Obesity paradox in patients with non‐ST‐elevation myocardial infarction undergoing percutaneous coronary intervention: A nationwide retrospective observational study

Abstract

Introduction

The prevalence of obesity in the United States is high, with over 35% of the population classified as obese. Obesity, along with the lifestyle factors contributing to it, is considered a major risk factor for heart disease. However, research studies have reported controversial findings regarding its impact on morbidity and mortality.

Methods

Data was obtained from the Nationwide Inpatient Sample database from January 2016 to December 2020. The study included adult patients with a principal discharge diagnosis of NSTEMI who underwent PCI during hospitalization. They were divided into obese (BMI > 30) and not obese (BMI < 30). The primary outcome was inpatient mortality. Secondary outcomes were cardiogenic shock, cardiac arrest, arrhythmias, and acute renal failure, as well as the need for transfusions, pressors, ventilators, and ECMO.

Results

We analyzed 448,424 hospitalizations for NSTEMI, of which 172,665 patients underwent PCI. Among these, 39,548 (22.9%) were obese (BMI >30). Obesity was associated with lower in-hospital mortality compared to non-obese patients (1.2% vs. 1.8%; adjusted odds ratio [aOR]: 0.82; 95% CI: 0.74–0.90; p < 0.001), after adjustment for age, sex, race, and Charlson Comorbidity Index. Obese patients also had lower rates of cardiogenic shock (2.7% vs. 3.3%; aOR: 0.84; 95% CI: 0.78–0.90; p < 0.001), cardiac arrest (1.3% vs. 1.6%; aOR: 0.86; 95% CI: 0.78–0.95; p < 0.005), and ventricular arrhythmias (5.2% vs. 5.7%; aOR: 0.93; 95% CI: 0.88–0.98; p < 0.005). They also required fewer blood transfusions (1.9% vs. 2.3%; aOR: 0.90; 95% CI: 0.83–0.98; p < 0.05) and mechanical ventilation (2.9% vs. 3.2%; aOR: 0.91; 95% CI: 0.85–0.98; p < 0.05). No significant differences were observed in dialysis, vasopressor use, or ECMO. However, obese patients had higher rates of acute kidney injury (15.0% vs. 13.5%; aOR: 1.25; 95% CI: 1.21–1.29; p < 0.001) and acute respiratory failure (9.1% vs. 8.2%; aOR: 1.17; 95% CI: 1.13–1.23; p < 0.001).

Conclusion

The “obesity paradox” suggests that despite obesity's role in cardiovascular disease risk, individuals with obesity may have a survival advantage during acute cardiovascular events and certain surgeries. Our study aligns with these findings, underscoring the need for further research to comprehend the underlying pathophysiological mechanisms.

1. Introduction

Among the best-established risk factors for cardiovascular disease is obesity, as it causes hypertension, diabetes, and dyslipidemia [1], [2]. However, the “obesity paradox” suggests that among patients with acute cardiovascular circumstances, obese patients are more likely to survive than non-obese patients [3]. The paradox challenges conventional beliefs about the health impact of obesity and questions its influence on clinical outcomes [1], [4].

Obesity remains a prevalent condition in the United States, with over 35% of the population classified as obese, and is well-established as a major risk factor for cardiovascular disease. However, the relationship between obesity and clinical outcomes in patients with acute coronary syndrome, specifically those diagnosed with non-ST-elevation myocardial infarction (NSTEMI) undergoing percutaneous coronary intervention (PCI), remains controversial. This nationwide retrospective observational study was undertaken to explore the impact of obesity on a range of outcomes in NSTEMI patients. We hypothesized that, despite the higher prevalence of comorbidities among obese individuals—such as hypertension, diabetes, and chronic kidney disease—obesity might be associated with improved in-hospital outcomes in this clinical context, a phenomenon known as the “obesity paradox.” By analyzing 172,665 NSTEMI patients, including 39,548 classified as obese, we aimed to provide further insight into the potential effects of obesity in acute cardiovascular care, as well as clarify its influence on in-hospital mortality, length of stay, hospitalization costs, and adverse events (Table 1).

Table 1.

