An Echocardiographic Evaluation of Patients With Idiopathic Heart Failure

An Echocardiographic Evaluation of Patients With Idiopathic Heart Failure* Bert Andersson, MD, PhD; Kenneth Caidahl, MD, PhD; and Finn Waagstein, MD , PhD The primary myocardial disease idiopathic dilated cardiomyopathy (IDCM) is not clearly defined in the literature. The description is both morphologic and etiologic. We examined consecutive patients with congestive heart failure (CHF) of unknown cause to identify possible cases of IDCM and to give a detailed description of echocardiographic data and possible diastolic dysfunction in this group. The hospital records of patients aged 16 to 65 years hospitalized due to CHF or IDCM during a 6-year period (N=2,711) were evaluated in a defined region of western Sweden. Twenty-two percent (584/2,711) of these records contained no plausible cause of CHF or IDCM, and among patients being alive, obvious cause was lacking in 411 of 1,516 (27%). These 411 patients were offered a diagnostic investigation, including echocardiography, and they were compared with a randomly selected control group (n=103) from the general population. Of 411 patients, 293 accepted investigation. From the control group, we defined the reference level for left ventricular (LV) dilatation to be >32 mm/m 2 , and reduced ejection fraction according to Teichholz formula to be <50%. Applying these borderlines, we identified LV dilatation and systolic dysfunction to be present in 30%, either dilatation or systolic dysfunction in 36%, and neither in 34%. In patients without any signs of systolic dysfunction 44% (26/59) showed signs of diastolic dysfunction. In a mul-

Jdiopathic dilated cardiomyopathy (IDCM) is a commonly used but poorly defined term.l· Al2

though the clinical entity of a dilated left ventricle (LV) with poor contractility is well recognized by every cardiologist, the term IDCM is often used without consideration of the fact that it is both descriptive and etiologic. 3 In the latter respect, it constitutes an exclusion diagnosis. No particular measurement defines IDCM. A thorough investigation is necessary to exclude other etiologies. To our knowledge, there has been no previous report on echocardiographic findings of IDCM among consecutive patients with congestive heart failure (CHF). The primary objective of the present study was to *From the Wallenberg Laboratory for Cardiovascular Research, the Division of Cardiology (Drs. Andersson and Waagstein), the Division of Clinical Physiology (Dr. Caidahl), Sahlgrenska University Hospital, Goteborg (Sweden) University. Manuscript received June 15, 1994; revision accepted August 29. Reprint requests: Dr. Andersson, Kardiolo_gdivisionen , Sahlgrenska Sjukhuset, S-41345, Goteberg, Sweden

680

tivariate analysis, LV dimension was not independently correlated to disease, although LV dimension was univariately correlated to ejection fraction (EF) (r=-0.59; p<0.0001). However, EF(p<0.0001), left atrial dimension (p<0.0001), and the first third filling fraction (p<0.0001) were the constellation of parameters that most accurately separated patients from controls. By using these three parameters, a positive and negative predictive accuracy of 98% and 61%, respectively, was achieved. Thus, in a consecutive group of patients with idiopathic CHF recruited from a nonselected group of hospitalized patients with CHF, all grades of ventricular function were found. In this group, 30% were identified as having IDCM. We give reference values for the diagnosis of idiopathic IDCM and a simple tool to identify patients with systolic and diastolic dysfunction. (Chest 1995; 107:680-89) BSA=body surface area; CHF=congestive heart failure; EF=ejection fraction; EPSS=E-point septal separation; FrFF=first third filling fraction; IDCM=idiopathic dilated cardiomyopathy; LA= left atrial; LV= left ventricle; NDCM =non dilated cardiomyopathy

Key words: congestive heart failure; dilated cardiomyopathy; diastolic function ; echocardiography; epidemiology; ventricular function

determine the frequency of idiopathic CHF subcategorized into those with echocardiographic IDCM and those with isolated diastolic dysfunction as a cause of heart failure symptoms. Secondary objective was to give a detailed echocardiographic description of functional abnormalities associated with a previous clinical diagnosis of idiopathic CHF. METHODS

The present study recruited patients from the world's largest population of unselected patients with CHF. 4

Primary Medical Record Investigation A survey of all 16- to 65-year-old patients hospitalized due to CHF or IDCM in 1980 through 1987 in the western region of Sweden was performed in 1985 to 1988. Among 1.64 million people in the region , 1.05 million were 16 to 65 years of age. All hospitals in the area participated in the study. During the study period, 2,711 patients were hospitalized due to CHF or IDCM. After exclusion of possible causes of CHF, 584 patients with CHF of unknown origin remained, 202 of whom had been given a diagnosis of IDCM. Of the 584 patients 173 were reported dead at the time of screening, and the remaining 411 patients were

Echocardiographic Evaluation of Idiopathic Heart Failure (Andersson, Caidah/, Waagstein)

Total population

1.64 million

16 - 65 years old

1.05 million

Hospital records CHF/IDCM

2711 (2461)

Possible etiology

1105 +

Alive

Accepted investigation

Adequate echocardiogram

293 (278)

286 (271) +

118 (113)

