Association between single-point insulin sensitivity estimator and heart failure in older adults: A cross-sectional study

Background: Heart failure (HF) is a condition caused by a malfunction of the heart's pumping function. The single-point insulin sensitivity estimator (SPISE) index is a novel indicator for assessing insulin resistance in humans. However, the connection between the SPISE index and the risk of HF in the elderly is unknown. Therefore, our study aims to evaluate the connection between the SPISE index and HF in older adults. Methods: The study was based on data collected from the 1999 – 2020 National Health and Nutrition Examination Survey database and included 6165 participants aged ≥ 60 years. The multivariable linear regression model and the smooth fitting curve model were applied to investigate the connection between the SPISE index and HF in the elderly. Furthermore, the subgroup analysis was performed to investigate the interactive factors. Results: In this study, the mean age of the population was 69.38 years. After adjusting for all covariates, we observed that the SPISE index was inversely related to the prevalence of HF (OR = 0.87, 95 % CI = 0.80 – 0.94, P < 0.001) in older adults. The interaction analysis showed that the association might be affected by diabetes mellitus and smoking status. Additionally, an inflection point between the SPISE index and HF was found among older women. Conclusions: An inverse correlation was detected between the SPISE index and HF in the elderly. This could provide new insight into the prevention and management of HF in the elderly population.


Introduction
As a complicated heterogeneous syndrome, heart failure (HF) takes place invariably at the end stage of a variety of cardiovascular diseases.There are many causes of HF, which are summarized as including damage to the heart muscle itself (myocardial ischemia, carditis, etc.), insufficient ventricular preload (Mitral stenosis, constrictive pericarditis, etc.), and overloading of the heart (hypertension, pulmonary arterial hypertension, etc).Among almost all chronic cardiovascular diseases, HF is the most common consequence of structural or functional abnormalities of the heart, which has become a major global health problem (Heidenreich et al., 2022).With the growing and aging population, there are increasing quantities of HF patients, and the morbidity and mortality of HF are increasing annually worldwide (Groenewegen et al., 2020;Metra and Teerlink, 2017).Globally, the number of deaths due to chronic HF is increasing, causing a huge social and economic burden on humanity (Ambrosy et al., 2014).Previous studies have indicated that approximately 64.3 million people worldwide suffer from HF (Disease et al., 2018), the prevalence of HF is predicted to rise to 3.0 % by 2030 (Heidenreich et al., 2013), and approximately 50 % of patients die within five years following diagnosis (Metra and Teerlink, 2017).Therefore, finding the relevant factor and the interrelationship with HF is critical to opening new avenues of prevention.
Insulin resistance (IR), features in target cells or whole organisms that reduce responsiveness to insulin concentrations, is associated with chronic diseases, including metabolic syndrome (MetSyn), cardiovascular disease, and type 2 diabetes mellitus (DM) (Ruderman et al., 2013).Among them, a strong association was observed between IR and HF in several clinical studies (Horwich and Fonarow, 2010).The increased susceptibility of the heart in insulin-resistant states has been attributed to various mechanisms, including increased utilization of free fatty acids, diminished cardiac efficiency, oxidative stress, increased levels of bio-active lipids, decreased mitochondrial oxidative capacity, altered calcium metabolism and signaling, and myocardial fibrosis (Aroor et al., 2012;McGuire and Gore, 2013;Riehle and Abel, 2016).PI3-kinase/Akt signaling is blocked by cardiac IR, which also affects calcium management (Lebeche et al., 2008).Additionally, in response to IR, cluster differentiation protein 36 (CD36) redistributes to the plasma membrane and systemic/adipose tissue IR leads to increased fatty acid flow into cardiac cells, which in turn causes enhanced fatty acid oxidation (Falcao-Pires and Leite-Moreira, 2012).Those changes eventually lead to heart fibrosis, myocardial cell death, and reduced cardiac energy efficiency.Furthermore, some age-related conditions in HF including hypertension, DM and chronic kidney disease (CKD) are associated with IR, which may aggravate HF through the mechanisms mentioned above (Reaven, 1991;Landau et al., 2011).Heart hypertrophy and fibrosis are caused by hypertension, whereas myocardial ischaemia and harmful remodeling are brought on by CAD and atherosclerosis.Diabetes-related metabolic disorders exacerbate the profibrotic cardiac milieu, which is exacerbated by insulin resistance, inflammation, and dyslipidemia linked to obesity.Fluid overload, electrolyte abnormalities, and uraemic toxins associated with CKD aggravate HF by inducing systemic inflammation and activating the renin-angiotensin-aldosterone system (RAAS).
The hyperinsulinemic-euglycemic clamp (HEC) has been utilized to evaluate IR; however, it is expensive and invasive and has limited clinical use (Marcovecchio et al., 2010).Recently, a single-point insulin sensitivity estimator (SPISE) index based on triglycerides (TG), body mass index (BMI), and high-density lipoprotein cholesterol (HDL-c), has been proposed as a biomarker to evaluate insulin sensitivity (Paulmichl et al., 2016).Compared to the HEC method, the SPISE index is much easier to obtain by physicians.Also, it has just been confirmed in adults and teenagers and has shown superior accuracy in predicting MetSyn (Correa-Burrows et al., 2020).Additionally, SPISE has also been shown to be able to predict Swedish participants' long-term risk of DM, or CHD (Cederholm and Zethelius, 2019).
It  (Riehle and Abel, 2016), obesity (Koliaki et al., 2019), CHD (Dunlay et al., 2009), and other factors are intimately related to HF through direct or indirect modalities.However, as a biomarker of insulin sensitivity, the study of the correlation between SPISE and HF in the elderly is limited and lacks adequate research.Therefore, aiming to investigate the association between SPISE and HF in older adults, the research was conducted using data from 1999 to 2020 from the National Health and Nutrition Examination Survey (NHANES), which will benefit the prevention of HF in the elderly in the clinical context.

