Correlation Between Prognostic Nutritional Index and Heart Failure in Adults with Diabetes in the United States: Study Results from NHANES (1999–2016)

Abstract

Background:

The relationship between diabetes and heart failure significantly impacts public health. This study assessed the prognostic nutritional index (PNI) as a predictor of heart failure risk in adult diabetic patients.

Methods:

An analysis was performed on 1823 diabetic adults using data collected from the National Health and Nutrition Examination Survey (NHANES) between 1999 and 2016. Serum albumin levels and lymphocyte counts were combined to calculate the PNI. We used descriptive statistics categorized by PNI quartiles and performed multivariate logistic regression to adjust for variables including age, gender, ethnicity, and coexisting medical conditions.

Results:

The median age (mean ± SD) was 59.942 ± 12.171 years, and the mean value ± SD of the PNI was 52.412 ± 5.430. The prevalence of heart failure was 7.405%. In the fully adjusted model, for each 1-unit increase in PNI, the risk of heart failure decreased by 8.2% (odds ratio (OR), 0.918; 95% confidence interval (CI) 0.884, 0.953). Participants in the highest PNI quartile (Q4) had a 63% reduced risk of heart failure compared to those in the lowest quartile (Q1). Tests for interactions did not reveal any statistically significant differences among these stratified subgroups (p for interaction > 0.05).

Conclusions:

This study demonstrated that a higher PNI was significantly associated with a decreased prevalence of heart failure in adults with diabetes.

Graphical abstract

Keywords

prognostic nutritional index / diabetes / heart failure / NHANES

1. Introduction

Diabetes, a prevalent chronic illness, is an endocrine condition defined by elevated blood glucose levels [1]. The main types include type 1 diabetes mellitus (T1DM), type 2 diabetes mellitus (T2DM), and gestational diabetes [2, 3], with T2DM accounting for over 90% of all cases [4]. Diabetes poses a significant threat to human health [5], as per data from the International Diabetes Federation (IDF), with approximately 537 million people worldwide receiving a diagnosis of diabetes in 2021. The projected estimate for 2030 is 643 million, and by 2045, it could increase to 783 million, thus potentially burdening global public health systems [6]. The numerous complications associated with diabetes, including cardiovascular diseases, retinopathy, kidney diseases, and neurological disorders, significantly increase the level of difficulty in patient care and the potential for diabetes-lined mortality [7]. Heart failure has become a significantly serious public health problem, and its incidence is increasing, resulting in considerable expenses related to hospitalization on a worldwide scale [8]. Research currently indicates a significant association between heart failure and diabetes, whereby patients diagnosed with diabetes are at a much greater risk of developing heart failure than non-diabetics. Moreover, there is a noticeable relationship between heart failure and an increase in the incidence of newly diagnosed diabetes [9]. Although the molecular mechanisms connecting these conditions have not been fully elucidated [10], it is evident that heart failure increases the mortality rate among diabetic patients, significantly affecting their quality of life and prognosis [11].

Buzby et al. [12] proposed the prognostic nutritional index (PNI), which utilizes serum albumin levels along with peripheral blood lymphocyte counts to evaluate an individual’s nutritional status and immune function [13]. The simplicity and objectivity of this method have led to a surge in research on the PNI, with it now commonly employed to evaluate preoperative nutritional status, postoperative outcomes, and levels of inflammation [14, 15]. Proper nutrition is essential in preventing and treating long-term health conditions such as diabetes [16]. Adequate dietary control and nutritional supervision may enhance the outlook for individuals with diabetes. A study has demonstrated that the PNI serves as a stand-alone predictor of unfavorable results in individuals with cardiac issues [17], and its prognostic precision exceeds that of albumin or lymphocyte counts. Therefore, PNI provides a new monitoring index for clinical research with potential clinical applications.

There is currently a paucity of research on heart failure and the PNI in diabetic adults. This research aimed to explore the relationship among these factors in adult individuals with diabetes by examining the National Health and Nutrition Examination Survey (NHANES) dataset. This study intended to assess the effectiveness of PNI as a predictive tool, providing clinical evidence that may help improve the prognosis and treatment management of diabetic patients.