Baseline characteristics of patients with NSTEMI undergoing PCI stratified by BMI.

Characteristic	Total population (N = 172,665)	Obese (n = 39,548 [22.9%])	Not obese (n = 133,117 [77.1%])	P-value
Demographics
Age, mean, y	65.72	61.88	66.87	<0.001
Sex
Male,%	65.8	61.6	67.0	<0.001
Female %	34.2	38.4	33.0
Race, %
White	72.8	73.4	72.7	<0.001
Black	10	11.2	9.6
Hispanic	7.9	7.9	7.9
Others	9.3	7.5	9.8
Household income category for patient's zip code by percentile, median, %
0–25th	30.1	30.9	29.8	<0.001
26th–50th	27.9	28.6	27.6
51th–75th	23.9	24.2	23.8
76th–100th	18.1	16.3	18.7
Primary expected payer
Medicare	54.8	47.2	57.1	<0.001
Medicaid	8.9	10.5	8.5
Private insurance	28.4	33.8	26.8
Self-pay	4.5	5.0	4.3
No charge	0.4	0.5	0.4
Other	2.9	3.1	2.9
Admission day, %
Weekdays	74.7	74.5	74.8	0.18
Weekends	25.3	25.5	25.2

Despite obesity being linked to an increased risk of severe comorbidities, our findings align with previous studies suggesting that obesity paradoxically results in decreased in-hospital mortality among critically ill patients [3], [4]. Furthermore, obese patients exhibited lower incidences of cardiogenic shock, cardiac arrest, and arrhythmias, as well as a reduced need for critical lifesaving procedures such as mechanical ventilation and transfusions [3]. However, obesity was associated with a higher incidence of acute renal failure and respiratory failure during hospitalization [4], [5] (Table 2).

Table 2.

Comorbidities of patients with NSTEMI undergoing PCI stratified by BMI.

Comorbidities	Total population (N = 172,665)	Obese (n = 39,548 [22.9%])	Not obese (n = 133,117 [77.1%])	P-value
HLD	74.5	80.1	72.8	<0.001
Hypertension	38.1	43.1	36.7	<0.001
Diabetes mellitus
Uncomplicated	18.5	22.6	17.2	<0.001
Complicated	25.1	35.6	22.0	<0.001
Hypothyroidism	12.1	12.9	11.9	<0.001
History of COPD	20.1	22.4	19.4	<0.001
Peripheral vascular disease	11.0	9.0	11.6	<0.001
Chronic kidney disease	22.1	23.8	21.6	<0.001
Anemia	13.9	13.7	14.0	0.8
Pulmonary vascular disease	4.4	4.8	4.3	<0.001
History of smoking cigarettes	19.8	19.2	20.0	<0.001
History of alcohol use	2.8	2.3	3.0	<0.001
History of drug use	2.2	2.0	2.2	0.006
Family history of CAD	17.4	19.6	16.7	<0.001

The mechanisms underlying the obesity paradox remain unclear. One hypothesis suggests that greater metabolic reserves in obese patients may provide a protective effect during acute illness, mitigating the catabolic impact of critical conditions [6]. Another explanation involves altered immune and inflammatory responses in obesity, which may help reduce myocardial injury and enhance recovery following cardiovascular events [5]. Although short-term survival appears to be improved in some obese patients, obesity remains a major risk factor for the development of long-term cardiovascular morbidity and mortality [1], [2] (Table 3).

Table 3.

Hospital characteristics where PCI was performed stratified by BMI.

Hospital characteristic	Total population (N = 172,665)	Obese (n = 39,548 [22.9%])	Not obese (n = 133,117 [77.1%])	P-value
Hospital bed size, %
Small	55.1	60.9	53.0	0.07
Medium	20.2	4.3	25.8
Large	24.7	34.8	21.2
Hospital location, %
Rural	50.5	56.5	48.5	0.19
Urban nonteaching	22.5	30.4	19.7
Urban teaching	27.0	13.1	31.8
Hospital region, %
Northeast	5.6	0.0	7.6	0.47
Midwest	16.9	5.6	11.2
South	47.2	52.2	45.5
West	30.3	26.1	31.8

With the growing prevalence of obesity, understanding its impact on cardiovascular outcomes is paramount. Future research should focus on refining risk assessment models, exploring pathophysiological mechanisms, and identifying patient subgroups who may benefit from targeted interventions [2], [3]. While the obesity paradox remains a significant clinical observation, it is critical to continue promoting prevention strategies aimed at improving overall cardiometabolic health [1] (Table 4).

