Introduction

Coronary heart disease (CHD) remains a leading cause of morbidity and mortality worldwide, significantly impacting public health and healthcare systems1,2. Despite advances in diagnostic and therapeutic strategies, the prognosis of CHD patients varies widely, necessitating reliable prognostic biomarkers to guide clinical management and improve outcomes3,4,5. Serum sodium, an essential electrolyte, plays a crucial role in maintaining fluid balance, regulating blood pressure, and ensuring cellular function6,7. Sodium homeostasis is tightly regulated by physiological mechanisms involving the kidneys8, endocrine system9, and cardiovascular system10. Understanding the impact of serum sodium concentrations on CHD could provide valuable insights into patient stratification and risk management.

Sodium intake10,11,12,13 and serum sodium concentrations14,15,16,17 have been associated with cardiovascular risk, particularly in the general population15,18,19,20 and in heart failure patients15,21,22,23. A recent large randomized controlled trial24 that contradicted the traditional belief that heart failure patients should strictly limit dietary sodium intake has sparked widespread discussion11,25,26. The Heart Failure Association of the European Society of Cardiology (ESC) recently recommended tailored sodium intake management for different populations22. However, the relationship between serum sodium concentrations and cardiovascular risk in CHD patients without heart failure remains poorly understood, with studies yielding inconsistent conclusions. Some studies suggest that low serum sodium19 is harmful to cardiovascular outcomes, others indicate risks associated with high levels7,27, and some report that both high and low levels are detrimental16. There are no authoritative reports specifically addressing baseline serum sodium level in CHD patients, particularly those without concomitant heart failure.

Therefore, this study aimed to investigate the relationship between serum sodium concentrations and the long-term prognosis of patients with newly diagnosed CHD through coronary angiography and had no prior systematic treatment. By categorizing patients based on serum sodium concentrations and analysing their clinical outcomes, we sought to determine whether serum sodium concentrations at admission can serve as a predictive biomarker for major adverse cardiovascular events (MACEs). Furthermore, in this study, we explored the potential dose‒response relationship between serum sodium concentrations and cardiovascular risk, aiming to better understand electrolyte management in CHD patients.

Methods

Study design and population

In this study, we utilized data from a prospective observational cohort conducted at the First Hospital of Hebei Medical University from August 1, 2018, to March 30, 2020. The study was approved by the Ethics Committee of the First Hospital of Hebei Medical University (ethical approval number: 2018003101), and informed consent for anonymous analysis was obtained from all participants. The study population consisted of patients with chest pain undergoing coronary angiography who were newly diagnosed with CHD, specifically defined as angina or acute myocardial infarction (AMI). Patients were excluded if they had coronary artery stenosis less than 50%, known heart failure, confirmed infectious diseases, severe renal insufficiency (stage 5 chronic renal failure), nonischaemic cardiac conditions (severe heart valve disease, acute myocarditis, non-ischemic malignant arrhythmias, primary dilated cardiomyopathy, and hypertrophic cardiomyopathy) or cancer, suspected Conn’s syndrome, Cushing’s syndrome, hypothyroidism, serum sodium concentrations less than 130 mmol/L, or incomplete data for important variables. All included patients were not taking sodium-dependent glucose transporters 2 (SGLT-2) inhibitors and diuretics at the time of enrollment. These exclusion criteria were chosen to ensure a homogeneous study population and to avoid confounding factors that could affect serum sodium concentrations or cardiovascular outcomes. The participant selection process is outlined in a flowchart (Fig. 1).

Figure 1
figure 1

Flowchart of the study participants.

Full size image

Regular follow-up was conducted via outpatient visits or phone calls at intervals of one month, three months, six months, one year, and annually up to five years. Electronic health records were also consulted to supplement the follow-up data. To enhance data accuracy, follow-up data were cross-verified with electronic health records. In this study, the primary focus was on MACEs, including cardiovascular death, nonfatal myocardial infarction, reperfusion therapies, and readmission for heart failure or severe angina. The final analysis included 681 patients after excluding those lost to follow-up.