Study population
NHANES, a nationwide initiative overseen by the National Center for Health Statistics (NCHS), serves as a comprehensive assessment of the health and nutritional well-being of both children and adults across the United States.The study was based on data collected from the 1999-2020 NHANES database, and a total of 6165 individuals were finally included.The research procedures for NHANES received approval from the NCHS Ethical Review Board, with all participants willingly providing written informed consent.Due to NHANES being a publicly accessible database, this particular study was exempt from further ethical review.Additional information regarding the NHANES study and its associated statistical methods can be found online at https: //www.cdc.gov/nchs/.

Inclusion criteria and exclusion criteria
The inclusion and exclusion criteria are as follows: Inclusion Criteria: (1) Age ≥ 60 years; (2) Individuals who participated in the NHANES from 1999 to 2020.Exclusion Criteria: (1) Participants with missing data on triglycerides, HDL-c, and BMI; (2) Unavailable information on HF status; (3) Lack of accessible data on demographic and socioeconomic variables; (4) Missing information on other disease histories; (5) Missing data on laboratory inspection indicators; (6) Unavailable information on dietary habits, lifestyle, and history of medication use (Fig. 1).

Definitions of exposure and outcome variables
Professional laboratory personnel conducted the measurement of HDL-c and TG, with these results accessible in the Blood Biochemistry section.BMI is determined during the physical examination.The SPISE formula is given by: SPISE = [600 × HDL-c (mg/dL)^0.185]/ [TG (mg/ dL)^0.2× BMI (kg/m^2)^1.338](Paulmichl et al., 2016).
Participants were considered to have HF if they answered yes to the following question: "Has a doctor or other health professional ever told you that you have HF?" Previous studies have validated the utility of self-reported HF (Chen et al., 2023;Tao et al., 2023;Zhang et al., 2023).