2. Methods

2.1 Data Source and Participants

The NHANES dataset used in this study was derived from a nationwide cross-sectional investigation by the National Center for Health Statistics (NCHS). A stratified multistage probability sampling technique was utilized to acquire a representative sample of the non-institutionalized civilian U.S. population. The NCHS Research Ethics Review Board approved the research activities, and all participants provided their informed consent prior to the study. The data collected included demographic information, physical exams, lab tests, health questionnaires, and prescribed medication records, all managed using a sophisticated computer system. This study analyzed the NHANES dataset from 1999 to 2016, which included 92,062 participants. In the data screening process, cases under the age of 20 and those with missing data on diabetes, PNI-related indicators, heart failure, and other relevant covariates were excluded. Ultimately, 1823 adults diagnosed with diabetes were included (Fig. 1). The diagnosis of diabetes was derived from questionnaires and laboratory tests. Participants were eligible if they answered affirmatively to any of the following inquiries: “Have you been diagnosed with diabetes by a medical professional?”, “Do you currently take insulin?”, or “Do you currently take medication for blood sugar control?”, or if their lab results met the diagnostic criteria for diabetes, which included glycated hemoglobin (%) \ge6.5% or a fasting blood glucose (mg/dL) \ge126 mg/dL. This approach integrated self-reported and biomarker information to improve the precision and dependability for the detection and diagnosis of diabetes.

2.2 Exposure and Outcomes

The PNI was utilized as the independent variable in this investigation, using the formula PNI = 5 \times lymphocyte count (measured in units of 10⁹/L) + serum albumin levels (measured in units of g/L) [18]. Lymphocyte counts were derived from a complete blood count (CBC), utilizing Beckman Coulter technology to analyze cell count and size [19]. This technology is broadly acknowledged for assessing dietary nutritional status and immune function. Serum albumin levels, indicative of nutritional status, were measured utilizing the bromocresol purple dye-binding technique as documented in the NHANES database [20]. Higher PNI values typically indicate better nutritional status. The outcome variable was defined as heart failure. Data on heart failure in the NHANES study were mainly gathered via personal interviews that relied on self-reporting. This study used the “MCQ160B” variable from the NHANES questionnaire to diagnose heart failure. This variable specifically asks participants: “Has a doctor or other health professional ever told you that you have congestive heart failure?”. Participants who confirmed receiving information about their heart failure diagnosis from a healthcare professional (such as a doctor) were classified as having heart failure. Although the NHANES dataset lacks key diagnostic markers (such as B-type natriuretic peptide (BNP), N-terminal pro B-type natriuretic peptide (NT-proBNP), troponin, electrocardiograms (EKGs), and cardiac imaging), and reliance on questionnaire information may introduce some ambiguity, previous studies have shown that self-reported data are valid for diagnosing heart failure among NHANES participants [21, 22, 23]. Across different racial and age groups, self-reported data have previously effectively described overall trends and racial differences [24]. Additionally, A study has shown that while self-reported heart failure data may have lower sensitivity, they exhibit high specificity (96–97%) and have significant application value in large-scale epidemiological studies [25].

2.3 Covariables

This study examined various covariates, encompassing demographic information including gender, age, race, marital status, educational attainment, and income level. In addition, it integrated medical history co-morbidities such as hypertension, stroke, coronary heart disease, angina, and myocardial infarction. Specific survey questionnaires were used to collect lifestyle characteristics; within this framework, individuals classified as smokers were those who had consumed a minimum of 100 cigarettes throughout their lifetime [26]. Similarly, individuals who had consumed alcohol on at least 12 distinct occasions were classified as drinkers [27]. NHANES grouped race and ethnicity by the responses provided to survey questions. Categories included Mexican American, non-Hispanic white, non-Hispanic black, and other ethnicities. Marital status was divided into two groups: unmarried (comprising never married, divorced, separated, and widowed) and married (including married and cohabiting individuals). Education level was categorized as less than high school, high school, and more than high school.

Blood pressure and body mass index (BMI) were assessed through laboratory examinations, with BMI divided into three groups: normal (BMI ranging from 18.5 to 25 kg/m²), overweight (BMI between 25 and 30 kg/m²), and obese (BMI exceeding 30 kg/m²) [28].

Details and information regarding these covariates can be accessed on the official website of the Centers for Disease Control and Prevention at https://www.cdc.gov/nchs/nhanes/.