Baseline data collection

Baseline demographic and clinical characteristics were collected at the time of hospital admission. These included age, sex, body mass index (BMI), smoking status, alcohol consumption status, medical history (including hypertension, diabetes, and family history of cardiovascular disease), and current medications. Laboratory data were obtained from fasting blood samples collected immediately upon admission or within 24 h of admission. The collected blood biomarkers included total cholesterol (TC) concentrations, low-density lipoprotein cholesterol (LDL-C) concentrations, high-density lipoprotein cholesterol (HDL-C) concentrations, triglyceride (TG) concentrations, and other relevant indicators, such as the estimated glomerular filtration rate (eGFR), blood urea nitrogen (BUN) concentrations, and serum electrolyte (Na, K, Cl, Ca) concentrations. Serum sodium concentrations were measured using the Beckman Coulter Chemistry Analyser AU5800 series following the manufacturer's instructions.

The severity of coronary artery lesions was assessed using the Gensini score based on coronary angiography findings28. The Gensini score provides a more comprehensive assessment of coronary heart disease severity than simply counting the number of stenotic vessels. Detailed information on the calculation method for the Gensini score and definitions for terms such as AMI, hypertension, diabetes, and smoking status are provided in the Appendix.

Statistical analysis

The baseline characteristics of the study population were compared across the five serum sodium level groups. Categorical data are presented as n (%), while continuous data are reported as the mean ± standard deviation or the median (interquartile range), depending on their distribution. Distribution normality and variance homogeneity were evaluated using the Kolmogorov‒Smirnov test and Levene’s test, respectively. Categorical variables were compared using the chi-square test, while continuous variables were analysed using one-way ANOVA for normally distributed data and the Kruskal‒Wallis test for nonnormally distributed data.

Kaplan–Meier survival curves were generated for different serum sodium concentrations, and the log-rank test was used to compare survival probabilities. Multivariate Cox proportional hazards models were constructed to evaluate the association between serum sodium concentrations and the risk of experiencing MACEs, adjusting for potential confounders selected based on their clinical importance and preliminary results from univariate Cox regression analyses. Hazard ratios (HRs) and 95% confidence intervals (CIs) were reported.

The predictive value of serum sodium concentrations was assessed by drawing receiver operating characteristic (ROC) curves. Improvements in predictive accuracy were quantified using the C-statistic (calculated through 1000 bootstrap samples), net reclassification index (NRI), and integrated discrimination improvement (IDI). Subgroup analyses were conducted to examine the consistency of the associations across different patient subgroups, including sex, diabetes status, hypertension status, and type of CAD. Interaction terms were tested to identify potential effect modifications. Nonlinear relationships between serum sodium concentrations and the risk of experiencing MACEs were explored using restricted cubic spline (RCS) regression models. Statistical analyses were performed using R version 4.2.3 (R Foundation for Statistical Computing, Vienna, Austria), and P < 0.05 was considered to indicate statistical significance.

Ethics approval and consent to participate

This study was approved by the Ethics Committee of the First Hospital of Hebei Medical University (ethical approval number: 2018003101), and was conducted in line with the Declaration of Helsinki.

Results

Baseline characteristics

The total cohort consisted of 681 patients, including 226 females (33.2%), with an average age of 60.78 ± 10.05 years. Among the patients, 293 (43.0%) were diagnosed with AMI, 188 (27.6%) had diabetes, and 417 (61.2%) had hypertension. The mean serum sodium level was 140.05 ± 2.97 mmol/L. The baseline characteristics of each group are shown in Table 1. Patients were categorized into five groups by their serum sodium concentration: L1 (130.0–135.9 mmol/L), L2 (136.0–138.9 mmol/L), L3 (139.0–140.9 mmol/L), L4 (141.0–142.9 mmol/L), and L5 (143.0–147.0 mmol/L). Significant differences were observed across these groups for various clinical parameters. Patients in the L1 group were significantly older (mean age 63.47 ± 9.08 years) than those in the other groups, and the youngest patients were in L2 (mean age 59.30 ± 10.23 years) (P = 0.022). The proportion of females was greatest in the L5 group (45.6%) and lowest in the L1 group (17.5%) (P < 0.001). The prevalence of AMI was highest in the L4 group (65.4%) and lowest in the L1 group (29.8%) (P < 0.001). Notably, the L1 group had the highest proportion of smokers (47.4%) and drinkers (33.3%).

