Abstract
Background
Obesity is a major risk factor for incident heart failure (HF). Paradoxically, in HF with reduced left ventricular ejection fraction (HFREF) a high body mass index (BMI) appears to be beneficial. Approximately 50% of HF patients have a preserved LVEF (HFPEF). However, there are few data regarding the relationship between BMI and outcomes in HFPEF.
Methods and Results
Baseline characteristics and cardiovascular (CV) outcomes were assessed in the 4,109 patients (mean age 72 years, mean follow-up 49.5 months) in the irbesartan in HF with preserved ejection fraction (I-PRESERVE) trial. Based on the BMI distribution 5 BMI categories were defined: < 23.5, 23.5–26.4, 26.5–30.9, 31–34.9, and ≥ 35 kg/m2. Most patients (71%) had a BMI ≥ 26.5, 21% had a BMI between 23.5 and 26.4, and 8% had a BMI < 23.5 kg/m2. Patients with higher BMI were younger, more often women, and more likely to have hypertension and diabetes, and higher LVEFs. Patients with BMI of 26.5–30.9 kg/m2 had the lowest rate for the primary composite outcome (death or CV hospitalization) and were used as reference group. After adjustment for 21 risk variables including age, gender, and NT-proBNP, the hazard ratio (HR) for the primary outcome was increased in patients with BMI < 23.5 (HR 1.27; CI 1.04–1.56; p = 0.019), and in those with BMI ≥ 35 (HR 1.27; CI 1.06–1.52; p = 0.011), compared to the referent group. A similar relationship was found for all-cause mortality and for HF hospitalization.
Conclusions
Obesity is common in HFPEF patients and is accompanied by multiple differences in clinical characteristics. Independent of other key prognostic variables, there was a U-shaped relationship, with the greatest rate of adverse outcomes in the lowest and highest BMI categories.
Keywords: heart failure, body mass index, diastolic dysfunction, obesity, prognosis
Observational studies in the general population indicate that body mass index (BMI) is an independent predictor of prognosis with a J-shaped relation between BMI and risk of death, with the lowest risk at a BMI of 23–26.5 kg/m2.1–3 Overweight is part of the metabolic syndrome, a major risk factor for coronary artery disease and stroke. Indeed, in middle-aged indviduals risk of death from atherosclerotic cardiovascular (CV) causes increases steadily with increasing BMI.2,4 In addition obesity is a major risk factor for the development of HF.5 However, even though increased BMI is associated with an increased incidence of coronary artery disease, a common cause of HF with reduced left ventricular ejection fraction (HFREF) and for incident HFREF, the prognosis in patients with established HFREF improves with increasing BMI. This phenomenon has also been called the “obesity paradox”.6–8
Patients with HF and a preserved ejection fraction (HFPEF) comprise approximately 50% of the overall HF population.9 Compared to patients with HFREF patients with HFPEF are usually older, are more frequently women, and have a higher incidence of diabetes and hypertension.10 However, relatively few data are available regarding the relationship between BMI and clinical events in patients specifically with HFPEF. In a post hoc analysis of the CHARM (Candesartan in Heart Failure: Assessment of Reduction in Mortality and morbidity) study a BMI of 30–35 kg/m2 was found to be associated with the lowest overall mortality not only in HFREF, but also in HFPEF, whereas BMI appeared to have no impact on the rate of hospitalization for worsening HF.11 The recently completed I-PRESERVE trial (Irbesartan in HF and PRESERVed Ejection fraction) is the largest randomized, controlled treatment trial in HFPEF.12 In the present analysis, we examined the relationship between BMI and CV morbidity and death in I-PRESERVE. We hypothesized that - similar to HFREF - a higher BMI would be associated with a lower rate of CV events and mortality in HFPEF.
Methods
This post hoc analysis of the I-PRESERVE trial dataset was performed in all patients with a BMI value available at baseline. BMI was derived from body weight and height at the time of randomization. A BMI value was missing in only 19 out of 4,128 patients, leaving 4,109 patients for further analysis. The design and primary outcome of the I-PRESERVE trial have been previously published.12,13 Briefly, I-PRESERVE was a randomized, double-blind, placebo-controlled trial that evaluated the effect of irbesartan, an angiotensin AT1 receptor-antagonist, in older HFPEF patients (age ≥ 60 years; LVEF > 45%). The mean follow-up was 49.5 months and the annual all-cause mortality rate was 5.2%. Irbesartan had no effect on the primary composite endpoint (death from any cause or hospitalization for a CV cause, i.e. HF, myocardial infarction, unstable angina, arrhythmia, or stroke) or two secondary endpoints (death from any cause and hospitalization for worsening HF, the latter being defined as hospitalization with signs and symptoms of HF requiring a significant augmentation of HF therapy). The present subanalysis focused on the primary composite endpoint and these two secondary endpoints.
