Redfield MM, Jacobsen SJ, Burnett, Jr JC, Mahoney DW, Bailey KR, Rodeheffer RJ. Burden of Systolic and Diastolic Ventricular Dysfunction in the CommunityAppreciating the Scope of the Heart Failure Epidemic. JAMA. 2003;289(2):194-202. doi:10.1001/jama.289.2.194
Author Affiliations: Division of Cardiovascular Diseases, Department of Internal Medicine (Drs Redfield, Burnett, and Rodeheffer) and the Divisions of Clinical Epidemiology (Dr Jacobsen) and Biostatistics (Mr Mahoney and Dr Bailey), and Department of Health Science Research, Mayo Clinic and Foundation, Rochester, Minn.
Context Approximately half of patients with overt congestive heart failure (CHF)
have diastolic dysfunction without reduced ejection fraction (EF). Yet, the
prevalence of diastolic dysfunction and its relation to systolic dysfunction
and CHF in the community remain undefined.
Objectives To determine the prevalence of CHF and preclinical diastolic dysfunction
and systolic dysfunction in the community and determine if diastolic dysfunction
is predictive of all-cause mortality.
Design, Setting, Participants Cross-sectional survey of 2042 randomly selected residents of Olmsted
County, Minnesota, aged 45 years or older from June 1997 through September
Main Outcome Measures Doppler echocardiographic assessment of systolic and diastolic function.
Presence of CHF diagnosis by review of medical records with designation as
validated CHF if Framingham criteria are satisfied. Subjects without a CHF
diagnosis but with diastolic or systolic dysfunction were considered as having
either preclinical diastolic or preclinical systolic dysfunction.
Results The prevalence of validated CHF was 2.2% (95% confidence interval [CI],
1.6%-2.8%) with 44% having an EF higher than 50%. Overall, 20.8% (95% CI,
19.0%-22.7%) of the population had mild diastolic dysfunction, 6.6% (95% CI,
5.5%-7.8%) had moderate diastolic dysfunction, and 0.7% (95% CI, 0.3%-1.1%)
had severe diastolic dysfunction with 5.6% (95% CI, 4.5%-6.7%) of the population
having moderate or severe diastolic dysfunction with normal EF. The prevalence
of any systolic dysfunction (EF ≤50%) was 6.0% (95% CI, 5.0%-7.1%) with
moderate or severe systolic dysfunction (EF ≤40%) being present in 2.0%
(95% CI, 1.4%-2.5%). CHF was much more common among those with systolic or
diastolic dysfunction than in those with normal ventricular function. However,
even among those with moderate or severe diastolic or systolic dysfunction,
less than half had recognized CHF. In multivariate analysis, controlling for
age, sex, and EF, mild diastolic dysfunction (hazard ratio, 8.31 [95% CI,
3.00-23.1], P<.001) and moderate or severe diastolic
dysfunction (hazard ratio, 10.17 [95% CI, 3.28-31.0], P<.001) were predictive of all-cause mortality.
Conclusions In the community, systolic dysfunction is frequently present in individuals
without recognized CHF. Furthermore, diastolic dysfunction as rigorously defined
by comprehensive Doppler techniques is common, often not accompanied by recognized
CHF, and associated with marked increases in all-cause mortality.
Congestive heart failure (CHF) is a clinical syndrome defined by characteristic
symptoms and physical findings. Echocardiography is often performed in patients
with CHF to measure the ejection fraction (EF) and determine if systolic function
is reduced, systolic CHF or preserved, diastolic CHF. Comprehensive Doppler
echocardiography can now characterize diastolic function directly in addition
to measurement of the EF.
Cardiovascular diseases (CVDs) such as hypertension, coronary artery
disease, and cardiomyopathies often lead to systolic and diastolic ventricular
dysfunction. Nearly all patients with systolic dysfunction have some degree
of concomitant diastolic dysfunction, specifically, impaired relaxation and
variable decreases in ventricular compliance.1 However,
it is now recognized that patients with normal EF can display marked impairment
in diastolic function (isolated diastolic dysfunction).2
Clinically, it has been recognized that some patients with advanced
systolic dysfunction remain free of symptoms of CHF. Thus, individuals may
have systolic dysfunction without receiving a diagnosis of or treatment for
CHF. This has been termed preclinical systolic dysfunction and may be common.3- 5 However,
the prevalence of diastolic dysfunction and its relation to systolic dysfunction
and recognized CHF in the community is unclear.
