Construction of a predictive model and prognosis of left ventricular systolic dysfunction in patients with sepsis based on the diagnosis using left ventricular global longitudinal strain

BACKGROUND: Cardiac dysfunction, a common complication of sepsis, is associated with increased mortality. However, its risk factors are poorly understood, and a predictive model might help in the management of cardiac dysfunction. METHODS: A monocentric prospective study of patients with sepsis was performed. Left ventricular global longitudinal strain (LV GLS) was measured using echocardiography within 72 h of the patients diagnosed with sepsis, and the patients were categorized into two groups: LV GLS > -17%, left ventricular systolic dysfunction group (LVSD group); and LV GLS ≤ -17%, non-left ventricular systolic dysfunction group (Non-LVSD group). The baseline characteristics and prognosis of the two groups were analyzed. Based on the results of the multivariate logistic regression analysis, a predictive model of LVSD was established and a nomogram was drawn. RESULTS: Fifty-one left ventricular systolic dysfunction in patients with sepsis and 73 non-LVSD sepsis patients were included. Prognostic analysis showed that patients with LVSD had higher ICU mortality, in-hospital mortality, the incidence of atrial fibrillation (P < 0.05), and risk of death (HR = 3.104, 95% CI = 1.617–5.957, P < 0.001) compared to patients with non-LVSD. There were no significant differences in the rate of tracheal intubation, the incidence of acute kidney injury (AKI), the proportion of continuous renal replacement therapy (CRRT), length of ICU stay, and length of hospital stay between the 2 groups (P > 0.05). High sensitive troponin I (Hs-TnI) ≥ 0.131 ng/ml, procalcitonin (PCT) ≥ 40 ng/ml, lactate (Lac) ≥ 4.2 mmol/L, and N-terminal pro-brain natriuretic peptide (NT-proBNP) ≥ 3270 pg/ml were found to be the best cut-off values for the prediction of LVSD. CONCLUSION: Sepsis patients with left ventricular systolic dysfunction had a higher risk of death and atrial fibrillation. Hs-TnI, PCT, Lac, and NT-proBNP were independent risk factors of LVSD, and the LVSD predictive model constructed using these factors showed good diagnostic performance. Trial registration: Chinese Clinical Trial Registry No: ChiCTR2000032128. Registered on 20 April 2020, http://www.chictr.org.cn/showproj.aspx?proj=52531.