Statistical analysis
BMI was evaluated as both a continuous and a categorical variable. Before defining the BMI categories, a density plot analysis of the BMI was performed (Figure 1). To create the BMI ranges, we strived to have clinically relevant categories, which excluded quartiles and quintiles for this particular sample because the resultant middle range would be too narrow. To allow for a sufficient number of patients in each category the BMI cut-offs used were necessarily slightly different from those recommended by the World Health Organization (WHO).14 Based on the density plot and the two previous criteria, five BMI categories were defined: < 23.5 (“underweight”), 23.5–26.4 (“normal” weight), 26.5–30.9 (overweight), 31.0–34.9 (“mild” obesity), and ≥ 35.0 kg/m2 (“severe” obesity). The association of BMI with outcomes was examined in unadjusted and fully adjusted Cox Proportional Hazards models that contained age, sex, NYHA-class, heart rate, systolic blood pressure, left ventricular hypertrophy, LVEF, cause of HF, hospitalization for HF within the last 6 months, history of hypertension, myocardial infarction, stroke, COPD and/or diabetes, use of diuretics, digoxin, a calcium-channel blocker, lipid-lowering agents, an ACE-inhibitor, or a beta-blocker, and NT-proBNP as covariates. The covariates were chosen for these multivariate analyses because they were found to be significant independent predictors of CV outcomes in previous publications.The BMI category with the lowest rate of the primary composite outcome (death or CV hospitalization) was selected and used as reference group for comparison regarding the relationship between BMI and all other predefined endpoints.
Figure 1.
Density plot of the BMI distribution of the total patient population (n = 4,109) and their distribution into 5 distinct BMI categories (group 1, BMI < 23.5 kg/m2, n= 336; group 2, BMI 23.5–26.4 kg/m2, n = 858; group 3, BMI 26.5–30.9 kg/m2, n = 1,506; group 4, BMI 31.0–34.9 kg/m2, n = 813; group 5, BMI ≥ 35.0 kg/m2, n = 596).
Results
Baseline Characteristics
The mean age (± SD) of the cohort was 71.6 (6.9) years, 29.4% were older than 75 years, and 60.4% were women. The mean LVEF was 59.4 (9.2) % and the median NT-proBNP was 339 pg/ml (range: 5 – 28,670 pg/ml). The baseline characteristics by BMI categories are shown in Table 1. The mean (± SD) BMI was 29.6 (5.3) kg/m2 in the entire group. In respect to WHO-defined BMI ranges14 16.5% of the patients had a BMI < 25 kg/m2, 42.4% between 25 and 30 kg/m2, and 41.3% greater than 30 kg/m2. Age was inversely related to BMI, whereas average LVEF tended to increase with BMI. The percentage of women was highest in the BMI categories < 23.5 kg/m2 and ≥ 31 kg/m2. The prevalence of arterial hypertension as an investigator-assigned etiology of HF increased with BMI whereas that of coronary artery disease declined. The serum creatinine and hemoglobin did not significantly differ between the BMI subgroups. However, median NT-proBNP level markedly declined with increasing BMI (Table 1). The use of diuretics, ACE-inhibitors, beta-blockers, calcium channel blockers, and lipid lowering agents increased with BMI, whereas the use of digoxin declined with BMI, and use of antiplatelets and spironolactone were unrelated to BMI.
Table 1.
Baseline Characteristics of the five BMI Categories
| BMI, kg/m2 | ||||||
|---|---|---|---|---|---|---|
| < 23.5 (n=336) | 23.5 to 26.4 (n=858) | 26.5 to 30.9 (n=1506) | 31.0 to 34.9 (n=813) | ≥ 35.0 (n=596) | p-value | |
| Age, yr | 75±7 | 73±7 | 72±7 | 70±6 | 70±7 | < 0.0001 |
| Female, % | 65 | 52 | 57 | 64 | 73 | < 0.0001 |
| NYHA class, % | ||||||
| II | 22 | 24 | 22 | 18 | 17 | 0.007 |
| III | 75 | 73 | 76 | 80 | 79 | 0.021 |
| IV | 3 | 3 | 2 | 2 | 4 | 0.194 |
| LVEF, % | 58±10 | 59±9 | 59±9 | 59±9 | 61±9 | 0.005 |
| Heart rate, bpm | 73±10 | 71±1 | 71±10 | 72±10 | 73±11 | < 0.001 |
| Blood pressure, mmHg | ||||||
| Systolic | 134±17 | 135±15 | 136±14 | 138±15 | 137±16 | < 0.001 |
| Diastolic | 77±10 | 78±8 | 71±10 | 72±10 | 73±11 | |
| Cause of HF, % | ||||||
| Hypertension | 52 | 60 | 63 | 69 | 68 | < 0.001 |
| Coronary artery disease | 29 | 29 | 26 | 22 | 19 | < 0.001 |
| History, % | ||||||
| Hypertension | 76 | 87 | 88 | 92 | 92 | < 0.001 |
| Myocardial infarction | 23 | 29 | 24 | 20 | 19 | < 0.001 |
| Diabetes mellitus | 18 | 20 | 27 | 31 | 41 | < 0.001 |
| Atrial fibrillation | 33 | 28 | 29 | 29 | 29 | 0.45 |