The efficacy of therapy to abort or delay the progression of preclinical
systolic dysfunction to CHF is recognized by CHF practice guidelines.6 Studies indicate that even simple Doppler evidence
of diastolic dysfunction is an independent risk factor for the future development
of CHF and cardiac death.7,8 Thus,
if common, early recognition and treatment of preclinical systolic and diastolic
dysfunction represent a potentially powerful strategy to reduce the incidence
Our objective was to establish the prevalence of preclinical systolic
and diastolic dysfunction and the prevalence of CHF in randomly selected residents
of Olmsted County, Minnesota,3 aged 45 years
or older. Furthermore, we sought to determine whether the presence of diastolic
dysfunction is independently predictive of all-cause mortality.
In 1990, 96% of the 106 470 residents of Olmsted County were white.
Other characteristics of this population have been previously described.9- 11 The Mayo Foundation
institutional review board approved this study.
Using the resources of the Rochester Epidemiology Project,10 a
random sample of residents who were at least 45 years old as of January 1,
1997, was identified. Participants were enrolled and studied during a 3-year
period, ending September 30, 2000. Of the 4203 eligible residents invited,
2042 (47%) participated. Analysis of the medical records of 500 randomly selected
residents who did not participate revealed similar age and sex distribution
to that observed in participants and a similar prevalence of hypertension,
coronary artery disease, previous myocardial infarction, diabetes, previous
cardiovascular hospitalization, and CHF.
Community medical records for each participant were reviewed by trained
nurse abstractors using established criteria for hypertension12 or
myocardial infarction.13 In addition, clinical
diagnoses of coronary artery disease and diabetes mellitus were recorded.
Each participant underwent a focused physical examination that included measurement
of blood pressure, height, and weight. Body mass index (BMI) was calculated
as weight in kilograms divided by the square of height in meters. Each participant's
medical records were reviewed to determine if any diagnosis of CHF had been
made. If so, each medical encounter was reviewed to determine whether the
documented clinical information fulfilled Framingham criteria9 (validated
Box). Participants with no CHF diagnosis but with either diastolic or
systolic dysfunction at echocardiography were considered to have preclinical
diastolic or systolic dysfunction. Such designation does not imply that the
participant would definitely develop CHF or did not have symptoms, only that
the participant had not sought evaluation or had not had an evaluation that
resulted in a diagnosis of CHF.
Major CriteriaParoxysmal nocturnal dyspnea
Elevated jugular venous pressure
Third heart sound
Cardiomegaly on chest radiograph
Pulmonary edema on chest radiograph
Minor CriteriaPeripheral edema
Dyspnea on exertion
Heart rate >120/min
Weight loss ≥4.5 kg in 5 days*
*Weight loss ≥4.5 kg in 5 days is considered a major criterion if
it occurred in response to therapy for congestive heart failure (CHF). A patient
was considered to have validated CHF if 2 major criteria were present or 1
major and 2 minor criteria were present concurrently.
All echocardiograms were performed by 1 of 3 registered diagnostic cardiac
sonographers who used the same echocardiographic instrument (HP-2500, Palo
Alto, Calif) according to a standardized protocol and interpreted by a single
echocardiologist (M.M.R.) who was masked to clinical data. Two-dimensional
(2-D) and color Doppler imaging were performed to screen for valvular disease.
In each participant, measurement of EF was performed by M-mode echocardiography
using the modified Quinones formula, by the quantitative 2-D (biplane Simpson)
method, and by the semiquantitative 2-D visual estimate method.11,14- 17
Each participant underwent pulsed-wave Doppler examination of mitral
inflow before and during Valsalva maneuver and of pulmonary venous inflow
and Doppler tissue imaging of the mitral annulus. Diastolic function was categorized
according to the progression of diastolic dysfunction: normal; mild, defined as impaired relaxation without evidence of increased
filling pressures; moderate, defined as impaired
relaxation associated with moderate elevation of filling pressures or pseudonormal
filling, and severe, defined as advanced reduction
in compliance or reversible or fixed restrictive filling as previously described
and validated (Figure 1).1,18 Participants were required to have
2 Doppler criteria consistent with moderate or severe diastolic dysfunction
to be so classified. Subjects with 1 criterion for moderate or severe diastolic
dysfunction or those whose parameters were borderline and suggestive of but
not definitive for diastolic dysfunction were classified as indeterminate
rather than as normal.
Left ventricular mass and left atrial volume were calculated from M-mode
and 2-D measurements, respectively, and were indexed to body surface area
as previously described.19,20
As part of the Rochester Epidemiology Project infrastructure, mortality
data on Olmsted County residents are routinely collected by reviewing community
medical records, death certificates, and obituary notices. Participants were
followed up until death or October 1, 2002, at which time they were censored.