7

(7)

offered a diagnostic investigation (Fig 1). The patients were contacted by letter and telephone and were offered an echocardiographic investigation. By that time, another 17 patients had died, 19lived outside the region, 14 patients could not be reached, 68 patients were unwilling to participate, and 293 accepted the invitation. Of those 293 accepting investigation, 206 (70%) were men, and the mean age of the group was 60±9.9 years (range, 24 to 73 years). To achieve a control group of approximately one third the size of the study group and with similar age and gender composition, we selected from the County Census Bureau Register a random sample of 92 men and 37 women living in Goteborg and its surroundings. By interview or examination, we excluded controls with diseases that might cause CHF to obtain a group comparable to those with IDCM. Thus, 26 control subjects were excluded due to the following reasons: hypertension in 15, ischemic heart disease in 4, aortic regurgitation in 4, previously undiscovered p~lmonary malignancy in 3, thyroid disease in 3, and diabetes Q:tellitus in l. The remaining control group comprised 77 men and 26 women (mean age, 58± 7.8 years; range, 26 to 67 yearsj. I

1022

Secondary Clinical and Ech,ocardiographic Investigation The western region of Swel:len is served by 19 hospitals, and the patients were invited to investigation at their local hospital. All patients were examined by obe investigator (B.A.) and one assistant. The patients had a phtsical examination and answered a questionnaire concerning symptoms, medical history, treatment, smoking, and alcohol consumption. Experience of dyspnea and

FIGURE l. The flow scheme of the screening procedure. The total number of patients is given in the boxes, with patients who experienced their first hospitalization during the study period in parentheses. chest pain was registered according to the questionnaire of Rose and Blackburn. 5 The Minnesota Classification was used to classify the resting ECG 6 The local echocardiographic equipment was used, which included machines of several manufacturers (Acuson, ATL, Irex III, and Vingmed). Most examinations were performed on one of two scanners (Acuson or Irex III). M-mode and two-dimensional examination was performed with the patient in the left lateral position. M-mode measurements were made according to the recommendations of the American Association of Echocardiography7 Evaluations and calculations of M-mode tracings were performed by two independent investigators who were blinded from clinical data of the patients. Ejection fraction (EF) according to the Teichholz et al 8 formula and LV mass9 were calculated from the LV M-mode tracings. To evaluate LV diastolic function, the first third filling fraction (FTFF) (atrial emptying index) was calculated from the echocardiogram of the aortic root. 9 •10 Aortic root motion, wall dimensions, mitral valve E-point septal separation (EPSS), 11 and LV dimensions were performed from stripchart recordings. However, in some centers where the echocardiographic equipment lacked a strip-chart recorder, LV measurements were performed by means of calipers on the video screen. The two-dimensional investigations were stored on videotape and subsequently analyzed by a blinded observer. The degree of chamber dilatation, gross myocardial hypertrophy, general and regional LV function, and valvular disorders were examined. The regional wall motion was interpreted to determine the presence of severe hypokinesia or akinesia as previously described.12 CHEST 1107131 MARCH, 1995

681

Statistical Analysis The normal range of each echocardiographic parameter was defined as the range between the 2.5th and the 97.5th percentile in the control group. Measurements are presented as mean± SD. The covariance of various echocardiographic measurements was tested by linear regression. A stepwise multivariate selection was used to identify echocardiographic variables that could separate diseased from nondiseased hearts. The ability of different echocardiographic variables to identify heart disease was determined from the combined group of patients and controls. The nosologic sensitivity and specificity were determined, as well as the positive predictive accuracy and the negative predictive accuracy. 13 Differences between means were tested by a two-tailed unpaired Student t test for parametric data and the Wilcoxon test for nonparametric data. Differences between proportions were tested by the x2 test. RESULTS

The investigation of the 293 patients with CHF of unknown cause was performed 44 ± 23 months (range, 5 to 99 months) after their first admission to the hospital due to heart failure. A previous echocardiographic recording was available in 87 of 293 (30%) patients. Satisfactory M-mode or two-dimensional recordings could be obtained in 254 of 293 (87%) and 273 of 293 (93%) patients, respectively, whereas either recording could be adequately assessed in 286 of 293 (98%) . Three patients were confined to wheelchair or bed, which resulted in lack of reliable weight measurements. Because of this, M-mode values adjusted for body surface area could only be calculated in 251 patients. The two-dimensional echo revealed signs of regional LV dysfunction (akinesia or dyskinesia) in 17 of 273 patients (6.2%).

Secondary Cardiomyopathies Patients with previously known hypertrophic cardiomyopathy, valvular disorders, or alcoholism had been excluded at the primary medical record investigation. The investigation did not reveal any patient with classic hypertrophic cardiomyopathy. Echocar-

diography or interview disclosed possible causes of heart failure in 11 patients. Thus, previously undiscovered valvular disease was found in seven patients: mitral insufficiency in one, mitral stenosis in two, and combined aortic valvular dysfunction in four patients. Excessive alcohol consumption was revealed in another three patients. One patient had a history of irradiation against a breast tumor, in which case there might have been a causal relationship. Exclusion of these 11 patients with secondary CHF from further analysis reduced the patient group to 275 patients with idiopathic CHF or IDCM.