| COPD | 10 | 10 | 8 | 9 | 13 | 0.03 |
| Stroke/TIA | 9 | 11 | 9 | 10 | 10 | 0.92 |
| HF hospitalization within the past 6 months | 49 | 49 | 43 | 39 | 43 | < 0.001 |
| Creatinine, mg/dl | 1.0±0.3 | 1.0±0.3 | 1.0±0.3 | 1.0±0.4 | 1.0±0.3 | 0.742 |
| eGFR, ml/min/1.73 m2 | 68±23 | 73±23 | 73±22 | 72±23 | 71±24 | 0.20 |
| Hemoglobin, mg/dl | 13±2 | 14±2 | 14±2 | 14±2 | 14±2 | 0.339 |
| NT-proBNP, pg/ml (median) | 694 | 404 | 337 | 248 | 254 | < 0.0001 |
| Medication, % | ||||||
| Diuretic | 78 | 81 | 82 | 83 | 91 | < 0.0001 |
| Spironolactone | 20 | 14 | 15 | 14 | 17 | 0.982 |
| ACE-inhibitor | 23 | 23 | 25 | 28 | 28 | < 0.0001 |
| Beta-blocker | 50 | 59 | 59 | 61 | 63 | 0.023 |
| Digoxin | 20 | 15 | 13 | 14 | 10 | < 0.0001 |
| Calcium channel blocker | 31 | 36 | 39 | 45 | 46 | < 0.0001 |
| Lipid lowering agent | 25 | 27 | 33 | 29 | 35 | 0.003 |
| Antiplatelet agent | 52 | 63 | 61 | 58 | 53 | 0.09 |
If not otherwise indicated mean ± SEM or percentage for each BMI category
Primary Composite Endpoint
The unadjusted Kaplan-Meier curves (Figure 2) indicated that the risk (the cumulative probability of adverse outcomes), correlated nonlinearly with BMI. The event rates for the primary outcomes in the 5 BMI groups are shown in Table 2. The lowest event rate (33.1 %) was seen for BMI category 26.5–30.9 kg/m2 and the highest event rate for BMI category < 23.5 kg/m2 (49.1 %) with the five groups showing a U-shaped relationship. The components of the primary composite endpoint in general had a similar U-shaped relationship with BMI as the composite endpoint except for all-cause mortality which had a linear, inverse relationship with BMI (Table 2).
Figure 2.
Kaplan-Meier curves for the unadjusted primary composite endpoint by baseline BMI category.
Table 2.
Major Unadjusted Outcomes for the five BMI Categories
| BMI, kg/m2 |
|||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| < 23.5 (n=336) | 23.5 to 26.4 (n=858) | 26.5 to 30.9 (n=1506) | 31.0 to 34.9 (n=813) | ≥ 35.0 (n=596) | |||||||||||
| n1 | event rate2 | event rate3 | n1 | event rate2 | event rate3 | n1 | event rate2 | event rate3 | n1 | event rate2 | event rate3 | n1 | event rate2 | event rate3 | |
| Primary composite outcome | 165 | 49.1 | 152 | 329 | 38.4 | 109 | 499 | 33.1 | 92 | 280 | 34.4 | 95 | 225 | 37.8 | 108 |
| All-cause mortality (death) | 108 | 32.1 | 83 | 205 | 23.9 | 60 | 287 | 19.1 | 47 | 151 | 18.6 | 45 | 123 | 20.6 | 50 |
| HF hospitalization and death | 6 | 22.6 | 66 | 126 | 14.7 | 40 | 236 | 15.7 | 41 | 136 | 16.7 | 44 | 137 | 23.0 | 63 |
| Cause specific mortality | |||||||||||||||
| - Sudden death | 28 | 8.3 | 22 | 61 | 7.1 | 18 | 74 | 4.9 | 12 | 43 | 5.3 | 13 | 24 | 4.0 | 10 |
| - HF death | 25 | 7.4 | 19 | 18 | 2.1 | 5 | 42 | 2.8 | 7 | 17 | 2.1 | 5 | 21 | 3.5 | 9 |
| - All CV deaths | 68 | 20.2 | 52 | 124 | 14.5 | 36 | 168 | 11.2 | 27 | 93 | 11.5 | 28 | 74 | 12.4 | 30 |
| - Non-CV deaths | 26 | 7.7 | 20 | 64 | 7.5 | 19 | 101 | 6.7 | 16 | 43 | 5.3 | 13 | 32 | 5.4 | 13 |
| Hospitalization for CV causes | |||||||||||||||
| - Worsening HF | 69 | 20.5 | 60 | 124 | 14.5 | 39 | 217 | 14.4 | 38 | 134 | 16.5 | 43 | 135 | 22.7 | 62 |
| - Myocardial infarction | 14 | 4.2 | 11 | 31 | 3.6 | 9 | 62 | 4.1 | 10 | 25 | 3.1 | 8 | 20 | 3.4 | 8 |
| - Unstable angina | 20 | 6.0 | 16 | 25 | 2.9 | 7 | 65 | 4.3 | 11 | 33 | 4.1 | 10 | 20 | 3.4 | 8 |
| - Stroke | 22 | 6.5 | 17 | 44 | 5.1 | 13 | 69 | 4.6 | 11 | 43 | 5.3 | 13 | 22 | 3.7 | 9 |
| - Atrial arrhythmia | 14 | 4.2 | 11 | 31 | 3.6 | 9 | 55 | 3.7 | 9 | 42 | 5.2 | 13 | 32 | 5.4 | 14 |
| - Ventricular arrhythmia | 0 | 0 | 0 | 2 | 0.2 | <1 | 5 | 0.3 | <1 | 2 | 0.2 | <1 | 3 | 0.5 | 1 |
n = number of events
event rate = crude event rate, %
event rate = per 1,000 patient years
With the BMI category 26.5–30.9 kg/m2 as the reference group, the unadjusted HR for the primary composite endpoint was mildly increased in patients with BMI 23.5–26.4 kg/m2 (HR 1.18, CI 1.03–1.36, p = 0.019) and even more so in patients with BMI < 23.5 kg/m2 (HR 1.65, CI 1.38–1.97, p < 0.001). However, after adjustment for the 21 pre-selected clinical variables (see methods), a significantly increased HR was only observed in the 2 most extreme BMI categories < 23.5 kg/m2 (HR 1.27, CI 1.04–1.56, p = 0.019) and ≥ 35 kg/m2 (HR 1.27, CI 1.06–1.52 p = 0.011), indicating a U-shaped relationship between BMI and the primary composite endpoint (Figure 3).