2.4 Missing Covariables

In handling missing covariables in the study data, this research adopted the strategy of directly deleting samples containing missing data [29]. This approach offers several significant advantages. First, it avoids the potential biases introduced by data imputation, thus ensuring the accuracy and reliability of the analysis results. Second, this method simplifies the data processing workflow, eliminating the need for complex statistical imputation techniques and thereby enhancing the robustness of the study findings. Moreover, removing samples with missing data reduces the errors that incorrect data imputation could cause, making the study results more credible. This approach is particularly helpful in maintaining the overall quality of the research, especially when addressing issues with missing data, which are difficult to interpret.

2.5 Statistical Analysis

All data analyses followed the protocols outlined by the Centers for Disease Control and Prevention (CDC), available at https://wwwn.cdc.gov/nchs/nhanes/tutorials/default.aspx. The research included descriptive analyses of the data from all participants. Percentages were used to represent categorical variables, while distribution properties detailed continuous variables. This analysis was conducted using either the mean and standard deviation (SD) or the median and interquartile range (IQR) [30].

Continuous variables were analyzed using the Student’s t-test to evaluate differences in clinical features, and categorical variables were assessed with the chi-square test. This approach identified notable differences among variables, providing a methodologically sound basis for subsequent analysis. This study used logistic regression models with weights to assess the relationship between the continuous PNI and its quartiles, alongside the risk of developing heart failure. The study calculated odds ratios (ORs) and 95% confidence intervals (CIs) using three distinct regression models: The initial model, Model 1, incorporated only the PNI variable. Model 2 accounted for additional variables such as gender, age (as a continuous variable), race (classified as Mexican American, non-Hispanic white, non-Hispanic black, and other), and body mass index (BMI categorized as normal (18.5 BMI 25 kg/m²), overweight (25 \le BMI \le 30 kg/m²), and obese (BMI > 30 kg/m²)). Model 3 included further refinements to adjust for education level (less than high school, high school, more than high school), marital status (unmarried, married), income relative to the poverty threshold (continuous variable), smoking status (yes, no), alcohol consumption status (yes, no), and cardiovascular conditions, such as stroke, coronary heart disease, angina, myocardial infarction, and hypertension.

Analyses were conducted on subgroups to investigate differences among variables, including gender, race, marital status, level of education, BMI, smoking habits, alcohol consumption, and cardiovascular health. These analyses employed weighted stratified linear regression models to test for interaction terms between subgroups, evaluating differences in effects. Significant statistical results were identified through the study of statistics, with a p-value less than 0.05.

Analysis was conducted using R version 3.4.3 (found at http://www.R-project.org, given by The R Foundation), Empower program (found at http://www.empowerstats.com/, created by X&Y Solutions, Inc., Boston, MA, USA), and DecisionLinnc 1.0 program (found at https://www.statsape.com/). DecisionLinnc is a system that combines several coding environments, making it easier to handle data, conduct analyses, and utilize machine learning through a graphical interface.

3. Results

3.1 Baseline Characteristics of Participants

This study included 1823 adults with diabetes, all meeting specific inclusion and exclusion criteria. The mean age of the participants was 59.94 years, with a standard deviation of 12.17 years. Among the participants, 30.55 were males, and 69.45% were females. The ethnical composition of the participants was diverse, with 28.85% identified as non-Hispanic whites, 27.04% as non-Hispanic blacks, 24.19% as Mexican Americans, and 19.91% belonging to other races. The mean PNI among the participants was 52.41, with a standard deviation of 5.43. Additionally, 7.41% of the subjects had heart failure, according to the data presented in Fig. 2.

Based on the PNI quartiles, the clinical characteristics of participants were compared across different factors such as age, race, marital status, BMI, hypertension, myocardial infarction, and heart failure. Significant differences were found in these factors across the PNI quartiles, with p-values less than 0.05 (Table 1). Participants in the Q4 group were generally younger, more likely to be Mexican American, married or cohabiting, had a BMI between 18.5 and 30 kg/m², and were less likely to have hypertension, myocardial infarction, or heart failure than those in the Q1 group. These findings suggest that there may be correlations between the PNI quartiles and certain clinical characteristics of the participants. Overall, the results indicate notable disparities in age, race, marital status, BMI, and various cardiovascular health conditions across different PNI quartiles. Participants in the Q4 group exhibited specific characteristics such as being younger, Mexican American, married or cohabiting, and having a healthier BMI range, along with lower rates of hypertension, myocardial infarction, and heart failure. These findings shed light on the potential associations between the PNI quartiles and the clinical profiles of individuals participating in the study. Additional research could further explore these relationships to improve understanding of the implications for healthcare interventions and strategies targeting specific demographic groups.