Echocardiographic Data The findings of the M-mode investigations are shown in Table 1, as well as the values from the control group. The distribution histograms of LV end-diastolic diameter and EF are depicted in Figure 2. Since LV end-diastolic dimension was significantly correlated to body surface area (BSA) in the patients with CHF (r=0.29; p<0.0001) as well as in the control group (r=0.37; p=0.0004), we have corrected for BSA in dimensional data. By univariate correlation analysis, EF was significantly correlated to the indices of LV dimension (Fig 3). New York Heart Association (NYHA) showed weak, but significant, correlation to FTFF (r= -0.21; p=0.002), to LV end-systolic diameter (r=0.17; p=0.02), and EF (r= -0.16; p=0.02). The 2.5th to the 97.5th percentile of the measurements in the control group was used as reference values to determine extent of cardiac dysfunction in patients with symptomatic heart failure of unknown cause. We subdivided the patients into two groups, 83 patients with IDCM (30% of 275), and 192 patients without IDCM, nondilated cardiomyopathy (NDCM) (70% of 275) . The IDCM group was defined as those having LV dilatation and reduced EF. The NDCM group thus comprised patients with idiopathic myo-

Table 1-M-Mode Echocardiographic Data (Mean±SD) in 275 Patients With Idiopathic Heart Failure and in 103 Healthy Control Subjects* Healthy Controls (n=l03) Study Group, (n =275) Mean±SD Left ventricular end-diastolic diameter, mm / m 2 Left ventricular end-systolic diameter, mm / m 2 Left atrial diameter, mm / m 2 Left ventricular wall diameter, mm Left ventricular wall to end-diastolic diameter ratio Left ventricular mass, g/ m 2 E-point septal separation, mm First third filling fraction Ejection fraction, %

32±5.8 25±7 24±4.2 10±2 0.32±0.08 109±31 14±9.6 0.72±0.23 44± 17

p Value

Mean±SD

2.5-97.5 Percentile

<0.0001 <0.0001 <00001 0.64 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001

27±2.7 17±2.4 20±2.2 9± 1.5 0.37±0.o7 87±19 3±2.8 0.92±0.12 65±7

23-32 13-22 16-25 7-12 0.26-0.51 60-122 0-10 0.54-1.00 50-78

*The normal values of the control groups are defined as the range between the 2.5th and the 97.5th percentile. The p values denote the statistical difference between the two groups.

682

Echocardiographic Evaluation

of Idiopathic Heart Failure (Andersson, Caidah/, Waagstein)

0.4

32

•

0.3 >.

0

c

(I)

:::>

0"

~

u..

0.2

Pa1ients

II

(I)

> ~

Q)

a:

Controls

0.1

60

A ~ Q)

E

r =- 0.59 p < 0.0001

50


'6

•!.1

£i

•

40

-o"E ~E

30

~~


::;

•

20

E

0

• •

.~:::--

"OE

0 ·;:: Q)

>

.:::: Q)

...J

Left Ventricular End-Diastolic Diameter (mrnlm'}

0.4

A

50

10

0

8

0

25

50

75

100

Ejection fraction (%) ~ c(I)

0.3

:::>

~

u..

-~ 1ij

c:

• •

50

0

~
0.2

a. Q)

Q)

40

rJl

a:

~

0.1

Q)

rJl~

-E .!:E o~

a.

0 84 78 72 66 60 54 48 42 36 30 24 18 12 6 Ejection Fraction (%}

FIGURE 2. Distribution histograms of LV end-diastolic diameter (panel A) and ejection fraction (EF) (panel B) in the idiopathic CHF patient group and in the control group. The limit of LV dilatation and reduced EF as defined by the values in the control group are shown by black vertical bars.

cardia! dysfunction but not fulfilling the criteria of IDCM with both dilatation and systolic dysfunction. Among the 83 patients with IDCM, 52 had a previous diagnosis of IDCM while 31 were revealed by our investigation. In the NDCM group, 98 patients (36% of 275) had signs of abnormal ventricular function , either LV dilatation with normal EF (18 patients) or low EF with normal dimensions (80 patients). Further, 94 of 275 (34%) had normal LV end-diastolic diameter as well as normal EF. In 30 of 87 patients with a former echocardiographic evaluation, the ventricular function had improved as defined by an increase in EF by 0.10 U or more, whereas the conditions of 14 patients had deteriorated. Diastolic function, as measured by an increased left atrial diameter and reduced FTFF, was significantly decreased among the patients with CHF as compared with the controls. In the NDCM group, LV mass was lower than in the IDCM group (Table 2). The LV ventricular wall to end-diastolic diameter ratio was normal among patients with NDCM, but decreased in patients with IDCM. Despite normal LV

w Qi

30

8

• ••• ••

r = - 0.71

p < 0.0001

20

15

~

~

10

~

0 0

25

50

75

100

Ejection fraction (%)

FIGURE 3. Scatter plots showing the relations between ejection fraction and LV dimensions. Panel A (top ), LV end-diastolic diameter vs EF. Panel B (bottom), Anterior mitral leaflet E-point septal separation vs EF. The limit of reduced EF and LV dilatation or E-point septal separation, respectively, as defined by the values in the control group, are shown by dotted lines.