This provided 7000 person-years of follow-up, with a median (25th, 75th percentile)
of 3.5 (2.9, 4.2) person-years of follow-up. Active surveillance of the first
41% (n = 974) of the cohort recruited to participate in our study for a follow-up
visit identified no additional deaths to those identified via the above mechanisms.
For each EF method, the corresponding distribution was summarized as
the empirically estimated cumulative distribution function. The overall prevalence
of systolic dysfunction was estimated for each method among participants from
whom EF was obtained by that method with the corresponding 95% confidence
interval (CI) based on the exact binomial distribution. Similar methods were
used to estimate the overall prevalence of diastolic dysfunction. The association
between the prevalence of systolic dysfunction with clinical variables was
investigated using the χ2 test for univariate associations
and logistic regression when controlling for potential confounding variables.
The Mantel-Haenszel χ2 test of trend was used to investigate
the association between the ordinal scale of diastolic dysfunction with dichotomous
clinical variables and the Spearman correlation coefficient for continuous
variables. Ordinal logistic regression was used to adjust the association
of clinical variables with diastolic dysfunction for age and sex. The survival-free
(of any cause) status was estimated using the Kaplan-Meier method and the
association with diastolic dysfunction was assessed using the log-rank test.
The Cox proportional hazards regression model was used to adjust the association
of diastolic dysfunction with all-cause mortality for age, sex, and EF. The
proportional hazards assumption was evaluated and not rejected using methods
developed by Grambsch and Therneau.21All analyses
were done using SAS version 8 (SAS Institute, Cary, NC) except for the test
of proportional hazards which was done using the survival analysis software
in S-Plus Version 6.1.2 (Seattle, Wash).
The mean (SD) age of study participants was 62.8 (10.6) with 29.4% aged
45 through 54, 30.6% aged 55 through 64, 25.4% aged 65 through 74 years, and
14.6% aged 75 years or older. The mean (SD) BMI was 28.4 (5.41). Of the participants,
8.9% were current and 50.1% were former smokers, 4.5 had diabetes, 12.2% had
a history of coronary artery disease, and 4.8 % had a previous myocardial
The prevalence of any CHF diagnosis was 2.6% (95% CI, 1.9%-3.3%) with
21 participants (41%) having an EF higher than 50%. A validated diagnosis
of CHF was present in 45 participants (2.2%; 95% CI, 1.6%-2.8%) with 20 (44%)
of those having an EF higher than 50%. The mean (SD) time between CHF diagnosis
and the echocardiogram was 4.9 (4.1) years (range, 0.1-16.2 years). The prevalence
of validated CHF increased with age groups: 0.7% for those aged 45 through
54; 1.3% in those aged 55 through 64; 1.5% for those aged 65 through 74; and
8.4% for those aged 75 years or older. The P value
with or without adjustment for sex was P<.001.
The prevalence of validated CHF was 2.7% in men vs 1.7% in women (P = .11, without adjustment for age). When adjusted for age, CHF was
slightly more common in men (P = .03).
Diastolic function was classified as normal or abnormal in 1779 participants
(87.1%) and as indeterminate in 263 participants (12.9%). Among the 45 with
validated CHF, only 1 met Doppler criteria for normal diastolic function.
Twenty-one met Doppler criteria for diastolic dysfunction. The remaining 23
participants were classified as indeterminate, 13 for atrial fibrillation
with mitral inflow deceleration time greater than 140 milliseconds, 1 for
other atrial arrhythmia, 1 for mild mitral stenosis, 3 for E–A fusion,
and 5 who had borderline parameters that were suggestive of diastolic dysfunction
but did not meet all criteria required for designation as diastolic dysfunction.
Overall, 20.8% (95% CI, 19.0%-22.7%) had mild, 6.6% (95% CI, 5.5%-7.8%)
had moderate, and 0.7% (95% CI, 0.3%-1.1%) had severe diastolic dysfunction
(Table 1) with 5.6% (95% CI, 4.5%-6.7%)
having moderate or severe diastolic dysfunction with normal EF. The prevalence
of diastolic dysfunction increased with age, was more common in participants
with CVD, diabetes, or systolic dysfunction, and was equally common in men
and women (Table 1 and Table 2). A greater percentage of particpants
with mild diastolic dysfunction (38.8%) were obese (BMI >30) compared with
those with normal (30.3%) diastolic function or moderate (29.7%) or severe
(23.1%) diastolic dysfunction.