3.2 Association between PNI and Heart Failure

Table 2 illustrates the results of the multivariable regression analysis evaluating the link between PNI and the risk of heart failure across three models, each with distinct levels of adjustment. Across all models, PNI exhibited a notable negative correlation with the risk of heart failure. The OR for PNI in the unadjusted Model 1 was 0.900 (95% CI: 0.871–0.930). For the preliminarily adjusted Model 2, the OR was 0.908 (95% CI: 0.877–0.940), and for the fully adjusted Model 3, the OR was 0.918 (95% CI: 0.884–0.953). This indicates that as PNI increases, the risk of developing heart failure progressively declines.

Additional analysis of the PNI quartiles further confirms this pattern. When considering all variables in Model 3, individuals in the lowest quartile (Q1, OR = 1.00) had a higher risk of heart failure compared to those in the second quartile (Q2, OR = 0.538, 95% CI: 0.319–0.909), third quartile (Q3, OR = 0.480, 95% CI: 0.285–0.808), and fourth quartile (Q4, OR = 0.370, 95% CI: 0.202–0.676), all of whom showed a significantly decreased risk of heart failure. Particularly in the fourth quartile, compared to the baseline, the risk of heart failure was reduced by 63.0%, demonstrating strong statistical significance (p trend 0.001), highlighting the clear association between higher PNI values and the reduced risk of heart failure (Table 2).

Additionally, we explored the relationship between PNI and heart failure status using a smoothing curve fitting method. The results showed a clear nonlinear negative correlation between the two, as depicted in Fig. 3. This underscores the complex association pattern between PNI and the risk of heart failure, revealing dynamic trends in their relationship.

3.3 Subgroup Analysis

Subgroup analyses were performed to evaluate the consistency of the relationship between PNI and heart failure across various demographic groups. As shown in Table 3, the analysis results indicate that the correlation between PNI and heart failure was inconsistent across different subgroups. In particular, significant associations were noted in subgroups categorized by ethnicity, marital status, tobacco use, alcohol consumption, cerebrovascular accident, coronary artery disease, and myocardial infarction (p 0.05). However, tests for interactions did not reveal any statistically significant differences among these stratified subgroups. This indicates that factors such as gender, race, educational level, marital status, smoking status, alcohol consumption, BMI, hypertension, history of stroke, coronary heart disease, angina, and myocardial infarction do not influence the inverse relationship between PNI and heart failure risk (p for interaction > 0.05) (Fig. 4). This emphasizes the universality of the relationship between higher PNI and reduced risk of heart failure, indicating certain stability and broad applicability of this link across different subgroups.

4. Discussion

Based on the NHANES public database, this study is the first to clearly demonstrate a negative correlation between the PNI and heart failure incidence among adults with diabetes. Using three models to gradually adjust for confounding factors yielded consistent results, indicating that a higher PNI is associated with a reduced risk of heart failure, with statistically significant differences. The coexistence of diabetes and heart failure poses a serious clinical challenge [31]. Patients with diabetes are in a state of chronic hyperglycemia, which leads to metabolic disturbances in cardiac cells and directly damages cardiomyocytes [32, 33]. Additionally, microvascular disease, structural and functional changes in the myocardium, and the development of atherosclerosis further increase the complexity and risk of heart failure [34, 35]. No studies have directly explored the relationship between these two factors in the diabetic population. This finding supports the potential of PNI as an effective biomarker for heart failure in diabetic patients, suggesting that PNI may become a valuable clinical tool for assessing the risk of heart failure.

The results of this study are consistent with previous research on the relationship between PNI and heart failure prognosis. Zhang et al. [36] followed 1048 patients with metabolic syndrome and heart failure, and their findings indicated that the PNI is an independent predictor of all-cause mortality and cardiovascular death in these patients, with a negative correlation between the PNI and adverse outcomes. In older patients with heart failure, low PNI values have also been associated with both short and long-term mortality [37]. Kawata et al. [38] explored the relationship between changes in PNI during hospitalization and outcomes in patients with acute heart failure, concluding that higher PNI levels are independently associated with better outcomes in heart failure patients. Other studies have also confirmed similar conclusions [39, 40].