Table 4.

In-hospital outcomes when adjusted for age, sex, race and Charlson's Comorbidity Index.

Outcome	Obese	Non-obese	aOR (95% CI)	P-value
In-hospital mortality, no. (%)	1.2	1.8	0.82 (0.74–0.90)	<0.001
Length of stay, mean	3.70	3.61		<0.001
Cost of hospitalization	109,051	110,035		0.09
Complications
Acute respiratory failure	9.1	8.2	1.17 (1.13–1.23)	<0.001
Acute renal failure	15.0	13.5	1.25 (1.21–1.29)	<0.001
Cardiac arrest	1.3	1.6	0.86 (0.78–0.95)	0.002
Ventricular arrhythmias	5.2	5.7	0.93 (0.88–0.97)	0.003
Interventions
New requirement for dialysis	3.8	3.7	0.95 (0.89–1.02)	0.13
Mechanical ventilation	2.9	3.2	0.92 (0.86–0.98)	0.012
ECMO	0.1	0.1	0.58 (0.37–0.90)	0.016
Pressor use	0.6	0.7	0.87 (0.75–1.01)	0.073
Blood transfusion	1.9	2.3	0.91 (0.82–0.99)	0.02
Tracheostomy	0.1	0.1	0.98 (0.71–1.36)	0.92

2. Methods

2.1. Study design and data source

We conducted a retrospective cohort study using data from the Nationwide Inpatient Sample (NIS) database for the period spanning January 1, 2016, to December 31, 2020. The NIS, maintained by the Healthcare Cost and Utilization Project (HCUP), represents the largest publicly available, all-payer inpatient healthcare database in the United States. It provides a stratified sample encompassing approximately 20% of all hospital discharges nationwide, enabling weighted estimates for national trends and outcomes in hospitalized populations.

2.2. Study population

We have identified adults (aged ≥18 years) with the principal discharge diagnosis of non-ST-elevation myocardial infarction (NSTEMI) who underwent percutaneous coronary intervention (PCI) during their hospitalization. Patients were identified using the International Classification of Diseases, 10thRevision (ICD-10) diagnosis and procedure codes. Patients were stratified into two groups based on obesity status, as determined by the presence or absence of an ICD-10 diagnosis code for obesity. Continuous BMI values were not available in the NIS database making non-linear modeling analysis difficult and an acknowledged limitation of study.

2.3. Exclusion criteria

To enhance cohort specificity and reduce confounding, we excluded patients younger than 18 years, those with missing or incomplete data on key variables such as mortality, obesity status, or PCI status, and patients with concurrent STEMI diagnoses to ensure NSTEMI-specific outcomes. Additionally, individuals who had undergone coronary artery bypass grafting (CABG) during the same hospitalization were excluded to isolate PCI-related outcomes.

2.4. Study outcomes

The primary outcome of interest was inpatient mortality and total length of stay (LOS)–associated hospitalization costs. Secondary outcomes included the development of cardiogenic shock, cardiac arrest, and arrhythmias, such as ventricular fibrillation, ventricular tachycardia, and atrial fibrillation. In addition, we examined assessed for development of acute renal failure, the need for blood transfusions, vasopressor support, mechanical ventilation, and the use of extracorporeal membrane oxygenation (ECMO). To ensure accurate classification all outcomes were identified using their respective ICD-10 diagnosis and procedure codes.

2.5. Statistical analysis

Statistical analyses were performed using Statistical Package for Social Sciences (SPSS). Categorical variables were compared using the chi-square test or Fisher's exact test, as appropriate. Event frequencies were calculated for each outcome and stratified by obesity status. Weighted analyses were performed to obtain nationally representative estimates. Comparative analyses between obese and non-obese cohorts were conducted to assess differences in clinical outcomes and resource utilization. Multivariable logistic regression models were constructed to adjust for potential confounder, including age, sex, comorbidities, etc. A two-tailed p-value <0.05 was considered statistically significant.