Table 1 Clinical characteristics of individuals by serum sodium concentration.
Full size table

Serum sodium concentrations and the risk of experiencing MACEs

Patients were followed up for a median duration of 51.04 months (interquartile range: 40.88–53.80 months). In this time, 131 (19.2%) patients experienced MACEs, including cardiovascular death (16, 2.3%), nonfatal myocardial infarction (20, 2.9%), reperfusion therapy (27, 4.0%), and readmission due to heart failure or severe angina (68, 10.0%). Kaplan–Meier survival curves for different concentrations of serum sodium are presented in Fig. 2A. The log-rank test indicated significant differences in survival probabilities among the groups (P = 0.002), with the L2 group having the lowest survival rates. The incidence of MACEs varied significantly between the serum sodium groups (P = 0.005), as illustrated in Fig. 2B. Survival outcomes for cardiovascular death and myocardial infarction as serious events, and reperfusion therapy and readmission due to heart failure or severe angina as secondary serious events, are shown in Figs. 3 and 4, respectively. The L1 group had the highest incidence of serious events (8.8%, P = 0.034) and the lowest incidence of secondary serious events (3.5%, P = 0.007).

Figure 2
figure 2

Kaplan–Meier curves and percentages of patients who suffered MACE at different serum sodium concentrations. Patients were categorized into five groups by their serum sodium concentrations: L1 (130.0–135.9 mmol/L), L2 (136.0–138.9 mmol/L), L3 (139.0–140.9 mmol/L), L4 (141.0–142.9 mmol/L), and L5 (143.0–147.0 mmol/L).

Full size image
Figure 3
figure 3

Kaplan–Meier curves and percentages of patients who suffered serious events (cardiovascular death or nonfatal myocardial infarction) at different serum sodium concentrations. Patients were categorized into five groups by their serum sodium concentrations: L1 (130.0–135.9 mmol/L), L2 (136.0–138.9 mmol/L), L3 (139.0–140.9 mmol/L), L4 (141.0–142.9 mmol/L), and L5 (143.0–147.0 mmol/L).

Full size image
Figure 4
figure 4

Kaplan–Meier curves and percentages of patients who suffered secondary serious events (reperfusion therapy and readmission due to heart failure or severe angina) at different serum sodium concentrations. Patients were categorized into five groups by their serum sodium concentrations: L1 (130.0–135.9 mmol/L), L2 (136.0–138.9 mmol/L), L3 (139.0–140.9 mmol/L), L4 (141.0–142.9 mmol/L), and L5 (143.0–147.0 mmol/L).

Full size image

Independent associations between serum sodium concentrations and the risk of experiencing MACEs

To explore these associations in more detail, HRs were analysed to determine the relationship between serum sodium concentrations and the risk of experiencing MACEs. The HRs for unadjusted factors related to the risk of experiencing MACEs are presented in appendix Table S1. The independent effects of serum sodium concentrations on the risk of experiencing MACEs, adjusted for multiple covariates, are shown in Table 2. Model 1 was adjusted for age and sex; Model 2 was further adjusted for diabetes status, hypertension status, type of CHD, and severity of coronary artery lesions; and Model 3 was additionally adjusted for the treatment (medical therapy alone, medical therapy + percutaneous coronary intervention, or medical therapy + coronary artery bypass grafting), eGFR, LDL-C concentrations, and TG concentrations. According to Model 3, patients in the L1 had a significantly lower risk of MACE (HR = 0.31, 95% CI 0.14–0.70; P = 0.005) than L2. Similarly, the HRs of L3, L4, and L5 were 0.48 (95% CI 0.30–0.78, P = 0.003), 0.56 (95% CI 0.34–0.92, P = 0.022), and 0.37 (95% CI 0.22–0.64, P < 0.001), respectively. Figure 5 shows the forest plot of the subgroup analysis for the risk of experiencing MACEs by serum sodium concentrations. The interaction test did not reveal any significant interactions between serum sodium concentrations and subgroups defined by sex, diabetes status, hypertension status, or type of CHD, suggesting that the effect of serum sodium on the risk of experiencing MACEs was consistent across these subgroups.

