Article Text
PDF
Abstract
BACKGROUND Left ventricular enlargement with normal systolic function is common in asymptomatic relatives of patients with familial dilated cardiomyopathy, many of whom progress to overt dilated cardiomyopathy at follow up.
OBJECTIVE To examine maximal and submaximal gas exchange variables of cardiopulmonary exercise testing in asymptomatic relatives with left ventricular enlargement.
DESIGN AND SETTING Controlled evaluation of metabolic exercise performance of patients with dilated cardiomyopathy and asymptomatic relatives with left ventricular enlargement identified through prospective family screening in a cardiomyopathy outpatient clinic.
METHODS 23 relatives with left ventricular enlargement, 33 normal controls, 29 patients with dilated cardiomyopathy, and 10 elite athletes with echocardiographic criteria of left ventricular enlargement (“physiological” enlargement) underwent symptom limited upright cycle ergometry using a ramp protocol.
RESULTS Peak oxygen consumption (pV̅ o 2; mean (SD)) was significantly reduced in relatives with left ventricular enlargement (78 (16.3)%) vnormal controls (96%, p < 0.01) and athletes (152%, p < 0.001), but was higher than in patients with dilated cardiomyopathy (69%, p < 0.01). pV̅ o 2was less than 80% of predicted in 75% of patients, 58% of relatives, 22% of controls, and none of the athletes. Oxygen pulse (pV̅ o 2/heart rate) was less than 80% of predicted in 69% of patients, 35% of relatives, 6% of controls, and none of the athletes. The slope of minute ventilation v CO2production (ΔVE/ΔV̅ co 2) was > 30 in 68% of patients, 50% of relatives, and in none of the controls or athletes. Anaerobic threshold, occurring in relatives at 37 (14)% of the predictedV̅ o 2, was higher than in the patients (32%, p < 0.01) and lower than in the controls (45%, p < 0.05) or in the athletes (55%, p < 0.001).
CONCLUSIONS Maximal and submaximal cardiopulmonary exercise test variables are abnormal in asymptomatic relatives with left ventricular enlargement, in spite of normal systolic function. This provides further evidence that left ventricular enlargement represents subclinical disease in relatives of patients with dilated cardiomyopathy. Metabolic exercise testing can complement echocardiography in identifying relatives at risk for the development of dilated cardiomyopathy.
- cardiomyopathy
- exercise
- diagnosis
Statistics from Altmetric.com
Idiopathic dilated cardiomyopathy usually develops insidiously and is often advanced at the time of diagnosis. Thus prognosis among patients referred to hospital is poor, with a 25–30% one year mortality and a 50% five year mortality.1 The ability to identify early disease would be a significant advance in the current management of this disease and facilitate studies of its pathogenesis.
Evaluation of families of patients with dilated cardiomyopathy has identified a subset of relatives who have left ventricular enlargement (defined as a left ventricular diastolic dimension (LVEDD) of > 112% of that predicted for age and body surface area2 in the presence of normal systolic function, and in the absence of an underlying cause such as hypertension or athletic training.3 ,4) Left ventricular enlargement in a relative of a patient with dilated cardiomyopathy may be a marker of early or mild disease. Relatives with left ventricular enlargement have a higher than expected prevalence of circulating heart specific antibodies, as well as raised circulating cytokines and cardiac creatine kinase isoforms.5 A recent three year follow up study showed that overt dilated cardiomyopathy develops in 27% of patients with left ventricular enlargement.6
Patients with dilated cardiomyopathy have abnormal maximal and submaximal responses to cardiopulmonary exercise testing,7involving both central and peripheral factors.8 We hypothesised that asymptomatic left ventricular enlargement in the relatives of such patients would be accompanied by abnormalities of cardiopulmonary exercise capacity, in contrast to cardiopulmonary performance in individuals with physiological left ventricular enlargement caused by athletic training. Our aim in this study was thus to examine maximal and submaximal responses to cardiopulmonary exercise testing in those relatives of cardiomyopathy patients found to have left ventricular enlargement, and to compare them with patients with overt dilated cardiomyopathy, normal controls, and athletes with “physiological” dilatation of the left ventricle as a result of training.
