Association between body mass index and clinical outcomes in patients with acute myocardial infarction and reduced systolic function: Analysis of PARADISE‐MI trial data

Abstract

Aims

The relationship between body mass index (BMI) and clinical outcomes in patients with cardiovascular disease, including acute heart failure (AHF) and acute myocardial infarction (AMI), remains debated. This study investigates the association between BMI and clinical outcomes within the PARADISE‐MI cohort, while also evaluating the impact of angiotensin receptor–neprilysin inhibitor (ARNI) versus angiotensin‐converting enzyme inhibitor (ACE‐I) treatment on this relationship.

Methods and results

The analysis included 5589 patients from the PARADISE‐MI study with available baseline BMI data. The cohort comprised patients with AMI and pulmonary congestion and/or left ventricular ejection fraction ≤40%. Patients were categorized into six World Health Organization BMI subgroups. The primary outcome of interest was the composite endpoint of cardiovascular death, heart failure (HF)‐associated hospitalization, and outpatient symptomatic HF episodes. The mean baseline BMI of the cohort was 28.1 ± 5.0 kg/m². The lowest rate of the primary composite endpoint (6.2/100 patient‐years) was observed in overweight patients (BMI 25–29.9 kg/m²), while the highest rates were found in the lowest and highest BMI subgroups (8.4/100 patient‐years for BMI <18.5 kg/m² and 9.7/100 patient‐years for BMI >40 kg/m²). There was no significant interaction between BMI and the treatment effect of ARNI versus ACE‐I on the primary composite outcome (p = 0.73). Additionally, no significant differences in the incidence of adverse events or serious adverse events were noted across the BMI subgroups.

Conclusions

In AMI with AHF patients, the relationship between BMI and the primary composite outcome is non‐linear, with the lowest event rates observed in overweight individuals. Outcomes and safety profiles for ARNI and ACE‐I treatments were similar across BMI subgroups.

Association between body mass index (BMI) and clinical outcomes in PARADISE‐MI. (A) Histogram for BMI (kg/m²), (B) adverse events for BMI subgroups, and spline model curves for (C) the primary composite outcome and (D) cardiovascular (CV) death by BMI subgroups. Covariates were adjusted for age (years), pulmonary congestion, percutaneous coronary intervention, left ventricular ejection fraction (%) and hypertension.

Introduction

A global obesity epidemic has been raging for decades, with increases in obesity prevalence documented in most countries since the 1980s. ¹ According to a recent report by the World Health Organization (WHO), the prevalence of obesity has tripled since 1975, and in 2016, more than 1.9 billion adults were considered overweight and obese. ² If current trends persist, up to 58% of the world's adult population is expected to be overweight or obese by the year 2030. ³

Body mass index (BMI) is routinely used to assess weight status in epidemiological and clinical research. ⁴ , ⁵ , ⁶ According to the WHO classification, patients are considered underweight when they have a BMI <18.5 kg/m², normal weight with BMI levels 18.5–25 kg/m², overweight with BMI 25–30 kg/m², class I obese with a BMI of 30–35 kg/m², class II obese at BMI 35–40 kg/m², and class III extremely obese at BMI ≥40 kg/m². ⁷ Despite its poor discrimination between muscle mass and fat tissue, BMI is an independent risk factor for various cardiovascular (CV) conditions, including acute coronary syndrome, stroke, sudden cardiac death, atrial and ventricular arrhythmia and heart failure (HF), with either reduced or preserved ejection fraction. ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² Mendelian randomization analysis found a direct link between obesity and the development of HF, whereas other associations, e.g. diabetes mellitus, hypercholesterolemia and hypertension, were indirect, mediated through other pathways, and possibly linked to HF through coronary artery disease. ¹³

Reports on the specific association between BMI and outcomes of acute myocardial infarction (AMI) have been inconsistent. Some studies showed a U‐shaped relationship between BMI and mortality in myocardial infarction patients, with lower risk in overweight and obese patients and higher risk in underweight, normal weight and morbidly obese patients. ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ This ‘obesity paradox’ persisted after adjusting for other prognostic variables. ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ Conversely, other works showed that patients with morbid obesity (BMI ≥40 kg/m²) ²¹ had lower odds of mortality compared to non‐obese patients. ²¹ Stienen et al. ²² found that a mean BMI <26 kg/m² and a BMI decrease during follow‐up were independently associated with CV death in patients with myocardial infarction and left ventricular systolic dysfunction. Of note, most of the studies neither used the full WHO classification nor included underweight or extremely obese patient subgroups.

