1
|
Liu X, Dumontier C, Hu P, Liu C, Yeung W, Mao Z, Ho V, Pj T, Kuo PC, Hu J, Li D, Cao D, Mark RG, Zhou FH, Zhang Z, Celi LA. Clinically Interpretable Machine Learning Models for Early Prediction of Mortality in Older Patients with Multiple Organ Dysfunction Syndrome (MODS): An International Multicenter Retrospective Study. J Gerontol A Biol Sci Med Sci 2022; 78:718-726. [PMID: 35657011 DOI: 10.1093/gerona/glac107] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/28/2022] [Indexed: 11/14/2022] Open
Abstract
BACKGROUND Multiple organ dysfunction syndrome (MODS) is associated with a high risk of mortality among older patients. Current severity scores are limited in their ability to assist clinicians with triage and management decisions. We aim to develop mortality prediction models for older patients with MODS admitted to the ICU. METHODS The study analyzed older patients from 197 hospitals in the US and one hospital in the Netherlands. The cohort was divided into the young-old (65-80 years) and old-old (≥80 years), which were separately used to develop and evaluate models including internal, external and temporal validation. Demographic characteristics, comorbidities, vital signs, laboratory measurements, and treatments were used as predictors. We used the XGBoost algorithm to train models, and the SHAP method to interpret predictions. RESULTS 34,497 young-old (11.3% mortality) and 21,330 old-old (15.7% mortality) patients were analyzed. Discrimination AUROC of internal validation models in 9,046 U.S. patients was as follows: 0.87 and 0.82, respectively; Discrimination of external validation models in 1,905 EUR patients was as follows: 0.86 and 0.85, respectively; and of temporal validation models in 8,690 U.S. patients: 0.85 and 0.78, respectively. These models outperformed standard clinical scores like SOFA and APSIII. The GCS, Charlson Comorbidity Index, and Code Status emerged as top predictors of mortality. CONCLUSIONS Our models integrate data spanning physiologic and geriatric-relevant variables that outperform existing scores used in older adults with MODS, which represents a proof of concept of how machine learning can streamline data analysis for busy ICU clinicians to potentially optimize prognostication and decision making.
Collapse
Affiliation(s)
- Xiaoli Liu
- School of Biological Science and Medical Engineering, Beihang University, 100191, Beijing, China.,Laboratory for Computational Physiology, Institute for Medical Engineering and Science, Massachusetts Institute of Technology, Cambridge, 02139, Massachusetts, USA.,Center for Artificial Intelligence in Medicine, The General Hospital of PLA, 100853, Beijing, China
| | - Clark Dumontier
- New England, GRECC (Geriatrics Research, Education and Clinical Center), VA Boston Healthcare System, 02130, Massachusetts, USA.,Division of Aging, Brigham and Women's Hospital, Boston, 02115, Massachusetts, USA
| | - Pan Hu
- Department of anesthesiology, The 920 Hospital of Joint Logistic Support Force of Chinese PLA, 650032, Kunming Yunnan, China.,Department of Critical Care Medicine, The First Medical Center, The General Hospital of PLA, 100853, Beijing, China
| | - Chao Liu
- Department of Critical Care Medicine, The First Medical Center, The General Hospital of PLA, 100853, Beijing, China
| | - Wesley Yeung
- Department of Medicine, National University Hospital, 119228, Singapore.,Laboratory for Computational Physiology, Institute for Medical Engineering and Science, Massachusetts Institute of Technology, Cambridge, 02139, Massachusetts, USA
| | - Zhi Mao
- Department of Critical Care Medicine, The First Medical Center, The General Hospital of PLA, 100853, Beijing, China
| | - Vanda Ho
- Division of Geriatric Medicine, Department of Medicine, National University Hospital, 119074, Singapore
| | - Thoral Pj
- Department of Intensive Care Medicine, Amsterdam UMC, 22660, Amsterdam, The Netherlands
| | - Po-Chih Kuo
- Laboratory for Computational Physiology, Institute for Medical Engineering and Science, Massachusetts Institute of Technology, Cambridge, 02139, Massachusetts, USA.,Department of Computer Science, National Tsing Hua University, 300044, Hsinchu, Taiwan