Table 2 Association between serum sodium levels and the risk of experiencing MACEs among patients with newly diagnosed CAD.
Full size table
Figure 5
figure 5

Associations between serum sodium concentrations and the risk of experiencing MACEs among patients with newly diagnosed CAD in different subgroups. Each subgroup was adjusted for age, sex, diabetes status, hypertension status, type of CAD (angina or AMI), Gensini score, treatment (medical therapy alone, medical therapy + percutaneous coronary intervention, or medical therapy + coronary artery bypass grafting), eGFR (estimated glomerular filtration rate), LDL-C (low-density lipoprotein cholesterol) and triglycerides.

Full size image

Improvement of cardiovascular risk prediction

The predictive value of serum sodium concentrations for cardiovascular risk was further assessed using the C-statistic, net reclassification index (NRI), and integrated discrimination improvement (IDI). The receiver operating characteristic (ROC) curves for the MACE prediction models with and without serum sodium concentrations are shown in Fig. 6A. Incorporating serum sodium concentrations into the model significantly improved the C-statistic from 0.647 (95% CI 0.598–0.696) to 0.679 (95% CI 0.632–0.726) (P = 0.022). Additionally, significant improvements were observed in the NRI (0.338, P < 0.001) and IDI (0.026, P < 0.001), as shown in Fig. 6B.

Figure 6
figure 6

Improvement in MACE risk reclassification and discrimination with the addition of serum sodium concentrations to the traditional model. The traditional model (basic model) included age, sex, diabetes status, hypertension status, type of CAD (angina or AMI), Gensini score, treatment (medical therapy alone, medical therapy + percutaneous coronary intervention, or medical therapy + coronary artery bypass grafting), eGFR (estimated glomerular filtration rate), LDL-C (low-density lipoprotein cholesterol) and triglycerides. Improvement was assessed by changes in the C-statistic, net reclassification improvement (NRI), and integrated discrimination improvement (IDI).

Full size image

Nonlinear relationship between serum sodium concentrations and MACE risk

Finally, the RCS curve depicted in Fig. 7A demonstrates a nonlinear relationship between serum sodium concentrations and the risk of experiencing MACEs (P-nonlinear = 0.014). Therefore, we further explored the trend of risk changes within different ranges of serum sodium concentrations by performing a segmented Cox regression analysis using the value with the highest risk as the cut-off. The hazard ratio (HR) for the risk of experiencing MACEs increased with increasing serum sodium concentrations up to approximately 138 mmol/L, beyond which the risk decreased after adjusting for the same covariates as in Model 3 (Fig. 7B). Within the 130–138 mmol/L sodium range, MACE risk gradually increased with higher sodium levels (HR 1.39, 95% CI 1.09–1.76, P = 0.008); whereas within the 138–147 mmol/L range, the risk gradually decreased (HR 0.88, 95% CI 0.80–0.98, P = 0.014).

Figure 7
figure 7

(A) Restricted cubic spline curve for the association between the risk of experiencing MACEs and serum sodium concentrations; (B) Hazard ratios for the risk of experiencing MACEs with serum sodium as a continuous variable (using 138 mmol/L as a cut-off point). Model 1: adjusted for age, sex. Model 2: adjusted for Model 1 + diabetes, hypertension, types of CHD (angina or AMI), and severity of coronary artery lesion (Gensini score). Model 3: adjusted for Model 2 + treatment (medical therapy alone, medical therapy + percutaneous coronary intervention, or medical therapy + coronary artery bypass grafting), the eGFR (estimated glomerular filtration rate), LDL-C (low-density lipoprotein cholesterol) and triglycerides.