Figure 3.
Adjusted HR for the primary composite endpoint for the five BMI categories. The BMI category 26.5 and 30.9 kg/m2 had the lowest event rate and was used as a referent group and assigned a HR 1.0. p vs. referent group.
Predefined Secondary Endpoints and Mode of Death
Similar results were observed for all-cause mortality (Figure 4A). Compared to the reference group, the adjusted HR for all-cause mortality was increased in patients with BMI < 23.5 kg/m2 (HR 1.44, CI 1.12–1.84, p = 0.005) and in those with BMI ≥ 35 kg/m2 (HR 1.31, CI 1.03–1.67, p = 0.029) in the fully adjusted model. However, mode of death differed between the five BMI categories. The unadjusted rate of sudden death and non-CV death tended to decrease with increasing BMI. However, on the contrary, the rate of HF death was highest in the lowest and highest BMI categories (BMI < 23.5 kg/m2 and BMI ≥ 35 kg/m2), where it equalled the rate of sudden death. In patients with a BMI between 23.5 – 34.9 kg/m2 rate of sudden death and non-CV death were twice to three times higher than the rate of HF death (Table 2, Figure 5).
Figure 4.
Adjusted HRs for all-cause mortality (A) and HF hospitalization (B). p vs. referent group (BMI category 26.5 and 30.9 kg/m2). p vs. referent group.
Figure 5.
Unadjusted rates of sudden death, HF death and non-CV death for the five BMI categories
In the fully adjusted model, the risk of HF hospitalization was substantially increased with BMI ≥ 35 kg/m2 (HR 1.52, CI 1.19–1.94 p = 0.001) (Figure 4B), even though these patients had the highest LVEF and the lowest NT-proBNP (Table 1). HF hospitalization also tended to increase in the BMI category 31.0–34.9 kg/m2 (HR 1.14, CI 0.89–1.44, p = 0.299). The lowest risk was seen with the BMI category 23.5–26.5 kg/m2 (HR 0.82, CI 0.65–1.05, p = 0.12). However, in the lowest BMI category (< 23.5 kg/m2) the risk for HF hospitalization was not significantly increased (HR 1.11, CI 0.82–1.50, p = 0.49) (Figure 4B).
Effect of Irbesartan
Irbesartan had no significant effect on the relationships between BMI and the primary composite endpoint, all-cause mortality, or HF hospitalization.
Discussion
The major findings of the present study are: 1) In patients with HFPEF overweight and obesity are highly prevalent and are accompanied by multiple differences in clinical and demographic characteristics; 2) there is a U-shaped relationship between BMI and both the primary composite endpoint and all-cause mortality, with the highest event rates in those patients with the lowest BMI (< 23.5 kg/m2) and those with the highest BMI (> 35 kg/m2). These relationships were not significantly altered by adjustment for 21 pre-selected clinical variables, including age, gender, NT-proBNP, and medication. Interestingly, the lowest incidence of adverse outcomes was observed in those patients with HFPEF who were moderately overweight (BMI between 26.5 and 30.9 kg/m2); 3) HF hospitalization was only significantly increased in patients with a BMI ≥ 35 kg/m2.
Clinical and Demographic Characteristics
More than 83% of the patients were overweight (BMI 25–30 kg/m2) or obese (BMI > 30 kg/m2). This rate of obesity is similar to the population-based studies that have specifically examined older age patients with HF, including HFPEF specifically.15 In the present study patients in the lowest BMI category (< 23.5 kg/m2) were older, had a slightly lower LVEF, a higher NT-proBNP, and a lower rate of arterial hypertension and diabetes mellitus. NT-proBNP has been shown to be an independent marker of prognosis in both HFPEF and HFREF and to be lower in obese HF patients.16 Indeed, patients with the highest BMI (≥ 35 kg/m2) had a median NT-proBNP of only 254 pg/ml and those with the lowest BMI (< 23.5 kg/m2) had an almost three-times higher median NT-proBNP level. Furthermore, patients with BMI < 23.5 kg/m2 had the lowest and those with a BMI ≥ 35.0 kg/m2 the highest likelihood of ACE-inhibitor, beta-blocker, calcium channel blocker, diuretic, and lipid lowering agents usage, whereas in patients with a BMI < 23.5 kg/m2 digoxin was most frequently administered. However, none of these agents has so far been shown substantially influence the prognosis in HFPEF.17
Despite marked differences in the key characteristics of patients with a BMI < 23.5 kg/m2 and those with a BMI ≥ 35.0 kg/m2, both BMI categories had a similar risk of CV events after adjustment for 21 variables including age, LVEF, NT-proBNP, and medication, indicating an independent prognostic relevance of BMI. However, this does not exclude that the pathophysiological mechanisms related to the prognostic impact of a very low and very high BMI may differ. This is supported by the relatively low NT-proBNP and relatively high LVEF in the highest BMI category.
