A CD36-based prediction model for sepsis-induced myocardial injury

Abstract

Background

Sepsis-induced myocardial injury (SIMI) is a prevalent form of organ dysfunction with a significant impact on the mortality rate among sepsis patients. This study aims to develop a predictive model for SIMI using plasma CD36 levels.

Methods

A prospective study was conducted from January 1, 2023, to December 1, 2023, involving sepsis patients admitted to the Department of Intensive Care Medicine at Shanghai General Hospital. Plasma CD36 levels were measured within 48 h of ICU admission, prior to the diagnosis of sepsis-associated myocardial injury. Myocardial damage was assessed using troponin levels.

Results

Two significant risk factors for SIMI were identified: age and elevated CD36 levels. CD36, THBS1, and BNP were determined to be independent mortality risk factors. The myocardial injury group exhibited higher plasma CD36 levels compared to the non-injury group. Additionally, the deceased group had higher plasma CD36 levels than the survivors. No significant differences in CD36 levels were observed between groups with lung and stomach infections or between Gram-positive and Gram-negative infection groups. Similarly, there was no statistically significant difference in CD36 levels between surgical and medical patients. A predictive model for SIMI was formulated as follows: ln [P/(1-P)] = -0.000818age + 0.4975756CD36 − 5.400293. The model’s quality of fit was tested with a P-value of 0.4682, indicating a good degree of discrimination and calibration, as evidenced by the area under the ROC curve (0.7724).

Conclusion

The prognosis of individuals with sepsis is closely associated with elevated CD36 levels. Elevated CD36 is identified as an independent risk factor for both SIMI and mortality in sepsis patients. The predictive model suggests that high CD36 levels are indicative of SIMI and are associated with a poor prognosis.

1. Introduction

Sepsis, with approximately 18 million severe cases annually and a fatality rate ranging from 30 % to 70 %, is a complex disease characterized by high morbidity and mortality [1], [2], [3], [4]. Cardiac injury is one of the most frequent complications of sepsis, with up to 70 % of cases involving sepsis-induced myocardial injury [2], [3]. Cardiac dysfunction due to myocardial damage is a leading cause of death in sepsis patients [3]. Early identification of sepsis-induced myocardial injury (SIMI) is crucial for reducing morbidity and mortality rates.

CD36, a key receptor involved in lipid transport on the surface of cardiomyocytes, is essential for lipid uptake [15]. Studies have shown that cardiomyocytes from patients with CD36 deficiency have impaired long-chain fatty acid uptake [15], and inhibition of CD36 reduces cardiac lipotoxicity [18]. Targeting class B scavenger receptors and CD36 has been shown to improve survival and acute outcomes in septic mice [7]. Modulation of CD36 has also been effective in reducing sepsis-induced oxidative stress and inflammation in alveolar epithelial cells [8]. Despite its importance in sepsis, research on the relationship between CD36 and myocardial injury is limited.

Macrophage migration inhibitory factor has been shown to predict reperfusion myocardial injury in patients with ST-segment elevation myocardial infarction (MI) [5]. Aortic systolic blood pressure is a predictor of perioperative myocardial injury following transcatheter aortic valve implantation [6]. Single biomarkers have limitations in predicting disease and prognosis due to varying sensitivity and specificity. There is a suggestion in the literature to develop predictive models for MI diagnosis [9], [10], [11], [12], [13]. However, few models for SIMI have been reported. This study aims to develop a predictive model for SIMI using a combination of plasma CD36 levels and age.

2. Materials and methods

2.1. Selection of cases

This observational study included adult patients aged 18 years and older with sepsis, as defined by the Sepsis 3.0 criteria [12]. The study period was from January 1, 2023, to December 1, 2023, with follow-up continuing until December 31, 2023. Eligible patients were those diagnosed with sepsis, and their hospital stays were monitored. Informed consent was obtained from all participants, and the study was approved by the Ethics Committee of Shanghai General Hospital.

2.2. Inclusion criteria

The criteria included age between 18 and 85 years, ICU admission, clinically diagnosed or confirmed infection, an Acute Physiology and Chronic Health Evaluation (APACHE) II score ≥ 2 or a Sequential Organ Failure Assessment (SOFA) score ≥ 2 due to organ failure (with a baseline SOFA score of 0 for patients with pre-existing organ dysfunction), and exclusion of patients previously admitted to the ICU.