| | - Jie Hu
- Department of Critical Care Medicine, The First Medical Center, The General Hospital of PLA, 100853, Beijing, China
| | - Deyu Li
- School of Biological Science and Medical Engineering, Beihang University, 100191, Beijing, China
| | - Desen Cao
- Department of Biomedical Engineering, The General Hospital of PLA, 100853, Beijing, China
| | - Roger G Mark
- Laboratory for Computational Physiology, Institute for Medical Engineering and Science, Massachusetts Institute of Technology, Cambridge, 02139, Massachusetts, USA
| | - Fei Hu Zhou
- Department of Critical Care Medicine, The First Medical Center, The General Hospital of PLA, 100853, Beijing, China.,Elderly Center, The General Hospital of PLA, 100853, Beijing, China
| | - Zhengbo Zhang
- School of Biological Science and Medical Engineering, Beihang University, 100191, Beijing, China.,Center for Artificial Intelligence in Medicine, The General Hospital of PLA, 100853, Beijing, China
| | - Leo Anthony Celi
- Laboratory for Computational Physiology, Institute for Medical Engineering and Science, Massachusetts Institute of Technology, Cambridge, 02139, Massachusetts, USA.,Department of Medicine, Beth Israel Deaconess Medical Center, Boston, 02215, Massachusetts, USA.,Department of Biostatistics, Harvard T.H. Chan School of Public Health, Boston, 02115, Massachusetts, USA
| |
Collapse
|
2
|
Yang YJ, Wei X, Zou G, Zhou FH, Sun LM. [Feasibility and safety of fetal intravascular transfusion via the intrahepatic vein in the treatment of fetal anemia]. Zhonghua Fu Chan Ke Za Zhi 2021; 56:244-250. [PMID: 33902235 DOI: 10.3760/cma.j.cn112141-20201026-00798] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Subscribe] [Scholar Register] [Indexed: 11/05/2022]
Abstract
Objective: To investigate the feasibility and safety of fetal intravascular transfusion via the intrahepatic vein in the treatment of fetal anemia. Methods: This was a retrospective analysis of all fetuses requiring intrauterine transfusion (IUT) in the Shanghai First Maternity and Infant Hospital between January 2010 and December 2019. According to the different ways of IUT, they were divided into intrahepatic venous transfusion group and umbilical venous transfusion group, fetal outcomes and the incidence of procedure-related complications between the two groups were compared. Results: A total of 97 IUTs were performed on 48 fetuses. Among them, 16 cases were performed in the intrahepatic vein (31 transfusions), 32 cases were performed in the cord of the umbilical vein (66 transfusions).There were no significant differences between the two groups in age, labor history and the proportion of fetal hydrops before the first transfusion. In the intrahepatic venous transfusion group, the posterior placenta was 14/16, which was significantly higher than 78% (25/32) in the umbilical venous transfusion group (P<0.01). The live-birth rates of the two groups were 13/16 and 75% (24/32). There was no significant difference between the two groups (P>0.05). Before intrahepatic venous transfusion, the proportion of fetal hydrops was significantly higher than that of umbilical venous transfusion [55% (17/31) vs 24% (16/66), P<0.05]. Puncture success rate of intrahepatic venous transfusion and umbilical venous transfusion were both 100%. In the umbilical venous transfasion group, the incidence of needle slippage (5%, 3/66) and the abnormality of fetal heart rate (11%, 7/66) were higher than those in the intrahepatic venous transfasion group [0 and 3% (1/31)], but there were no significant differences between the two groups (all P>0.05). There were no cases of fetal loss within 24 hours, premature rupture of membranes, infection within 7 days and emergency cesarean section after IUT in both groups. Conclusions: Fetal intravascular transfusion via the intrahepatic vein is safe and feasible in the treatment of fetal anemia. But the requirements of puncture technique are relatively high, so it is recommended to be carried out in experienced fetal treatment center.
