based on the use of Lopinavir/Ritonavir, baseline eGFR, use of dexamethasone and AKI severity. The three clusters expressed distinct characteristics in terms of AKI severity and recovery, metabolic patterns and hospital mortality.ConclusionWe propose here a new method to phenotype AKI patients according to their most important individual risk factors for AKI development. When applied to an ICU cohort of COVID-19 patients, we were able to differentiate three groups of patients. Each expressed specific AKI characteristics and outcomes, which probably reflect a distinct pathophysiology. |
| first_indexed | 2024-04-10T05:27:56Z |
| format | Article |
| id | doaj.art-615c72c69c1d48b3abcf2358276d4cea |
| institution | Directory Open Access Journal |
| issn | 2296-858X |
| language | English |
| last_indexed | 2024-04-10T05:27:56Z |
| publishDate | 2022-10-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Medicine |
| spelling | doaj.art-615c72c69c1d48b3abcf2358276d4cea2023-03-07T12:16:53ZengFrontiers Media S.A.Frontiers in Medicine2296-858X2022-10-01910.3389/fmed.2022.980160980160Unsupervised clustering reveals phenotypes of AKI in ICU COVID-19 patientsDavid Legouis0David Legouis1Gilles Criton2Benjamin Assouline3Christophe Le Terrier4Sebastian Sgardello5Jérôme Pugin6Elisa Marchi7Frédéric Sangla8Division of Intensive Care, Department of Acute Medicine, University Hospital of Geneva, Geneva, SwitzerlandLaboratory of Nephrology, Department of Medicine and Cell Physiology, University Hospital of Geneva, Geneva, SwitzerlandGeneva School of Economics and Management, University of Geneva, Geneva, SwitzerlandDivision of Intensive Care, Department of Acute Medicine, University Hospital of Geneva, Geneva, SwitzerlandDivision of Intensive Care, Department of Acute Medicine, University Hospital of Geneva, Geneva, SwitzerlandDepartment of Surgery, Center Hospitalier du Valais Romand, Sion, SwitzerlandDivision of Intensive Care, Department of Acute Medicine, University Hospital of Geneva, Geneva, SwitzerlandDivision of Intensive Care, Department of Acute Medicine, University Hospital of Geneva, Geneva, SwitzerlandDivision of Intensive Care, Department of Acute Medicine, University Hospital of Geneva, Geneva, SwitzerlandBackgroundAcute Kidney Injury (AKI) is a very frequent condition, occurring in about one in three patients admitted to an intensive care unit (ICU). AKI is a syndrome defined as a sudden decrease in glomerular filtration rate. However, this unified definition does not reflect the various mechanisms involved in AKI pathophysiology, each with its own characteristics and sensitivity to therapy. In this study, we aimed at developing an innovative machine learning based method able to subphenotype AKI according to its pattern of risk factors.MethodsWe adopted a three-step pipeline of analyses. First, we looked for factors associated with AKI using a generalized additive model. Second, we calculated the importance of each identified AKI related factor in the estimated AKI risk to find the main risk factor for AKI, at the single patient level. Lastly, we clusterized AKI patients according to their profile of risk factors and compared the clinical characteristics and outcome of every cluster. We applied this method to a cohort of severe COVID-19 patients hospitalized in the ICU of the Geneva University Hospitals.ResultsAmong the 248 patients analyzed, we found 7 factors associated with AKI development. Using the individual expression of these factors, we identified three groups of AKI patients, based on the use of Lopinavir/Ritonavir, baseline eGFR, use of dexamethasone and AKI severity. The three clusters expressed distinct characteristics in terms of AKI severity and recovery, metabolic patterns and hospital mortality.ConclusionWe propose here a new method to phenotype AKI patients according to their most important individual risk factors for AKI development. When applied to an ICU cohort of COVID-19 patients, we were able to differentiate three groups of patients. Each expressed specific AKI characteristics and outcomes, which probably reflect a distinct pathophysiology.https://www.frontiersin.org/articles/10.3389/fmed.2022.980160/fullAKIclusteringmachine learningCOVID-19critical care |
| spellingShingle | David Legouis David Legouis Gilles Criton Benjamin Assouline Christophe Le Terrier Sebastian Sgardello Jérôme Pugin Elisa Marchi Frédéric Sangla Unsupervised clustering reveals phenotypes of AKI in ICU COVID-19 patients Frontiers in Medicine AKI clustering machine learning COVID-19 critical care |
| title | Unsupervised clustering reveals phenotypes of AKI in ICU COVID-19 patients |
| title_full | Unsupervised clustering reveals phenotypes of AKI in ICU COVID-19 patients |
| title_fullStr | Unsupervised clustering reveals phenotypes of AKI in ICU COVID-19 patients |
| title_full_unstemmed | Unsupervised clustering reveals phenotypes of AKI in ICU COVID-19 patients |
| title_short | Unsupervised clustering reveals phenotypes of AKI in ICU COVID-19 patients |
| title_sort | unsupervised clustering reveals phenotypes of aki in icu covid 19 patients |
| topic | AKI clustering machine learning COVID-19 critical care |
| url | https://www.frontiersin.org/articles/10.3389/fmed.2022.980160/full |
| work_keys_str_mv | AT davidlegouis unsupervisedclusteringrevealsphenotypesofakiinicucovid19patients AT davidlegouis unsupervisedclusteringrevealsphenotypesofakiinicucovid19patients AT gillescriton unsupervisedclusteringrevealsphenotypesofakiinicucovid19patients AT benjaminassouline unsupervisedclusteringrevealsphenotypesofakiinicucovid19patients AT christopheleterrier unsupervisedclusteringrevealsphenotypesofakiinicucovid19patients AT sebastiansgardello unsupervisedclusteringrevealsphenotypesofakiinicucovid19patients AT jeromepugin unsupervisedclusteringrevealsphenotypesofakiinicucovid19patients AT elisamarchi unsupervisedclusteringrevealsphenotypesofakiinicucovid19patients AT fredericsangla unsupervisedclusteringrevealsphenotypesofakiinicucovid19patients |