2.3. Exclusion criteria

These included other cardiovascular diseases (e.g., coronary heart disease, non-sepsis-related cardiomyopathy, myocardial infarction, myocardial sepsis, myocarditis, chronic heart failure, chronic obstructive pulmonary disease, valvular disease), heart transplant recipients, ICU patients admitted within 24 h, history of myocardial infarction at ICU admission, certain chronic and progressive neoplastic diseases associated with CD36, refusal of emergency care, and history of allergy to antibiotics or other drugs.

2.4. Patient selection criteria

The selection was based on demographic characteristics (age, sex, weight), disease status, severity scores (APACHE II or SOFA), ICU admission data within 24 h (mechanical ventilation and continuous renal replacement therapy), initial laboratory data, and outcomes (28-day mortality, ICU, and hospital length of stay).

2.5. Sepsis definition and diagnostics

Sepsis 3.0 was defined as life-threatening organ dysfunction due to infection with a SOFA score ≥ 2 compared to baseline [12]. Troponin T levels ≥ 0.01 ng/mL after 24 h were indicative of sepsis-induced myocardial injury (SIMI) [14].

2.6. Ethical standards

The study adhered to medical ethical standards and was approved by the Medical Ethics Committee of Shanghai General Hospital[Approval No. [2021]KY037)]. Patient confidentiality was maintained by anonymizing data, using only hospitalization numbers for data validation. Informed consent was obtained from all participants and/or their legal guardians. The study was performed in accordance with the Declaration of Helsinki.

2.7. Clinical follow-up

Clinical data were collected for 28 days post-enrollment, including basic patient information and disease conditions, comorbidities, and primary and secondary outcomes.

2.8. Blood collection and analysis

Blood samples were collected in sterile heparin tubes within 48 h post-hospitalization, and CD36 levels were determined using ELISA kits (ab267614, abcam).

2.9. Statistical analysis

Data analysis was performed using SPSS 22.0. Normally distributed data were presented as mean ± standard deviation (SD), while non-normally distributed data were shown as median (interquartile range, IQR). Categorical data were expressed as number and percentage. Propensity Matching Score was used to control for the variables. Univariate and multivariate logistic regression models were used to assess independent effects, and odds ratios (ORs) with 95 % confidence intervals (CIs) were calculated for myocardial injury risk. The area under the receiver operating characteristic curve (AUC ROC) was used to evaluate the diagnostic value of CD36 in predicting myocardial injury and death. AUC ROC analyses compared patients with myocardial injury to those without and survivors to non-survivors. The optimal critical value was determined by maximizing the Youden index for sensitivity and specificity. A p-value < 0.05 was considered statistically significant for all tests, which were two-sided.

3. Results

3.1. Demographic and baseline characteristics between the two groups

The study started in January 2023 and ended in December 2023 A total of 120 patients were selected for the study during the one-year period. According to the exclusion criteria of this study, 36 patients were excluded (21 previously diagnosed with severe cardiovascular disease, 8 with advanced tumors, 5 who refused all treatments, and 2 with a history of drug allergy. A total of 84 patients (70 %) were included. Of these, 39 patients were diagnosed with MI, 12 patients died of MI, and 11 patients died of non-MI (Fig. 1). Table 1 lists the demographic and baseline characteristics of the patients (based on MI subgroups). There were 51 (60.7 %) males and 33 (39.3 %) females, and all 84 patients were included in the final data analysis (Fig. 1). The median age was 62.5 (IQR: 48.25, 70.75) years, and hypertension, diabetes mellitus and COPD disease were the most common comorbidities. The median APACHE II score was 17.0 (IQR: 12.25, 20.75) and the median SOFA score was 9 (IQR: 7, 13). Mechanical ventilation and norepinephrine utilization within 24 h of ICU admission were 63.1 % and 53.6 %, respectively. The main source of infection was abdominal infection in 44 (52.4 %) followed by pulmonary infection in 40 (47.6 %).39 (46.4 %) had MI.27.4 % of patients received acute renal replacement therapy. The mortality rate was 27.4 %. The age, sex, and comorbidity rates were similar in the non-MI (n = 45) and MI (n = 39) groups.

Fig. 1.

Flowchart of the research.

Table 1.