The Prospective ARNI versus ACE Inhibitor Trial to Determine Superiority in Reducing HF Events after Myocardial Infarction (PARADISE‐MI) ²³ was a double‐blind, active‐controlled, randomized clinical trial designed to determine whether sacubitril/valsartan was superior to ramipril in improving the outcomes of patients with pre‐defined ‘high‐risk’ AMI characteristics (online supplementary Table  S1 ). The current work is a post‐hoc analysis of PARADISE‐MI trial data ²³ designed to characterize the BMI distribution and its relationship with clinical outcomes in patients with AMI and pulmonary congestion and/or systolic left ventricular ejection fraction (LVEF) <40%. The second aim was to assess whether BMI impacted the effect of sacubitril/valsartan versus ramipril treatment.

Methods

Trial design

The design, baseline characteristics, and primary results of the trial were previously published. ²³ , ²⁴ PARADISE‐MI enrolled patients between December 2016 and March 2020. ²³ The trial received institutional review board/ethics committee approval from all participating institutions and each patient provided written informed consent. Overall, 5661 patients with AMI and with LVEF ≤40% and/or pulmonary congestion were randomly assigned to receive either sacubitril/valsartan or ramipril (angiotensin receptor–neprilysin inhibitor [ARNI] or angiotensin‐converting enzyme inhibitor [ACE‐I], respectively) in addition to recommended therapy. Inclusion and exclusion criteria are detailed in online supplementary Table  S1 .

Patients in each cohort were classified based on the six WHO BMI categories. The primary composite endpoint was incidence of CV death, HF‐associated hospitalization or outpatient HF events and secondary endpoint events were incidence of HF hospitalization or outpatient HF.

Statistical analysis

Baseline characteristics were summarized using means and standard deviations or counts and percentages for continuous and categorical variables, respectively. Characteristics were compared across BMI subgroups using ANOVA and Pearson's Chi‐squared tests. Time‐to‐event data were analysed using Cox proportional hazards models (to estimate hazard ratios [HR]) and Poisson models (to estimate incidence rates and rate ratios) and are displayed graphically using Kaplan–Meier curves. Potentially non‐linear associations were assessed by modelling BMI using restricted cubic spline terms with 3 knots. The models were adjusted for age (years), pulmonary congestion, percutaneous coronary intervention, left ventricular ejection fraction (%), hypertension, UN Region.

Additional analyses were conducted using the approaches described above, but with the key exposure variable defined BMI normalized for five pre‐specified geographic regions, i.e. Z‐score by region (i.e. each patient's BMI, subtracted from the region's mean BMI, and then divided by the standard deviation of BMI within that region): Asia/Pacific, Central Europe, Latin America, North America and Western Europe. No adjustments were made for multiple comparisons. A p‐value <0.05 was considered statistically significant. All analyses were conducted using STATA version 16 (Stata Corp., College Station, TX, USA).

Results

Study cohort

After the exclusion of 72 patients with missing baseline BMI data, the final cohort included 5589 patients. Of these, 2802 were assigned to receive sacubitril/valsartan and 2787 patients received ramipril. The median duration of follow‐up was 22 months.

Patient characteristics, comorbidities and geographical origin by body mass index subgroups

Baseline and clinical characteristics by BMI subgroup are presented in Table  1 .

Table 1.