Full size image

Discussion

To our knowledge, the present study is the first in which the association between baseline serum sodium concentrations and long-term disease prognosis was investigated in patients with newly diagnosed CHD who do not have concomitant heart failure. The results of this study indicated that the concentration of serum sodium is significantly associated with disease prognosis among patients who had not received systematic treatment for CHD. Specifically, patients with serum sodium concentrations in the range of 130–135.9 mmol/L or 139–147 mmol/L had a lower incidence of MACE than those with relatively low (but normal) serum sodium concentrations (136.0–138.9 mmol/L). Notably, RCS analysis revealed an inverted U-shaped relationship, where the risk of experiencing MACEs peaked at approximately 138 mmol/L and declined at both lower and higher sodium concentrations.

Serum sodium concentrations are tightly regulated to maintain fluid and electrolyte balance6,29,30,31. Sodium concentrations are influenced by dietary intake and complex metabolic systems and play a crucial role in extracellular fluid osmolality7,32,33,34. Inappropriate serum sodium concentrations can activate the renin–angiotensin–aldosterone system (RAAS), which, while helping to maintain blood pressure, may exert harmful effects on the cardiovascular system7,9,35. RAAS activation is associated with vasoconstriction, increased blood pressure, and inflammatory processes that can exacerbate atherosclerosis and overall cardiovascular health10,32,35. This association between serum sodium concentrations and the risk of developing CHD is likely due to the combined effects of RAAS activation, fluid imbalance, and other metabolic disturbances7,36,37.

The optimal management of sodium metabolism for cardiovascular health, including sodium intake strategies and management of serum sodium concentrations, remains contentious36,38,39,40. Although current medical guidelines generally recommend low sodium intake for most cardiovascular patients10,36,37,41, these recommendations are based on expert consensus rather than robust clinical trial evidence22. Clinical studies of heart failure patients have provided mixed results22, with epidemiological data and clinical trials of varying designs highlighting the beneficial40,42,43, neutral13, or potentially harmful effects of low sodium intake21,39,44. For instance, the SODIUM-HF trial24, a high-quality randomized controlled trial, did not reveal benefits of sodium restriction for heart failure patients. Recently, the Heart Failure Association of the ESC emphasized individualized treatment for different patients rather than a one-size-fits-all approach to sodium restriction22. High-quality research on sodium management specifically for CHD patients without heart failure is scarce, highlighting the urgent need for comprehensive studies. Many studies on sodium intake lack data on serum sodium concentrations12,38,40,45, and the effects of strict sodium restriction on CHD patients with low or low-normal sodium concentrations remain unknown.

Our study revealed that within the normal range, lower serum sodium concentrations are associated with poorer prognosis. This finding aligns with previous studies showing that lower serum sodium is linked to greater cardiovascular mortality in various populations, including those without cardiovascular disease16,19,20,39,45, heart failure patients23,24,25,32,44, and patients with suspected or confirmed CHD14,46,47. Conversely, one study revealed that higher sodium concentrations within the normal range increased the risk of heart failure15, but this study focused on a different population (individuals without diabetes or heart failure aged 45–66 years) and investigated heart failure only as an outcome.

Interestingly, our RCS analysis found an inverted U-shaped relationship between serum sodium concentrations and prognosis in the RCS analysis. Despite having the most severe coronary stenosis, as indicated by the highest Gensini scores18, patients in the mildly hyponatraemic group (L1) had a lower risk of experiencing MACEs than did those in the L2 group. Although the L1 group had the lowest overall incidence of MACE, they had the highest rates of cardiovascular death and myocardial infarction. This suggests that the lower overall MACE rate in the L1 group may be due to a reduction in less severe events such as readmission for heart failure or angina, likely attributable to several factors. First, the impact of serum sodium concentrations on CHD prognosis is complex, and no single outcome is definitive, which is why further research is necessary32,40. Second, an individual's sodium metabolism is dynamic20,22, complicating the situation further, as baseline sodium concentrations may not reflect long-term concentrations. Third, patients with hyponatraemia are likely to receive more attention from healthcare providers44 because of the need for better management of electrolytes and more thoughtful treatment plans, which can improve outcomes32. Additionally, patients in the L1 group, due to their severe condition and higher rates of PCI and CABG, were likely more compliant with medical advice, including taking medications on time, attending follow-up appointments, quitting smoking and drinking, and maintaining a healthy diet, which contributed to their lower risk of non-severe MACE41,48.

