In this cohort study involving adult intensive care CS patients, we stratified the analysis according to age and BMI. Elderly patients with a BMI ≥ 30 kg/m2 had the best prognostic outcome at 28 days, with a survival rate of 66.75%, and, compared with younger patients and elderly patients with a BMI < 30 kg/m2, elderly patients with a BMI ≥ 30 kg/m2 had a lower in-hospital mortality and ICU mortality were lower, despite longer ICU stays. This is despite the fact that underweight (BMI < 18.5 kg/m2) was demonstrated to be a confirmed risk factor for death at 28 d in elderly critically ill patients with CS by univariate and multivariate logistic analyses [OR 2.02, 95%CI (1.06, 3.87), P = 0.033]. However, overweight or obesity, to some extent, does not have a significant effect on the 28-day prognosis of elderly CS patients.

The obesity paradox is not a new discovery. As early as 2002, Gruberg [12] et al. first observed that the 1-year mortality rate of normal weight patients with coronary artery disease after percutaneous coronary intervention was significantly lower than that of overweight and obese patients (P < 0.05). Subsequent studies covered by the umbrella term 'reverse epidemiology' have further confirmed this phenomenon [13], particularly in cardiovascular diseases [14,15,16]. For example, in a study of 7767 patients with stable heart failure, Curtis [17] et al. found that overweight and obese patients had a significantly lower risk of death than healthy weight patients (risk ratios of 0.88 and 0.81, respectively). In a study of BMI and mortality in 64,436 patients with acute coronary syndromes (ACS), Oska [18] et al. found that patients who were underweight (BMI < 18.5 kg/m2) had the highest risk of death, whereas moderately overweight (BMI 26.5–28 kg/m2) patients had the lowest risk of death when receiving medical therapy and coronary interventions, and there was a U-shaped relationship between BMI and risk of death, with overweight or obese patients having the lowest risk of death and normal-weight and underweight patients having the highest risk. Despite this, the theory of the obesity paradox remains controversial in patients with CS [19]. And according to our findings, overweight or obesity did not significantly affect the 28-day mortality of young and elderly CS patients.

Excess adipose tissue leads to an increase in the metabolic demands of the body [20]. And this extra tissue in turn increases the total circulating blood volume through increased output per beat [21]. As blood circulates back to the heart through the venous system, the increased cardiac load leads to a corresponding increase in wall tension and stress in the left and right ventricles [22]. With this hyperdynamic circulation, hemodynamic overload, and increased cardiac output, the patient's biventricles become hypertrophied, and in the long run, complete heart failure develops and cardiac output decreases [23, 24]. Whereas CS is an acute process involving pump failure leading to myocardial and systemic underperfusion with compensatory physiologic mechanisms, it spirals into a vicious cycle leading to multiorgan dysfunction [25]. Previous literature elucidates that cardiac power is equal to cardiac index multiplied by mean arterial pressure and is closely related to mortality in patients with CS [26]. The higher mortality in patients with a low cardiac power index is due to the fact that the cardiac index in this equation has cardiac output as the numerator and BMI as the denominator, so patients with higher BMI values may have a lower cardiac index, which in turn leads to lower cardiac power [25]. This may well explain why the prognosis of obese patients in the event of CS remains controversial.

Our study showed that there were significant differences in in-hospital mortality, ICU mortality, and 28-day mortality between patients with a BMI < 30 kg/m2 and those with a BMI ≥ 30 kg/m2 in both the younger and older age groups, with a significantly higher 28-day survival rate for patients with a BMI ≥ 30 kg/m2 in the older age group. Yet based on previous research, the obesity paradox seems to be contradictory. In the latest research examining mechanical circulatory aids for CS and obesity, Sreenivasan and his colleagues noted a consistent increase in mortality within the hospital among patients with moderate to severe obesity compared to those without obesity [27]. In contrast, in an analysis by Kwon et al., the obesity paradox was found to exist in patients with CS does exist and apparently occurs in males with CS, with a significant reduction in in-hospital mortality in obese male patients compared with non-obese male patients (24.1% vs. 34.2%, P = 0.004) [8]. In addition to the effects of obesity on cardiac function, the presence of these two outcomes may also be related to the body mass index itself, which is not a direct indicator of body fat content or of the potential harm that may result from obesity. BMI represents the total of the fat mass index (composed of peripheral and visceral fat tissue) and the lean mass index (responsible for skeletal muscle mass, bones, and organs) [28]. Goyal suggests that the increased fat tissue may play a protective role ("healthy obesity"), but more commonly, fat tissue is harmful ("unhealthy obesity"), leading to metabolic abnormalities and a low inflammatory state, both of which are important components of metabolic syndrome [29]. Therefore, it is difficult to determine true obesity solely by using BMI. In addition, due to the presence of obesity, obese patients may be diagnosed with diseases (e.g., cardiovascular disease) earlier than normal-weight patients, which may be a source of diagnostic time bias, which in turn may affect the patient's prognosis [30].