Frequency distribution of baseline characteristics by body mass index subgroup patient

	BMI <18.5 (n = 44)	18.5 ≤  BMI <24.9 (n = 1516)	25 ≤  BMI <29.9 (n = 2313)	30 ≤  BMI <34.9 (n = 1196)	35 ≤  BMI <39.9 (n = 390)	BMI ≥40 (n = 130)	p‐value
Characteristic
Age (years)	67.0 ± 11.2	65.6 ± 11.7	64.0 ± 11.2	62.2 ± 11.3	60.6 ± 11.5	59.2 ± 11.8	< 0.001
Female sex, n (%)	14 (31.8)	384 (25.3)	480 (20.8)	295 (24.7)	114 (29.2)	53 (40.8)	< 0.001
Race, n (%)							< 0.001
Caucasian	24 (54.5)	918 (60.6)	1794 (77.6)	1033 (86.4)	340 (87.2)	114 (87.7)
Other	20 (45.5)	598 (39.5)	519 (22.4)	163 (13.7)	50 (12.8)	16 (12.3)
Region, n (%)							< 0.001
Asia/Pacific	26 (59.1)	504 (33.2)	384 (16.6)	113 (9.4)	37 (9.5)	11 (8.5)
Central Europe	4 (9.1)	296 (19.5)	639 (27.6)	408 (34.1)	114 (29.2)	35 (26.9)
Latin America	2 (4.5)	164 (10.8)	301 (13.0)	132 (11.0)	57 (14.6)	12 (9.2)
North America	4 (9.1)	86 (5.7)	187 (8.1)	143 (12.0)	62 (15.9)	38 (29.2)
Western Europe	8 (18.2)	465 (30.7)	801 (34.6)	400 (33.4)	120 (30.8)	34 (26.2)
Vital signs and inclusion criteria at presentation
Heart rate, bpm	77.2 ± 14.6	76.1 ± 11.9	75.1 ± 11.7	75.9 ± 11.7	76.3 ± 11.6	77.2 ± 10.5	0.04
SBP, mmHg	117.4 ± 12.4	119.0 ± 12.5	120.8 ± 13.3	122.3 ± 13.9	123.2 ± 13.5	126.7 ± 13.8	< 0.001
DBP, mmHg	70.0 ± 12.2	72.6 ± 9.2	73.6 ± 9.4	75.0 ± 10.3	75.4 ± 10.4	76.9 ± 12.0	< 0.001
BMI, kg/m2	17.4 ± 1.1	22.9 ± 1.6	27.4 ± 1.4	32.1 ± 1.4	37.0 ± 1.4	43.6 ± 4.1	< 0.001
LVEF, %	37.1 ± 9.4	36.0 ± 9.8	36.8 ± 9.4	36.3 ± 8.9	36.9 ± 8.7	38.2 ± 9.5	0.03
Pulmonary congestion, n (%)	29 (65.9)	781 (51.5)	1260 (54.5)	658 (55.0)	218 (55.9)	65 (50.0)	0.14
Killip class ≥II, n (%)	27 (64.3)	868 (60.2)	1316 (58.5)	661 (56.4)	20 (54.9)	71 (56.3)	0.25
Medical history, n (%)
Prior MI	5 (11.4)	209 (13.8)	389 (16.8)	205 (17.1)	77 (19.7)	22 (16.9)	0.03
Prior CABG or PCI	4 (9.1)	205 (13.5)	390 (16.9)	215 (18.0)	83 (21.3)	26 (20.0)	< 0.001
Prior stroke	3 (6.8)	62 (4.1)	105 (4.5)	62 (5.2)	23 (5.9)	4 (3.1)	0.82
Hypertension	26 (59.1)	794 (52.4)	1536 (66.4)	864 (72.2)	308 (79.0)	105 (80.8)	< 0.001
Diabetes	10 (22.7)	491 (32.4)	961 (41.5)	589 (49.2)	237 (60.8)	82 (63.1)	< 0.001
Current smoking	9 (20.5)	320 (21.1)	525 (22.7)	237 (19.8)	73 (18.7)	20 (15.4)	0.13
Atrial fibrillation/flutter	7 (15.9)	170 (11.2)	340 (14.7)	184 (15.4)	58 (14.9)	19 (14.6)
Lab test
eGFR, ml/min/1.73 m2	75.5 ± 21.9	72.9 ± 22.4	71.8 ± 22.5	70.4 ± 21.7	71.4 ± 22.8	72.1 ± 24.6	0.08
Clinical and treatment approach, n (%)
STEMI	31 (70.5)	1190 (78.5)	1769 (76.5)	876 (73.2)	281 (72.1)	93 (71.5)	0.007
Reperfusion	39 (88.6)	1340 (88.4)	2062 (89.1)	1075 (89.9)	346 (88.7)	113 (86.9)	0.82
STEMI without reperfusion within 24 h	7 (15.9)	162 (10.7)	187 (8.1)	86 (7.2)	36 (9.2)	8 (6.2)	0.006
Thrombolytic therapy a	2 (4.5)	90 (5.9)	101 (4.4)	38 (3.2)	13 (3.3)	8 (6.2)	0.025
PCI	37 (84.1)	1319 (87.0)	2046 (88.5)	1062 (88.8)	346 (88.7)	110 (84.6)	0.44
Drug‐eluting stent	34 (77.3)	1173 (77.4)	1836 (79.4)	957 (80.0)	307 (78.7)	99 (76.2)	0.55
Location of MI, n (%)							0.80
Anterior	31 (70.5)	1061 (70.0)	1569 (67.8)	808 (67.6)	258 (66.2)	86 (66.2)
Inferior	9 (20.5)	270 (17.8)	428 (18.5)	232 (19.4)	71 (18.2)	26 (20.0)
Other	4 (9.1)	185 (12.2)	316 (13.7)	156 (13.0)	61 (15.6)	18 (13.8)
>1 risk augmenting factors b , n (%)	27 (61.4)	767 (50.6)	1180 (51.0)	656 (54.8)	218 (55.9)	70 (53.8)	0.08
Time to randomization, days	4.6 ± 1.7	4.4 ± 1.7	4.3 ± 1.8	4.2 ± 1.8	4.3 ± 1.7	4.2 ± 1.8	0.20
Medical treatment at randomization, n (%)
DAPT	40 (90.9)	1398 (92.2)	214 (92.6)	1105 (92.4)	356 (91.3)	118 (90.8)	0.93
Beta‐blocker	24 (54.5)	1235 (81.5)	1978 (85.5)	1063 (88.9)	352 (90.3)	122 (93.8)	< 0.001
MRA	20 (45.5)	580 (38.3)	984 (42.5)	498 (41.6)	164 (42.1)	60 (46.2)	0.11
Diuretics	14 (31.8)	653 (43.1)	987 (42.7)	564 (47.2)	197 (50.5)	71 (54.6)	< 0.001
Statin	40 (90.9)	1438 (94.9)	2199 (95.1)	1137 (95.1)	371 (95.1)	124 (95.4)	0.89
ACE‐I/ARB	26 (59.1)	1074 (70.8)	1841 (79.6)	988 (82.6)	341 (87.4)	112 (86.2)	< 0.001