3. Results

3.1. Patient characteristics

Among 172,665 patients with non–ST-elevation myocardial infarction (NSTEMI) undergoing percutaneous coronary intervention (PCI), 39,548 (22.9%) were classified as obese. The overall mean age was 65.7 years; obese patients were significantly younger than non-obese patients (61.9 vs. 66.9 years, p < 0.001). The majority of patients were male (65.8%), though a higher proportion of females was observed in the obese group compared to the non-obese group (38.4% vs. 33.0%, p < 0.001). Significant differences were noted in racial distribution, income by ZIP code quartile, and primary expected payer between obese and non-obese patients (p < 0.001 for all), reflecting sociodemographic variation.

3.2. Comorbidities

Obese patients had a greater burden of comorbid conditions. Rates of hyperlipidemia (80.1% vs. 72.8%), hypertension (43.1% vs. 36.7%), uncomplicated diabetes mellitus (22.6% vs. 17.2%), and complicated diabetes mellitus (35.6% vs. 22.0%) were all significantly higher in obese patients (p < 0.001 for all). Additionally, obesity was associated with increased prevalence of hypothyroidism, chronic obstructive pulmonary disease (COPD), and chronic kidney disease. In contrast, peripheral vascular disease was more prevalent in non-obese patients (11.6% vs. 9.0%, p < 0.001).

3.3. Hospital characteristics

Hospital-level data showed a trend toward more obese patients being treated in large-bed (bariatric) hospitals (34.8% vs. 21.2%, p = 0.07), though this did not reach statistical significance. Obese patients were more frequently admitted to rural and urban non-teaching hospitals, compared to non-obese patients. No significant differences were found in geographic distribution across hospital regions (Northeast, Midwest, South, and West).

3.4. In-hospital outcomes

Following multivariable adjustment, obese patients demonstrated lower in-hospital mortality compared to non-obese patients (adjusted odds ratio [aOR] 0.82; 95% confidence interval [CI], 0.74–0.90; p < 0.001). The mean length of hospital stay was slightly longer in the obese group (3.70 vs. 3.61 days, p < 0.001), although total hospitalization costs were not significantly different (p = 0.09).

Obese patients had higher adjusted rates of acute respiratory failure (aOR 1.17; 95% CI, 1.13–1.23; p < 0.001) and acute renal failure (aOR 1.25; 95% CI, 1.21–1.29; p < 0.001). However, they had significantly lower rates of cardiac arrest and ventricular arrhythmias. The use of mechanical ventilation and extracorporeal membrane oxygenation (ECMO) was associated with more favorable outcomes among obese patients. In contrast, interventions such as dialysis, vasopressor administration, and blood transfusion demonstrated mixed associations, highlighting the complexity of obesity's impact on procedural outcomes and post-PCI complications.

4. Discussion

In this nationwide retrospective observational study, we conducted an assessment encompassing 172,665 individuals diagnosed with NSTEMI who underwent PCI, with 39,548 categorized as obese. Our investigation primarily centered on the obese cohort vs non-obese classified by BMI. We examined various outcomes including in-hospital mortality, length of hospital stays, and hospitalization expenses. Furthermore, we scrutinized the occurrence of adverse events such as acute respiratory failure, acute renal failure, cardiac arrest, and ventricular arrhythmias alongside monitoring medical interventions such as mechanical ventilation, extracorporeal membrane oxygenation (ECMO), dialysis, vasopressor administration, and blood transfusions. Our analysis notably unveiled a notable phenomenon identified as the “obesity paradox” within NSTEMI patients subjected to PCI. Despite the higher prevalence of comorbidities among the obese subset, including hypertension, hyperlipidemia, diabetes mellitus, COPD, and CKD, which conventionally portend unfavorable clinical outcomes, our findings contribute to the emerging body of evidence suggesting that obesity, notwithstanding its recognized role as a significant cardiovascular risk factor, may confer a survival benefit within specific clinical contexts [1], [2].