Collapse
Affiliation(s)
- Y J Yang
- Fetal Medicine Unit and Prenatal Diagnosis Center, Shanghai First Maternity and Infant Hospital, Tongji University, Shanghai 201204, China
| | - X Wei
- Fetal Medicine Unit and Prenatal Diagnosis Center, Shanghai First Maternity and Infant Hospital, Tongji University, Shanghai 201204, China
| | - G Zou
- Fetal Medicine Unit and Prenatal Diagnosis Center, Shanghai First Maternity and Infant Hospital, Tongji University, Shanghai 201204, China
| | - F H Zhou
- Fetal Medicine Unit and Prenatal Diagnosis Center, Shanghai First Maternity and Infant Hospital, Tongji University, Shanghai 201204, China
| | - L M Sun
- Fetal Medicine Unit and Prenatal Diagnosis Center, Shanghai First Maternity and Infant Hospital, Tongji University, Shanghai 201204, China
| |
Collapse
|
3
|
Yang CR, Zhang XY, Liu Y, Du JY, Liang R, Yu M, Zhang FQ, Mu XF, Li F, Zhou L, Zhou FH, Meng FJ, Wang S, Ming D, Zhou XF. Antidepressant Drugs Correct the Imbalance Between proBDNF/p75NTR/Sortilin and Mature BDNF/TrkB in the Brain of Mice with Chronic Stress. Neurotox Res 2020; 37:171-182. [PMID: 31493120 DOI: 10.1007/s12640-019-00101-2] [Citation(s) in RCA: 25] [Impact Index Per Article: 6.3] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/18/2019] [Revised: 08/16/2019] [Accepted: 08/20/2019] [Indexed: 02/08/2023]
Abstract
Depression is a worldwide problem with a great social and economic burden in many countries. In our previous research, we found that the expression of proBDNF/p75NTR/sortilin is upregulated in patients with major depressive disorder. In addition, the treatment of proBDNF antibodies reversed both the depressive behaviors and the reduced BDNF mRNA detected in our rodent chronic stress models. Antidepressant drugs are usually only effective in a subpopulation of patients with major depression with a delayed time window of 2-4 weeks to exert their efficacy. The mechanism underlying such delayed response is not known. In this study, we hypothesize that antidepressant drugs exert their therapeutic effect by modulating proBDNF/p75NTR and mature BDNF/TrkB signaling pathways. To test the hypothesis, C57 mice were randomly divided into normal control, chronic unpredictable mild stress (CUMS), vehicle (VEH), fluoxetine (FLU), and clozapine (CLO) groups. Behavioral tests (sucrose preference, open field, and tail suspension tests) were performed before and after 4 weeks of CUMS. The gene and protein expression of proBDNF, the neurotrophin receptor (p75NTR), sortilin, and TrkB in the cortex and hippocampus were examined. At the protein level, CUMS induced a significant increase in proBDNF, p75NTR, and sortilin production while the TrkB protein level was found to be lower in the cortex and hippocampus compared with the control group. Consistently, at the mRNA level, p75NTR expression increased with reduced BDNF/TrkB mRNA in both cortex and hippocampus, while sortilin increased only in the hippocampus after CUMS. FLU and CLO treatments of CUMS mice reversed all protein and mRNA expression of the biomarkers in both cortex and hippocampus, except for sortilin mRNA in the cortex and proBDNF in the hippocampus, respectively. This study further confirms that the imbalance between proBDNF/p75NTR/sortilin and mBDNF/TrkB production is important in the pathogenesis of depression. It is likely that antidepressant FLU and antipsychotic CLO exert their antidepressant-like effect correcting the imbalance between proBDNF/p75NTR/sortilin and mBDNF/TrkB.
