Kitzman DW, Little WC, Brubaker PH, Anderson RT, Hundley WG, Marburger CT, Brosnihan B, Morgan TM, Stewart KP. Pathophysiological Characterization of Isolated Diastolic Heart Failure in Comparison to Systolic Heart Failure. JAMA. 2002;288(17):2144–2150. doi:10.1001/jama.288.17.2144
Author Affiliations: Sections of Cardiology (Drs Kitzman, Little, and Hundley, and Ms Stewart) and Gerontology (Dr Marburger), Department of Internal Medicine; Department of Public Health Sciences (Drs Anderson and Morgan); and Division of Surgery (Dr Brosnihan), Wake Forest University School of Medicine; and Department of Health and Exercise Sciences (Dr Brubaker), Wake Forest University, Winston-Salem, NC.
Context Many older patients with symptoms of congestive heart failure have a
preserved left ventricular ejection fraction (LVEF). However, the pathophysiology
of this disorder, presumptively termed diastolic heart failure (DHF), is not well characterized and it is unknown whether it represents
true heart failure.
Objective To assess the 4 key pathophysiological domains that characterize classic
heart failure by systematically performing measurements in older patients
with presumed DHF and comparing these results with those from age-matched
healthy volunteers and patients with classic systolic heart failure (SHF).
Design and Setting Observational clinical investigation conducted in 1998 in a general
community and teaching hospital in Winston-Salem, NC.
Participants A total of 147 subjects aged at least 60 years. Fifty-nine had isolated
DHF defined as clinically presumed heart failure, LVEF of at least 50%, and
no evidence of significant coronary, valvular, or pulmonary disease. Sixty
had typical SHF (LVEF ≤35%). Twenty-eight were age-matched healthy volunteer
Main Outcome Measures Left ventricular structure and function, exercise capacity, neuroendocrine
function, and quality of life.
Results By echocardiography, mean (SE) LVEF was 60% (2%) in patients with DHF
vs 31% (2%) in those with SHF and 54% (2%) in controls. Mean (SE) LV mass-volume
ratio was markedly increased in patients with DHF (2.12 [0.14] g/mL) vs those
with SHF (1.22 [0.14] g/mL) (P<.001) and vs controls
(1.49 [0.17] g/mL) (P = .002). Peak oxygen consumption
by expired gas analysis during cycle ergometry was similar in the DHF and
SHF groups (14.2 [0.5] and 13.1 [0.5] mL/kg per minute, respectively; P = .40) and in both was markedly reduced compared with
healthy controls (19.9 [0.7] mL/kg per minute) (P =
.001 for both). Ventilatory anaerobic threshold was similar in the DHF and
SHF groups (9.1 [0.3] and 8.7 [0.3] mL/kg per minute, respectively; P<.001) and in both was reduced compared with healthy
controls (11.5 [0.4] mL/kg per minute) (P<.001).
Norepinephrine levels were similar in the DHF (306  pg/mL) and SHF (287
 pg/mL) groups (P = .56) and in both were markedly
increased vs healthy controls (169  pg/mL) (P =
.007 and .03, respectively). Brain natriuretic peptide was substantially increased
in both the DHF (56  pg/mL) and the SHF (154  pg/mL) groups compared
with healthy controls (3  pg/mL) (P = .02 and
.001, respectively). Quality-of-life decrement score as assessed by the Minnesota
Living with Heart Failure Questionnaire was substantially increased from the
benchmark score of 10 in both groups (SHF: 43.8 [3.9]; DHF: 24.8 [4.4]).
