The Dual Impact of Obesity and Diabetes on Heart Failure Readmissions: Results from the Congestive Heart Failure Registry at Harlem Hospital Center Open Access

Congestive heart failure (CHF) affects 6.5 million individuals in the USA [1] and is the leading cause of hospitalization for those aged 65 and older, contributing to over $31 billion in annual healthcare costs [2, 3]. Despite advancements in care, CHF prevalence continues to rise [4, 5], with key risk factors including diabetes, myocardial infarction, age, smoking, dyslipidemia, and obesity [6]. Obesity, affecting 42.5% of US adults [7], increases the risk of CHF [8‒10] but is paradoxically linked to improved survival rates in CHF patients, known as the “obesity paradox” [11‒13]. Similarly, type II diabetes (T2D) affects 40% of CHF patients and increases CHF risk through hyperglycemia-induced cardiac damage [14, 15]. The relationship between hemoglobin A1c (HbA1c) levels and mortality in patients with diabetes and CHF is debated, with some studies suggesting an inverse correlation [16‒18] and others finding no significant impact on short-term mortality [19]. Despite evidence that obesity may benefit CHF patients in terms of morbidity and mortality [20], these advantages do not consistently reduce hospital readmission rates, as obese patients often exhibit higher 30-day and 6-month readmissions [20]. Interestingly, patients who are morbidly obese may not experience increased readmissions [20]. Limited research also indicates that T2D prolongs hospital stays and raises 30-day readmissions [21‒23]. This study investigated the interactions between obesity, diabetes, and CHF readmissions with an aim to improve treatment strategies and patient outcomes.

This study was conducted as project of CHF Registry at Harlem Hospital Center (HHC) under the auspices of Department of Cardiology and was approved by the Institutional Review Board (IRB). Data were collected from patients diagnosed with CHF between January 1, 2019, and January 1, 2021.

Patients with a primary diagnosis of CHF were identified through multiple databases at HHC, including the electronic medical record system (EPIC), a code list managed by the Care Management Department, and the monthly billing list obtained from the “Get with the Guidelines-Heart Failure (GWTG-HF)” program. Diagnosis confirmation was achieved through a comprehensive review of cardiology consultations, medical team notes, laboratory test results, and echocardiograms. Inclusion criteria comprised all patients with a new diagnosis of CHF or CHF-related complications during the study period. Exclusion criteria included underweight patients (body mass index [BMI] <18.5 kg/m²), as this subgroup comprised only 7 patients, limiting statistical analysis. After applying the inclusion and exclusion criteria, eligible patients were included in the analysis.

The data from patients were extracted from EPIC and entered into a standardized Google Form-based CHF registry encompassing 312 variables categorized into five domains: demographics, comorbidities, laboratory and imaging results, miscellaneous variables, and COVID-19 readmissions. This study specifically focused on the relationship between BMI, T2D mellitus, and 30-day hospital readmission rates. Patients were stratified into BMI categories based on CDC guidelines: normal (18.5 to <25 kg/m²), overweight (25 to <30 kg/m²), and obese (≥30 kg/m²). HbA1c levels were used to assess glycemic control and categorized as normal (≤5.6%), prediabetic (5.7–6.4%), or diabetic (≥6.5%). The primary outcome was 30-day readmission after discharge. The independent variables were BMI and diabetes diagnosis, with a secondary focus on HbA1c levels as an indicator of glycemic control.

Baseline demographic and clinical characteristics were summarized using descriptive statistics. Continuous variables were presented as mean ± standard deviation (SD) and compared using Wilcoxon rank-sum tests, while categorical variables were expressed as percentages and compared using chi-square tests. Logistic regression models were employed to evaluate the association between BMI categories, diabetes status, and 30-day readmission. Multicollinearity among independent variables was assessed using variance inflation factors, with a variance inflation factors >10 indicating high multicollinearity. Odds ratios (ORs) and 95% confidence intervals (CIs) were reported for all logistic regression models. Statistical significance was set at p < 0.05. Analyses were performed using SPSS version 25 (SPSS Inc., Chicago, IL, USA).