PNI reflects an individual’s nutritional status [41, 42] and immune function [43, 44]. Malnutrition can lead to hypoproteinemia and weakened immunity, which in turn can trigger heart failure. Moreover, chronic hypoperfusion, congestion, and inflammatory responses in heart failure patients can impair liver and kidney function, leading to reduced albumin production and exacerbating malnutrition [45, 46, 47]. Inflammation activation and immune infiltration play critical roles in the pathological process of heart failure [48, 49]. Abnormal immune function can promote the progression of heart failure [50], and anti-myocardial autoreactivity by the adaptive immune system has been implicated in structural remodeling, functional decline, and the development of heart failure [51]. Individuals with good nutritional status typically possess strong metabolic capacity and immune responses, which are crucial for defending against infections and other stressors that may lead to heart failure [52]. Additionally, individuals with higher PNI levels generally have better overall health, healthier BMIs, and a lower likelihood of developing hypertension, myocardial infarction, and heart failure. Conversely, individuals with lower PNIs may face an increased risk of heart failure due to the combined effects of these factors.

This study is the first to focus on the diabetic population, exploring the relationship between PNI and heart failure and confirming that PNI remains negatively correlated with heart failure in this high-risk group. Although previous studies have demonstrated the significant value of using the PNI in general heart failure patients, caution is needed when directly extrapolating these findings to the general population, given the specific metabolic and pathophysiological characteristics of diabetic patients.

5. Limitations

This study has several limitations. First, due to the observational design of the study, the causality between the PNI and heart failure risk cannot be definitively established. Second, despite the adjustment for numerous confounding variables, there remains a possibility of undisclosed confounders that might influence the outcomes. Furthermore, the research predominantly relies on a sole assessment of PNI, failing to investigate how alterations in PNI levels over time could impact the likelihood of developing heart failure.

6. Conclusions

This research proposes a significant inverse relationship between the PNI and the risk of heart failure. This finding implies that PNI could be a valuable predictor of heart failure. However, further research is necessary to validate these results and to evaluate the influence of time-related variations in PNI on the occurrence of heart failure. Future research should also investigate the potential biological mechanisms that underlie the association between PNI and heart failure, aiming to enhance our understanding of these mechanisms for use in clinical practice.

Availability of Data and Materials

Publicly available datasets were analyzed in this study. This data can be found here: https://www.cdc.gov/nchs/nhanes/index.htm.

References

[1] Li X, Chen H, Jia Y, Peng J, Li C. Inhibitory Effects against Alpha-Amylase of an Enriched Polyphenol Extract from Pericarp of Mangosteen (Garcinia mangostana). Foods (Basel, Switzerland). 2022; 11: 1001.

[2] Li M, Zhu P, Wang SX. Risk for Cardiovascular Death Associated With Waist Circumference and Diabetes: A 9-Year Prospective Study in the Wan Shou Lu Cohort. Frontiers in Cardiovascular Medicine. 2022; 9: 856517.

[3] Song BR, Alam MB, Lee SH. Terpenoid-Rich Extract of Dillenia indica L. Bark Displays Antidiabetic Action in Insulin-Resistant C2C12 Cells and STZ-Induced Diabetic Mice by Attenuation of Oxidative Stress. Antioxidants (Basel, Switzerland). 2022; 11: 1227.

[4] Ofosu FK, Elahi F, Daliri EBM, Chelliah R, Ham HJ, Kim JH, et al. Phenolic Profile, Antioxidant, and Antidiabetic Potential Exerted by Millet Grain Varieties. Antioxidants (Basel, Switzerland). 2020; 9: 254.

[5] Zhang D, Zhang X, Bu Y, Zhang J, Zhang R. Copper Cobalt Sulfide Structures Derived from MOF Precursors with Enhanced Electrochemical Glucose Sensing Properties. Nanomaterials (Basel, Switzerland). 2022; 12: 1394.

[6] Zhang J, Qu X, Li J, Harada A, Hua Y, Yoshida N, et al. Tissue Sheet Engineered Using Human Umbilical Cord-Derived Mesenchymal Stem Cells Improves Diabetic Wound Healing. International Journal of Molecular Sciences. 2022; 23: 12697.

[7] Zhang Y, Yang R, Hou Y, Chen Y, Li S, Wang Y, et al. Association of cardiovascular health with diabetic complications, all-cause mortality, and life expectancy among people with type 2 diabetes. Diabetology & Metabolic Syndrome. 2022; 14: 158.