end-diastolic diameter and EF, 13 of the 85 patients (with interpretable M-mode recordings) in the NDCM group presented signs of reduced systolic function, by either enlarged LV end-systolic diameter or an increased EPSS. When these 13 patients were excluded from analysis, among the remaining 72 patients, left atrial (LA) diameter, LV wall diameter, FTFF, and LV mass could all be evaluated in 59 patients. In this group with no signs of reduced LV systolic function, 26 of 59 (44%) had at least one diastolic parameter outside the normal range. In the group with isolated diastolic dysfunction, the correlation analysis revealed significant correlations between FTFF and heart rate (r = -0.59; p=0.002), between LA diameter and dyspnea (r=0.46; p=0.02), CHEST / 107 / 3 / MARCH, 1995

683

Table 2-Echocardiographic M-Mode Data in the Two Subgroups, the IDCM, the NDCM, and the Healthy Control Group* IDCM (n=83) Mean±SD [p vs Controls] Left ventricular end-diastolic diameter, mm/m 2 Left ventricular end-systolic diameter, mm/m 2 Left atrial diameter, mm/m 2 Left ventricular wall diameter, mm Left ventricular wall to end-diastolic diameter ratio Left ventricular mass, g/ m 2 E-point septal separation, mm First third filling fraction Ejection fraction, %

38±3.9 [<0.0001) 33±4.6 [<0.0001) 26±4 [<0.0001] 9± 1.7 [0.51) 0.26±0.007 [<0.0001] 132±32 [<0.0001) 24±7.6 [<0.0001) 0.7±0.24 [<0.0001) 27±9.7 [<0.0001)

p IDCM vs NDCM

1\DCM (n=192) Mean±SD [p vs Controls] 29±3.6 [0.0008) 21 ±3.8 [<0.0001] 23±3.5 [<0.0001] 10±1.7 [0.33] 0.35±0.075 [0.056] 98±24 [0.0004) 10±6.3 [<0.0001] 0.74±0.22 [<0.0001) 52±14 [<0.0001)

<0.0001 <00001 <0.0001 0.12 <0.0001 <00001 <0.0001 0.22 <0.0001

Healthy Controls (n=103) Mean±SD 27±2.7 17±2.4 20±2.2 9±1.5 0.37±0.07 87± 19 3±2.8 0.92±0.12 65±7

*For definition of the groups, see text. The p values denote the statistical differences in comparing each subgroup with the control group, as well as comparison between subgroups.

only EF (p<0.0001), LA diameter (p<0.0001), and FTFF (p<0.0001) were significant identifiers of diseased vs nondiseased. Setting disease as l.O and nondiseased as 0, the regression formula was 0.993-0.01034XEF+0.266XLA -0.478XFTFF. Using this formula as a diagnostic test, with disease defined as >0.50, sensitivity became 92%, specificity 84%, positive predictive accuracy 93%, and negative predictive accuracy 80%. A simpler and more useful way would be to use the normal values of EF (<50%), LA diameter (>25 mm/m 2), and FTFF (<0.54) to identify disease (Table 3).

and between heart rate and LV wall diameter (r= -0.43; p=0.03). Markers of Disease vs Nondisease The identification of "LV disease" by application of reference limits from the control group was evaluated by various noninvasive measures. Reduced EF and increased LV diameter as markers of heart disease identified 76 patients and 1 control subject in the combined group of patient and controls with adequate M-mode echocardiograms (Table 3). Using only EF as marker of disease identified 138 patients and 3 controls. In a multivariate stepwise procedure,

Table 3-Sensitivity and Specificity of Different Echocardiographic M-Mode Variables Used as Instruments to Identify Patients With Heart Failure*

Combined parameters EF (<50%)+ L VEDD (>32 mm/m 2) EF (<50%) or LAD (>25 mm/m 2 ) or FTFF ( <0.54) Isolated parameters EF (<50%) LVEDD (>32 mm/m 2) LVESD (>22 mm/m 2 ) LAD (>25 mmjm 2 ) EPSS (> 10 mm) FTFF ( <0.54)

Nosologic Sensitivity, %

Nosologic Specificity, %

Positive Predictive Accuracy, %

Negative Predictive Accuracy, %

32

99

99

34

74

97

98

61

57 39 56 31 56 22

97 97 99 100 100 100

98 97 99 100 100 96

44 36 45 37 50 34

*LAD=left atrial diameter; LVEDD=Ieft ventricular end-diastolic diameter; LVESD=Ieft ventricular end-systolic diameter. The different variables were used in a combined group of patients and controls to differentiate disease vs nondisease.

684

Echocardiographic Evaluation of Idiopathic Heart Failure (Andersson, Caidahl, Waagstein)

Table 4-Symptoms and Signs in the Two Subgroups, the IDCM, the NDCM, and the Healthy Control Group* IDCM (n=83) Mean±SD [p vs Controls] Age, yr Male,% NYHA Functional class, 1-5 Dyspnea, 0-5 Chest pain, 0-2 Severe chest pain , 0-1 Body mass index Systolic BP, mm Hg Diastolic BP, mm Hg H eart rate, bpm Alcohol consumption, g/ d Smoker,% ECG Q waves, % Atrial fibrillation , % Left ventricular akinetic segments, %

58±11 [0.82] 80 [0.55] 2.90±1.10 [<0.0001] 2.10± 1.80 [<0.0001] 0.74±0.79 [<0.0001) 0.29±0.46 [<0.0001) 25±4.2 [0.98) 124±22 [0.0004) 76±11 [0.01) 78±16 [0.0001) 5.6±10 [0.79] 18 [0.43) 7.5 [0.01) 19 [<0.0001] 17 [<0.0001)

p!DCM vs NDCM 0.28 0.04 0.13 0.34 0.50 0.28 0.0001 <0.0001 0.004 0.14 0.93 0.13 0.17 0.22 <0.0001