Demographic and clinical information of patients (n = 84).

Characteristics	No-MI (N = 45)	MI (N = 39)	p	Total (N = 84)
Sex, % male	27(60)	24(61.5)	0.886	51(60.7)
Age, years	58（51, 70）	64（46, 71）	0.648	62.5（48.25, 70.75）
Comorbidities, n (%)
Hypertension	18(40)	19(48.7)	0.422	37(44.0)
Diabetes	7(15.5)	7(10.1)	0.769	14(16.7)
Immune diseases	5(11.1)	3(7.7)	0.719	8(9.5)
CKD	7(15.5)	3(7.7)	0.327	10(11.9)
Liver disease	4(8.9)	3(7.7)	0.581	7(8.3)
COPD	5(11.1)	6(15.4)	0.748	11(13.1)
SOFA	9（12, 16）	9（7, 14）	0.677	9（7, 13）
APACHE II	21（14, 23）	17（13, 23）	0.233	17.0（12.25, 20.75）
28 days mortality	11(24.4)	12(30.8)	0.517	23（27.4）
28 days survival time（days）	28（28, 28）	28（13, 28）	0.266	28（19.25, 28）
1 years mortality	11(24.4)	15(38.5)	0.237	26(30.9)
Noradrenaline use, (%)	24(53.3)	23(51.1)	0.384	45（53.6）
Noradrenaline use, (day)	1(0,5)	1(0,2)	0.107	6(0,6)
Mechanical ventilation, (%)	26(49.1)	27(50.9)	0.278	53（63.1）
Mechanical ventilation, (day)	4(0,9)	4(0,10)	0.955	9(0,9)
Need for dialysis, (%)	15(65.2)	8(34.8)	0.189	23（27.4）
Need for dialysis, (day)	0(0,3)	0(0,0)	0.686	3(0,7)

CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; APACHE, Acute Physiology And Chronic Health Evaluation; SOFA, sequential organ failure assessment. EVLW, extravascular lung water, Values are expressed as the mean ± standard deviation or the median and interquartile range. *P < 0.05.**p＜0.01.For normally distributed data, the mean ± standard deviation is used, and for non-normal distribution, the median, lower quartile and upper quartile are used.

The mortality rate was higher in patients with MI compared to the non-MI group, but not statistically significant (24.4 % vs. 30.8 %, p = 0.517). Table 1 shows the demographic and baseline characteristics of the groups according to MI. There was no statistically significant difference between the two groups with respect to mechanical ventilation (49.1 % vs. 50.9 %, p = 0.278); the need for norepinephrine (53.3 % vs. 51.1 %, p = 0.007), APACHE II score median 21 (IQR: 14, 23) vs. 17 (IQR: 13, 23), p = 0.233, and SOFA score median 9 (IQR: 12, 16) vs 9 (IQR:7, 14),p = 0.677, and there was no statistical difference between the two groups (Table 1).

3.2. Laboratory findings of the patients

The levels of IL-6, TNF-α, cTnI, and BNP in the septic myocardial injury group and septic non-myocardial injury group were, respectively, (938.0 ± 240.4vs 201.6 ± 40.83) mmol/l; (32.41 ± 5.147 vs 15.94 ± 1.245) mmol/l;(2.041 ± 0.3076 vs. 0.2734 ± 0.2104) μg/l; (765.7 ± 170.0 vs.

286.7 ± 76.37) pg/ml, IL-6, TNF-α, cTnI, and BNP were significantly higher in septic myocardial injury group, p = 0.0018, 0.0014, < 0.0001, and 0.0087 the differences were statistically significant. (Fig. 2A-D, Table3).

Fig. 2.

Inflammation in patients with sepsis, and CD36 in myocardial injury in sepsis: A. Comparison of peripheral serum IL6 levels in patients with sepsis-induced myocardial injury (SIMI) and patients without SIMI. B. Comparison of peripheral serum TNF-α levels in patients with SIMI and without SIMI. C. Comparison of peripheral serum TNF-α levels in patients with SIMI and without SIMI. D. Comparison of peripheral serum BNP levels in patients with SIMI and without SIMI. E. Comparison of peripheral serum CD36 levels in patients with SIMI and without SIMI. F. Comparison of peripheral serum CD36 levels in patients who survived sepsis and in patients who died of sepsis. G. ROC plots of CD36 prediction of septic myocardial injury. H. ROC plots of CD36 prediction of septic mortality.