We recommend more research on long-term monitoring of serum sodium levels, as these levels can easily fluctuate with changes in the body's condition. For instance, endocrine disorders and diarrhea can significantly impact serum sodium levels32,49. Conditions such as adrenal insufficiency, hypothyroidism, and hyperthyroidism can disrupt sodium homeostasis, leading to hyponatremia or hypernatremia, thus influencing cardiovascular risk31. Diarrhea, through increased gastrointestinal losses, can cause electrolyte imbalances, including hyponatremia, which may affect cardiovascular outcomes49. Although we excluded patients with known endocrine disorders at baseline, these conditions could have developed during follow-up. This variability underscores the importance of continuous sodium monitoring and comprehensive management of underlying conditions to better understand their effects on cardiovascular outcomes in CHD patients.

Although the exact mechanisms underlying this association remain unclear, our study has potential significant implications. Serum sodium concentrations are often overlooked in CHD patients. We found that baseline serum sodium concentrations are clearly associated with disease prognosis for CHD patients without heart failure and that incorporating sodium concentrations into predictive models can significantly enhance risk stratification accuracy. Clinicians and researchers should give attention to CHD patients with low-normal sodium concentrations, as these concentrations are often ignored. Future research should focus on the relationship between long-term serum sodium concentrations and the risk of experiencing CHD outcomes, as well as the effects of sodium intake, to develop personalized health management strategies for CHD patients and improve their prognosis.

Strengths and limitations

This study has several strengths. First, the prospective cohort design allowed us to establish a reliable relationship between the baseline serum sodium concentration and long-term cardiovascular outcomes. As one of the first studies to investigate these associations in newly diagnosed CHD patients without heart failure, this study covers a significant gap in the literature. The use of advanced statistical techniques, such as Cox proportional hazards models, RCS analysis, and improvements in predictive models through the C-statistic, NRI, and IDI, adds rigor and depth to the findings. Importantly, the identification of an inverted-U-shaped relationship between serum sodium concentrations and the risk of experiencing MACEs provides new insights into sodium management in CHD patients, potentially paving the way for more extensive research interest in this area.

This study also had several limitations. First, the subjects were from a single centre, which may have resulted in selection bias. We did not measure urinary sodium concentrations or sodium intake, limiting our ability to assess the impact of sodium intake. We also focused solely on baseline sodium concentrations without multiple or long-term sodium measurements. Although we excluded patients with heart failure at baseline and adjusted for hypertension status and other conditions that may require diuretics or RAAS inhibitors, we cannot completely rule out the potential impact of the use of these medications during follow-up. However, the proportion of patients using them may be very small. Unfortunately, we do not have accurate and detailed data on the exact rates of diuretics and RAAS inhibitors usage during the follow-up period, which represents a limitation of this study. Therefore, the clinical significance of the findings in the present study requires further investigation.

Conclusion

In this study, we demonstrated that baseline serum sodium concentrations exhibit an inverted U-shaped relationship with long-term cardiovascular risk in newly diagnosed CHD patients without heart failure. However, low sodium levels may increase the risk of severe events such as cardiovascular death and myocardial infarction. Furthermore, incorporating serum sodium concentrations into predictive models significantly improves their accuracy. These findings highlight the potential importance of considering baseline serum sodium concentrations for risk stratification in newly diagnosed CHD patients. Further research is warranted to explore the underlying mechanisms of this relationship and the impact of long-term serum sodium concentrations and sodium intake on prognosis to establish optimal sodium management strategies for this patient population.