NDCM (n=192) Mean±SD [p vs Controls]

Healthy Controls (n=103) Mean±SD

60±8 [0.054] 66 [0.16] 2.70±1.10 [<0.0001] 1.80± 1.80 [<0.0001] 0.63±0.75 [<0.0001) 0.23±0.42 [<0.0001] 28±5.00 [0.0004) 140±23 [0.24) 80±11 [0.74] 74±16 [0.0001) 5.5±8.9 [0.81) 27 [0.46) 12 [<0.0001] 28 [<0.0001] 2 [0.31)

58±7.8 75 1.30±0.81 0.38±0.82 0.17±0.44 0.04±0.19 25±3.5 136± 16 81±8.5 63±9.1 5.5±8.9 23 0 0 0

*For definition of the groups, see text. The p values d enote the statistical differences in comparing each subgroup with the control group, as well as comparison between subgroups.

Symptoms Symptoms as assessed by questionnaires disclosed significant findings of dyspnea and chest pain in both subgroups (Table 4). The IDCM group as well as the NDCM group also reported a higher frequency of severe chest pain, defined as severe chest pain across the chest lasting for 30 min or more. There were no significant differences between the two groups with regard to symptom grading. The patients with NDCM had higher blood pressure compared with the IDCM group, and increased body mass index. A larger proportion of the patients with IDCM showed LV akinetic segments, predominantly in the septal region. DISCt:SSI0:'\1

We describe the echocardiographic findings from 293 of 411 survivors among 586 consecutive patients hospitalized due to idiopathic CHF or IDCM in a well-defined region of Sweden. In this group, 30% presented signs of both reduced EF and LV dilatation consistent with the description of IDCM. Left ventricular dimensions showed a close relationship to EF.

To our knowledge, this is the first investigation showing that there is a continuous spectrum of LV dysfunction present among patients with a history of CHF. In patients without LV dilatation and with no signs of systolic dysfunction , 44% had impaired diastolic function and/ or LV hypertrophy. Reduced EF, increased LA diameter, or reduced first third FTFF could identify patients with CHF with good sensitivity and specificity.

Etiologic Definition of Dilated Cardiomyopathy The term IDCM comprises the meaning of unknown etiology, and therefore the frequently used idiopathic prefix. Alot of effort has been laid into the search of the primary origin of IDCM, and the entity has been reduced somewhat owing to improved diagnostic tools. In the future, several hitherto unknown causes will probably be identified. Almost all heart disorders, such as ischemic heart disease, hypertensive heart disease, valvular disease, alcohol or drug abuse, cytotoxic treatment, and several systemic diseases can lead to CHF and a morphologic picture CHEST / 107/3/ MARCH, 1995

685

indistinguishable from IDCM. 14 In the present study, we tried to exclude all known causes of CHF , including alcohol-related CHF. 4 The remaining 411 patients in the study thus had idiopathic CHF according to diagnostics of today. Coronary artery disease is probably, besides alcohol, the strongest confounding factor. We believe that this is the first study in which an attempt has been made to estimate the proportion of coronary heart disease in a consecutive population of patients with apparently idiopathic CHF. In the substudy, we were able to show that only 16% of patients diagnosed as having IDCM had significant coronary artery obstruction on angiocardiography (unpublished results). However, neither noninvasive investigations nor symptoms, such as chest pain, Q waves/ 5·16 LV regional akinesia, 17 or exercise-induced ST depressions could distinguish these patients with ischemia from the patients with IDCM.

LV Dilatation and Systolic Function There is no exact morphometric definition of IDCM in the literature. Most studies have used only a general description of ventricular function or arbitrary chosen limits of EF and LV dilatation. 18·19 In a recent report, Manolio et at2° proposed guidelines as how to define IDCM. They suggested EF to be <45 % and LV end-diastolic diameter to be >2.7 cm / m 2, however, without references. It is noteworthy that in our control group, the average value of LV end-diastolic diameter was 2.7 cm/m 2, and the upper normal range was as high as 3.2 cm/m 2. No study has used a proper control group, which has been investigated by the same method as the patient group. This is essential, as different methods, ie, one- and twodimensional echocardiography, angiography, nuclear angiography, give different values of EF and LV dimensions. It is also apparent from the results of this study that LV dimension and EF are completely continuous variables, and the delineation of an abnormally dilated heart is highly dependent on proper reference values. Furthermore, an appropriate correction for BSA is necessary, as LV size is significantly correlated to BSA both in patients and in controls. It has been suggested that systolic dysfunction does not exist until the EF falls below 40%. 21 More than half of our patients with past or present CHF had an EF above 40%, which is in accordance with previous studies. 22·23 It has been suggested that LV end-systolic diameter would be a more sensitive measurement with regard to ventricular dilatation and prognosis, 24 although others have come to conflicting results.25 ·26 In our study, EF was one of the parameters that separated patients from controls, whereas LV dilatation did not. There was a considerable overlap in LV diameter measurements be686

tween patients and controls, and most of the patients had normal LV diameter. Additionally, it is important to stress that echocardiographic findings of reduced EF and LV dilatation are not specific for IDCM, and do not exclude an ischemic, alcoholic, or viral etiology .