Table 3.

Laboratory findings of patients (n = 84).

Laboratory results	No-MI (N = 45)	MI (N = 39)	p	Total (N = 84)
IL6 (mmol/l)	201.6 ± 40.83	938.0 ± 240.4	0.0018	543.51 ± 119.9
TNFα(mmol/l)	15.94 ± 1.245	32.41 ± 5.147	0.0014	23.58 ± 2.62
cTNI(μg/l)	0.2734 ± 0.2104	2.041 ± 0.3076	< 0.0001	1.04 ± 0.17
BNP (pg/ml)	286.7 ± 76.37	765.7 ± 170.0	0.0087	509.08 ± 92.12
CD36 (ng/ml)	9.650 ± 0.2965	11.68 ± 0.2999	＜0.0001	10.09 ± 0.18
18–39 years old CD36 (ng/ml)	10.1847 ± 2.372	11.8261 ± 1.628	＜0.0001	11.0685 ± 2.096
40–59 years old CD36 (ng/ml)	9.4792 ± 1.886	11.7613 ± 1.783	＜0.0001	10.5165 ± 2.1399
60–80 years old CD36 (ng/ml)	9.6068 ± 2.006	11.6014 ± 2.052	＜0.0001	10.5023 ± 2.2421

IL-6, Interleukin 6；TNF-α, Tumor Necrosis Factor-α ；*p＜0.05, **p＜0.01, ***p＜0.001；For normally distributed data, the mean ± standard deviation is used, and for non-normal distribution, the median, lower quartile and upper quartile are used.

CD36 levels in the septic myocardial injury group and septic non-myocardial injury group were (11.68 ± 0.2999 vs 650 ± 0.2965) ng/ml, and CD36 levels in the septic death group and septic non-death group were(12.19 ± 0.3238 vs 9.994 ± 0.2668) ng/ml, in which both CD36 were significantly higher, p < 0.0001, and the difference was statistically significant, (Fig. 2E-F, Table 3).

After Propensity Matching Score, The levels of IL-6, TNF-α, cTnI, BNP and CD36 in the septic myocardial injury group and septic non-myocardial injury group were, respectively, (936.3 ± 246.8vs.226.7 ± 46.78) mmol/l; (32.13 ± 5.277 vs.15.98 ± 1.055) mmol/l;(2.039 ± 0.3157vs. 0.185 ± 0.042) μg/l; (771.4 ± 174.4 vs. 316.7 ± 89.27) pg/ml; (9.666 ± 0.324 vs. 11.68 ± 0.308)ng/ml,IL-6, TNF-α, cTnI, BNP and CD36 were significantly higher in septic myocardial injury group, p = 0.007, 0.005, < 0.0001, 0.024, and ＜0.0001 the differences were statistically significant (Table4).

Table 4.

Laboratory findings of patients after Propensity Matching Score (n = 76).

Laboratory results	No-MI (N = 38)	MI (N = 38)	p	Total (N = 76)
IL6 (mmol/l)	226.7 ± 46.78	936.3 ± 246.8	0.007	581.5 ± 131.3
TNFα(mmol/l)	15.98 ± 1.055	32.13 ± 5.277	0.005	24.05 ± 2.831
cTNI(μg/l)	0.185 ± 0.042	2.039 ± 0.3157	< 0.0001	1.112 ± 0.191
BNP (pg/ml)	316.7 ± 89.27	771.4 ± 174.4	0.024	544.0 ± 100.8
CD36 (ng/ml)	9.666 ± 0.324	11.68 ± 0.308	＜0.0001	10.67 ± 0.251

IL-6, Interleukin 6；TNF-α, Tumor Necrosis Factor-α ；*p＜0.05, **p＜0.01, ***p＜0.001；For normally distributed data, the mean ± standard deviation is used, and for non-normal distribution, the median, lower quartile and upper quartile are used.

3.3. CD36 predicts septic myocardial injury and prognosis/prognosis

CD36 predicted sepsis myocardial injury in ROC plot AUC = 0.7724, p < 0.0001, statistically significant, the results surface that CD36 can predict sepsis myocardial injury. (Fig. 2G) ROC plot AUC = 0.7862 for CD36 predicting sepsis death, p < 0.0001, which is statistically significant and results in surface CD36 predicting sepsis prognosis. (Fig. 2H).