Values are given as n (%) or mean ± standard deviation.

ACE‐I, angiotensin‐converting enzyme inhibitor; ARB, angiotensin receptor antagonist; BMI, body mass index; CABG, coronary artery bypass grafting; DAPT, dual antiplatelet therapy; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; LVEF, left ventricular ejection fraction; MI, myocardial infarction; MRA, mineralocorticoid receptor antagonist; PCI, percutaneous coronary intervention; SBP, systolic blood pressure; STEMI, ST‐elevation myocardial infarction.

^a Thrombolytic therapy above 10% missing data.

^b At least one of eight pre‐specified risk‐augmenting factors: (1) age ≥70 years, (2) diabetes mellitus, (3) previous MI, (4) eGFR <60 ml/min/1.73 m² of body surface area at screening, (5) atrial fibrillation, (6) LVEF <30% associated with the index myocardial infarction, (7) Killip class III or IV, (8) STEMI without reperfusion within 24 h after presentation.

The majority of patients met the criteria for normal weight (n = 1516), overweight (n = 2313) or class I obesity (n = 1196) subgroups. A minority of patients were included in the underweight (n = 44), class II obesity (n = 390) or class III obesity (n = 130) subgroups. Most patients in each BMI subgroup were Caucasian and, in each group, more than 50% were male.