Adverse Outcomes
The rate of the primary composite endpoint was highest in patients with in the lowest BMI category (< 23.5 kg/m2) (Table 2) with 152 events per 1,000 patient years. This high event rate was driven mainly by the fact that this BMI category had the highest all-cause mortality rate and a high rate of worsening HF and stroke. However, there were no differences between the five BMI categories in the rates of myocardial infarction, unstable angina, and major arrhythmias. Likewise, a recent meta-analysis of 40 studies of patients with coronary artery disease found no difference in CV mortality between normal weight patients and overweight or mildly obese patients, whereas patients with severe obesity had a 1.88 fold higher rate of major CV events.18 In the present study the rate of CV events may have been too low to allow for further differentiation.
Similar to the primary endpoint, a U-shaped relationship was found for all-cause mortality. Again patients with in the lowest and highest BMI categories (< 23.5 kg/m2 and > 35 kg/m2, respectively) appeared to have the highest adjusted risk of death (Figure 4A). This finding is noteworthy, since we previously reported in the I-PRESERVE patient cohort that the cause of death is twice as often non-CV as compared to HFREF (30% vs. approx. 14% of total mortality).19 In contrast to all-cause mortality, the rate of HF hospitalization was only significantly increased in severely obese patients (i.e. BMI ≥ 35 kg/m2). It was lowest in patients with a BMI between 23.5 – 26.4 kg/m2 and only tended to increase in patients with a BMI < 23.5 kg/m2 (Figure 4B). In the present study the rate of both sudden death and non-CV death declined with increasing BMI when focusing on the unadjusted event rates (Table 2). In contrast, HF death was relatively common, equaling rates of sudden death and non-CV death only in those patients with a BMI < 23.5 kg/m2 and in those with a BMI ≥ 35 kg/m2 (Figure 5).
Comparison to Population-Based Studies
Three recently published large cohort studies based on general, unselected populations reported a J-shaped relationship between BMI and all-cause mortality. In US men aged 66–71 years the optimum BMI was 25.0–26.4 kg/m2 and in women aged 66–71 years 23.5–24.9 kg/m2, with a diminished impact of BMI at older age.1 Although the mean age in the present study was almost 72 years BMI still had a significant influence on outcome, and although many patients in I-PRESERVE suffered form coronary artery disease, the relationship between BMI and adverse CV outcomes rather resembled that between BMI and cancer death than that between BMI and death form atherosclerotic causes.4,7
Comparison to other Databases with HFPEF
Data on the relationship between BMI and prognosis in HFPEF have been sparse. The results of the present study both confirm and extend the results of a previous report on BMI and prognosis in the CHARM-Preserved cohort.11 In that cohort, the lowest risk of all-cause death was observed in those with a BMI between 30.0 and 34.9 kg/m2 and the highest risk (HR 1.99) in those with a BMI < 22.5 kg/m2. However, no significant relationship was found between BMI and hospitalization for HF.11 There are significant differences between the CHARM-Preserved and the I-PRESERVE cohorts, including: a) The LVEF inclusion criteria were higher in I-PRESERVE (>45% vs. > 40%) such that it that some patients with signficantly reduced LVEF may have been included in CHARM-Preserved; b) the I-PRESERVE cohort included a higher proportion of women (> 60% vs. < 40%) and the patients were significantly older (mean age 72 vs. 65 years). This gender and age distribution of I-PRESERVE is more similar to that reported in population-based studies of HFPEF9; c) the inclusion criteria for I-PRESERVE required either prior HF-related hospital admission within the past year or NYHA functional class III symptoms plus echo abnormalities, whereas CHARM only required symptomatic HF (i.e. NYHA functional classes II-IV).
Comparison to HFREF
Data regarding the relationship between BMI and prognosis are more abundant in HFREF. The results of the present study support that the relationship between BMI and prognosis appears qualitatively similar in HFPEF and HFREF.20 In particular, they suggest that underweight has an adverse effect on prognosis in HF independent from the LVEF, as was shown in the CHARM study which included both HFPEF and HFREF.11 In HFREF, this appears due to catabolism, inflammation, and activation of stress hormone systems.21 These processes result in tissue wasting and lead to adverse prognosis, apparently independent from the underlying disease.Overweight and moderately obese patients may have more metabolic reserve, adipocytes and lipoprotein pools to serve as effective scavengers to bind with and neutralize circulating lipopolysaccharides including endotoxins.22 However, these potentially beneficial mechanisms do not explain the signficantly worse prognosis in severe obesity (BMI ≥ 35 kg/m2) shown in our study and in others. They are also in contradiction of the significantly increased risk attributable to obesity for the development of HF and coronary artery disease.4,5
Potential Limitations
While the data in the present study are consistent with a relatively large number of publications in HF, coronary disease, and other disorders, they are not sufficient to promote weight gain in underweight patients or to discourage weight loss in obese HF patients. In our study, similar to other published observational studies, we are unable to exclude occult, pre-existing cancers or other chronic illnesses that affect BMI and prognosis. 23 Furthermore, the significantly worse prognosis seen in the highest BMI category indicates the lack of a uniform mechanism to explain this phenomenon.