Collapse
Affiliation(s)
- C R Yang
- Department of Pathology, The Second Hospital of Tianjin Medical University, Tianjin, People's Republic of China
- Academy of Medical Engineering and Translational Medicine, Tianjin University, Tianjin, People's Republic of China
| | - X Y Zhang
- Department of Pathology, The Second Hospital of Tianjin Medical University, Tianjin, People's Republic of China
| | - Y Liu
- Tianjin Mental Health Center, Tianjin, People's Republic of China
| | - J Y Du
- Tianjin Mental Health Center, Tianjin, People's Republic of China
| | - R Liang
- Department of Pathology, The Second Hospital of Tianjin Medical University, Tianjin, People's Republic of China
| | - M Yu
- Department of Pathology, The Second Hospital of Tianjin Medical University, Tianjin, People's Republic of China
| | - F Q Zhang
- Department of Neurology, Second Hospital of Tianjin Medical University, Tianjin, People's Republic of China
| | - X F Mu
- Department of Clinical Laboratory, Qingdao Central Hospital, Qingdao, Shandong Province, People's Republic of China
| | - F Li
- Department of Anatomy and Neurobiology, School of Basic Medical Science, Central South University, Changsha, Hunan Province, People's Republic of China
| | - L Zhou
- The Mental Hospital of Yunnan Province, Kunming, Yunnan Province, People's Republic of China
| | - F H Zhou
- School of Pharmacology and Medical Science, University of South Australia, Adelaide, 5000, Australia
| | - F J Meng
- Department of Thoracic Surgery, Second Hospital of Tianjin Medical University, Tianjin, People's Republic of China
| | - S Wang
- Department of Thoracic Surgery, Second Hospital of Tianjin Medical University, Tianjin, People's Republic of China
| | - D Ming
- Academy of Medical Engineering and Translational Medicine, Tianjin University, Tianjin, People's Republic of China.
| | - X F Zhou
- School of Pharmacology and Medical Science, University of South Australia, Adelaide, 5000, Australia.
| |
Collapse
|
5
|
Abstract
BACKGROUND and aims: In neonates the gastrointestinal tract is exposed to food and bacterial antigens at a time when the gut mucosal immune system has not developed the ability to induce oral tolerance. This increases the risk for an inappropriate immune response to oral antigens. Transforming growth factor beta (TGF-beta) is an immunoregulatory cytokine present in high concentration in maternal milk. Interleukin 18 (IL-18) is a cytokine that mediates early immune events, and drives T cell development. We assessed the role of TGF-beta in mediating mucosal immune development and specifically the effect on endogenous IL-18. METHODS Rat pups were randomly assigned to the following groups, naturally suckled, maternal milk via cannula, and formula fed with and without physiological levels of TGF-beta2. A comparison of the immune response profile was then carried out. Cytokine profiles, dendritic cell, intestinal mast cell, and eosinophil numbers were assessed. RESULTS We show that feeding formula deficient in TGF-beta2 resulted in accumulated IL-18 protein release from intestinal epithelial cells and IL-18 mRNA up regulation. A proinflammatory cytokine profile resulted in the gut, along with increased numbers of activated dendritic cells, eosinophils, and mast cells. Supplementation of the formula with TGF-beta2 down regulated the proinflammatory cytokine mRNA as well as the number of activated lymphocytes, eosinophils, mast cells, CD80, and CD86 positive dendritic cells. CONCLUSION The data suggests an important role for maternal milk, in regulating immune responses after exposure to food antigens, which might otherwise induce deleterious immune responses in the intestine of suckling neonates. This regulation is potentially mediated by milk TGF-beta2, as well as endogenous IL-18.
Collapse
Affiliation(s)
- I A Penttila
- Child Health Research Institute, North Adelaide, SA 5006, Australia.
| | | | | | | | | | | | | |
Collapse
|