3.4. Correlation of THBS1, CD36, IL6, TNF-α and myocardial injury in sepsis

Peripheral serum THBS1 in sepsis patients was positively correlated with TNI, a marker of myocardial injury, R2 = 0.1273, p = 0.0009. Peripheral serum CD36 in sepsis patients was positively correlated with TNI, R2 = 0.1314, p = 0.0007. Peripheral serum IL6 in sepsis patients had no significant correlation with TNI in myocardial injury, R2 = 0.02355, p = 0.1634. peripheral serum TNF-α was positively correlated with myocardial injury TNI in patients with sepsis, R2 = 0.06985, p = 0.0151 (Fig. 3A-D).

Fig. 3.

Correlation of THBS1, CD36, IL6, and TNF-α with myocardial injury in sepsis, Differences in CD36 between etiology, site of infection, medical / surgical sources, and single/mixed infections. A. Peripheral serum THBS1 correlates with myocardial injury in sepsis patients. B. Peripheral serum CD36 correlates with myocardial injury in sepsis patients. C. Peripheral serum IL6 correlates with myocardial injury in sepsis patients. D. Peripheral serum TNF-α correlates with myocardial injury in sepsis patients. E. Difference in CD36 between Gram-positive bacteria (G+) and Gram-negative bacteria (G-) infections. F. Difference in CD36 among patients with respiratory, abdominal, and urinary tract infections. G. Difference in CD36 between the surgical and the medical groups. H. Difference in CD36 between the single and mixed infections.

3.5. Multiple regression analysis of MI and risk factors for mortality

Multiple regression analysis of MI and mortality risk showed that age (OR = 1.031, 95 % CI 0.993–1.071, p = 0.110), and CD36 level (OR = 1.787, 95 % CI1.330–2.400, p < 0.001) were sepsis-associated MI THBS1 (OR = 1.021, 95 % CI 1.008–1.034, p = 0.002), BNP (OR = 1.001, 95 % CI 1.000–1.002, p < 0.001) and CD36 levels (OR = 1.456, 95 % CI 1.087–1.400, p < 0.001) were independent risk factors for sepsis-related MI. % CI 1.087–1.950, p = 0.012) were identified as independent risk factors for death (Table 2).

Table 2.

Multivariate analysis of MI and risk of death (n = 84).

SIMI	OR	CI (95 %)	p
Age	1.031	(0.993–1.071)	0.110
CD36	1.787	(1.330–2.400)	＜0.001*
Death	OR	CI (95 %)	p
CD36	1.456	(1.087–1.950)	0.012*
THBS1	1.021	(1.008–1.034)	0.002*
BNP1	1.001	(1.000–1.002)	＜0.001*

SIMI, sepsis-induced myocardial injury; Or, odds ratio; CI, Confidence interval* p < 0.05.

3.6. Differences in CD36 between pathogens, sources of infection, medical and surgical sources, single and co-infected

There were no statistically significant differences between G + -sexual and G--sexual bacterial infections in CD36. There were no statistically significant differences in CD36 among patients with respiratory tract infections, abdominal infections, and urinary tract infections. There was no significant difference in CD36 between the surgical and medical groups. There was no significant difference in CD36 in the monoinfection and co-infection groups (Fig. 3E-H).

3.7. ROC curve

ROC analysis of SIMI based on CD36/ IL6 / TNFα / BNP. Area under the ROC curve for CD36 > IL6 > TNFα > BNP (Fig. 4A).

Fig. 4.

ROC analysis of SIMI. A. ROC analysis of SIMI based on CD36/ IL6 / TNFα / BNP. Area under the ROC curve for CD36 > IL6 > TNFα > BNP,B. ROC analysis of the CD36-based SIMI prediction model. The model fit goodness-of-fit test showed p = 0.000 and the area under the ROC curve for the model = 0.9862, C. ROC analysis of the CD36-based sepsis prognostic model. The goodness-of-fit test for the model showed p = 0.2035 and the area under the ROC curve for the model = 0.8730.