Covariates
Based on clinical relevance, potential covariates that may influence the relationship between the SPISE index and HF in the elderly were included in this study.Demographic covariates were assessed at the individual interviews.Race was categorized into five distinct categories: Mexican American, non-Hispanic black, non-Hispanic white, other Hispanic, and other races.Marital status was categorized as either married/living with a partner, widowed/divorced/separated, or never married.Education level was stratified into three groups: less than high school, high school graduate, and more than high school.Energy intake was assessed by two-day dietary recall.Participants' alcohol use status was classified into never, former, and current according to the previous study (Hicks et al., 2021;Rattan et al., 2022).Histories of DM, hypertension, stroke, heart attack, CHD, angina, and rheumatoid arthritis were collected by the health questionnaire.The information about the use of anti-diabetic agents was obtained from the questionnaire on drug use.An individual who has smoked fewer than 100 cigarettes during his or her lifetime is considered to have never smoked; a person who has smoked >100 cigarettes during his or her lifetime and does not now smoke at all is considered to have ever smoked, and a person who has smoked >100 cigarettes during his or her lifetime and smokes on several days or every day is considered to be a current smoker.Uric acid, hemoglobin, and fasting glucose levels were obtained from the laboratory analysis results provided by the NHANES project.The estimated glomerular filtration rate (eGFR) was calculated using serum creatine (Levey et al., 2009).Urinary albumin/creatinine ratio (UACR) was calculated based on the laboratory data provided by the NHANES (Warsame et al., 2023).Participants with an estimated glomerular filtration rate (eGFR) of <60 ml/min/1.73m2 were considered to have CKD (Liu et al., 2024).In addition, BMI ≥30.0 kg/m 2 was considered as obesity (Hong et al., 2024).

Statistical analysis
Appropriate sampling weights, clustering, and stratification were considered according to the statistical analysis guidelines provided by the NHANES database.Two groups were categorized according to the presence or absence of HF.Continuous variables were expressed as mean ± standard error using weighted linear regression models, and categorical variables were expressed as percentages using the Rao-Scott chisquare test.Univariate logistic regression analysis was applied to identify latent confounders.Multivariate logistic regression models were used to explore the relationship between the SPISE index and HF in the elderly, with quartiles used to categorize levels of the SPISE index.Model 1 was not adjusted for any covariates.Model 2 was adjusted for age, gender, and race.Model 3 was adjusted for all confounding factors which presented P < 0.05 in the univariate logistic analysis.Moreover, the receiver operating (ROC) curve was used to compare the efficacy between different models.Based on model 3, the smooth fitting curve model was utilized to evaluate the dose-response relationship between the SPISE index and HF in older adults further.Additionally, subgroup analyses and interaction tests were performed to investigate whether the relationship between the SPISE index and HF was robust across older populations, and the factors included gender, race, marital status, education level, smoking status, alcohol use, DM, and hypertension were stratified.Sensitivity analysis was also performed to validate our results.Statistical analyses were performed using the R package (www.R-pr oject.org)and EmpowerStats software.Statistical significance was set at P < 0.05 for all analyses.

Baseline characteristics of the study participants
Table 1 listed the baseline characteristics of the study sample with HF as a column-stratification factor.The mean age of the study population was 69.38 ± 6.84 years and 45.11 % of them were male.Among them, 401 older participants had HF.Compared with the older participants without HF, the HF group was more likely to be older, male, poorer, and less educated (all P < 0.001).Those with HF had lower energy intake, eGFR, hemoglobin, higher UACR, fasting glucose, and uric acid (all P < 0.05).Moreover, we observed a higher proportion of non-alcohol users, DM, hypertension, obesity, stroke, heart attack, CHD, angina, CKD, rheumatoid arthritis, and the use of diabetic agents (all P < 0.05) in HF individuals.Nonetheless, no significant disparities were observed between the HF and control groups for race and smoking status (P > 0.05).