Our data analysis was based on a retrospective subanalysis of carefully characterized patients from a controlled trial (I-PRESERVE). Therefore, there was an unequal number of patients in each BMI category and the most commonly used BMI ranges (WHO criteria) had to be slightly adjusted to allow for a sufficient number of subjects in each BMI category. As HFPEF is mainly a disease of the elderly, the average age was higher than in previous studies in HFREF. However, if anything, this bolsters our key findings, since most variables loose their prognostic power at a higher age.2–4
Although 21 variables were used in the adjusted model, which all by themselves have been shown to influence adverse CV outcomes in the I-PRESERVE cohort,24 other residual confounders cannot completely be ruled out. The adjusted model included a previous myocardial infarction and a HF hospitalization 6 months prior to randomization, which was identical to the day of assessment of BMI. However, no information regarding the exact time intervals of these events and their potential influence on BMI and CV outcomes are available. Furthermore, the I-PRESERVE data base does not contain any information on weight changes prior to and after randomization, which might have influenced the event rate.
Obese patients may have signs and symptoms, resembling those of HF. This makes the ability to diagnose HFPEF in obese patients more difficult. However, the I-PRESERVE trial required that patients had a clinical diagnosis of HF along with either HF hospitalization or pertinent echocardiographic or Doppler criteria consistent with HFPEF. Obese patients fulfilled these criteria and had in fact the highest HF hospitalization rate.
Waist circumference, waist-to-hip ratio, and percent body fat may be better indicators of prognosis than BMI, 3,4,25,26 but these data were unfortunately not assessed in I-PRESERVE. Furthermore, body composition was not analyzed, although fluid retention (e.g. edema) not only increases BMI but also influences prognosis.27 However, only patients in stable clinical conditions at randomization were included into the I-PRESERVE trial.
Independent from these potential limitations, we believe this to be the largest data set analysed for the impact of BMI on adverse outcomes during long-term follow-up in HFPEF.
Conclusions
Obesity is highly prevalent in elderly HFPEF patients and is accompanied by multiple differences in clinical characteristics. Consistent with reports in patients with HFREF, HFPEF patients in the lowest BMI category (BMI < 23.5 kg/m2) had the highest risk of subsequent events. The event rate was also significantly increased in HFPEF patients with severe obesity (BMI ≥ 35 kg/m2). This finding of a U-shaped relationship between BMI and adverse CV outcomes persisted even after adjustment for baseline imbalances for all endpoints.
Acknowledgments
The thorough statistical analyses performed by Scott Hetzel (Department of Biostatistics and Medical Informatics, University of Wisconsin, Madison, WI) are greatly acknowledged.
Sources of Funding
The I-PRESERVE trial was funded by Bristol-Myers Squibb and Sanofi-Aventis, as was the present subanalysis. Also supported in part by NIH grants 37AG18915 and P30AG21222 (Dalane W. Kitzman).
Abbreviations and Acronyms
- BMI
Body mass index
- CHARM
Candesartan in Heart Failure: Assessment of Reduction in Mortality and morbidity
- CI
Confidence interval
- COPD
Chronic obstructive pulmonary disease
- CV
Cardiovascular
- eGFR
Estimated glomerular filtration rate (using the MDRD equation)
- EPIC
European Prospective Investigation into Cancer and nutrition
- HF
Heart failure
- HFPEF
Heart failure with preserved ejection fraction
- HFREF
Heart failure with reduced ejection fraction
- HR
Hazard ratio
- I-PRESERVE
Irbesartan in heart failure with PRESERVed Ejection fraction
- LVEF
Left ventricular ejection fraction
- NT-proBNP
N-terminal pro brain natriuretic peptide
- WHO
World Health Organization
Footnotes
Disclosures
The members of the steering committee (Dalane W. Kitzman, Michael Zile, Barry Massie, and Peter E Carson) and of the endpoint committee (Markus Haass, Inder S. Anand, Michael Zile, and Alan Miller,) of I-PRESERVE received honoraria from Bristol-Myers Squibb for providing these services.