Methods
PATIENTS
Twenty three consecutive asymptomatic relatives with left ventricular enlargement, from 19 families attending a tertiary referral centre between July 1997 and March 1998, underwent metabolic exercise testing. They were compared with 29 unrelated patients with dilated cardiomyopathy being followed up during the same period. Evaluation of asymptomatic relatives was performed with local ethics committee approval and has been described in detail elsewhere.6 Left ventricular enlargement was defined as an unexplained left ventricular end diastolic dimension (LVEDD) of more than 112% of predicted in the presence of a shortening fraction of greater than 25%.6Predicted left ventricular end diastolic dimension (LVEDDc) was calculated according to the formula of Henry and colleagues2: LVEDDc = [45.3 × BSA0.3] − [0.03 × age] − 7.2, where BSA = body surface area. The measured LVEDD was then expressed as a percentage according to the ratio: LVEDD% = LVEDD/LVEDDc. Dilated cardiomyopathy was defined according to World Health Organization criteria9; in addition, all patients with dilated cardiomyopathy had an LVEDD of > 112% of predicted, with a shortening fraction of < 25%. Patients with coronary disease, hypertension, valvar disease, or a regular alcohol intake of > 21 units/week in men and > 14 units/week in women were excluded. All relatives with left ventricular enlargement filled in a questionnaire about their daily physical activity.
CONTROLS
Forty five volunteers with a normal 12 lead ECG and echocardiogram underwent metabolic exercise testing under identical conditions. All volunteers were required to fill in a questionnaire about their daily physical activity. Individuals performing more than two hours of organised physical training a week were excluded as controls.
Ten healthy elite endurance athletes who fulfilled echocardiographic criteria for left ventricular enlargement were also evaluated. An elite athlete was defined as an individual who had attained a sufficiently high standard in his or her sport to compete at international level.
ECHOCARDIOGRAPHY
All subjects underwent conventional two dimensional echocardiography less than six weeks before the exercise test. End diastolic and end systolic diameters were measured from the short axis views at the level of the tips of the mitral valve leaflets, and shortening fraction was derived from these measurements.10All echocardiograms were analysed blind to the clinical diagnosis. All patients with dilated cardiomyopathy and left ventricular enlargement underwent clinical assessment, 12 lead ECG, and 24 hour Holter monitoring.
EXERCISE
All patients and controls were fasted for two hours before exercise testing. Exercise was performed in the upright position on a Sensormedics ergometrics 800 S cycle ergometer (Ergoline, Bitz, Germany), using a ramp protocol ranging from 10–15 watts/min (selected to ensure adequate stress and avoid premature fatigue) in a quiet air conditioned room with an average temperature of 21°C and full resuscitation facilities. Each test was supervised by an experienced cardiologist, nurse, and senior technician.
Before the test all patients underwent a three minute practice run at zero work rate. A respiratory exchange ratio below 0.85 was required before starting the test. Simultaneous breath by breath gas exchange analysis was performed using a dedicated Sensormedics metabolic cart (V Max 29 Console, Sensormedics, Yorba Linda, California, USA). Respiratory gas was sampled continuously from a mouthpiece and analysed using a 1111D/000 paramagnetic transducer for oxygen and a 2900 MMC non-dispersive infrared sensor for carbon dioxide. The signals underwent analogue to digital conversion for the calculation of oxygen consumption (V̅ o 2) and carbon dioxide production (V̅ co 2) using an established technique.11 Measurements includedV̅ o 2(l/min),V̅ co 2(l/min), minute ventilation (VE), heart rate (beats/min), work rate (WR; watts), and respiratory quotient. Graphs ofV̅ o 2 v V̅ co 2to calculate anaerobic threshold,V̅ o 2 v work rate to calculate aerobic work efficiency, VE v V̅ o 2, oxygen pulse v V̅ o 2, and heart rate v V̅ o 2were generated by an IBM computer (Sensormedics Vmax/Vision software) using on line gas exchange data from the metabolic cart, each variable being plotted at 10 second intervals. Signals from a 12 lead ECG were displayed continuously and recorded at two minute intervals using a Marquette ECG recorder (Marquette Electronics Inc, Milwaukee, Wisconsin, USA). Blood pressure was measured by auscultation at the brachial artery at one minute intervals during exercise and for the first three minutes after exercise, using a mercury sphygmomanometer.