Relationship between SPISE and HF in the elderly
Based on the univariate logistic regression model, we included and adjusted variables presented with P < 0.05 in the further analysis (Table S1).According to the multivariate logistic regression model, the relationship between the SPISE index and HF in the elderly was shown in Table 2.After adjusting for all covariates (including age, gender, race, marital status, DM, hypertension, education level, poverty income ratio (PIR), eGFR, UACR, uric acid, fasting glucose, obesity, CKD, rheumatoid arthritis, use of diabetic agents, hemoglobin, stroke, heart attack, CHD, and angina), we found that a higher SPISE index was associated with a lower risk of developing HF in older adults, with a negative association (OR = 0.87, 95 % CI = 0.78-0.96,P < 0.001).Using the Q1 group as the reference, older participants in the Q4 group (OR = 0.55, 95 % CI = 0.34-0.91,P < 0.001) had a lower prevalence of HF.Moreover, based on the results of ROC analysis, we observed that model 3 showed the highest efficacy and the area under curve was 0.873 (Table S2).According to the smooth-fitting curve model, the SPISE index exhibited a significant inverse association with the risk of HF in older adults (P for log-likelihood ratio > 0.05), after accounting for the latent covariates (Fig. 2).

Subgroup analysis
Subgroup analysis was applied to assess the potential interactive factor on the relationship between the SPISE index and HF in the elderly (Table 3).We observed that this association remained stable across a number of subgroups (all P for interaction >0.05).Nevertheless, a significant interaction was observed in smoking status and DM subgroups (P for interaction <0.05).The SPISE index was not associated with current elderly smokers (OR = 1.01, 95 % CI =0.87-1.17,P = 0.864), and was more pronounced in diabetics (OR = 0.73, 95 % CI = 0.62-0.87,P < 0.001).Furthermore, based on the gender-specific population, a smooth-fitting curve was performed to examine the latent dose-response relationship (Fig. 3).We investigated that a significant U-shaped curve was presented in the relationship between the SPISE index and HF in the older women, while the same was not observed in the elderly men.The threshold effect analysis was also conducted to explore the potential infection point and its effect on the relationship (Table 4).Judging from the results, considering the P for log-likelihood ratio, the relationship between the SPISE index and HF in elderly men should be explained by the linear regression model (OR = 0.85, 95 % CI = 0.74-0.98,P = 0.024).Also, for the connection between the SPISE index and HF in older women, the two-segment piecewise linear regression model was chosen.We observed that the SPISE was significantly connected with the risk of HF in elderly women (OR = 0.35, 95 % CI = 0.18-0.66,P = 0.001) when the SPISE index was lower than 3.58.However, no statistically significant connection was observed when the SPISE index (OR = 1.01, 95 % CI = 0.86-1.18,P = 0.951) was >3.58.

Sensitivity analysis
After excluding the participants with stroke, angina, CHD, and heart attack, we performed the multivariate logistic analysis to examine the robustness of the result (Table S3).After controlling the potential confounder, the SPISE was still significantly inversely associated with the risk of HF (OR = 0.78, 95 % CI = 0.66-0.93,P < 0.001).Also, the participants in the Q4 group had a lower risk of HF when the Q1 group was used as the reference (OR = 0.36, 95 % CI = 0.15-0.86,P = 0.022).