We obtained EF from 78.0% of participants by M-mode, 79.2% by biplane
Simpson method, and 99.7% by 2-D visual methods. Of the 2042 subjects, 1888
(92.5%) had quantitative assessment of EF measured by M-mode or biplane Simpson
method if no M-mode was possible. The mean (SD) EF among participants without
CD was similar by M-mode (63.5% [6.5%]), biplane Simpson (63.9% [6.7%]), and
2-D visual (63.3% [5.4%]) methods.
The cumulative distribution of EF within the population as assessed
by the 3 techniques is displayed in Figure
2, which illustrates the prevalence of systolic dysfunction according
to the EF level and method of measuring it.
In 1888 participants with a quantitative EF assessment, the prevalence
was 6.5% (95% CI, 5.4%-7.6%) for those with an EF of 50% or less and was 1.8%
(95% CI, 1.2%-2.4%) for those with an EF of 40% or less. Among 2036 participants
whose EF was measured by the 2-D visual method, the prevalence was 6.0% (95%
CI, 5.0%-7.1%) for those with an EF of 50% or less and was 2.0% (95% CI, 1.4%-2.5%)
for those with an EF of 40% or less. The prevalence of systolic dysfunction
was lower when restricting the population by using only 1 quantitative method
(Figure 2), was higher in men than
women (Table 3) with and without
controlling for age (P<.001 for all), and increased
with age with and without adjustment for sex (P<.001
for all). Systolic dysfunction determined by the 2-D visual method was more
common in participants with CD (Table 3). All associations were similar when only participants with quantitative
EF assessment were examined.
An EF of 50% or less was present in 1.4% of participants with normal,
10.5% with mild, 19.5% with moderate, 61.5% with severe, and 13.6% with indeterminate
diastolic function. Overall, 100 participants (5.6%; 95% CI, 4.5%-6.7%) had
moderate or severe diastolic dysfunction but had normal EF (isolated diastolic
Among subjects with normal EF and no CHF diagnosis, increasing severity
of diastolic dysfunction was associated with a higher mean (SD) left ventricular
mass index (normal, 92.2 [17.7]; mild, 105.9 [24.8]; moderate, 104.4 [25.1];
and severe, 107.6 [44.2] g/m2; Spearman r =
0.23; P<.001) and mean (SD) left atrial volume
index (normal, 22.9 [6.0]; mild, 24.6 [7.6]; moderate, 30.5 [8.1]; and severe,
46.3 [12.6] mL/m2; Spearman r = 0.20; P<.001).
Figure 3 shows the percentage
of participants with any or validated CHF diagnosis according to the level
of systolic or diastolic dysfunction. The percentage of participants with
recognized CHF increased according to the severity of systolic or diastolic
dysfunction, clearly indicating that diastolic as well as systolic dysfunction
is associated with CHF. However, even when only participants with moderate
or severe diastolic dysfunction or with an EF of 40% or less are considered
less than half had any or validated CHF diagnosis. Of participants with an
EF of 40% or less, 47.5% were taking angiotensin-converting enzyme (ACE) inhibitors
and 22.5% were taking β-blockers. Of subjects with moderate or severe
diastolic dysfunction, 14.2% were taking ACE inhibitors and 40.2%, β-blockers.
The prevalence of preclinical systolic and diastolic dysfunction in
a high-risk group defined by simple clinical characteristics greatly exceeded
that observed in the general adult population (Table 4).
All-cause mortality (n = 48 deaths) was increased among those with diastolic
dysfunction (Figure 4). Multivariate
analysis showed that diastolic dysfunction was predictive of all-cause mortality
even when controlling for age, sex, and EF (Table 5).
This study provides the first estimates of the prevalence of diastolic
dysfunction in the community as assessed by rigorous and hemodynamically validated
Doppler criteria. Diastolic dysfunction was common and isolated diastolic
dysfunction was as common as systolic dysfunction. The frequency of CHF increased
dramatically with increasing severity of diastolic dysfunction. However, even
severe diastolic dysfunction was often preclinical with no recognized CHF
diagnosis. When controlling for age, sex, and EF both mild and moderate or
severe diastolic dysfunction were predictive of all-cause mortality. Systolic
dysfunction was also common. Although the frequency of CHF also increased
with worsening systolic function, even among those with an EF of 40% or less,
fewer than 50% of participants had a diagnosis of CHF. Simple clinical characteristics
allow identification of individuals at highest risk for preclinical diastolic
or systolic dysfunction.
