Classification and Regression Tree Analysis in Adult Hemophagocytic Syndrome: Identifying Acute Heart Failure Predictors in ICU Patients

Abstract

Sukriye Miray Kilincer Bozgul,1 Ilkce Akgun Kurtulmus,2 Figen Yargucu Zihni,3 Nur Akad Soyer,4 Burcu Yagmur,5 Ajda Gunes,4 Gorkem Koymen,6 Devrim Bozkurt1 1Ege University, Faculty of Medicine, Department of Internal Medicine, Division of Intensive Care, Izmir, Turkey; 2Torbalı State Hospital, Izmir, Turkey; 3Ege University, Faculty of Medicine, Department of Internal Medicine, Division of Rheumatology, Izmir, Turkey; 4Ege University, Faculty of Medicine, Department of Internal Medicine, Division of Hematology, Izmir, Turkey; 5Ege University, Faculty of Medicine, Department of Cardiology, Izmir, Turkey; 6Izmir City Hospital, Department of Internal Medicine, Division of Medical Oncology, Izmir, TurkeyCorrespondence: Figen Yargucu Zihni, Ege University, Faculty of Medicine, Department of Internal Medicine, Division of Rheumatology, Izmir, Turkey, Email [email protected]: Hemophagocytic syndrome (HPS) is a rare but life-threatening condition often complicated by heart failure (HF). This study aimed to identify predictors of acute heart failure in HPS patients using classification and regression tree (CART) analysis.Methods: A retrospective analysis was performed on 146 HPS patients without a diagnosis of HF. Variables such as age, cardiothoracic ratio (CTR), etiology, N-terminal pro-brain natriuretic peptide (NT-proBNP), C-reactive protein (CRP), ferritin, triglyceride, and lactate dehydrogenase (LDH) levels at diagnosis were included. CART analysis was employed to develop models predicting heart failure status. The model’s performance was evaluated using sensitivity, specificity, and overall accuracy.Results: This study identified several key predictors of acute heart failure in HPS patients. CTR emerged as the most significant predictor, with patients exhibiting a higher ratio being at a greater risk of developing heart failure. NT-proBNP levels and CRP levels were also significant predictors, indicating cardiac stress and systemic inflammation. Age and etiology played crucial roles, with older patients and those with rheumatological causes showing a higher susceptibility to heart failure. The CART models demonstrated good accuracy, with CTR being the most important predictor.Conclusion: This study highlights critical factors, such as CTR, NT-proBNP, CRP levels, age, and etiology in predicting acute heart failure in HPS patients. Early identification of these predictors can facilitate timely interventions, potentially improving outcomes and reducing mortality rates. These findings provide valuable insights for clinical practice and pave the way for further research on acute heart failure management in HPS.Plain Language Summary: HPS is a serious condition that can lead to heart failure, a life-threatening complication. This study aimed to identify factors that predict heart failure in HPS patients to help doctors diagnose and treat it early. By analyzing data from 146 patients, we found that a higher cardiothoracic ratio, elevated NT-proBNP levels, and increased CRP levels were strong predictors of heart failure. These findings mean that doctors can monitor these factors in HPS patients to catch heart failure early and start treatment sooner, potentially saving lives. The study also found that women with HPS might be more likely to develop heart failure, suggesting they need closer monitoring. Overall, this research provides valuable information that can improve patient care and outcomes in HPS.Keywords: hemophagocytic syndrome, acute heart failure, CART analysis, predictive factors, inflammation

Keywords

• hemophagocytic syndrome
• acute heart failure
• cart analysis
• predictive factors
• inflammation