Discussion
In this study, we analyzed the negative relationship between SPISE and HF in older populations based on the NHANES 1999-2020 data in older adults, revealing that the SPISE index may contribute to identifying HF in older people.Furthermore, the subgroup analyses and interaction analyses showed that this negative association was stable in most demographic contexts, except for smokers or diabetics.Additionally, an inflection point was observed among older female individuals.
Several existing studies have found a relationship between many factors, such as metabolic disorders, and HF.Researchers found that the SPISE index was associated with the risk of cardiovascular diseases among 10,190 participants with DM (Deng et al., 2024).In a prospective cohort study comprising 61,113 Chinese individuals, the combination of proteinuria and a higher BMI level was found to be positively linked to the incidence of HF in the Chinese population (Wang et al., 2023).Moreover, a Mendelian randomization study based on European individuals revealed a causal connection between IR and the risk of cardiovascular diseases from the genetic aspect (Zhang and Yu, 2024).Similarly, based on an observational method, researchers from Germany demonstrated that IR is an independent risk factor for a decline in left ventricular function (Dinh et al., 2010).
In addition, as a vital component of the SPISE index, high BMI is usually related to being overweight or obese in most people, while obesity is expected in the HF population (Horwich et al., 2018).Generally, obesity is a primary risk factor for hypertension and myocardial hypertrophy, which further leads to HF as well (Kenchaiah et al., 2002;Carbone et al., 2020).In the Framingham Heart Study, which involved 5881 patients, BMI was found to have a dose-dependent relationship with the risk of HF.Each one-unit increase in BMI was associated with a significant 5 % increase in HF risk in men and a 7 % increase in HF risk in women.This effect remained statistically significant even after accounting for demographic factors and other established risk factors, including DM, hypertension, and cholesterol levels (Trullas et al., 2021).Another significant factor considered in the SPISE index, HDL-c probably repairs the heart via energy, inflammation, and cardiomyocyte protection mechanisms in HF patients (Jackson et al., 2021).Generally speaking, several clinical factors correlate with HF, showing that it may be biased to evaluate the incidence of HF via a single indicator.SPISE, a new index proposed in the European population, based on TG, HDL-c, and BMI to estimate insulin sensitivity, was demonstrated to probably be a predictor of adolescents with DM and insulin sensitivity for MetSyn (Ha et al., 2022).Among 7837 Korean adults in a large cross-sectional study, regardless of gender, the SPISE index has demonstrated a powerful predictive accuracy of ≥ 0.89 for diagnosing MetSyn, showing the perfect diagnostic value of the novel index (Seo et al., 2023).Nevertheless, the relationship between SPISE and HF in the elderly from the US has not been studied in previous research.As a result, we investigated the association between the two in this study, benefiting clinical prevention and treatment of HF in the older population in America.
A growing corpus of research has demonstrated a robust correlation between IR and HF (Rutter et al., 2003).IR is a major factor in the pathophysiology of HF since insulin signaling regulates both glucose and fatty acid metabolism in the heart (Ingelsson et al., 2005).Changes in the metabolism of myocardial substrates have been linked to the development of HF and contractile dysfunction (Witteles and Fowler, 2008;Neubauer, 2007).Indeed, two of the main variables influencing the evolution of HF are impaired insulin signaling and the emergence of IR, which typically occurs before cardiac dysfunction (Ingelsson et al., 2005).Insulin stimulates the fatty acid translocase CD36's translocation, thereby increasing the absorption of FA that might be allocated to lipid synthesis and storage (Glatz et al., 2010).In hyperinsulinemic states, which are commonly associated with enhanced accessibility to triglycerides and FA to the heart, this insulin-mediated upregulation and redistribution of CD36 can be mediated via PI3K/Akt signaling pathways and could lead to enhanced FA oxidation and myocardial oxygen consumption (Chabowski et al., 2004).Besides, it has been found that the human diabetic heart (transplanted from non-diabetic patients) could overexpress inflammatory/oxidative stress pathways (Sardu et al., 2020a) and the Sodium/glucose cotransporter 2 (SGLT2) mediated pathways.Overdistension of the cardiomyocytes during volume overload compromises contractility and lowers cardiac output even further.
In such circumstances, the diuresis and natriuresis linked to SGLT2 inhibition may increase contractility by lowering myocardial stretch and preload.Beyond lowering blood pressure, SGLT2 inhibitors may also enhance afterload by reducing arterial stiffness, thus resulting in the reduction of systolic and diastolic function (Liu et al., 2022a).In this case, SGLT2i may lead to improvements in these pathways as well as systolic and diastolic function, which would enhance HF (Liu et al., 2022b).Furthermore, in pressure overload hypertrophy, hyperactive insulin signaling quickens the unfavorable remodeling of the left ventricle (Rutter et al., 2003), (Sundstrom et al., 2000;Lopez-Izquierdo et al., 2014).Hepatic IR was brought on by chronic pressure overload, which also raised plasma insulin levels.Stretching the heart muscle activate-dInsulin Receptor (Insr), and prolonged high pressure elevated both the expression of Insr and Receptor Substrate-1(Irs1) protein as well as the activity of insulin signaling (pIrs1 and pAkt levels) (Benito-Vicente et al., 2020).This in turn made it easier for hyperinsulinemia to activate cardiac insulin signals.This activation led to an increase in cardiomyocyte mortality and a mismatch between the vascularity and size of the cardiomyocytes.The increase was then linked to systolic dysfunction and

