Stable Stratified Correlation and Linear Negative Correlation Between Serum Albumin and Risk of All-Cause Death in Patients with Ischaemic Stroke

Abstract

Aims

To assess the correlation between serum albumin and all-cause death in patients with ischaemic stroke.

Methods

In this prospective cohort study, a total of 1000 patients with ischemic stroke were included. Multivariate Cox regression analysis, subgroup analysis, receiver operating characteristic (ROC) curve, Kaplan–Meier survival analysis, and restricted cubic spline (RCS) plot were used to assess the association between serum albumin and all-cause death in patients with ischaemic stroke.

Results

A total of 272 patients with ischemic stroke died during a median follow-up of 25 months. Multivariate Cox regression analysis showed that the risk of all-cause death in patients with albumin at T2 (40–43 g/L) and T3 (>43 g/L) was 33.6% and 74.4% lower than that in patients with T1 (<40 g/L), respectively, and the risk of all-cause death decreased by 13.2% for every unit increase in albumin (P < 0.001). Subgroup analysis showed that the correlation between higher levels of albumin and lower risk of all-cause death was significant in all subgroups (P < 0.05). ROC analysis indicated that albumin could predict all-cause death in both the overall population and different sex groups (P < 0.001). Kaplan–Meier survival analysis revealed that with the extension of follow-up time, the higher the albumin, the slower the survival probability decreased (P < 0.001). RCS also established a linear negative correlation between albumin and the risk of all-cause death (P for nonlinearity = 0.092).

Conclusion

In patients with ischemic stroke, there was a stable stratified correlation and linear negative correlation between serum albumin and all-cause death risk.

1 Introduction

The Global Burden of Disease (GBD) study showed that the incidence, prevalence, and mortality of stroke have increased by 70.0%, 85.0%, and 43.0%, respectively, over the past 30 years, of which in 2019, the incidence, prevalence and mortality of stroke were as high as 12 million, 101 million and 6.55 million, respectively, indicating that stroke continues to be one of the most important causes of disability and mortality worldwide and is one of the most important public health problems at present [1]. Furthermore, this GBD study not only revealed that the incidence of ischaemic stroke was as high as 62.4% of total stroke events in 2019, but also confirmed that high systolic blood pressure (SBP), high body mass index (BMI), high fasting plasma glucose (FPG), ambient particulate pollution, and smoking were the five most important risk factors for stroke [1]. Another GBD study for China reported 3.94 million new stroke cases and 2.19 million deaths in China in 2019, an increase of 86.0% and 32.3%, respectively, compared with 30 years ago, with a whopping 24.18 million ischaemic strokes prevalent in 2019, and they also found that high SBP, exposure to ambient particulate matter pollution, smoking, and a high-nutrient diet were the main four attributing factors for the disease burden of stroke [2]. A GBD study in 2022 demonstrated that despite a decline in age-standardized prevalence and mortality rates of ischaemic stroke globally over the past 30 years, age-standardized incidence rate has continued to increase, particularly in women over 50 years of age [3]. In addition, a recent GBD study in a younger age group demonstrated that global age-standardised prevalence and mortality rates of ischaemic stroke in young adults have declined over the past 30 years, while the age-standardised prevalence rate of ischaemic stroke in 2019 was still higher in young women than in young men [4]. These data indicate that the burden of stroke and ischaemic stroke remains high both globally and in China, and that timely prevention and treatment strategies are needed to reduce the high incidence, prevalence and mortality of stroke.

Current evidence suggests that serum albumin is the most abundant essential protein in the circulation, which can combine and transport some drugs and substances, and maintain blood osmotic pressure and normal physiological functions of human body [5]. Serum albumin, as a protein reflecting the nutritional status of the body, not only has anti-inflammatory, anti-thrombotic and antioxidant properties, but has also been shown to independently predict the risk of death in several patient groups [5,6,7,8,9]. However, the role of serum albumin in predicting the risk of all-cause death in patients with ischaemic stroke requires further discussion. In this study, to explore the predictive role of serum albumin on the prognostic risk of cerebrovascular disease, we aimed to assess the correlation between serum albumin and the risk of all-cause death in patients with ischaemic stroke, as well as potential nonlinear association and stratified association.

2 Methods

2.1 Study Population

In this prospective cohort study, all patients were from ischaemic stroke patients who were admitted to the Huai’an Huai’an Hospital, from December 2017 to December 2020. Inclusion criteria: (1) diagnosis of definite ischaemic stroke; (2) voluntary participation in this study. Exclusion criteria: (1) age < 40 years; (2) lack of serum albumin data at baseline; (3) malignant tumour; (4) severe haematological disease; (5) hepatic and renal failure; (6) cardiorespiratory failure; (7) mental retardation or severe psychiatric disorders; and (8) loss of follow-up. A total of 1000 patients were selected for entry into this study after enforcing the inclusion and exclusion criteria. The study protocol adhered to the Declaration of Helsinki and was approved by the hospital ethics committee, and informed consent was obtained from all patients for this study.