Many prominent observations from this study included the lower in-hospital mortality rate, lower risk of cardiogenic shock, lower risk of cardiogenic arrest, and lower risk of ventricular arrythmias among obese patients compared to their non-obese counterparts. They also required less lifesaving treatments such as mechanical ventilation and blood transfusion during hospitalization. This is contrasted with the observation of more acute presentations such renal and respiratory failure.

These findings align with prior studies that have reported similar trends in various cardiovascular conditions, including heart failure, acute coronary syndrome, and cardiac arrest [3]. Similar paradoxical trends have been observed in research of obesity within stroke patients, where elevated BMI was associated with reduced in-hospital mortality and shorter hospital stays [4]. Research has shown a paradoxical U-shaped relationship between obesity and mortality across cardiovascular diseases, notably including myocardial infarction and heart failure. Suggesting that individuals with mild to moderately higher body mass index may exhibit similar or even lower short- and long-term mortality rates [1], [4]. In line with this perspective, a nationwide data analysis of stroke patients revealed a reverse J-shaped association between BMI and in-hospital mortality. Specifically, elevated BMI emerged as an independent predictor of reduced in-hospital mortality and shorter hospital stays among patients admitted for stroke during the study period [4]. These findings underscore the complexity of the relationship between obesity and clinical outcomes, suggesting nuanced implications for patient management and risk stratification. The inverse relationship between obesity and mortality raises intriguing questions about the underlying mechanisms that may contribute to this phenomenon.

Several hypotheses have been proposed to explain the obesity paradox. One possible explanation is the presence of greater metabolic reserves in obese individuals compared to non-obese individuals, which may provide a protective effect during acute illness [6]. Adipose tissue, particularly visceral fat, acts as an energy reservoir, potentially mitigating the catabolic effects of critical illness. Additionally, obese patients may exhibit distinct hemodynamic responses, such as higher circulating catecholamine levels and increased cardiac output, which could improve tissue perfusion during ischemic events [1], [5]. Furthermore, altered inflammatory and immune responses observed in obesity may play a role in the observed outcomes. Adipose tissue secretes adipokines and cytokines, which modulate inflammation and immune function. Interestingly, chronic low-grade inflammation in obese individuals has been hypothesized to provide a protective effect against acute inflammatory insults, such as those seen in myocardial infarction [6].

Despite the various mechanisms postulated to explain the obesity paradox, inconsistencies remain across studies. Obesity stands as a recognized independent risk factor for chronic health conditions, such as hypertension, type 2 diabetes, and cardiovascular diseases. Nevertheless, emerging research suggests heightened survival rates in specific clinical contexts concerning adverse events associated with these conditions [1], [2]. However, primary prevention remains paramount, emphasizing the importance of reducing obesity rates to improve long-term health outcomes. The role of BMI as an independent predictor of mortality remains controversial, with some studies suggesting that factors such as age at presentation and the intensity of treatment interventions may bias results in favor of the paradox [3]. Another possibility is that the paradox is driven by worse outcomes in low BMI groups, something that prior studies have analyzed. For example, low BMI patients—particularly older individuals—may be more vulnerable to complications from antithrombotic therapy, increasing cardiovascular mortality and potentially inflating favorable mortality data among obese populations [7], [8].

The obesity paradox reflects limitations of BMI-based classification, as BMI does not incorporate body composition or fat distribution—factors strongly associated with cardiovascular risk and mortality. Previous studies have used A Body Shape Index (ABSI) to overcome this limitation, defined as waist circumference (m) / (BMI^(2/3) × height (m)^(1/2)) [9]. Studies using ABSI demonstrate a more linear association between obesity and mortality rather than the U-shaped relationship observed with BMI [9], [10]. This suggests that the U-shaped curve seen in obesity paradox studies relying on BMI may be partially attributable to measurement limitations rather than a true protective effect. Furthermore, ABSI may more accurately reflect central adiposity, which is strongly linked to inflammation, diabetes, and adverse cardiovascular outcomes [10], [11].