Despite their young age, patients with class III obesity had a higher prevalence of diabetes mellitus and hypertension, while normal weight and underweight patients had the lowest prevalence of these comorbidities. Underweight patients had lower systolic and diastolic blood pressures and higher rates of Killip class ≥II and pulmonary congestion at presentation.

More than 70% of participants in each BMI subgroup were admitted due to ST‐elevation myocardial infarction, commonly with anterior wall infarct. Almost 90% of patients underwent reperfusion within 24 h (Table  1 ). When compared to all other BMI subgroups, underweight patients had a less common medications history involving beta‐blockers, diuretics and ACE‐I/angiotensin receptor blockers.

In the Asia/Pacific region, a greater proportion of patients were underweight or normal weight in contrast to Central Europe, Latin America, North America, and Western Europe, where more patients were overweight or had class II or class III obesity.

Clinical outcomes

A relationship between BMI subgroups and the primary composite endpoint was observed, with the lowest event rates in overweight patients (6.2/100 patient‐years) (Figure  1 , online supplementary Table  S2 ). Class III obese patients experienced significantly higher incidences of primary composite events (9.7/100 patient‐years; HR 1.69, 95% confidence interval [CI] 1.09–2.63, p = 0.018) and secondary endpoint events, including HF hospitalization or outpatient HF (6.2/100 patient‐years; HR 1.79, 95% CI 1.03–3.10, p = 0.039) compared to other subgroups. However, no overall statistical significance in event rates was found across BMI subgroups when considering the overall p‐value (p = 0.12 and p = 0.63 for the primary and secondary endpoints, respectively).

Figure 1.

Incidence of primary and secondary endpoints across body mass index subgroups. (A) Primary endpoints, (B) death from cardiovascular (CV) causes, (C) hospitalization for heart failure (HF), and (D) all‐cause death, by body mass index (kg/m²). Covariates were adjusted for number of CV risk factors, age (years), pulmonary congestion, percutaneous coronary intervention, left ventricular ejection fraction (%) and hypertension.

The higher event rates in higher BMI subgroups were primarily attributed to HF‐associated hospitalizations, indicating a positive correlation between BMI and HF hospitalization. In contrast, the higher rate of primary events in the lower BMI subgroup was largely due to an increased incidence of CV death (Figure  1 , online supplementary Table  S2 ). Additionally, the Z‐score by region‐adjusted overall analysis revealed that geographical subgroups with the lowest and highest BMI had significantly higher rates of both primary and CV death events (p = 0.021 and p = 0.027, respectively; online supplementary Table  S2 ).

Drug safety, efficacy, and side effects across body mass index categories

In each BMI subgroups, a ratio close to 1:1 of patients was treated with ARNI versus ACE‐I

(20:24 for BMI <18.5 kg/m²; 750:766 for BMI 18.5–24.9 kg/m²; 1176:1137 for BMI 25–29.9 kg/m²; 590:606 for BMI 30–34.9 kg/m²; 190:200 for BMI 35–39.9 kg/m² and 66:64 for BMI ≥ 40 kg/m²). No significant differences in primary and secondary clinical outcomes were detected within the BMI subgroups for patients treated with ARNI versus ACE‐I (Figure  2 , online supplementary Table  S3 ). Additionally, no significant differences in the incidence of adverse events or serious adverse events were noted across the BMI subgroups (online supplementary Table  S4 ). With the exception of higher blood potassium levels in underweight and normal weight patients, no other clinically abnormal lab results were noted in the BMI subgroups (online supplementary Table  S5 ).

Figure 2.

Effect of sacubitril/valsartan versus ramipril on primary and secondary endpoints across body mass index subgroups. (A) Primary endpoints, (B) death from cardiovascular (CV) causes, (C) hospitalization for heart failure (HF), and (D) all‐cause death, by body mass index (kg/m²).

Discussion

In our study of 5589 patients with AMI and pulmonary congestion and/or an LVEF <40%, enrolled in the PARADISE‐MI trial, we assessed potential associations between BMI and clinical events in these ‘high‐risk’ AMI patients, focusing on HF and mortality. Additionally, we explored potential differential effects of the randomized therapies ARNI versus ACE‐I across different BMI subgroups.