Predictive model:

SIMI prediction model based on CD36

The SIMI prediction model based on this CD36 is as follows: ln [ P /(1-P)] = -0.000818*age + 0.4975756*CD36-5.400293

Model fit goodness-of-fit test P = 0.4682,Area under the curve of the model ROC curve = 0.7724 (Fig. 4B).The results suggest that the model has a high degree of differentiation and calibration, and has good clinical utility. The model can be used to determine the prognosis of patients at the early stage of hospitalization. The predictive efficacy of the model can be further improved by expanding the sample size and optimizing the predictive indexes.

The prognostic model of sepsis based on CD36 is as follows:

ln [ P /(1-P)] = 0.5048321*CD36 + 0.0058275*THBS1 + 0.0002047*BNP1--8.287575

Model fit goodness-of-fit test P = 0.7143,Area under the curve of the model ROC curve = 0.8054 (Fig. 4C).The results suggest that the model has a high degree of discrimination and calibration, and has good clinical utility. The model can be used to determine the prognosis of patients at the early stage of hospitalization. The predictive efficacy of the model can be further improved by expanding the sample size and optimizing the predictive indexes.

3.8. Validation

The internal validation model bootstrap self-sampling method was used to validate the predictive effectiveness of the model (test, n = 50). The area under the ROC curve for the model = 0.71 (Fig. 5A). External Validation Model A was re-selected with a new sample dataset (test, n = 38) and the ROC curve for SIMI was validated using SPSS software. The area under the curve was 0.625 (Fig. 5B). Validation ROC curve of the internal validation of the death prediction model, area under the curve is 0.831 (Fig. 5C). Validation ROC curve for external validation of the SIMI prediction model with an area under the curve of 0.707 (Fig. 5D).

Fig. 5.

Validation prediction model. A. Validation ROC curve for internal validation of SIMI prediction model with area under the curve of 0.710, B. Validation ROC curve for external validation of SIMI prediction model with area under the curve of 0.625, C.Validation ROC curve for internal validation of death prediction model with area under the curve of 0.831,D.Validation ROC curve for external validation of death prediction model with area under the curve of 0.707.

4. Discussion

The pathogenesis of Sepsis-Induced Myocardial Injury (SIMI) is multifaceted. On one hand, it can directly induce myocardial cell death via the TLR4 receptor under the influence of Pathogen-Associated Molecular Patterns (PAMPs) and Lipopolysaccharides (LPS). On the other hand, an excessive immune response during the early stages of sepsis is also a significant cause of myocardial cell damage. Despite the routine clinical use of Continuous Renal Replacement Therapy (CRRT) and glucocorticoids, some patients still experience SIMI, indicating a need for improved early recognition of the condition [2], [3].

The primary objective of this study was to develop a predictive model for SIMI using plasma CD36 levels in conjunction with age. This model not only aimed to predict SIMI but also to elucidate the relationship between reduced serum CD36 levels and mortality. The model's predictive accuracy was validated using both internal and external data sets, demonstrating good discriminatory power. The novel model, incorporating age and plasma CD36 levels, identified age and elevated plasma CD36 as the main independent risk factors for SIMI in adult patients. Furthermore, BNP, high levels of CD36, and THBS1 were identified as independent risk factors for death. Notably, CD36 showed superior predictive discrimination and utility for septic myocardial injury compared to IL6, TNFα, and BNP.

A retrospective cohort study identified predictors of in-hospital mortality in sepsis-induced myocardial injury, including diabetes risk factors, APACHE II score, mechanical ventilation, vasoactive support, troponin T, and creatinine [16]. However, this study did not predict the occurrence of myocardial injury, and our model using CD36 demonstrated a higher AUC for predicting death than the aforementioned study. This may be attributed to the significant role of CD36 in myocardial injury during sepsis.

CD36 is known to play a crucial role in various cardiac processes, including lipid metabolism, calcium homeostasis, inflammation, and repair [17]. Studies have shown that cardiomyocyte-specific ablation of CD36 improves functional recovery after ischemia [19], and inhibition of CD36 has been shown to attenuate myocardial injury in post-ischemic treated diabetic rats [20] and in diabetic cardiomyopathic rats by down-regulating CD36-mediated iron death [21]. Additionally, soluble CD36 assessed by a novel monoclonal antibody-based sandwich ELISA has been shown to predict cardiovascular mortality in dialysis patients [22], and plasma soluble CD36 has been associated with cardiovascular risk factors in patients with early-onset coronary artery disease [23]. Further mechanistic studies on the role of CD36 in myocardial injury during sepsis are warranted.