Interpreting these findings within the context of existing literature is crucial, as prior work has demonstrated that the obesity paradox represents a non-linear relationship. The paradoxical benefit appears most evident in patients with mild to moderate obesity, while it is attenuated or absent in those with severe obesity [12]. Due to the structure of the Nationwide Inpatient Sample (NIS), which lacks continuous BMI data, we were unable to perform non-linear analyses such as BMI quartiles or spline-based modeling. As a result, we cannot conclude that obesity across all severity levels confers a protective paradoxical effect. Therefore, further robust and well-controlled studies are needed to better characterize this phenomenon, particularly in populations with advanced cardiovascular disease or those undergoing invasive cardiac procedures [2].

Chronic inflammation in obesity is characterized by elevated cytokines and inflammatory markers such as tumor necrosis factor–α and C-reactive protein [13], [14]. Persistent low-grade inflammation may prime the immune system for acute physiological stress. This adaptive response may confer short-term benefits by facilitating improved responses to high-impact stressors such as sepsis or myocardial infarction, potentially limiting tissue injury and excessive immune activation [15], [16]. Adipose-derived hormones, including adiponectin and leptin, may further contribute to vascular repair and attenuation of oxidative stress [17]. Additionally, immune cells such as macrophages may undergo a shift toward a reparative (M2) phenotype in obese individuals, which has been associated with reduced ischemia-reperfusion injury during acute cardiovascular events [18].

Although inflammation is a major driver of chronic disease in obesity, it may also provide short-term protective effects by buffering against acute metabolic stress [12], [19]. A modest inflammatory response during myocardial infarction may promote more effective healing while limiting myocardial damage. These mechanisms may help explain why obese patients, despite a higher burden of comorbidities, often demonstrate favorable outcomes in critical care settings, including lower rates of cardiogenic shock and ventricular arrhythmias [15].

It is essential to recognize that obesity is a heterogeneous and multifactorial condition with diverse metabolic and physiological consequences. While our study demonstrated a lower risk of select adverse outcomes among obese patients—including cardiogenic shock, cardiac arrest, and ventricular arrhythmias—we also observed higher rates of acute renal failure and acute respiratory failure. These findings highlight the nuanced relationship between obesity and cardiovascular outcomes and emphasize the importance of contextual interpretation.

This study has several limitations. First, its retrospective observational design using administrative data precludes causal inference and is susceptible to residual confounding despite multivariable adjustment. Second, reliance on ICD coding for obesity and comorbid conditions may introduce misclassification bias. The definition of obesity was based on ICD codes rather than direct anthropometric measurements, precluding stratification by obesity class and limiting assessment of heterogeneity within obese subgroups.

Administrative undercoding represents an additional limitation, as national estimates suggest an obesity prevalence of approximately 42% among U.S. adults, whereas only 23% of patients in our cohort were coded as obese [20]. Despite this potential bias, our findings continued to demonstrate evidence supporting the obesity paradox, reinforcing the robustness of the observed association between obesity classification and in-hospital mortality.

Additionally, underweight individuals were not included, potentially limiting insight into the full spectrum of weight-related risk. Key clinical variables—including left ventricular ejection fraction, infarct size, medication use, and laboratory data—were unavailable, and post-discharge outcomes could not be assessed. Despite these limitations, the large, nationally representative nature of the dataset enhances the generalizability of our findings.

In conclusion, our study demonstrated an association between obesity classification and lower in-hospital mortality among patients with NSTEMI undergoing PCI, supporting the existence of the obesity paradox, though this relationship may vary by obesity severity. As global obesity rates continue to rise, understanding the complex interplay between adiposity, inflammation, and cardiovascular outcomes remains essential. Future studies should focus on mechanistic pathways and refined adiposity measures—such as ABSI—to better characterize risk and guide preventive and therapeutic strategies [1], [2], [4], [9], [10], [11].

CRediT authorship contribution statement

Sean Moore: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Aditya Thakkar: Writing – original draft, Visualization, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Shahnawaz Notta: Writing – review & editing, Writing – original draft. Jacob Snipp: Writing – review & editing. Colin McGuire: Writing – review & editing. Venkata Vedantam: Writing – original draft, Supervision. Manar H. Jbara: Writing – review & editing, Writing – original draft, Supervision.

Funding sources

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Declaration of competing interest

The authors declare no conflicts of interest.