In our study, most participants had a BMI between 20 and 25 kg/m², with fewer individuals in the higher BMI categories. The bar chart shows no significant difference in adverse event frequencies across BMI categories. Cox proportional hazards models indicated that a lower BMI (<18.5 kg/m²) was linked to a higher and earlier occurrence of both the primary outcome and CV death (Graphical Abstract). A non‐linear relationship was found between BMI and the primary composite endpoint, with the lowest rate documented for overweight patients with a BMI 25–29.9 kg/m², and higher rates reported for patients with a BMI below or above this range (Graphical Abstract). As in previous studies, a U‐shaped relationship between BMI and outcomes was observed, with overweight patients having a better outcome after AMI than those of normal weight and, while limited in number, compared to underweight participants as well. ²⁰ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³² , ³³ , ³⁴ However, after adjusting for other known prognostic variables, these associations no longer remained significant. Some authors suggested that the deleterious effects of cachexia, rather than the favourable influence of obesity, increase mortality rates. ³³

To our knowledge, although other large studies have been conducted to assess BMI changes in AMI patients, ²² the present large‐scale study is one of the few to examine the relationship between BMI and a composite of CV death and HF‐related hospitalization in patients with reduced systolic function including all BMI subgroups.

Morbid obesity and classic obesity are well‐known risk factors for myocardial hypertrophy, fibrosis and remodelling, which lead to overt clinical HF. ³⁵ , ³⁶ , ³⁷ Yet, contrary to the expected increase in mortality, the ‘obesity paradox’ was not observed here, with higher rates of primary events in obese as well as in low‐BMI subgroups. In the post‐hoc studies of the VALsartan In Acute myocardial iNfarcTion (VALIANT) trial, ³⁸ BMI >25 kg/m² was associated with a greater risk for HF hospitalization, whereas lower BMI was associated with increased risk of mortality. ³⁸ In the present study, increased rates of inpatient and outpatient hospitalizations and of CV death and all‐cause death were noted for patients with a BMI >30 kg/m².

Limited data are available with regard to the effect of BMI on mortality in HF patients with reduced ejection fraction receiving sacubitril/valsartan. In a retrospective multicentre study, Kido et al. ³⁹ found no significant associations between the clinical outcomes of normal weight, overweight or obese BMI patients diagnosed with acute HF treated with sacubitril versus valsartan. Recently, Butt et al. ⁴⁰ found no evidence for an ‘obesity‐survival paradox’ (BMI >25 kg/m²) in patients with HF with reduced ejection fraction after comprehensive adjustment for other prognostic variables. Similarly, the present work found no associations between BMI sub‐classes and mortality rates and/or HF readmission. Of note, while there were no significant differences in the rates of safety events across the BMI subgroups, lower BMI was associated with significantly higher rates of hyperkalaemia.

Demographics, race and ethnicity may impact adverse BMI‐associated CV outcomes. ⁴¹ There has been controversy regarding the possibility of a geographic region effect. ⁴¹ The current Z‐score analysis investigated a potential association between BMI and clinical outcomes in five UN‐defined geographic regions previously compared in the PARADISE‐MI, i.e. Asia/Pacific, Central Europe, Latin America, North America and Western Europe. ⁴¹ A statistically significant, generally U‐shaped, relationship was identified, with the lowest event rates observed in the groups that were most normal relative to their region.

In this study, we evaluated the efficacy and safety of ARNI compared to ramipril across various BMI categories in patients with AMI. Our findings revealed no significant interaction between BMI and the effectiveness of ARNI versus ramipril in reducing primary composite outcomes or secondary endpoints. Additionally, no significant differences in the incidence of adverse events or serious adverse events were noted across the BMI subgroups.