METABOLIC INDICES: DEFINITIONS AND PREDICTED VALUES
The peak oxygen consumption (pV̅ o 2), defined as the highest oxygen consumption achieved during exercise, was calculated by measuring the mean of the highest values over the last 10 seconds of exercise. To allow for age, body size, and sex differences,12 results were expressed as a percentage of the predicted maximum oxygen consumption (V̅ o 2max), which was calculated using established nomograms based on age, sex, height, and weight.13-15 Values below 80% fall below established 95% confidence limits and were considered abnormal.16 The anaerobic threshold was calculated non-invasively using the V slope method17 and expressed as a percentage of the predictedV̅ o 2max. Values under 40% were considered abnormal.18 The oxygen pulse (O2 pulse) was calculated as the quotient of the pV̅ o 2and peak heart rate. Values were expressed as a percentage of the maximum predicted O2 pulse calculated from formulae for predictedV̅ o 2max and predicted maximum heart rate.19 Values below 80% of predicted were considered abnormal. Values above 80% were regarded as normal provided the individual had achieved a maximum heart rate of at least 80%. The work rate (WR) in watts at peak exercise was determined and expressed as a percentage of the predicted maximum work rate calculated according to formulae based on age, sex, height, and weight.20 ,21
Expired volume v expired CO2 was plotted graphically, and the relation (ΔVE/ΔV̅ co 2) was calculated as the slope below the point of anaerobic threshold. The normal gradient of the slope is 26–30.22 A respiratory quotient of less than 1.1 was considered to indicate submaximal effort, and such subjects were excluded from the study. Six patients with dilated cardiomyopathy and two relatives with left ventricular enlargement were excluded on this basis.
STATISTICAL ANALYSIS
Statistical analysis was performed using SPSS for Windows (SPSS Inc, Chicago, Illinois, USA). All variables were tested for normality of distribution using the one sample Kolmogorov–Smirnov test. Unpaired Student's t test, analysis of variance, χ2 test, Mann–Whitney U test, Kruskall–Wallis test, and bivariate correlation tests were used where appropriate. A value of p < 0.05 was considered significant. Metabolic and echocardiographic results were expressed as absolute values and as percentages of predicted values to minimise differences resulting from age and sex disparities. Results are expressed as mean (SD).
Results
All 103 patients and controls underwent exercise testing without complication. Seventy per cent of athletes, 69% of patients with dilated cardiomyopathy, 60% of relatives with left ventricular enlargement, and 67% of normal controls were male. Mean (SD) age was 26 (12.2) years in athletes (range 15–54), 46 (12.6) years in patients (19–70), 32 (12.6) years in relatives (17–62), and 34 (12.8) years in controls (15–66). Two relatives with left ventricular enlargement who undertook more than two hours of structured physical activity a week were included in the analysis.
ECHOCARDIOGRAPHIC DATA
Baseline echocardiographic data for all four groups are shown in table 1. Mean (SD) LVEDD% in relatives with left ventricular enlargement was 118.4 (5). This did not differ significantly from LVEDD% of either athletes or patients with dilated cardiomyopathy, but was higher than in normal controls (p < 0.001). Fractional shortening (FS) in relatives with left ventricular enlargement was 31 (4.4)%; this was lower than in normal controls (37 (5.3)%, p < 0.001) and higher than in the patients (17 (4.4)%, p < 0.001).
Echocardiographic and metabolic exercise data (all groups)
CLINICAL FEATURES OF PATIENTS WITH LEFT VENTRICULAR ENLARGEMENT AND DILATED CARDIOMYOPATHY
Most of the patients with dilated cardiomyopathy were in New York Heart Association (NYHA) functional class I (10 (36)%) or II (16 (54)%); three patients (10%) were in class III. The patients were all symptomatic at presentation. All of them were on angiotensin converting enzyme inhibitors, 45% were on diuretics, 54% on digoxin, 55% on warfarin, 20% on amiodarone, 10% on β blockers, and 7% on angiotensin II antagonists. All the relatives were asymptomatic and none was receiving cardiac drugs.
Nineteen of the patients with cardiomyopathy (66%) were in sinus rhythm and 10 (34%) had atrial fibrillation; six (20%) had left axis deviation on the ECG, and five (17%) had left bundle branch block. Non-sustained ventricular tachycardia (three or more ventricular ectopic beats at a rate of ⩾ 120 beats/min) was documented at initial assessment in 11 (39%) of the patients. The asymptomatic subjects with left ventricular enlargement were all in sinus rhythm. The relatives with left ventricular enlargement all had normal ECGs.
EXERCISE VARIABLES
Metabolic exercise variables are shown for all four groups in table 1. Maximum heart rate (per cent predicted) in the relatives did not differ significantly from controls or athletes but was higher than in the patients (92 (9.8)% v 89 (15)%, p = 0.02). Relatives achieved 83 (20)% of their predicted maximum work rate v 63% (p < 0.004) in the patients, 104% (p = 0.001) in normal controls, and 156% (p < 0.001) in athletes. Similar significant differences were observed when absolute values (beats/min and watts respectively) were analysed (table 1).