A novel aspect of this study is our effort to describe the frequency
of diastolic dysfunction and its association with systolic dysfunction and
CHF in the population. These data are important because population-based studies9,22- 24 have
repeatedly demonstrated that 40% to 50% of individuals with CHF have normal
EF, a finding we again confirm. Recent studies have documented that individuals
with CHF and normal EF consistently demonstrate diastolic dysfunction when
subjected to hemodynamic study and document that patients presenting with
CHF and normal EF do not have transient systolic dysfunction.2,25 Furthermore,
using more rudimentary indices to assess diastolic function, Aurigemma et
al7 documented that mitral inflow patterns
suggesting mild or moderate or severe diastolic dysfunction were independently
predictive of future development of CHF in free-living older volunteers. Bella
et al8 report that a mitral inflow pattern
suggestive of diastolic dysfunction was associated with increased cardiac
mortality, independent of pertinent covariates. In our study, we used more
rigorous Doppler methods to characterize diastolic function, requiring that
2 indices proven as predictive of advanced diastolic dysfunction be present
for assignment to the moderate or severe diastolic dysfunction category. We
observed the previously described association between diastolic dysfunction
and age.26 These data are consistent with age-associated
increases in CVD and with studies suggesting that senescence itself may be
associated with impairment in diastolic function.27 The
presence of diastolic dysfunction was closely associated with the presence
of CVD, confirming the propensity of hypertension28,29 and
coronary artery disease28,30 to
produce diastolic dysfunction. Even confining the analysis to participants
with normal EF and no CHF, worsening diastolic dysfunction was associated
with increases in indexed left ventricular mass and left atrial volume. These
structural findings support the Doppler evidence of diastolic dysfunction
because the hypertrophied ventricle is more likely to display diastolic dysfunction
and chronic increases in left atrial pressure associated with diastolic dysfunction
would be expected to lead to atrial enlargement. Finally, we demonstrate that,
as rigorously defined in our study, both mild and moderate or severe diastolic
dysfunction was associated with marked increases in all-cause mortality, independent
of age, sex, and EF.
In our study, EF was measured by 3 different techniques to facilitate
comparison to previous studies that have used a variety of EF methods. The
prevalence of any systolic dysfunction was previously reported as 7.7% for
an urban population (aged 25-75 years; biplane Simpson method) in North Glasgow,
Scotland.3 When age, sex, and technique-specific
prevalence rates are compared, we found a somewhat lower prevalence of systolic
dysfunction in Olmsted County. Devereux et al31 reported
that 14.1% of American Indians (aged 45-74 years) had systolic dysfunction.31 Other population-based studies have reported prevalence
rates similar to ours when age, sex, and technique-specific rates were examined.4,32 Although the prevalence of systolic
dysfunction increased with increasing severity of diastolic dysfunction, most
participants with diastolic dysfunction had a normal EF. Indeed, moderate
or severe isolated diastolic dysfunction was as common as systolic dysfunction.
The frequency of any and validated CHF diagnoses among patients with
ventricular dysfunction increases with the severity of ventricular dysfunction,
but even among those with advanced diastolic or systolic dysfunction, more
than 50% have no CHF diagnosis or received no treatment. Although previous
studies also suggest that up to 50% of patients with systolic dysfunction
have preclinical systolic dysfunction,3- 5 the
current data extend previous studies by examining both types of ventricular
dysfunction known to be associated with the development of CHF, their relation
to each other, and the frequency of CHF diagnosis. Easily identified high-risk
groups have a higher prevalence of preclinical ventricular dysfunction.
The lifetime risk of developing CHF for those who have reached the age
of 40 years is 20% for both men and women33 and
exceeds the lifetime risk of many conditions commonly screened for in the
community. Prevention of CHF through the treatment of preclinical systolic
dysfunction is recommended in CHF guidelines.6 Although
we acknowledge the lack of therapies proven to modify disease course in diastolic
CHF, several clinical trials are underway. The current data are crucial if
we are to be poised to extend the paradigm of prevention of CHF through treatment
of preclinical systolic dysfunction to those with preclinical diastolic dysfunction.
Screening strategies to detect preclinical ventricular dysfunction include
Doppler echocardiography and potentially, measurement of plasma brain natriuretic
peptide concentration (BNP). However, studies vary as to the sensitivity and
specificity of BNP for the detection of systolic or diastolic dysfunction
and more data are needed.11,34- 39
The Olmsted County population is primarily white and may not be representative
of national demographics of diastolic and systolic dysfunction. Although comparison
of the clinical characteristics of participants and nonparticipants did not
reveal significant differences, preferential participation by subjects with
or without disease cannot be excluded.
In the community, systolic dysfunction is frequently present in subjects
without recognized CHF. Furthermore, diastolic dysfunction as rigorously defined
by comprehensive Doppler techniques is common, often not accompanied by recognized
CHF and associated with marked increases in all-cause mortality.