While ARNI has been shown to improve clinical outcomes in HF patients, particularly in the Prospective Comparison of ARNI With ACEI to Determine Impact on Global Mortality and Morbidity in Heart Failure (PARADIGM‐HF) trial, ⁴² evidence specific to BMI subgroups remains limited. The PARADIGM‐HF trial demonstrated that sacubitril/valsartan was superior to enalapril in reducing the risk of CV death and HF hospitalization, but the study did not focus explicitly on stratifying results by BMI categories. ⁴² Similarly, the PARADISE‐MI trial compared ARNI to ramipril in post‐AMI patients, but subgroup analysis based on BMI was not a primary focus. ⁴³ On the other hand, ramipril has been extensively studied in the context of CV risk reduction, notably in the Heart Outcomes Prevention Evaluation (HOPE) trial. ⁴⁴ However, the impact of BMI on ramipril efficacy remains underexplored. Most studies involving ramipril have not systematically evaluated its efficacy across varying BMI subgroups, limiting the ability to draw clear conclusions about the drug performance in patients with higher or lower BMI.

Given the growing prevalence of obesity worldwide and its well‐established association with adverse CV outcomes, the lack of robust data specifically addressing the influence of BMI on the efficacy of ARNI and ramipril is an important gap in the literature. Future studies should aim to stratify patients by BMI to better understand the differential effects of these treatments in obese versus non‐obese populations. Additionally, further research is needed to explore whether weight management interventions or BMI‐specific treatment strategies could optimize outcomes for patients at the extremes of the BMI spectrum.

This study had several limitations. Given that the PARADISE‐MI trial did not reach overall statistical significance for the primary outcome and did not identify a significant benefit of ARNI in the study population, it is difficult to tease out an interaction between BMI and the effect of ARNI. The dataset lacked comprehensive information regarding blood biomarkers, such as serum natriuretic peptide levels, which are associated with adverse CV events. Furthermore, there were no data regarding drug concentrations or pharmacokinetics in the different BMI subgroups. The small sample sizes in the extreme BMI categories (<18.5 and >40 kg/m²) was a third limitation. While these groups represent clinically significant populations with distinct risk profiles, the limited number of patients may reduce the statistical power to detect meaningful associations. Therefore, the findings related to these extreme BMI groups should be interpreted with caution. Further studies with larger sample sizes are needed to confirm these observations and explore their implications in more detail. BMI was only measured at baseline, with no data on BMI changes during hospitalization. This is important as BMI may have fluctuated due to factors such as fluid retention in HF patients, potentially affecting the observed associations. Finally, BMI is a crude anthropometric measure and may demonstrate a weaker correlation with CV outcomes than measures of abdominal obesity. Collider stratification bias could be a partial explanation for the apparent obesity paradox. Finally, while we examined the relationship between BMI and clinical outcomes in patients with AMI and reduced systolic function, the dataset did not include detailed anthropometric measures such as waist circumference measures, waist‐to‐length ratio or dual‐energy X‐ray absorptiometry‐based fat mass. These measures can provide more comprehensive insights into body composition and its impact on clinical outcomes. Therefore, future studies should incorporate these detailed anthropometric measures to enhance the understanding of the relationship between body composition and clinical outcomes in this patient population. Additionally, our study did not include measurements of muscle mass, which is an important factor in interpreting the results. Low muscle mass, particularly in the low BMI category, may contribute to higher mortality rates. Future research should include muscle mass measurements to further elucidate these findings.

In conclusion, this study showed a non‐linear correlation between the BMI of ‘high‐risk’ AMI patients presenting with pulmonary congestion and/or left ventricular dysfunction, and primary composite outcomes, which was maintained across geographical regions. While BMI is a widely used measure of body composition, our findings indicate that it is an inadequate indicator of cardiometabolic risk in this cohort of post‐AMI patients. More precise measures of adiposity, such as waist circumference, may offer better tools for assessing cardiometabolic risk at the bedside. Additionally, our data do not support using BMI as a deciding factor for prescribing ARNI therapy in the post‐AMI setting. The relationship between BMI and clinical outcomes is complex and varies based on the specific outcome. These findings highlight the need for a more sophisticated approach to risk stratification, one that goes beyond BMI and emphasizes individualized assessment in managing high‐risk post‐AMI patients.

Conflict of interest: none declared.

Supporting information

Appendix S1. Supporting Information.