[8] Kitai T, Miyakoshi C, Morimoto T, Yaku H, Murai R, Kaji S, et al. Mode of Death Among Japanese Adults With Heart Failure With Preserved, Midrange, and Reduced Ejection Fraction. JAMA Network Open. 2020; 3: e204296.

[9] Groenewegen A, Zwartkruis VW, Cekic B, de Boer RA, Rienstra M, Hoes AW, et al. Incidence of atrial fibrillation, ischaemic heart disease and heart failure in patients with diabetes. Cardiovascular Diabetology. 2021; 20: 123.

[10] Wang J, Bai J, Duan P, Wang H, Li Y, Zhu Q. Kir6.1 improves cardiac dysfunction in diabetic cardiomyopathy via the AKT-FoxO1 signalling pathway. Journal of Cellular and Molecular Medicine. 2021; 25: 3935–3949.

[11] Rhee EJ, Kwon H, Park SE, Han KD, Park YG, Kim YH, et al. Associations among Obesity Degree, Glycemic Status, and Risk of Heart Failure in 9,720,220 Korean Adults. Diabetes & Metabolism Journal. 2020; 44: 592–601.

[12] Buzby GP, Mullen JL, Matthews DC, Hobbs CL, Rosato EF. Prognostic nutritional index in gastrointestinal surgery. The American Journal of Surgery. 1980; 139: 160–167.

[13] Li N, Jiang M, Wu WC, Zou LQ. The value of prognostic nutritional index in nasal-type, extranodal natural killer/T-cell lymphoma. Annals of Hematology. 2022; 101: 1545–1556.

[14] Seo YJ, Kong YG, Yu J, Park JH, Kim SJ, Kim HY, et al. The prognostic nutritional index on postoperative day one is associated with one-year mortality after burn surgery in elderly patients. Burns & Trauma. 2021; 9: tkaa043.

[15] Hirahara N, Matsubara T, Kaji S, Uchida Y, Hyakudomi R, Yamamoto T, et al. Influence of nutrition on stage-stratified survival in gastric cancer patients with postoperative complications. Oncotarget. 2022; 13: 183–197.

[16] Kweon S, Park JY, Park M, Kim Y, Yeon SY, Yoon L, et al. Trends in food and nutrient intake over 20 years: findings from the 1998-2018 Korea National Health and Nutrition Examination Survey. Epidemiology and Health. 2021; 43: e2021027.

[17] Wang Z, Wang B, Fu G, He B, Chu H, Zhang S. The Prognostic Nutritional Index May Predict Left Atrial Appendage Thrombus or Dense Spontaneous Echo Contrast in Patients With Atrial Fibrillation. Frontiers in Cardiovascular Medicine. 2022; 9: 860624.

[18] Suzuki E, Kawata N, Shimada A, Sato H, Anazawa R, Suzuki M, et al. Prognostic Nutritional Index (PNI) as a Potential Prognostic Tool for Exacerbation of COPD in Elderly Patients. International Journal of Chronic Obstructive Pulmonary Disease. 2023; 18: 1077–1090.

[19] Luna DJ, R Pandian NK, Mathur T, Bui J, Gadangi P, Kostousov VV, et al. Tortuosity-powered microfluidic device for assessment of thrombosis and antithrombotic therapy in whole blood. Scientific Reports. 2020; 10: 5742.

[20] Niu W, Yang X, Yan H, Yu Z, Li Z, Lin X, et al. Peritoneal Protein Clearance Is Associated With Cardiovascular Events but Not Mortality in Peritoneal Dialysis Patients. Frontiers in Medicine. 2022; 9: 748934.

[21] Zhang X, Sun Y, Li Y, Wang C, Wang Y, Dong M, et al. Association between visceral adiposity index and heart failure: A cross-sectional study. Clinical Cardiology. 2023; 46: 310–319.

[22] Wu Z, Tian T, Ma W, Gao W, Song N. Higher urinary nitrate was associated with lower prevalence of congestive heart failure: results from NHANES. BMC Cardiovascular Disorders. 2020; 20: 498.

[23] Liu Z, Liu H, Deng Q, Sun C, He W, Zheng W, et al. Association Between Dietary Inflammatory Index and Heart Failure: Results From NHANES (1999-2018). Frontiers in Cardiovascular Medicine. 2021; 8: 702489.