Our study also found no significant difference in CD36 levels between patients with pulmonary infections and those with abdominal or urinary tract infections. Similarly, no significant differences were observed in CD36 levels across various atypical pathogens, including G+, G-, fungi, and viruses. There were also no significant differences in CD36 levels between medical and surgical patients or between those with monoinfections and mixed infections. This suggests that CD36 may not be influenced by these factors, indicating its potential for broad application in monitoring various sepsis types.

Our findings indicate that CD36 can predict not only myocardial infarction (MI) but also 28-day mortality. When combined with BNP and Thrombospondin 1 (THBS1), CD36 adds value in predicting mortality. Consequently, we developed a combined predictive model incorporating CD36, BNP, and THBS1. In patients with high plasma CD36 levels, clinicians should maintain an appropriate fluid status, avoid cardiac overload, and closely monitor cardiac function. Additionally, the specific role of plasma CD36 in the severity stratification of SIMI merits further investigation.

Our predictive model, which included CD36, THBS1, and BNP, demonstrated an AUC of 0.8054 for predicting death in patients with sepsis-related myocardial injury. This suggests a high degree of differentiation and calibration, with good clinical utility. The model allows for the early determination of a patient's prognosis upon hospital admission and showed better discrimination and utility than the SOFA score alone. It is expected to support standardized sepsis treatment and cardioprotection decision-making. Further research is needed to analyze the specific role of plasma CD36 in SIMI and to explore the potential of CD36-specific therapies in treating SIMI.

In real-world clinical settings, high CD36 levels in sepsis patients suggest a potential for increased myocardial injury. Treatment protocols should focus on early and appropriate antibiotic therapy, along with supportive care to address hemodynamic instability and organ dysfunction. Monitoring strategies should include regular echocardiography to assess myocardial function, with a focus on sensitive markers like global longitudinal strain. Management of septic cardiomyopathy should be tailored to the patient's hemodynamic profile, with careful fluid management and close monitoring of cardiac function. Finally, a multidisciplinary approach that integrates these protocols and strategies is essential for optimizing patient outcomes in sepsis-induced myocardial infarction.

4.1. Deficiencies

This study has several limitations. The sepsis patient cohort primarily consisted of individuals with pulmonary and abdominal infections, and the majority were older and more critically ill. Since the study relied solely on plasma CD36 measurements taken within 48 h of ICU admission, the prognostic implications of the temporal changes in CD36 levels throughout the sepsis course were not assessed. Further validation in larger samples and multi-center studies is necessary.

5. Conclusion

In summary, plasma CD36 levels measured within 48 h of ICU admission were found to be predictive of SIMI in our patient population. These levels were associated with MI and strongly correlated with MI severity. Initial plasma CD36 levels also served as a predictor of mortality within 28 days of a sepsis episode. Collectively, these findings suggest that blood biomarkers, physiologic parameters, and clinical risk scores may serve as potential predictors of myocardial injury in sepsis. Future research should aim to refine the accuracy and reliability of these indicators for clinical use, with the goal of providing more precise predictions and personalized management strategies for patients with septic myocardial injury.

Declarations

Consent for publication

All authors were involved in the final written approval of the manuscript.

Availability of data and material

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Ethics approval and consent to participate

The study adhered to medical ethical standards and was approved by the Medical Ethics Committee of Shanghai General Hospital[Approval No. [2021]KY037)]. Patient confidentiality was maintained by anonymizing data, using only hospitalization numbers for data validation.

Authors' contributions

YX and HL were involved in conception and design. RW, HCS and PJG were involved in administrative support; ZGZwere involved in the provision of study materials or patients; YX was involved in the collection and organization of data; DNC was involved in the analysis and interpretation of data; and all authors were involved in the writing of the manuscript. All authors were involved in the final approval of the manuscript.

Funding

This study was supported by the Youth Program of the National Natural Science Foundation of China (82202423).

CRediT authorship contribution statement

Yun Xie: Writing – review & editing, Writing – original draft, Methodology. Hui Lv: Formal analysis. Daonan Chen: Data curation. Peijie Huang: Data curation. Zhigang Zhou: Investigation. Ruilan Wang: Writing – review & editing.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.