2.2 Data Collection and Definitions

In this study, we extracted the required data from the hospital electronic medical record system, including demographic data, comorbidity and medication data, and biomarker data, according to the design of the study protocol. Smoking was defined as having smoked often in the past, regardless of current smoking status. The standards of the International Society of Hypertension (ISH) were used for the diagnosis of hypertension [SBP ≥ 140 mmHg or diastolic blood pressure (DBP) ≥ 90 mmHg] [10]. Blood pressure was measured at least three times at different time points, and the average was taken. Measurements were taken with the patient in a seated position, ensuring they had rested for at least 5 min before measurement. Follow-up visits were conducted after discharge for newly diagnosed cases to confirm whether target blood pressure was achieved. The standards recommended by the American Diabetes Association (ADA) were used for the diagnosis of diabetes [11]. For patients with abnormal blood glucose levels, multiple FPG tests (FPG ≥ 7.0 mmol/L) and oral glucose tolerance tests (OGTT) (2-h postprandial blood glucose ≥ 11.1 mmol/L) were conducted, combined with glycated haemoglobin (HbA1c) levels (HbA1c ≥ 6.5%) for comprehensive assessment. For patients who received cortisone treatment during hospitalization, their medical history was recorded in detail, and their blood glucose levels were re-evaluated after treatment. For patients requiring short-term insulin treatment, follow-up visits were conducted after discharge to confirm whether their blood glucose levels had returned to normal. Hyperlipidaemia was defined as fasting total cholesterol ≥ 5.72 mmol/L or triglyceride ≥ 1.70 mmol/L [12]. Ischaemic stroke was defined as a past diagnosis of ischaemic stroke by a physician or a new diagnosis of ischaemic stroke during this hospitalisation. Data on drug therapy included antihypertensive, antihyperglycaemic and lipid-lowering drugs. The values of height, weight, SBP and DBP were measured by professional nurses according to standard procedures when patients were admitted to hospital. The body mass index (BMI) = weight (kg)/height (m²). The blood concentrations of blood markers were extracted from patients’ peripheral veins by professional nurses according to standard procedures and sent to the laboratory for determination by laboratory professionals, including triglyceride, total cholesterol, low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), blood urea nitrogen, creatinine, uric acid, FPG, HbA1c, C-reactive protein (CRP) and serum albumin. Furthermore, estimated glomerular filtration rate (eGFR) is an important indicator for assessing kidney function. For the Chinese population, commonly used eGFR formulas are based on serum creatinine levels, combined with age and gender [13]. In this study, all patients were divided into three groups based on tertiles of serum albumin: T1 (<40 g/L), T2 (40–43 g/L) and T3 (>43 g/L).

In this study, all patients were followed from the time of discharge until death or December 2023. Follow-up methods included telephone follow-up, outpatient follow-up, or inpatient chart follow-up, which included events and their time points (in months). During follow-up, two persons unknown to the study protocol performed the follow-up and were checked by a third researcher. All-cause death was used as the endpoint event of the study and included both cardiovascular and cerebrovascular deaths and non-cardiovascular and cerebrovascular deaths.

2.3 Statistical Analysis

Categorical variables were expressed as frequencies (percentages) and differences between groups were assessed by chi-square test or Fisher’s exact test. Normally distributed continuous variables were expressed as mean ± standard deviation and differences between groups were assessed by one-way ANOVA, while non-normally distributed continuous variables were expressed as median (first quartile, third quartile) and differences between groups were assessed by the non-parametric test (Kruskal–Wallis H test). Cox regression analysis was used to assess the independent correlation between serum albumin and the risk of all-cause death in patients with ischaemic stroke. Stratified association between serum albumin and risk of all-cause death was assessed by dividing patients into 10 subgroups according to age, sex, hypertension, diabetes and hyperlipidemia. The receiver operating characteristic (ROC) curve analysis was used to assess the diagnostic ability of serum albumin for all-cause death in patients with ischaemic stroke. The Kaplan–Meier survival analysis was used to evaluate the difference of survival probability among the three groups of serum albumin with follow-up time. The restricted cubic spline (RCS) plot analysis was used to assess the nonlinear relationship between serum albumin and the risk of all-cause death in patients with ischaemic stroke. All statistical analyses were performed using SPSS 26.0 and R 4.1.0, where P < 0.05 was defined as statistically significant.