Conclusion Patients with isolated DHF have similar though not as severe pathophysiologic
characteristics compared with patients with typical SHF, including severely
reduced exercise capacity, neuroendocrine activation, and impaired quality
Congestive heart failure (HF) is a major cause of morbidity and mortality
in the United States and is the leading cause of hospitalization in older
patients.1 Several epidemiologic studies have
recently shown that more than 50% of older patients who present with symptoms
of HF have preserved left ventricular (LV) systolic function.1- 6 This
syndrome has been presumptively termed diastolic heart failure (DHF).2,7- 10 However,
clinical criteria for HF are not specific. There are other conditions, which
are particularly common in elderly patients, that cause similar signs and
symptoms. This is further confounded by normal age-related changes in cardiovascular
function, the absence of a practical, definitive diastolic function test,
and relative lack of information regarding the pathophysiology of DHF. These
issues have led to doubts regarding whether patients with DHF have "real"
HF11,12 and may have created barriers
to progress in characterization and treatment of this important disorder of
To address whether DHF represents real HF and to characterize the pathophysiology
of this syndrome, we compared 3 groups of subjects: older patients with symptoms
of HF, a normal ejection fraction, and no other identifiable cause for their
symptoms (isolated DHF); age-matched healthy controls; and age-matched typical
patients with classic systolic heart failure (SHF). Detailed physiologic testing
was performed in 4 domains pivotal to the pathophysiology of the classic HF
syndrome: LV structure and function, exercise performance, neuroendocrine
function, and quality of life.
As previously described,15- 17 SHF
was defined as HF with severely reduced systolic function (LV ejection fraction
≤35%), and isolated DHF was defined as HF with normal systolic function
(LV ejection fraction ≥50%, no segmental wall motion abnormalities) and
no evidence of significant coronary, valvular, infiltrative, pericardial,
or pulmonary disease. The clinical diagnosis of HF was based on previously
described criteria15 that included an HF clinical
score of 3 or greater from NHANES I (National Health and Nutrition Examination
Survey I),18 as well as those criteria used
by Rich et al19 (history of acute pulmonary
edema, or the occurrence of at least 2 of the following with no other identifiable
cause and with improvement following diuresis: dyspnea on exertion, paroxysmal
nocturnal dyspnea, orthopnea, bilateral lower extremity edema, or exertional
A total of 573 patients were identified by review of 1998 clinic visit
and hospital discharge records from the Wake Forest University Medical Center,
Winston-Salem, NC, to potentially fulfill inclusion/exclusion criteria. These
patients were contacted for a screening visit. Thirty-nine percent refused
to participate and 10% were deceased. The screening visit included, in the
following order, a history and physical examination by a board-certified cardiologist,
spirometry, electrocardiogram, rest echocardiogram, and bicycle exercise echocardiogram.15 If an exclusion criterion was found at any stage,
the screening visit was terminated. The echocardiogram and exercise test were
used only for excluding unsuspected valvular or ischemic heart disease and
for confirming that the subject met prescribed ejection fraction criteria.
The exercise test also served to familiarize the subjects with the testing
environment and did not include expired gas analysis. None of the primary
outcomes of the study (LV function, exercise function, neuroendocrine function,
and quality of life) were measured at this visit. This visit resulted in additional
exclusions due to conditions not apparent during the initial records review,
including significant ischemic heart disease (6%); valvular heart disease
(3%); other medical problems (chronic pulmonary disease, anemia, renal failure,
cancer, uncontrolled hypertension, debilitating stroke or arthritis; 12% for
all combined); or because they were found to not have symptoms of HF by the
above criteria (9%). Key demographic variables (age, sex, race) for excluded
patients were similar to those for included patients (70 [SD, 3] years; 52%
women; 13% African American). The remaining 21% met all criteria and entered
the study. There were 60 patients with SHF and 59 patients with isolated DHF.
Age-matched healthy controls (n = 28) were recruited from the community
and were screened to exclude those who had any chronic medical illness; were
receiving any chronic medication; had symptoms or abnormal physical examination
results; had abnormal results on the screening exercise, echocardiogram, or
spirometry test; or who were regularly exercising.
The study protocol was approved by the Wake Forest University Institutional
Review Board and written informed consent was obtained from all patients.