Of the 390 CHF registry patients with BMI data, 70.2% were above the recommended BMI range, with 24.6% overweight and 45.6% obese, while 28.4% had a normal BMI. Individuals with obesity had a younger mean age (60.7 years) compared to those with normal weight (67.8 years) and those classified as overweight (67.4 years), with a statistically significant difference (p < 0.001). Ethnic and socioeconomic disparities were observed, with Hispanic or Latino patients and individuals experiencing homelessness being more likely to have a normal BMI (p = 0.040 and p = 0.006, respectively). Overweight individuals had the lowest mean left ventricular ejection fraction at 34.5%, compared to 36.9% in individuals with normal weight and 38.3% in those with obesity (p = 0.0009).

Common comorbidities were examined by BMI category (Table 1). Obesity was associated with higher rates of T2D (55.1%) compared to overweight (67.7%) and normal-weight patients (49.1%, p = 0.055). Similarly, obstructive sleep apnea was more prevalent in obese patients (29.2%) compared to those with normal BMI (2.6%) and overweight groups (5.2%, p < 0.001).

Readmission rates within 30 days varied by BMI (Table 2). Overweight patients had the lowest readmission rates (15.6%), followed by obese (10.7%), and normal-weight patients (22.4%, p = 0.022). While individuals with obesity were less likely to be readmitted within 30 days for CHF-related issues, this difference approached but did not achieve statistical significance (p = 0.052). The influence of a diabetes diagnosis on CHF outcomes is presented in Table 3. Readmission rates within 30 days were similar across normal, prediabetic, and diabetic groups (17.6%, 14.3%, and 17.6%, respectively, p = 0.733). No significant differences were found in readmission rates or other outcomes by the diabetes status.

The combined effects of obesity and diabetes on CHF outcomes are shown in Table 4. Patients with obesity alone were less likely to experience 30-day readmissions (4.4%) compared to those with diabetes (18.8%) or both conditions (17.2%, p = 0.006). Similarly, CHF-related 30-day readmissions were lowest in the obesity-only group (1.1%) and highest in patients with both comorbidities (16.1%, p = 0.014). Patients with obesity alone also had fewer all-time hospitalizations (>5: 3.3%) compared to those with diabetes (14.1%) or both conditions (24.1%, p < 0.001).

Logistic regression models for 30-day readmissions, adjusted for confounders, are shown in Table 5. Compared to CHF-only patients, those with obesity, diabetes, or both did not exhibit significant differences in readmission odds (obesity: OR 0.86, 95% CI: 0.39–1.93; diabetes: OR 0.53, 95% CI: 0.44–1.26; obesity and diabetes: OR 0.90, 95% CI: 0.39–2.09).

The aim of this study was to evaluate the presence of the obesity paradox in HHCs CHF population and assess whether it is influenced by the presence of diabetes. While previous research has suggested an association between obesity and reduced morbidity and mortality risks, findings on readmissions remain inconsistent. Cox et al. [21] reported no link between increased BMI and all-cause readmissions despite lower mortality, whereas others observed reduced 30-day readmission rates among morbidly obese CHF patients [20]. Conversely, diabetes has been consistently linked to increased 30-day readmissions in CHF populations [19, 22]. Recent work underscores that the obesity paradox may reflect confounding by cardiorespiratory fitness or unmeasured metabolic reserves, rather than a true protective effect of BMI – particularly in the high-risk populations like ours, and that socioeconomic disparities further complicate these relationships [23, 24].

This study’s findings initially supported the obesity paradox, showing fewer overweight and obese patients readmitted compared to normal-weight individuals (p = 0.022, Table 2). These results contradict previous studies suggesting higher 30-day readmission risks in obese patients [20‒22, 25]. Additionally, no significant differences in long-term readmissions, CHF-specific readmissions, or outpatient follow-up metrics were observed across weight categories, indicating that the reduced 30-day readmissions were not attributable to better post-discharge care or prior hospitalization history. The study proposes that the short-term benefits observed in obese patients might be due to factors such as metabolic reserves, hemodynamic adaptations, or their clinical management strategies.

However, several alternative explanations may underlie these findings. Reverse causation (where lower BMI reflects frailty or advanced illness) could explain higher readmission rates in patients with normal-weight. BMI also fails to capture body composition, which may affect the recovery. Socioeconomic factors like healthcare access, nutrition, and support systems may confound the results. Patients with obesity might receive more intensive care due to perceived risk, improving short-term outcomes. Unmeasured variables such as health literacy and discharge planning may also influence readmission. These factors underscore the need for cautious interpretation of the obesity paradox.