3 Results

3.1 Baseline Characteristics

Table 1 shows that age, sex, smoking, diabetes, antihyperglycaemic drugs, lipid-lowering drugs, BMI, SBP, DBP, triglyceride, total cholesterol, LDL-C, blood urea nitrogen, eGFR, HbA1c, CRP and the risk of all-cause death differed significantly among the three groups of serum albumin (P < 0.05), with a gradual decrease in the incidence of all-cause death as serum albumin increased (P < 0.001).

Table 1 Baseline characteristics of participants stratified by level of serum albumin concentration

3.2 Association of Albumin with All-Cause Death

A total of 272 patients with ischemic stroke died during a median follow-up of 25 months. Multivariate Cox regression analysis showed that after adjusting for age, sex, hypertension, antihypertensive drugs, diabetes, antihyperglycemic drugs, hyperlipidaemia, lipid-lowering drugs, BMI, SBP, DBP, triglyceride, total cholesterol, HDL-C, HbA1c, eGFR and CRP, the risk of all-cause death in patients with albumin at T2 and T3 was 33.6% and 74.4% lower than that in patients with T1, respectively (HR: 0.664, 95% CI: 0.514–0.858, P = 0.002; HR: 0.256, 95% CI: 0.165–0.398, P < 0.001), and the risk of all-cause death decreased by 13.2% for every unit increase in albumin (HR: 0.868, 95% CI: 0.839–0.899, P < 0.001) (Table 2). Subgroup analysis showed that the correlation between higher levels of albumin and lower risk of all-cause death was significant in all subgroups (P < 0.05) (Table 3). ROC analysis indicated that albumin could predict all-cause death in both the overall population and different sex groups (Total population, cut-off value = 42 g/L, unadjusted AUC = 0.698, P < 0.001; male, cut-off value = 41 g/L, unadjusted AUC = 0.710, P < 0.001; female, cut-off value = 42 g/L, unadjusted AUC = 0.678, P < 0.001) (Fig. 1). Kaplan–Meier survival analysis revealed that with the extension of follow-up time, the higher the albumin, the slower the survival probability decreased (P < 0.001) (Fig. 2). RCS also established a linear negative correlation between albumin and the risk of all-cause death (P for nonlinearity = 0.092) (Fig. 3).

Table 2 Association of albumin with all-cause death

Table 3 Stratified correlation between albumin and all-cause death

Fig. 1

Diagnostic performance of albumin on the risk of all-cause mortality. A Total population, B male, C female. AUC, area under the curve

Fig. 2

Kaplan–Meier survival curve adjusted for other covariates

Fig. 3

Restricted cubic spline for evaluating the nonlinearity of albumin and all-cause death. HR, hazard ratio; CI, confidence interval

4 Discussion

In this inpatient-based cohort study, we not only confirmed an independent stratified correlation between serum albumin and the risk of all-cause death in stroke patients, but also identified a linear negative correlation between them, which reflects the importance of nutritional status in the prognostic management of stroke patients and suggests that we should incorporate monitoring of nutritional indices into the routine screening of stroke patients.