Table 2
The association between the SPISE index and risk of heart failure.Model 1: unadjusted.Model 2: age, gender, and race were adjusted.Model 3: age, gender, race, marital status, diabetes, hypertension, education level, PIR, eGFR, UACR, uric acid, fasting glucose, obesity, CKD, rheumatoid arthritis, use of diabetic agents, hemoglobin, stroke, heart attack, CHD, and angina were adjusted.might be a contributing factor to HF brought on by persistent pressure overload.Additionally, it has also been reported recently that hyperinsulinemia may reduce myocardial contractility through a pathway whereby G-protein receptor kinase triggers Gi skewed β2AR signaling, which in turn suppresses adenylate cyclase, the production of cAMP, and cardiac contractility (Fu et al., 2014).The subgroup and interaction analyses were conducted in this study, revealing the interaction effect in the different smoking status and DM populations.Studies have found that smokers tend to have more inflammatory factors, such as C-reactive protein and interleukin-6, than non-smokers, which trigger inflammatory responses that attack cardiomyocytes (Gottdiener et al., 2022).In addition, long-term smoking impairs endothelium-dependent vasodilation and further promotes the development of atherosclerosis, inducing various cardiovascular events (Watson et al., 2019).Cardiac structural changes caused by endothelial injury and oxidative stress eventually lead to HF (Lu et al., 2021).Moreover, people with DM are chronically in hyperglycemic, hyperinsulinic, or IR states that alter the steady state of vascular homeostasis (Di Carli et al., 2003).Microvascular and macrovascular lesions are both critical causes of death in diabetic patients, which can lead to vascular endothelial dysfunction and clinical myocardial ischemia (Gottdiener et al., 2022;Coats and Anker, 2000).Moreover, studies have shown that over-glycosylation can have some impact on insulin sensitization insulin homeostasis, and resistance (Gambardella et al., 2022).Deficient dicarbonyl detoxifying ability favors glycolate degradation of RyR2, leading to calcium leak and mitochondrial damage in the senescent myocardium, which in turn affects the outcome of HF.Additionally, the indices of blood glucose that we included were collected in epidemiologic community surveys of mostly asymptomatic healthy populations rather than hospital populations, and therefore, there were no extremely abnormal values in the blood glucose data and the effect of over-glycation was acceptable.
Our study has several strengths.First, this is the first time that the negative association between SPISE and HF has been investigated in older adults based on a large population dataset.Second, we conducted subgroup and interaction analyses to further clarify the association between SPISE and HF in the elderly in different populations.However, this study also suffers from certain limitations.The causal relationship between SPISE and HF in the elderly could not be established on account of the limitations of the cross-sectional study.Besides, our research was based on data from older adults, whose results may not be applicable to all populations.In addition, no information regarding the etiology of HF was available, so the relationships between the SPISE and particular kinds of HF were not investigated.Moreover, the information on the left ventricular ejection fraction (LVEF) was unavailable in the NHANES database and we could not include it in the analysis.However, considering the clinical significance of the LVEF, this should be taken into consideration in the future study.In addition, due to the limitations of the questionnaire, we could not access the information about anti-HF therapy, which may be confounding.Furthermore, the BMI used in our study does not allow for a more accurate assessment of abnormalities in body fat distribution than the waist-to-hip ratio (WHR).Since BMI does not take into consideration body composition, which includes fluid in the third space and fat distribution, patients with high BMIs may be incorrectly diagnosed with HF owing to dyspnea.In particular, BMI might overlook the impact of belly fat, which has been linked to mortality in the general population and has been recognized as a possible risk factor in the onset of HF (Aune et al., 2016).In future studies, we will consider including WHR in the study design to improve the rigor of the analysis.Moreover, DM patients' cardiac pumps may improve with CRT (and ICDs) (Sardu et al., 2020b;Sardu et al., 2017) as well as in older diabetic patients (Sardu et al., 2014).Due to its beneficial hemodynamic and clinical effects, automated vs. echo-guided CRTd optimization has the potential to considerably lower the levels of inflammatory biomarkers (CRP, IL6, TNFa), as well as BNP values in HF patients with T2DM.For HF patients, the GLP1RA (Sardu et al., 2018) and the ARNI (Sardu et al., 2022) may also produce the best clinical results.In addition to conventional hypoglycemic medications, GLP-1 RA therapy may modify the genesis, duration, and propagation of action potentials in the cardiac walls and heart chambers, as well as depolarizing and repolarizing cardiac activity.This could lessen the burden of arrhythmias in CRTd patients with diabetes (Sokos et al., 2006).The effects of ARNI may also have an impact on the epigenetic processes that modulate the levels of miRs, which are implicated in the primary pathways of heart dysfunction and negatively impact cardiac remodeling responses to CRTd (Sardu et al., 2022).However, we were unable to analyze them due to the limited data in the survey questionnaire for the community.Moreover, because of the large amount of missing information on some diseases, including osteoporosis and osteopenia, we are unable to include them.Additionally, regarding the nature of the cross-sectional study, sample size calculations were not required.However, considering the benefits of the sample size calculation, it is essential to calculate the sample size in the prospective study and we will focus on this point in our future research.
Finally, even though some potential confounders were adjusted, the effect of other latent confounding variables could not be completely excluded.Further prospective studies are required to corroborate our