Diastolic Dysfunction Isolated diastolic dysfunction has been shown to be part of the spectrum of CHF. 22 ·23 ·27 However , it is difficult to define a state of diastolic dysfunction solely based on an echocardiographic investigation. The diastolic function is very complex and is influenced by several factors, such as age, 28 systolic function,29·30 heart rate,31·32 atrioventricular pressure gradient,33 preload, and mitral regurgitation. 34·35 The additional apprehension of compliance and pressures is necessary to describe fully the diastolic performance of the ventricles. In addition to FTFF, increased LV wall thickness, and thickness to LV dimension ratio, we used LA dimension as a measure of diastolic dysfunction. 36 The Doppler technique was not available on all echocardiographic scanners, and we cannot exclude influence also from mitral regurgitation on LA dimension in some cases. Further, pulsed-wave transmitral Doppler has become a popular tool to evaluate diastolic function. However, there is a phenomenon of "pseudo-normalization" of the E:A ratio among patients with severe LV dysfunction,37·38 by which the method is very difficult to interpret in a larger population. The FTFF used in this study was clearly reduced in patients with severe LV dysfunction, as well as in patients with normal systolic function. There are other reports suggesting that analysis of LV motion may disclose abnormal diastolic function, where Doppler registration is still normaP 9·40 In the present study, 44% of the patients without LV dilatation and with normal systolic function had signs of impaired diastolic function or LV hypertrophy. If CHF is present with normal EF, it has been considered as evidence of diastolic dysfunction. 27 If LA dilatation or a reduced FTFF was added to EF, the efficacy of the diagnostic test improved substantially. This reflects the ability of these two latter variables to identify diastolic CHF.

Defining Dilated Cardiomyopathy In the NDCM group was a subgroup of patients with either dilatation or impaired EF. This type of cardiomyopathy has previously been characterized as "mildly dilated cardiomyopathy" in a small study, 41 and later studies have also paid attention to this condition.42 In these studies, however, the patients were subjected to selection based on LV function before entering the study, and the investigated groups have therefore probably not been represen-

Echocardiographic Evaluation of Idiopathic Heart Failure (Andersson, Caidahf, Waagstein)

tative of a larger population with CHF. In the present study, some investigated patients were selected by a previous diagnosis of IDCM, but most patients were selected on the basis of a previous episode of CHF. Although one third of the patients fulfilled the criteria of IDCM with both LV dilatation and reduced EF, most of the patients had less disturbed LV function. We therefore suggest the nomenclature to be NDCM. "Mildly dilated cardiomyopathy" is not an appropriate term to characterize these patients, because most of them have normal dimensions with reduced EF or diastolic dysfunction with normal EF and LV dimensions. In this study, there were 18 patients with LV dilatation but normal EF. These patients were "dilated" but nevertheless were classified as having NDCM, because they did not fulfill the criteria of IDCM with both reduced EF and dilatation. These patients are depicted in the upper right quadrant of panel A in Figure 3, and it is suggested that this type of cardiomyopathy is a part of the continuous spectrum of LV dysfunction in CHF. Doppler recordings were available only in 5 of these 18 patients, and significant mitral regurgitation could therefore not be excluded by certainty among the other patients. To diagnose IDCM, besides exclusion of specific causes of CHF (including alcohol and ischemic heart disease), it is implied that LV function is compared with a normal control group. In our study we used M-mode echocardiography and the Teichholz EF formula for evaluation of LV function, utilizing the 2.5th to the 97.5th percentile to delineate the normal ranges. Using the mean values ± 2 SD yields similar results. It is suggested that both EF (<50%) and LV dimension (>32 mm/ m 2) should be outside the normal range to assign a diagnosis of IDCM. If other methods are used to assess LV function, similar comparisons with normal control values should be performed .

Limitations of the Study The study population was confined to previously hospitalized patients, excluding mild forms of CHF. However, to test the possibility to extend the population further, we examined the only computerized outpatient record register in the region. No additional case of CHF or IDCM was revealed by this register. This reflects a tendency by the physicians to admit new patients with CHF aged 16 to 65 years to the hospital. Patients with CHF who were not hospitalized because of mild disease or patients dying outside the hospital before their conditions were diagnosed would not have been identified by this investigation. In this study, we chose to investigate the patients after the acute hospital admission, to be able to study them in a more stable phase of the dis-

ease. By this procedure, the most severely diseased patients died before investigation. However, most patients' conditions improve following institution of treatment with modern cardiovascular drugs,43,44 and investigations during the phase of acute CHF would not be representative of the total CHF population. The area of investigation was well defined and it is unlikely that more than a few patients from the area were hospitalized due to CHF outside the region only. Hospital care is almost free of charge in Sweden, but the patients are confined to visit the nearest county hospital, except in emergency cases. Even though our population was restricted to patients hospitalized below the age of 66 years, we were able to review the largest group of patients with idiopathic CHF, so far, in comparison with a randomly selected age- and sex-matched control group. There are limitations in using echocardiography as a diagnostic tool, expressed by the fact that 13% of the patients could not be adequately investigated by M-mode or two-dimensional echocardiography. However, as yet there is no other method that is suitable for screening procedures. The present data are limited by lacking quantative two-dimensional recordings and Doppler echocardiography, which are used routinely today. In the absence of LV aneurysm, M-mode echocardiography is a reliable tool to interpret LV function, and the time-resolution is superior to other noninvasive methods. Twodimensional echocardiography is widely used for evaluating EF, but the method has not been validated for use in larger unselected populations or for the assessment of volumes. Doppler was not generally available at the time of investigation, but should have been of value in evaluating valvular and diastolic function. However, patients with echocardiographic findings suggesting the possibility of valvular heart disease were referred for an extended noninvasive evaluation, including Doppler. CONCLUSIONS