Current evidence from epidemiological and clinical studies suggests that lower serum albumin concentrations are strongly associated with a higher risk of ischaemic heart disease, atrial fibrillation, heart failure, thromboembolism, and ischaemic stroke, independent of other traditional risk factors [14]. In addition, there is some evidence that lower serum albumin concentrations are associated with a worse prognosis for a variety of diseases [15]. For example, Jin et al. found that lower levels of serum albumin were strongly associated with a higher risk of all-cause mortality in a large prospective cohort study of 1834 older adults (>65 years), and they also revealed a negative linear correlation between serum albumin and the risk of all-cause mortality [16]. In addition, Shannon et al. in a community-based atherosclerosis risk study of 4947 community residents showed that a 1 g/dL reduction in baseline serum albumin was associated with a 67% increase in the risk of death [7]. Frenkel et al. in a retrospective analysis of 3967 patients with sepsis found that serum albumin levels 1 week after hospital admission were inversely associated with the risk of in-hospital death, and that the younger the age, the stronger the correlation, but no significant correlation was found between serum albumin on admission and the risk of in-hospital death [17]. And in a large cohort study involving 77,531 community-dwelling older adults, individuals with albumin levels of 4.2–4.3 g/dL had a 1.16 times greater risk of death than those with albumin levels of ≥4.4 g/dL, and these older adults had a higher risk of all-cause death, cancer death, and cardiovascular death when albumin levels were below 4.2 g/dL [6]. Furthermore, in a large meta-analysis and systematic review that included eight studies and 21,667 patients with acute coronary syndromes, Zhu et al. demonstrated that lower serum albumin levels were independently associated with a higher risk of all-cause mortality in patients with acute coronary syndromes, independent of traditional risk factors [18]. Moreover, Nouraie et al. in adult sickle cell patients also demonstrated that serum albumin was an independent predictor of disease severity and death in patients with sickle cell disease [19]. Although the above studies confirmed the correlation between serum albumin and all-cause death in the general population or in some disease populations, the correlation with the risk of all-cause death in patients with ischaemic stroke is still unclear. After analysing data from the Third China National Stroke Registry, Wang et al. found that lower serum albumin-to-globulin ratios were strongly associated with the risk of all-cause mortality at 3-month and 1-year follow-up in patients with acute ischaemic stroke, independently of other confounders [20]. In a cohort study of 605 patients with ischemic stroke treated by intravascular thrombectomy, Gao et al. also found that lower serum albumin level was independently related to the poor prognosis and death risk at 90 days after acute great vascular occlusion stroke in anterior circulation treated by intravascular thrombectomy [21]. In addition, in a study that included cohort analysis and meta-analysis, Zhou et al. found that low serum albumin levels in the cohort analysis were closely related to the 3-month and 1-year risk of all-cause death in patients with acute ischemic stroke or transient ischemic attack, and meta-analysis confirmed this correlation, that is, for every 1 g/L reduction in serum albumin, the risk of all-cause death increased by 3–7% [22]. Although these studies confirmed the correlation between serum albumin and the risk of all-cause death in patients with ischaemic stroke, they did not further explore the stratified correlation and nonlinear correlation between serum albumin and the risk of death. In contrast, our study not only revealed the correlation between lower levels of serum albumin and higher risk of all-cause death in patients with ischaemic stroke, but also further confirmed the stability of the relationship between albumin and the risk of death in the subgroups of <50 years old, ≥50 years old, males, females, without hypertension, with hypertension, without diabetes, with diabetes, without hyperlipidaemia, and with hyperlipidaemia. In addition, we also found a negative linear association between serum albumin and the risk of all-cause death in patients with ischaemic stroke.

Despite the favourable results obtained in this study, the mechanisms involved need to be further explored. Since serum albumin can carry a variety of exogenous and endogenous substances (such as inorganic ions, bilirubin, vitamins, and hormones), it has a variety of physiological properties [14]. There is some evidence that serum albumin has antioxidant properties due to its abundant sulfhydryl groups, which scavenge excess reactive oxygen species [23,24,25]. There is also evidence that serum albumin has anti-platelet aggregation, anticoagulant, and vascular osmotic pressure maintenance properties [26, 27]. In addition, serum albumin’s greatest property is that it serves as a biomarker of the body’s nutritional status, and hypo-serum albuminaemia often signifies malnutrition or a severe disease state that promotes the occurrence of a number of adverse events. The above data suggest that low levels of serum albumin may promote adverse events through oxidative stress and thrombosis and enhancement and vascular osmotic pressure imbalance, and malnutrition may aggravate this process.

In addition, there were several limitations of this study that need to be discussed. First, this study only assessed serum albumin as a nutritional indicator and did not comprehensively assess the correlation between other nutritional indicators and the risk of death in stroke patients, so the results might be limited. Second, although this was a longitudinal cohort study, a causal association between serum albumin and risk of death could not be determined because of the lack of genetic data. Third, the study population was limited to Chinese inpatients, so the results need to be validated in different countries, regions, or clinical centres before they can be further disseminated. Fourth, in this study, comorbidities were very prevalent in the albumin T1 group, including diabetes and hypertension, and the proportion of females in the T1 group was significantly higher compared to other groups (33.5% in T1 vs. only 15.9% in T3). Generally, high comorbidity scores are associated with higher all-cause mortality. Therefore, the high comorbidity score in the T1 group might significantly affect the study results, which could partly explain the relationship between albumin levels and all-cause mortality in the T1 group. Hence, these comorbidities need to be considered when interpreting our results. Additionally, the cut-off value of albumin might differ between sexes, suggesting that the threshold for predicting mortality with albumin could be different for males and females [28]. However, in our subgroup analysis, we validated the correlation between albumin and all-cause mortality separately for different sexes. Future research should further explore the biological effects of sex on the impact of albumin. Finally, this study confirmed the association between serum albumin and the risk of death, but did not explore the association between albumin and other adverse prognoses.

5 Conclusions

In conclusion, this study revealed the correlation between serum albumin and the risk of all-cause death in stroke patients, which provides some clues and theoretical basis for enhancing the monitoring of individual nutritional status in the prognostic management of stroke patients.