Stratifying by diabetes diagnosis (Table 3), no significant differences in 30-day or long-term readmissions emerged, diverging from studies linking diabetes to increased CHF readmissions [22, 25, 26]. However, analysis of combined obesity and diabetes status (Table 4) revealed significant differences in all-cause 30-day readmissions (p = 0.006). Notably, patients with obesity alone had lower 30-day readmissions compared to CHF-only patients (8.9% vs. 31.1%), aligning with the obesity paradox. Conversely, the benefit diminished in patients with both obesity and diabetes, where readmissions (33.3%) were similar to those with neither condition (31.1%). These findings are consistent with Adamopoulos et al. [19] who reported an absence of the obesity paradox in CHF patients with diabetes, albeit focused on mortality rather than readmissions.

Despite these trends, adjustments for confounders such as age, ethnicity, residence, and obstructive sleep apnea (Table 5) revealed no significant differences in 30-day readmission odds based on weight category or comorbidity status. These results challenge the obesity paradox in 30-day hospital readmissions, suggesting that the apparent relationship between weight and readmission risk is likely confounded by other factors. Our results emphasize the need for a comprehensive approach to readmission risk assessment, considering multiple patient characteristics rather than focusing solely on weight or individual comorbidities.

This study has several strengths, including a large sample size of 390 patients and the use of the CHF registry, a comprehensive database with 312 variables derived from electronic medical records of real-world patients admitted to HHC. The population, predominantly from lower income, underserved communities and largely composed of minority groups such as Black or African American and Hispanic individuals, provides valuable insights into the impact of obesity and diabetes on CHF readmissions in demographics heavily affected by chronic illnesses and often underrepresented in research [27].

However, the study’s retrospective design inherently limits the ability to establish causality and poses challenges in fully accounting for the confounding variables. Many patients presented with multiple chronic conditions, and despite efforts to adjust for known confounders, residual confounding remains a major limitation. Furthermore, the registry lacks data on key variables such as the duration of obesity or diabetes, body composition and fat distribution, intentionality of weight loss, and temporal weight changes. These omissions restrict a more nuanced understanding of how these factors may influence the CHF outcomes.

In summary, while this study observed indications of the obesity paradox in 30-day CHF readmissions within the HHC population, these associations did not remain statistically significant after adjusting for confounders. The presence of diabetes further attenuated the observed protective effect of obesity. Future studies with larger sample sizes and more comprehensive control of confounding variables are necessary to clarify these complex relationships.

We would like to express our gratitude to the Academic Affairs Department at Harlem Hospital Center for their support of this publication. We also extend our sincere appreciation to Damian Kurian, Anna Hughes, Lily McCann, Natalia Ionescu, Ivrose Janvier, Michelle Thomas, Cristina Bradley, Mohammad Islam, Yana Ruban, Nakia Diallo, and Yurika Brown from the Division of Cardiology for their invaluable contributions and assistance. Additionally, we thank the Institute of Human Nutrition at Columbia University for providing the opportunity to undertake this research as part of the master’s program.

The BRANY IRB determined that this research submission is exempt from IRB review under category 4(iii) of 45 CFR 46.104(d), confirming that it involves the use of identifiable health information regulated for healthcare operations or public health activities, provided all procedures comply with applicable state and federal laws.

The authors declare that they have no competing interests.

This research work was not funded.

Mariel Magdits developed the idea and contributed to the draft of the manuscript, registry development, and data collection. Asmaa AlShammari contributed to the idea development, served as the main writer, and participated in registry development and data collection. Rosemarie Majdalani contributed to the idea development and was involved in registry development and data collection. Sriraman Devarajan served as the biostatistician. Farbod Raiszadeh served as the principal investigator of CHF Registry and provided mentorship and guidance in all stages of the project. All authors contributed to the draft of the work and table preparation. All authors agree to be accountable for all aspects of this study in ensuring that questions related to the accuracy and integrity of any part of the study are appropriately investigated and resolved.

The data cannot be disclosed due to IRB regulations and concerns about protecting participants’ privacy and confidentiality.