Our study presents the largest echocardiographic study of patients with idiopathic CHF, and we were also able to compare the results with a random sample of normal subjects. To our knowledge, this is the first study to show that reductions of contractility and ventricular dilatation are truly continuous phenomena, and all grades were found in this large sample of patients. The continuous distribution made the proportion of the IDCM group highly dependent on the used control values. We presented reference values for the diagnosis of IDCM. Although one third of the patients had both LV dilatation and reduced EF consistent with IDCM, a simple combination of EF, LA dimension, and FTFF identified both systolic and CHEST / 107 / 3 / MARCH, 1995

687

diastolic heart failure in most of the remammg patients. We suggest the term NDCM to describe this group.

21

ACKNOWLEDGMENT: Eva Lavik Olofsson, Med. Technician., gave technical assistance.

22

REFERENCES

1 Goodwin JF, Oakley CM. The cardiomyopathies. Br Heart J 1972; 34:545-52 2 Abelmann WH. Classification and natural history of primary myocardial disease. Prog Cardiovasc Dis 1984; 27:73-94 3 Roberts WC. Defining idiopathic dilated cardiomyopathy: a courtroom discussion. Am J Cardiol 1989; 63:893-96 4 Andersson B, Waagstein F. Spectrum and outcome of congestive heart failure in a hospitalized population. Am Heart J 1993; 126:632-40 5 Rose GA, Blackburn H. Cardiovascular survey methods: WHO monograph series No 56. Geneva: World Health Organization, 1968; 172-78 6 Rose GA, Blackburn H, Gillum RF, et a!. Cardiovascular survey methods. 2nd ed. Geneva: World Health Organization, 1982; 123-42 7 Sahn DJ, DeMaria A, Kisslo J, eta!. Recommendations regarding quantitation in M-mode echocardiography: results of a survey of echocardiographic measurements. Circulation 1978; 58:1072-83 8 Teichholz LE, Kreulen T, Herman MV, et a!. Problems in echocardiographic volume determinations: echocardiographicangiographic correlations in the presence or absence of asynergy. Am J Cardiol1976; 37:7-11 9 Caidahl K, Eriksson H, Hartford M, et a!. Dyspnoea of cardiac origin in 67 year old men: II. Relation to diastolic left ventricular function and mass: the study of men born in 1913. Br Heart J 1988; 59:329-38 10 Wasserman AG, Meyer JF, Ross AM. The relationship of pulmonary artery wedge pressure to the posterior aortic wall echocardiogram in patients free of obstructive mitral valve disease. Am Heart J 1980; 100:500-05 11 Caidahl K, Sviirdsudd K, Eriksson H, eta!. Dyspnea of cardiac origin in 67-year-old men: relation to clinical and noninvasive measures of cardiac function. Am J Noninvas Cardiol 1988; 2:59-68 12 Andersson B, Caidahl K, Waagstein F. Recovery from left ventricular asynergy in ischemic cardiomyopathy following long-term beta-blockade treatment. Cardiology 1994; 85:14-22 13 Feinstein AR. Clinical epidemiology. Philadelphia: WB Saunders, 1985: 434-37 14 Wynne J, Braunwald E. The cardiomyopathies and m yocarditides: toxic, chemical and physical damage to the heart. In: Braunwald E, ed. Heart Diseases. Philadelphia: W BSaunders Company. 1992; 1394-1450 15 Gau GT, Goodwin JF, Oakley CM, eta!. Qwaves and coronary arteriography in cardiomyopathy. Br Heart J 1972; 34:1034-41 16 Feld H, PriestS, Denson M. Importance of pathologic Q-waves in patients with dilated cardiomyopathies. Am J Med 1993; 94:546-48 17 Wallis DE, O'Connell JB, Henkin RE, et al. Segmental wall motion abnormalities in dilated cardiomyopathy: a common finding and good prognostic sign. J Am Coli Cardiol 1984; 4:674-79 18 Bagger JP, Baandrup U, Rasmussen K, eta!. Cardiomyopathy in western Denmark. Br Heart J 1984; 52:327-31 19 Codd MB, Sugrue DD, Gersh BJ, et al. Epidemiology of idiopathic dilated and hypertrophic cardiomyopathy. Circulation 1989; 80:564-72 20 Manolio T A, Baughman KL, Rodeheffer R, et a!. Prevalence