[24] Rethy L, Petito LC, Vu THT, Kershaw K, Mehta R, Shah NS, et al. Trends in the Prevalence of Self-reported Heart Failure by Race/Ethnicity and Age From 2001 to 2016. JAMA Cardiology. 2020; 5: 1425–1429.

[25] Camplain R, Kucharska-Newton A, Loehr L, Keyserling TC, Layton JB, Wruck L, et al. Accuracy of Self-Reported Heart Failure. The Atherosclerosis Risk in Communities (ARIC) Study. Journal of Cardiac Failure. 2017; 23: 802–808.

[26] Thomson B, Emberson J, Lacey B, Lewington S, Peto R, Jemal A, et al. Association Between Smoking, Smoking Cessation, and Mortality by Race, Ethnicity, and Sex Among US Adults. JAMA Network Open. 2022; 5: e2231480.

[27] Upson K, O’Brien KM, Hall JE, Tokar EJ, Baird DD. Cadmium Exposure and Ovarian Reserve in Women Aged 35-49 Years: The Impact on Results From the Creatinine Adjustment Approach Used to Correct for Urinary Dilution. American Journal of Epidemiology. 2021; 190: 116–124.

[28] Sousa A, Sych J, Rohrmann S, Faeh D. The Importance of Sweet Beverage Definitions When Targeting Health Policies-The Case of Switzerland. Nutrients. 2020; 12: 1976.

[29] Tucker A, Wang Z, Rotalinti Y, Myles P. Generating high-fidelity synthetic patient data for assessing machine learning healthcare software. NPJ Digital Medicine. 2020; 3: 147.

[30] D’Amato G, Faienza MF, Palladino V, Bianchi FP, Natale MP, Christensen RD, et al. Red blood cell transfusions and potentially related morbidities in neonates under 32 weeks’ gestation. Blood Transfusion = Trasfusione Del Sangue. 2021; 19: 113–119.

[31] Lu L, Ma J, Liu Y, Shao Y, Xiong X, Duan W, et al. FSTL1-USP10-Notch1 Signaling Axis Protects Against Cardiac Dysfunction Through Inhibition of Myocardial Fibrosis in Diabetic Mice. Frontiers in Cell and Developmental Biology. 2021; 9: 757068.

[32] Thakur V, Alcoreza N, Delgado M, Joddar B, Chattopadhyay M. Cardioprotective Effect of Glycyrrhizin on Myocardial Remodeling in Diabetic Rats. Biomolecules. 2021; 11: 569.

[33] Chang X, Zhang T, Wang J, Liu Y, Yan P, Meng Q, et al. SIRT5-Related Desuccinylation Modification Contributes to Quercetin-Induced Protection against Heart Failure and High-Glucose-Prompted Cardiomyocytes Injured through Regulation of Mitochondrial Quality Surveillance. Oxidative Medicine and Cellular Longevity. 2021; 2021: 5876841.

[34] Lin N, Lin H, Yang Q, Lu W, Sun Z, Sun S, et al. SGLT1 Inhibition Attenuates Apoptosis in Diabetic Cardiomyopathy via the JNK and p38 Pathway. Frontiers in Pharmacology. 2021; 11: 598353.

[35] Ouerd S, Idris-Khodja N, Trindade M, Ferreira NS, Berillo O, Coelho SC, et al. Endothelium-restricted endothelin-1 overexpression in type 1 diabetes worsens atherosclerosis and immune cell infiltration via NOX1. Cardiovascular Research. 2021; 117: 1144–1153.

[36] Zhang X, Zhang J, Liu F, Li W, Zhang T, Fang B, et al. Prognostic Nutritional Index (PNI) as a Predictor in Patients with Metabolic Syndrome and Heart Failure. Diabetes, Metabolic Syndrome and Obesity: Targets and Therapy. 2023; 16: 2503–2514.

[37] Candeloro M, Di Nisio M, Balducci M, Genova S, Valeriani E, Pierdomenico SD, et al. Prognostic nutritional index in elderly patients hospitalized for acute heart failure. ESC Heart Failure. 2020; 7: 2479–2484.

[38] Kawata T, Ikeda A, Masuda H, Komatsu S. Changes in prognostic nutritional index during hospitalization and outcomes in patients with acute heart failure. Heart and Vessels. 2022; 37: 61–68.