Clinical Trial Registration. URL: http://www.clinicaltrials.gov. Unique identifier: NCT000095238
References
- 1.Adams KF, Schatzkin A, Harris TB, Kipnis V, Mouw T, Ballard-Barbash R, Hollenbeck A, Leitzman MF. Overweight, obesity, and mortality in a large prospective cohort of persons 50 to 71 years old. N Engl J Med. 2006;355:763–778. doi: 10.1056/NEJMoa055643. [DOI] [PubMed] [Google Scholar]
- 2.Jee SH, Sull JW, Park J, Lee SY, Ohrr H, Guallar E, Samet JM. Body-mass index and mortality in Korean men and women. N Engl J Med. 2006;355:779–787. doi: 10.1056/NEJMoa054017. [DOI] [PubMed] [Google Scholar]
- 3.Pischon T, Boeing H, Hoffmann K, Bergmann M, Schulze MB, Overvad K, van der Schouw YT, Spencer E, Moons KG, Tjonneland A, Halkaer J, Jensen MK, Stegger J, Clavel-Chapelon F, Boutron-Ruault MC, Chajes V, Linseisen J, Kaaks R, Trichopoulou D, Bamia C, Sieri S, Palli D, Tumino R, Vineis P, Panico S, Peeters PH, May AM, Bueno-de-Mesquita HB, van Duijnhoven FJ, Hallmans G, Weinehall L, Manjer J, Hedblad B, Lund E, Agudo A, Arriola L, Barricarte A, Navarro C, Martinez C, Quiros JR, Key T, Bingham S, Khaw KT, Boffetta P, Jenab M, Ferrari P, Riboli E. General and abdominal adiposity and risk of death in Europe. N Engl J Med. 2008;359:2105–2120. doi: 10.1056/NEJMoa0801891. [DOI] [PubMed] [Google Scholar]
- 4.Yusuf S, Hawken S, Ounpuu S, Bautista L, Franzosi MG, Commerford P, Lang CC, Rumboldt Z, Onen CL, Lisheng L, Tanomsup S, Wangai P, Jr, Razak F, Sharma AM, Anand SS INTERHEART Study Investigators. Obesity and the risk of myocardial infarction in 27,000 participants from 52 countries: a case-control study. Lancet. 2005;366:1640–1649. doi: 10.1016/S0140-6736(05)67663-5. [DOI] [PubMed] [Google Scholar]
- 5.Kenchaiah S, Evans JC, Levy D, Wilson PW, Benjamin EJ, larson MG, Kannel WB, Vasan RS. Obesity and the risk of heart failure. N Engl J Med. 2002;347:305–313. doi: 10.1056/NEJMoa020245. [DOI] [PubMed] [Google Scholar]
- 6.Oreopoulos A, Padwal R, Kanlantar-Zadeh K, Fonarow GC, Norris CM, McAlister FA. Body mass index and mortality in heart failure: a meta-analysis. Am Heart J. 2008;156:13–22. doi: 10.1016/j.ahj.2008.02.014. [DOI] [PubMed] [Google Scholar]
- 7.Lavie CJ, Milani RV, Ventura HE. Obesity and cardiovascular disease: risk factor, paradox, and impact of weight loss. J Am Coll Cardiol. 2009;53:1925–1932. doi: 10.1016/j.jacc.2008.12.068. [DOI] [PubMed] [Google Scholar]
- 8.Artham SM, Lavie CJ, Patel HM, Ventura HO. Impact of obesity on the risk of heart failure and its prognosis. J Cardiometab Syndr. 2008;3:155–161. doi: 10.1111/j.1559-4572.2008.00001.x. [DOI] [PubMed] [Google Scholar]
- 9.Bursi F, Weston SA, Redfield MM, Jacobsen SJ, Pakhomov S, Nkomo VT, Meverden RA, Roger VL. Systolic and diastolic heart failure in the community. JAMA. 2006;296:2209–2216. doi: 10.1001/jama.296.18.2209. [DOI] [PubMed] [Google Scholar]
- 10.Gaasch WH, Zile MR. Left ventricular diastolic dysfunction and diastolic heart failure. Annu Rev Med. 2004;55:373–934. doi: 10.1146/annurev.med.55.091902.104417. [DOI] [PubMed] [Google Scholar]
- 11.Kenchaiah S, Pocock SJ, Wang D, Finn PV, Zornoff LAM, Skali H, Pfeffer MA, Yusuf S, Swedberg K, Michelson EL, Granger CB, McMurray JJV, Solomon SD for the CHARM investigators. Body mass index and prognosis in patients with chronic heart failure. Insights from the candesartan in heart failure: assessment of reduction in mortality and morbidity (CHARM) program. Circulation. 2007;116:627–636. doi: 10.1161/CIRCULATIONAHA.106.679779. [DOI] [PubMed] [Google Scholar]
- 12.Massie BM, Carson PE, McMurray JJ, Komajda M, McKelvie R, Zile MR, Anderson S, Donovan M, Iverson E, Staiger C, Ptaszynska A. Irbesartan in patients with heart failure and preserved ejection fraction. N Engl J Med. 2008;359:2456–2467. doi: 10.1056/NEJMoa0805450. [DOI] [PubMed] [Google Scholar]
- 13.Carson P, Massie BM, McKelvie R, McMurray J, Komajda M, Zile M, Ptaszynska A, Frangin G for the I-Preserve investigators. The irbesartan in heart failure with preserved systolic function (I-PRESERVE) trial: rationale and design. J Card Fail. 2005;11:576–585. doi: 10.1016/j.cardfail.2005.06.432. [DOI] [PubMed] [Google Scholar]