The outcome measures of exercise capacity, LV function, neuroendocrine function,
and health-related quality of life were obtained during a single visit. Tests
were performed in the morning and participants had not eaten food or ingested
caffeine for more than 4 hours. Subjects with SHF and DHF were ambulatory
outpatients who had been stable and well-compensated for at least 6 weeks.
Testing was performed and results were analyzed by individuals blinded to
subject groups and other clinical information.
Exercise Testing Protocol. Exercise testing was performed with patients in the upright position
on an electronically braked bicycle, with expired gas analysis and venous
lactate measurement under continuous electrocardiographic and blood pressure
monitoring.15- 17 Participants
were encouraged to exercise to exhaustion. Peak values were averaged from
the final 30 seconds of the exercise test. Ventilatory anaerobic threshold
was assessed by standardized methods using ventilatory equivalents.15 A 6-minute walk test was performed as described by
Guyatt et al.20
Echocardiography. Echo-Doppler examinations were performed using a Sonos 5500 ultrasound
imaging system with a multiple frequency transducer (Hewlett-Packard, Palo
Alto, Calif).9,21,22 Standard
2-dimensional images were obtained in the parasternal long and short axes,
and in the apical 4- and 2-chamber views. Pulsed-wave Doppler tracings of
mitral valve inflow were recorded at the leaflet tips.21,22 Left
ventricular volumes and Doppler tracings were analyzed using a digital echocardiography
workstation as previously described.9,21,22
Neurohormones. Before exercise testing and after at least 15 minutes of quiet, supine
rest, venous blood samples were drawn into prepared, chilled EDTA vacutainers,
placed on ice, and then centrifuged; plasma was then separated. Aprotinin
(25 µL per milliliter of plasma) was added to the plasma tested for
natriuretic peptides and 100 µL of sodium metabisulfite was added to
the plasma tested for catecholamines. Samples were then stored at −70°C.
Commercially available radioimmunoassays (Phoenix Pharmaceuticals Inc, Mountain
View, Calif) were used for both C-terminal atrial natriuretic peptide (ANP)
and for brain natriuretic peptide-32 (BNP). Norepinephrine was separated and
purified by alumina solid-phase extraction and analyzed by high-pressure liquid
chromatography with electrochemical detection as previously described.23
Quality of Life. The standardized Medical Outcomes Study Short-Form 36-Item Health Survey
(SF-36) was administered to assess general health limitations. The Minnesota
Living with Heart Failure Questionnaire (MLHFQ), a condition-specific measure,
was administered to assess the impact of HF on the patients' well-being.24- 26
All outliers were included after verification of unadjusted data. The
study was designed a priori to compare DHF with normal controls and DHF with
SHF. Adjustments for sex and body surface area were made using analysis of
covariance when groups were compared for outcome variables. Logarithmic transformation
was used for nonnormally distributed variables that were highly skewed (eg,
norepinephrine, ANP, BNP). SAS v8.0 was used for all analyses. All reported P values are 2-sided, with .05 considered significant.
The 3 groups were well matched for age (Table 1). A higher percentage of the DHF group was women compared
with the control group and the SHF group; this is similar to the percentages
reported in the community from the Cardiovascular Health Study2 and
other population-based studies.1,3 All
statistical comparisons of physiologic data (LV function, exercise function,
and quality of life; Table 2, Table 3, and Table 4) were adjusted for sex, and LV volumes and mass (Table 2) also included adjustment for body
surface area. However, these adjustments did not alter any of the overall
results of intergroup comparisons for outcomes.
New York Heart Association (NYHA) class was similarly distributed between
DHF and SHF groups (median, class II; interquartile range, I-III). The majorities
of both the DHF and the SHF groups were NYHA class II at the time of testing.
The DHF group had a higher prevalence of hypertension and higher systolic,
diastolic, and pulse blood pressures compared with SHF and controls.