[39] Ju C, Zhou J, Lee S, Tan MS, Liu T, Bazoukis G, et al. Derivation of an electronic frailty index for predicting short-term mortality in heart failure: a machine learning approach. ESC Heart Failure. 2021; 8: 2837–2845.

[40] Çinier G, Hayıroğlu Mİ Pay L, Yumurtaş AÇ Tezen O, Eren S, et al. Prognostic nutritional index as the predictor of long-term mortality among HFrEF patients with ICD. Pacing and Clinical Electrophysiology: PACE. 2021; 44: 490–496.

[41] Prausmüller S, Heitzinger G, Pavo N, Spinka G, Goliasch G, Arfsten H, et al. Malnutrition outweighs the effect of the obesity paradox. Journal of Cachexia, Sarcopenia and Muscle. 2022; 13: 1477–1486.

[42] Wang K, Lian L, Chen C, Wang M, Chen C, Hu X. The change in nutritional status is related to cardiovascular events in patients with pacemaker implantation: A 4-year follow-up study. Frontiers in Nutrition. 2022; 9: 986731.

[43] Wang Y, Chen L, Zhang B, Song W, Zhou G, Xie L, et al. Pretreatment Inflammatory-Nutritional Biomarkers Predict Responses to Neoadjuvant Chemoradiotherapy and Survival in Locally Advanced Rectal Cancer. Frontiers in Oncology. 2021; 11: 639909.

[44] Chen L, Kong X, Huang S, Su Z, Zhu M, Fang Y, et al. Preoperative Breast Immune Prognostic Index as Prognostic Factor Predicts the Clinical Outcomes of Breast Cancer Patients Receiving Neoadjuvant Chemotherapy. Frontiers in Immunology. 2022; 13: 831848.

[45] Sunaga A, Hikoso S, Yamada T, Yasumura Y, Tamaki S, Yano M, et al. Change in Nutritional Status during Hospitalization and Prognosis in Patients with Heart Failure with Preserved Ejection Fraction. Nutrients. 2022; 14: 4345.

[46] Katano S, Yano T, Kouzu H, Ohori K, Shimomura K, Honma S, et al. Energy intake during hospital stay predicts all-cause mortality after discharge independently of nutritional status in elderly heart failure patients. Clinical Research in Cardiology: Official Journal of the German Cardiac Society. 2021; 110: 1202–1220.

[47] Lin Z, Liu X, Xiao L, Li Y, Qi C, Song S, et al. The MELD-XI score predicts 3-year mortality in patients with chronic heart failure. Frontiers in Cardiovascular Medicine. 2022; 9: 985503.

[48] Liu X, Xu S, Li Y, Chen Q, Zhang Y, Peng L. Identification of CALU and PALLD as Potential Biomarkers Associated With Immune Infiltration in Heart Failure. Frontiers in Cardiovascular Medicine. 2021; 8: 774755.

[49] Rivera AS, Sinha A, Ahmad FS, Thorp E, Wilcox JE, Lloyd-Jones DM, et al. Long-Term Trajectories of Left Ventricular Ejection Fraction in Patients With Chronic Inflammatory Diseases and Heart Failure: An Analysis of Electronic Health Records. Circulation. Heart Failure. 2021; 14: e008478.

[50] Elgebaly SA, Todd R, Kreutzer DL, Christenson R, El-Khazragy N, Arafa RK, et al. Nourin-Associated miRNAs: Novel Inflammatory Monitoring Markers for Cyclocreatine Phosphate Therapy in Heart Failure. International Journal of Molecular Sciences. 2021; 22: 3575.

[51] Forte E, Perkins B, Sintou A, Kalkat HS, Papanikolaou A, Jenkins C, et al. Cross-Priming Dendritic Cells Exacerbate Immunopathology After Ischemic Tissue Damage in the Heart. Circulation. 2021; 143: 821–836.

[52] Chien SC, Chandramouli C, Lo CI, Lin CF, Sung KT, Huang WH, et al. Associations of obesity and malnutrition with cardiac remodeling and cardiovascular outcomes in Asian adults: A cohort study. PLoS Medicine. 2021; 18: e1003661.

Funding

Gansu Provincial Natural Science Foundation Project(21JR7RA379)

Provincial Health Industry Plan Project(GSWSKY2020-49)