Oxygen pulse
The results are shown in fig 1. Oxygen pulse in the relatives was 89 (17)% of predicted, which was lower than that observed in normal controls (105%, p < 0.05) and athletes (167%, p < 0.001), and higher than in the patients (70%, p < 0.01). O2 pulse was less than 80% of predicted in 69% of the patients, in 35% of the relatives, in 6% of the controls, and in none of the athletes (table 2). One patient and one relative had an O2 pulse that appeared normal in spite of a low pV̅ o 2, because of a low maximum heart rate. In 40% of the patients and 46% of the relatives the oxygen pulse did not increase in the expected parabolic pattern23 but began to plateau below the anaerobic threshold. In these patients, heart rate increased more rapidly at the point of flattening.
Boxplot of oxygen pulse values for each group, showing median, quartiles, and 95th centiles. Dashed line represents 80% of predicted oxygen pulse.
Prevalence of abnormal results within each group
Peak oxygen uptake
The results are shown in fig 2. pV̅ o 2as a percentage of the predictedV̅ o 2max was significantly reduced in the relatives (78 (16.3)%) compared with the normal controls (96%, p < 0.01) and the athletes (152%, p < 0.001), and was higher than in the patients (69%, p < 0.01). Similar significant differences were observed when absolute values in ml/kg/min were analysed (table 1). Peak oxygen consumption was less than 80% in 76% of the patients, in 58% of the relatives, in 22% of the controls, and in none of the athletes (table 2).
Boxplot of p V̅ o 2per cent predicted for each group, showing median, quartiles, and 95th centiles. Dashed line represents 80% of predicted pV̅ o 2.
ΔVE/ΔV̅ co2
The results are shown in fig 3. The slope of the minute ventilation as a function ofV̅ co 2below anaerobic threshold (ΔVE/ΔV̅ co 2) was increased in the patients, with a mean value of 34 (6.2). The mean ΔVE/ΔV̅ co 2slope was normal in controls and at the upper limit of normal in the relatives. Sixty eight per cent of the patients and 50% of the relatives had a ΔVE/ΔV̅ co 2slope of > 30. No normal controls and no athletes had a value above 30 (table 2).
Boxplot of ΔVE/ΔV̅ co 2slope for each group, showing median, quartiles, and 95th centiles. Dashed line represents the normal slope of 30.
Anaerobic threshold
The results are shown in fig 4. Anaerobic threshold occurred at a mean of 37 (14)% of the predictedV̅ o 2max in the relatives, which was higher than in the patients (32%, p = 0.004) and lower than in the normal controls (45%, p = 0.02) or the athletes (55%, p < 0.001). Anaerobic threshold occurred at less than 40% of the predictedV̅ o 2max in 75% of the patients, 65% of the relatives, 22% of the controls, and none of the athletes (table 2).
CORRELATION BETWEEN ECHOCARDIOGRAPHIC DIMENSIONS, METABOLIC VARIABLES, AND FUNCTIONAL CLASS
When all 103 subjects were analysed together, modest correlations were observed between fractional shortening and O2 pulse (r = 0.49, p < 0.001),V̅ o 2max (r = 0.47, p < 0.001), VE/V̅ co 2(r = −0.49, p < 0.001), maximum heart rate (r = 0.29, p = 0.01), anaerobic threshold (r = 0.42, p < 0.001), NYHA class (r = −0.50, p < 0.001), and WR% (r = 0.53, p < 0.001). LVEDD% correlated poorly with metabolic measurements, even when athletes were excluded from analysis.
Discussion
The results show that asymptomatic relatives of patients with dilated cardiomyopathy, and who also have left ventricular enlargement, have significantly abnormal cardiopulmonary exercise test results. This adds to the serological and immunological evidence that left ventricular enlargement represents early disease in some individuals. While the mechanisms of abnormal responses remain speculative, this study suggests that both cardiac and peripheral mechanisms may be responsible.