LV Function. The LV mass/volume ratio was markedly increased in the DHF group compared
with the SHF group and controls (Table 2). Doppler early diastolic filling velocity was increased in the
SHF group compared with the DHF group and controls. Atrial filling velocity
was increased in both the DHF and SHF groups compared with controls. Early
deceleration time was decreased in the SHF group and was similar in the DHF
group compared with controls.
Exercise Performance. All subjects gave an exhaustive exercise effort and had a peak respiratory
exchange ratio of 1.05 or greater; in 87% of subjects it was 1.10 or greater
(Table 3). Peak workload, exercise
time, and oxygen consumption were markedly reduced in the patients with DHF
and SHF compared with healthy controls. In general, impairments were not quite
as severe in the DHF group as in the SHF group.
Ventilatory anaerobic threshold and peak lactate level, both measures
of submaximal exercise performance that are relatively independent of effort,
were also markedly abnormal in the DHF and SHF groups compared with controls.
While SHF patients appeared slightly more impaired than DHF patients, this
was not statistically significant. There were similar findings for 6-minute
walk distance (Table 3).
Heart rate was similar in patients with DHF and SHF at peak and at specific
submaximal workloads. Peak exercise pulse pressure was increased in the DHF
group compared with both the SHF group and controls.
Neuroendocrine Function. Mean (SE) norepinephrine level was similar in the DHF (306  pg/mL)
and SHF (287  pg/mL) groups (P = .56) and in
both was markedly increased compared with controls (169  pg/mL) (P = .007 and .03, respectively) (Figure 1). The mean (SE) BNP level was significantly increased in
both the DHF (56  pg/mL) and the SHF (154  pg/mL) groups compared
with controls (3  pg/mL) (P = .02 and .001, respectively).
The mean ANP level was also significantly increased in both the DHF and the
SHF groups (34  and 92  pg/mL, respectively) compared with controls
(14  pg/mL) (P = .02 and .001, respectively).
Both natriuretic peptides were increased significantly more in the SHF group
than in the DHF group. These data trends were not altered meaningfully after
adjustment for age, sex, and/or medications. Thus, patients with DHF had substantial
Quality of Life. General life functioning scores, as assessed by the standardized SF-36,
were not different between the DHF and SHF groups except for general health
status (Table 4). For both the
DHF and SHF groups, scores were lower (ie, worse quality of life) than benchmarks
for a general population sample of men and women aged 65 to 74 years.25 Benchmark values for the MLHFQ were obtained from
the SOLVD (Studies of Left Ventricular Dysfunction) Prevention Trial,24 based on 172 subjects without overt heart failure,
classified as NYHA class I (mean [SD] ejection fraction, 28% [5%]). Patients
with DHF and SHF both had mean scores that were substantially above the benchmark
of 10 (ie, reduced quality of life). Scores on the MLHFQ were worse in patients
with SHF than in those with DHF. Thus, patients with DHF had substantially
reduced general and symptom-specific quality of life.
In this study, we found that despite markedly different degrees of LV
systolic function, elderly patients with presumed DHF had key pathophysiologic
abnormalities that were qualitatively similar to those of patients with classic
SHF, including severely reduced exercise performance, substantial neuroendocrine
activation, and reduced quality of life. These data support the hypothesis
that isolated DHF is a form of "real" heart failure.
Data from the present study complement those from 2 other recent reports
in validating DHF as a true clinical HF syndrome.9,27 The
first showed that most patients with this syndrome have ejection fractions
within the normal range, even during acute exacerbations.9 The
second showed that most patients with HF symptoms and a normal ejection fraction
have abnormal diastolic function as assessed by invasive measurements.27
While numerous large trials have established specific therapies for
SHF, such trials are lacking for DHF. The finding of similar key pathophysiologic
abnormalities in DHF compared with SHF suggests the possibility that therapies
that have been successful for SHF may have a role in therapy for DHF.
