To our knowledge, few studies investigated the association between baseline LDL-C levels and long-term all-cause mortality in patients with CAD. The main finding of the current study is that the worse prognosis of patients with low LDL-C (< 1.8 mmol/L) is mainly mediated by their higher prevalence of malnutrition. Our study demonstrated that the cholesterol paradox also existed among people with CAD. After accounting for the marked differences in their age, sex and presenting comorbidities, the differences were not significant between patients with different LDL-C levels. After adjustment of malnutrition, CAD patients with low LDL-C showed low risk of long-term all-cause mortality.
Our study showed an inverse association between baseline LDL-C level and long-term outcomes in unadjusted analysis. The cholesterol paradox was also observed in other studies involving subgroups of CAD patients (e.g., ACS) [3–7], highlighting the impact of baseline confounders. Cho et al found that among patients with acute myocardial infarction (AMI), lower baseline LDL-C (< 1.8 mmol/L) was associated with higher 1-year mortality before adjusting for baseline confounders . In that study, age and rate of comorbidities decreased as LDL-C increased. In a Cox proportional-hazards model, LDL-C level was not an independent predictor in 1-year mortality, after adjusting for baseline confounders. Wang et al demonstrated that higher admission LDL-C level was associated with better in-hospital survival in patients with ACS in unadjusted analysis . Similarly, the results suggest cholesterol paradox may due to other clinical characteristics. Reddy et al, Al-Mallah et al and Nakahashi et al demonstrated that lower baseline LDL-C level was associated with higher mortality before and after adjusting for baseline confounders[5–7]. All five studies attempted to explore the mechanisms behind the paradox by taking baseline confounding factors into account, which included demographic information and comorbidities. There are noteworthy similarities between our study and the five previous studies. First, these study groups were all patients with CAD. Second, the LDL-C levels were all the baseline values collected at admission. Third, after adjusting for baseline confounders (age, gender, comorbidity, etc.), decreased baseline LDL-C concentration was not associated with declined long-term mortality.
However, the influence of nutritional status was not considered in depth. According to available data from the five studies, patients with lower LDL-C also had lower plasma albumin and total cholesterol concentrations, which may reflect underlying malnutrition status. CONUT score was an efficient tool for early detection and continuous control of in-hospital undernutrition . Low LDL-C level always combined malnutrition. In our study, prevalence of malnutrition was 56.30%, 51.75% and 90.30% in overall population, LDL-C ≥ 1.8mmol/L group and LDL-C < 1.8mmol/L group, respectively. The malnutrition rate of patients with low LDL-C level was much higher than that of patients with high LDL-C level. This was also observed in other studies, highlighting the importance of nutritional status [10, 11]. Roubín et al’s finding showed that in patients with ACS 38.5%, 10.4% and 0.9% patients were mildly, moderately and severely malnourished according to CONUT score, respectively . Roubín et al and Wada et al found that nutritional status assessed by the CONUT score was significantly associated with long-term clinical outcomes in patients with CAD [10, 11].
There may be several possible explanations for the cholesterol paradox. First, a plausible explanation for the absence of a positive correlation between baseline LDL-C and long-term all-cause mortality before adjustment is that these patients have a higher proportion of elder people (≥ 75years) and comorbidities according to baseline characteristics, which was associated with worse prognosis. Patients with LDL-C level < 1.8 mmol/L had significantly higher prevalence than those with LDL-C ≥ 1.8mmol/L in patients aged ≥ 75years (19.10% vs. 14.04%), diabetes mellitus (33.37% vs. 26.05%), anemia (43.18% vs. 30.30%) and atrial fibrillation (2.74% vs. 2.22%). Additionally, the increased long-term mortality may result from basic diseases to some extent. It is well known that smokers have better survival after AMI (the smoker paradox) . The smoker’s paradox was explained by the younger age and fewer cardiovascular risk factors in smokers compared with nonsmokers. Also, several studies have previously reported that patients with higher BMI have better clinical outcomes after PCI (the obesity paradox) [21, 22]. Furthermore, recent studies have suggested that the mechanism of the obesity paradox may be related to confounders of baseline characteristics associated with survival [23–25]. Similarly, in our study, multivariate Cox regression analysis revealed that baseline LDL-C level was not an independent predictor of mortality after adjusting for age, sex and comorbidities. Second, patients with lower LDL-C level were at poorer nutritional status. Malnutrition, in particular, significantly affected prognosis. Compared with unadjusted model, the correlation between low LDL-C level and prognosis changed from negative to positive after adjustment of malnutrition. Previous study has found that low total cholesterol represents a biological marker for concurrent cachexia, malnutrition, cancer, and other chronic diseases with a proven adverse impact on survival. Moreover, evidences are emerging that cholesterol is related to the regulation of immune cell function by improving their antitumor activity and activating immune signaling, which may provide novel insights into the role of cholesterol in the development of cancer[27–29]. Malnutrition has been singled out as the most common cause of secondary immunologic dysfunction . Failure to recognize or to anticipate the development of malnutrition can allow the needless presence of nutrition-related immunodeficiency and proneness to infection, with morbidity and mortality increased. Hence, the cholesterol paradox in patients with CAD may be mainly associated with malnutrition.
All these findings strongly support the need for physicians to integrate in their daily practice the identification of malnutrition. This may improve the risk stratification and guide subsequent interventions of secondary prevention. The effect of malnutrition should be considered when LDL-C is used to assess the risk of poor prognosis in CAD patients. Clinicians should not let down their guard when they meet CAD patients with low LDL-C level. Malnutrition screening in patients with CAD might identify patients at high risk of adverse outcomes who might benefit from tailored secondary prevention programs with nutritional supplements to improve their prognosis.
The study was conducted at Guangdong Provincial People’s Hospital which is the largest cardiovascular center in South China and the sample size was sizeable. However, there were some limitations among this analysis. First, this study was a retrospective single-center study. Second, a single value of LDL-C at admission was used, which made it difficult to assess the effect of changes in LDL-C level at follow-up on clinical outcomes. Therefore, the study focused on the clinical significance of initial LDL cholesterol level on clinical outcomes among patients with CAD.