- 14.Obesity: Preventing and Managing the Global Epidemic: Report of a WHO Consultation on Obesity, Geneva, June 3–5, 1997. Geneva, Switzerland: World Health Organization; 1998. [PubMed] [Google Scholar]
- 15.Senni M, Tribouilloy CM, Rodeheffer RJ, Jacobsen SJ, Evans JM, Bailey KR, Redfield MM. Congestive heart failure in the community: a study of all incident cases in Olmsted County Minnesota, in 1991. Circulation. 1998;98:2282–2289. doi: 10.1161/01.cir.98.21.2282. [DOI] [PubMed] [Google Scholar]
- 16.Krauser DG, Lloyd-Jones DM, Chae CU, Cameron R, Anwaruddin S, Baggish AL, Chena A, Tung R, Januzzi JL. Effect of body mass index on natriuretic peptide levels in patients with acute congestive heart failure: a proBNP investigation of dyspnea in the emergency department (PRIDE) substudy. Am Heart J. 2005;149:744–750. doi: 10.1016/j.ahj.2004.07.010. [DOI] [PubMed] [Google Scholar]
- 17.Kindermann M, Reil JC, Pieske B, van Veldhuisen DJ, Böhm M. Heart failure with normal ejection fraction: what is the evidence? Trends Cardiovasc Med. 2008;18:280–292. doi: 10.1016/j.tcm.2008.12.003. [DOI] [PubMed] [Google Scholar]
- 18.Romero-Corral A, Montori VM, Somers VK, Korinek J, Thomas RJ, Allison TG, Mookadam F, Lopez-Jiminez F. Association of bodyweight with total mortality and with cardiovascular events in coronary artery disease: a systematic review of cohort studies. Lancet. 2006;368:666–678. doi: 10.1016/S0140-6736(06)69251-9. [DOI] [PubMed] [Google Scholar]
- 19.Zile MR, Gaasch WH, Anand IS, Haass M, Little WC, Miller AB, Lopez-Sendon J, Teerlink JR, White M, McMurray JJ, Komajda M, McKelvie R, Ptaszynska A, Hetzel SJ, Massie B, Carson PE. Mode of death in patients with heart failure and a preserved ejection fraction: results from the Irbesartan in Heart Failure with Preserved Ejection Fraction Study (I-Preserve) trial. Circulation. 2010;121:1393–1405. doi: 10.1161/CIRCULATIONAHA.109.909614. [DOI] [PubMed] [Google Scholar]
- 20.Curtis JP, Selter JG, Wang Y, Rathore SS, Jovin IS, Jadabaie F, Kosiborod M, Portnay EL, Sokol SI, Bader F, Krumholz HM. The obesity paradox: body mass index and outcomes in patients with heart failure. Arch Intern Med. 2005;165:55–61. doi: 10.1001/archinte.165.1.55. [DOI] [PubMed] [Google Scholar]
- 21.Von Haehling S, Horwich TB, Fonarow GC, Anker SD. Tipping the scale – Heart failure, body mass index, and prognosis. Circulation. 2007;116:588–590. doi: 10.1161/CIRCULATIONAHA.107.716662. [DOI] [PubMed] [Google Scholar]
- 22.Anker SD, Rauchhaus M. Insights into the pathogenesis of chronic heart failure: immune activation and cachexia. Curr Opin Cardiol. 1999;14:211–216. doi: 10.1097/00001573-199905000-00004. [DOI] [PubMed] [Google Scholar]
- 23.Pocock SJ, McMurray JJ, Dobson J, Yusuf S, Granger CB, Michelson EL, Ostergren J, Pfeffer MA, Solomon SD, Anker SD, Swedberg KB. Weight loss and mortality risk in patients with chronic heart failure in the candesartan in heart failure: assessment of reduction in mortality and morbidity (CHARM) programme. Eur Heart J. 2008;29:2641–2650. doi: 10.1093/eurheartj/ehn420. [DOI] [PubMed] [Google Scholar]
- 24.Komajda M, Carson PE, Hetzel S, McKelvie R, McMurray J, Ptaszynska A, Zile MR, Demets D, Massie BM. Factors associated with outcome in heart failure with preserved ejection fraction: findings from the irbesartan in heart failure with preserved ejection fraction study (I-PRESERVE) Circ Heart Fail. 2011;4:27–35. doi: 10.1161/CIRCHEARTFAILURE.109.932996. [DOI] [PubMed] [Google Scholar]
- 25.Nicklas Barbara J, Cesari Matteo, Penninx Brenda WJH, Kritchevsky Stephen B, Ding Jingzhong, Newman Anne, Kitzman Dalane W, Kanaya Alka M, Pahor Marco, Harris Tamara B. Abdominal obesity is an independent risk factor for chronic heart failure in older people. J Am Geriatr Soc. 2006;54:413–420. doi: 10.1111/j.1532-5415.2005.00624.x. [DOI] [PubMed] [Google Scholar]
- 26.Lavie CJ, Milani RV, Ventura HO, Romero-Corral A. Body composition and heart failure prevalence and prognosis: getting to the fat of the matter in the “obesity paradox”. Mayo Clin Proc. 2010;85:605–608. doi: 10.4065/mcp.2010.0333. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Goldsmith SR, Brandimarte F, Gheorghiade M. Congestion as a therapeutic target in acute heart failure syndromes. Prog Cardiovasc Dis. 2010;52:383–392. doi: 10.1016/j.pcad.2009.11.005. [DOI] [PubMed] [Google Scholar]