Table 4
Threshold effect analysis of SPISE and heart failure among female and male participants, respectively.Adjusted for age, gender, race, marital status, education level, PIR, eGFR, UACR, uric acid, fasting glucose, obesity, CKD, rheumatoid arthritis, use of diabetic agents, hemoglobin, stroke, heart attack, CHD, and angina were adjusted.
is generally acknowledged that type 2 DM (S.M. Dunlay, M.M. Givertz, D. Aguilar, L.A. Allen, M. Chan, A.S. Desai, A. Deswal, V.V. Dickson, M.N.Kosiborod, C.L. Lekavich, R.G. McCoy, R.J. Mentz, I.L. Pina, F. American Heart Association Heart, C. Transplantation Committee of the Council on Clinical, C. Council on, N. Stroke, and A. the Heart Failure Society of, Type 2 Diabetes Mellitus and Heart Failure: A Scientific Statement From the American Heart Association and the Heart Failure Society of America: This statement does not represent an update of the, 2017), generalized IR

Fig. 1 .
Fig. 1.Flow chart of participant selection.Abbreviations: NHANES, National Health and Nutrition Examination Survey, SPISE, Single Point Insulin Sensitivity Estimator.

Fig. 2 .
Fig. 2. The smooth fitting curve of the correlation between the SPISE index and heart failure Abbreviations: SPISE, Single Point Insulin Sensitivity Estimator.

Fig. 3 .
Fig. 3.The smooth fitting curve about the correlation between the SPISE index and heart failure among female and male participants, respectively Abbreviations: SPISE, Single Point Insulin Sensitivity Estimator.

Table 1
Demographic and clinical characteristics of study participants.Mean ± SD for continuous variables: the P value was calculated by the weighted linear regression model (%) for categorical variables: the P value was calculated by the weighted chi-square test.

Table 3
Association between SPISE and risk of heart failure in different subgroups.Adjusted for age, gender, race, marital status, education level, PIR, eGFR, UACR, uric acid, fasting glucose, obesity, CKD, rheumatoid arthritis, use of diabetic agents, hemoglobin, stroke, heart attack, CHD, and angina were adjusted.