Diastolic heart failure is a heterogeneous disorder. As suggested by
Caruana et al11 and by Banerjee et al,12 many patients with clinically presumed DHF were found
during subsequent screening examinations to have potentially confounding medical
disorders or were found to not have HF. The patients with DHF included in
this study had isolated DHF, as we and others have defined it,8,15,16 a
more "pure" subset of DHF. In the population-based Cardiovascular Health Study,
isolated DHF accounted for 42% of older individuals with HF and a normal LV
ejection fraction.2 Furthermore, demographic
and other characteristics of the patients with DHF in the present study were
similar to those in population-based studies.2,5- 7 Thus,
the results of the present study may well be generalizable.
For most, but not all, variables, impairments were somewhat worse for
SHF than for DHF. This quantitative gradient of pathophysiological abnormality
from healthy controls to DHF to SHF parallels the mortality gradient in epidemiologic
Two variables that were abnormal in patients with DHF but not in those
with SHF were increased LV mass/volume ratio and exercise pulse pressure.
These findings, along with other recent reports,16,30 lend
support to a potential role of increased vascular stiffness in the pathogenesis
Exercise intolerance is the primary symptom of chronic HF, regardless
of etiology. The objective measurements in the present study confirm results
from prior smaller studies and show that elderly patients with DHF have severely
reduced peak and submaximal exercise performance.16,31 Submaximal
exercise performance is relatively independent of patient motivation and is
more relevant than peak performance to activities of daily living. In this
regard, impairments in ventilatory anaerobic threshold and 6-minute walk distance
were similar in patients with DHF and those with SHF.
The "neurohormonal hypothesis" has become central to our understanding
of the pathophysiology and therapy of classic SHF.32 The
severity of neuroendocrine activation is closely related to onset of SHF,33 symptomatic status,34,35 progression,33 survival,36 and response
to therapy. Norepinephrine and BNP are among the most widely studied neurohormones
in SHF. In the present study, both were increased in patients with DHF compared
with healthy controls. Norepinephrine was increased to a similar degree in
patients with DHF and in those with SHF. Level of BNP was not as severely
increased in patients with DHF as in those with SHF. These relationships were
unchanged after adjustment for medications as well as sex. These data indicate
that neuroendocrine activation, a pivotal feature of HF pathophysiology, is
present in patients with isolated DHF.
There are relatively few data available for cardiac peptides in patients
with DHF, and a variety of BNP analytic techniques have been used, complicating
interstudy comparisons.37- 40 The
BNP measurements in the present study were made in well-compensated, stable,
ambulatory outpatients. Thus, the absolute levels of BNP would be expected
to be lower than those reported from hospitalized patients with acute, decompensated
congestive HF. Some patients with DHF in the present study had BNP levels
near those of the controls. However, this was true for the patients with SHF
as well, as could be expected in well-compensated patients, and emphasizes
the importance of intrastudy controls for direct comparison. In addition,
mean BNP level was increased nearly 20-fold in patients with DHF compared
Patients with DHF had severe impairments on the MLHFQ compared with
standards from the SOLVD study24,25 for
nonovert HF, and on the SF-36 compared with benchmarks for the general population.
Scores for most items on the SF-36 were similar in patients with DHF and those
with SHF. However, on the MLHFQ, which emphasizes physical functioning and
symptoms, patients with DHF had less-severe impairments than did those with
This study has limitations. Although demographic variables were similar
to those reported in epidemiologic studies, our study was not population-based
and therefore the potential for selection bias cannot be excluded. Two newer
techniques for assessment of diastolic function—color M-mode and tissue
Doppler—were not widely available at the time of the study.
Patients with isolated DHF have similar key pathophysiologic characteristics
compared with patients with typical SHF, including severely reduced exercise
capacity, neuroendocrine activation, and impaired quality of life. Thus, such
patients have "real" HF. Furthermore, these findings suggest that therapies
that have been successful for SHF may have a role in the therapy for DHF.
Future randomized clinical trials will be needed to determine how best to
manage patients with this common and disabling disorder.