688

23 24

25 26

27 28

29 30

31

32

33

34 35 36

37 38 39

and etiology of idiopathic dilated cardiomyopathy (summary of a national heart, lung, and blood institute workshop). Am J Cardiol1992; 69:1458-66 Johnson RA, Palacious I. Dilated cardiomyopathies of the adult. N Eng! J Med 1982; 307:1051-58 Dougherty AH, Naccarelli GV, Gray EL, et al. Congestive heart failure with normal systolic function. Am J Cardiol 1984; 54:778-82 Soufer R, Wohlgelernter D, Vita NA, eta!. Intact systolic left ventricular function in clinical congestive heart failure. Am J Cardiol1985; 55:1032-36 White HD, Norris RM , Brown MA, et a!. Left ventricular end-systolic volume as the major determinant of survival after recovery from myocardial infarction. Circulation 1987; 76:44-51 LeeTH, Hamilton MA, Stevenson LW, eta!. Impact of left ventricular cavity size on survival in advanced heart failure. Am J Cardiol1993; 72:672-76 Wong M, Johnson G, Shabetai R, et a!. Echocardiographic variables as prognostic indicators and therapeutic monitors in chronic congestive heart failure-Veterans Affairs Cooperative Studies V-HEFT-1 and V-HEFT-11. Circulation 1993; 87(suppl6):65-70 Cohn JN, Johnson G. Heart failure with normal ejection fraction: the V-HeFT study. Circulation 1990; 81(suppl3):III-48-53 Kuo LC, Quinones A, Rokey R, et al. Quantification of atrial contribution to left ventricular filling by pulsed Doppler echocardiography and the effects of age in normal and diseased hearts. Am J Cardiol1987; 59:1174-78 Courtois M, Mechem CJ, Barzilai B, et a!. Factors related to end-systolic volume are important determinants of peak early diastolic transmitral flow velocity. Circulation 1992; 85:1132-38 Yamamoto K, Masuyama T, Tanouchi J, et al. Importance of left ventricular minimal pressure as a determinant of transmitral flow velocity pattern in the presence of left ventricular systolic dysfunction. J Am Coli Cardiol 1993; 21:662-72 Lavine SJ, Krishnaswami V, Levinson M, et al. Effect of heart rate alterations produced by atrial pacing on the pattern of diastolic filling in normal subjects. Am J Cardiol1988; 62:10981102 Galderisi M, Benjamin EJ, Evans JC, et al. Impact of heart rate and PR interval on doppler indexes of left ventricular diastolic filling in an elderly cohort (the Framingham Heart Study). Am J Cardiol1993; 72:1183-87 Takahashi T, lizuka M, Sato H, et al. Doppler echocardiographic-determined changes in left ventricular diastolic filling flow velocity during the lower body positive and negative pressure method. Am J Cardiol1990; 65:237-41 Lavine SJ, Arends D. Importance of the left ventricular filling pressure on diastolic filling in idiopathic dilated cardiomyopathy. Am J Cardiol1989; 64:61-6 Vanoverschelde J-LJ, Raphael DA, Robert AR, et al. Left ventricular filling in dilated cardiomyopathy: relation to functional class and hemodynamics. JAm Coli Cardiol1990; 15:1288-95 Hamby RI, Zeldis SM, Hoffman I, et al. Left atrial size and left ventricular function in coronary artery disease: an echocardiographic-angiographic correlative study. Cathet Cardiovasc Diagn 1982; 8:173-83 Plotnick GD. Changes in diastolic function-difficult to measure, harder to interpret. Am Heart J 1989; 118:637-41 DeMaria AN, Wisenbaugh TW, Smith MD, et al. Doppler echocardiographic evaluation of diastolic dysfunction. Circulation 1991; 84(suppl 1):1-288-95 Lee CH, Hogan JC, Gibson DG. Diastolic disease in left ventricular hypertrophy: comparison of M mode and Doppler echocardiography for the assessment of rapid ventricular filling. Br Heart J 1991; 65:194-200

Echocardiographic Evaluation of Idiopathic Heart Failure (Andersson, Caidah/, Waagstein)

40 Yoneda Y, Suwa M, Hanada H. eta!. Noninvasive detection of left ventricular diastolic dysfunction using M-rnode echocardiography to assess left ventricular posterior wall kinetics in hypertrophic cardiomyopathy. Am J Cardiol1992; 70:1583-88 41 Keren A, Billingham ME, Weintraub 0 , eta!. Mildly dilated congestive cardiomyopathy. Circulation 1985; 72:302-09 42 Gavazzi A, Demaria R, Renosto G, et al. The spectrum of left ventricular size in dilated cardiomyopathy-clinical correlates

and prognostic implications. Am Heart J 1993; 125:410-22 43 Cintron G, Johnson G, Francis G, eta!. Prognostic significance of serial changes in left ventricular ejection fraction in patients with congestive heart failure. Circulation 1993; 87(suppl 6):17-23 44 Waagstein F, Bristow MR, Swedberg K, et al. Beneficial effects of metoprolol in idiopathic dilated cardiomyopathy. Lancet 1993; 342:1441-46

8TH European Congress of Intensive Care Medicine October 18- 22, 1995; Athens, Greece Sponsored by the European Society of Intensive Care Medicine. For information, contact: Mary Constantinidou, Congress Secretariat, Triaena Congress, 24, Har. Trikoupi St., GR 106 79 Athens, Greece. Tel: +30+1+3609511-15 or Fax: +30+1+3607962.

CHEST / 107 / 3 / MARCH, 1995

689