In our study it was also found that underweight in elderly patients is increases the risk of death by 28 days. I Not only that, inflammation indicator (NLR) can affect the prognosis of these individuals. It is always known that the body changes with age. Aging is associated with significant decreases in energy expenditure, loss of skeletal muscle mass, and increased accumulation of visceral fat [31, 32], whereas in the elderly population sarcopenia is associated with weakness, overall functional impairment, and poor survival [33,34,35]. In addition, several adipokines that can be secreted by adipose tissue (e.g., lipocalin, apelin, and reticulin) have been shown to be cardioprotective and to exert a variety of beneficial effects on cardiovascular function [36], whereas underweight elderly patients lack certain fat reserves, and, therefore, the myocardial protection is correspondingly diminished. This conclusion of ours has been similarly reported previously, when a cross-sectional study in the United States showed that among people under 40 years of age, the risk of CVD was higher in low weight individuals, reaching 2.3 times that of normal weight individuals [37]. Overweight and obesity were associated with better survival in diabetic STEMI patients in a study that found that only underweight patients were more likely to experience CS (OR = 1.25) [38]. There are few studies on inflammatory indicators (NLR) on cardiovascular disease. Shah [39] et al. found that NLR > 4.5 independently predicted long-term mortality in patients with coronary heart disease in the general healthy population (HR 2.68, 95% CI 1.07–6.72, P = 0.035). In Basem et al.'s [40] study of the long-term prognosis of patients with non-ST-segment elevation myocardial infarction, it was found that the mean NLR level remained a significant predictor of hospitalization and 4-year mortality in patients, with an increased HR per unit increase in mean NLR (log) of 1.06 (P < 0.05) and 1.09 (P < 0.05), respectively. CS is a serious complication of cardiovascular disease and is closely related to systemic inflammation [41]. Cardiogenic shock is a serious complication of cardiovascular disease and is closely associated with systemic inflammation [41]. NLR can be considered as a robust prognostic marker for predictors of disease severity and mortality. It is closely associated with immune system disorders and can be used as a predictor of disease severity and mortality, especially in diseases characterized by systemic inflammation [42]. The importance of NLR for the diagnosis and management of cardiovascular disease and for determining the severity of the systemic inflammatory response, especially in elderly patients, was mentioned in the report by Buonacera and colleagues [43].

In a study by Regolo et al. [42], a significant negative correlation was found between NLR and PaO2/FiO2, suggesting that NLR plays a key role in patients with worsening PaO2/FiO2, but also implying that immune system dysfunction is present in such patients and may be strongly associated with mortality risk. In our study, we found that CS patients may have decreased PaO2/FiO2 due to systemic acute or chronic inflammation. PaO2/FiO2 was found to be an influential factor in patients' 28-day mortality by logistic multifactorial, both in younger and elderly patients. Combined with the fact that CS is strongly associated with systemic inflammatory response, we further explored the relationship between NLR and PaO2/FiO2, and came to a similar conclusion as Regolo et al. [42] that there was a significant correlation between NLR and PaO2/FiO2 in elderly patients. This reinforces the predictive value of NLR in elderly patients with CS.

There are several limitations to this study. First, this was a single-center retrospective cohort study based on the MIMIC-IV database, which includes primarily Western populations and may have challenges in representing diverse populations. In addition, the database only included patients admitted to the intensive care unit, i.e., those with more severe conditions. Second, because this study relied on database analysis, the correlation between time to BMI detection and patient admission to the ICU remains unclear. And we were unable to derive information on how long the patient had been in CS. Third, we only used BMI to assess obesity. We did not have other measures of obesity that are more relevant to cardiovascular prognosis, such as visceral fat. It would be helpful if more studies included different nutritional metrics, and although BMI is a widely used metric, it may not accurately reflect nutritional status given confounding factors such as fluid retention.