LEFT VENTRICULAR FUNCTION IN LEFT VENTRICULAR ENLARGEMENT
Although baseline systolic function was normal in the relatives, analysis of the oxygen pulse pattern suggests that abnormalities of cardiopulmonary exercise measurements may be caused by impaired augmentation of systolic function in response to exercise. Oxygen pulse is the product of stroke volume and arteriovenous oxygen difference.16 The peak oxygen pulse was abnormally low in 35% of the relatives. In 46% of the relatives (compared with 40% of the patients) the oxygen pulse started to plateau with increasing work rate instead of rising in its normal parabolic fashion. One possible mechanism for an early plateau of oxygen pulse is the failure of stroke volume to increase. In patients with dilated cardiomyopathy and left ventricular enlargement who had abnormally flat oxygen pulse responses, the plateau occurred at or shortly after the anaerobic threshold. This phenomenon has previously been described in patients with hypertrophic cardiomyopathy23 and may be caused by the early development of metabolic acidaemia because of a depressed stroke volume response.24
PERIPHERAL MECHANISMS
An alternative explanation for the reduced pV̅ o 2and oxygen pulse in the relatives would be a reduction in arteriovenous oxygen difference in these individuals. Alterations that could affect peripheral oxygen consumption—including changes in peripheral blood flow, ratios of type I to type II muscle fibres, and mitochondrial density, which are observed in patients with dilated cardiomyopathy and heart failure25 ,26—may also occur in left ventricular enlargement. However, as arteriovenous oxygen difference is generally increased in heart failure in spite of the changes described above, a reduction in the relatives is unlikely.
ΔVE/δV̅CO2
Analysis of the ΔVE/ΔV̅ co 2slope also raises the possibility of peripheral mechanisms for reduced exercise performance in relatives with left ventricular enlargement. While mean ΔVE/ΔV̅ co 2was normal in the relatives, 46% had a slope of > 30. A raised ΔVE/ΔV̅ co 2is well documented in patients with congestive cardiac failure,27-29 but the mechanisms of this increased ventilatory response are controversial. Ventilation–perfusion mismatch reflecting increased pulmonary dead space caused by pulmonary vascular abnormalities has been proposed.30 This is supported by the finding that end tidal CO2 was normal in all patients in this study. However, relatives with left ventricular enlargement who do not have heart failure are unlikely to have pulmonary vascular abnormalities. Other mechanisms that may be involved include augmented chemosensitivity,31 an enhanced ergoreflex causing direct stimulation of ventilation by muscle activity,32 and reduced ability to increase gas transfer on exercise.33Similar reflexes may be involved in the regulation of peripheral blood flow and could be responsible for the abnormal oxygen pulse patterns, anaerobic threshold, and ΔV̅ o 2/ΔWR relation seen in the relatives. This suggests that asymptomatic relatives may not only be developing subtle cardiac abnormalities but also have more widespread changes affecting cardiovascular performance.
COMPARISON WITH LEFT VENTRICULAR ENLARGEMENT IN ATHLETES
The contrast between results obtained in the relatives and in the athletes reinforces the view that left ventricular enlargement in relatives is an indicator of underlying pathology. Left ventricular enlargement in athletes is common and represents a normal physiological response to intensive physical training. Of 515 elite athletes evaluated at this institution, 32% had echocardiographic criteria of left ventricular enlargement.34 The high peak oxygen consumption observed in athletes with left ventricular enlargement who underwent cardiopulmonary exercise testing suggests that it is not the ventricular enlargement per se that is responsible for the cardiopulmonary abnormalities observed in the relatives, and furthermore that metabolic exercise testing may be a useful method of distinguishing physiological from pathological left ventricular enlargement.
LIMITATIONS
The definition of left ventricular enlargement as “enlargement with normal systolic function” should be qualified by the recognition that while systolic function is within normal limits in these patients it is lower than in normal controls (30% v35%). Therefore early or mild dilated cardiomyopathy is not manifested exclusively by dilatation but is accompanied by evidence of a decline in systolic function, which nonetheless remains within normal limits, such that left ventricular enlargement is the more striking finding. Isolated depressed fractional shortening is also a recognised finding when relatives are evaluated, but this is less common and is not clearly associated with risk of progression.
Maximum (and hence peak) oxygen consumption is normally distributed within the population. As a result, overlap between groups is evident from the data. However, statistical analysis shows that metabolic indices in each group lie on a continuum, with athletes at one end and patients with cardiomyopathy at the other. The intermediate location of relatives with left ventricular enlargement, between normal controls and patients, is consistent with the hypothesis that left ventricular enlargement represents mild or early dilated cardiomyopathy. Because of this tendency for pV̅ o 2to be normally distributed, metabolic exercise testing cannot in isolation identify abnormal relatives, but may provide useful additional information in the appropriate clinical context. Relatives with left ventricular enlargement who performed more than two hours of organised physical activity a week were not specifically excluded from our analysis. This could have biased this group towards better metabolic indices, in contrast to the controls in whom this was a specific exclusion criterion. The two relatives who exceeded this level of activity had a pV̅ o 2of 95% and 80.5%, respectively. In contrast, the lowest pV̅ o 2among the athletes with left ventricular enlargement was 114% of predicted.
Finally, although the use of cycle ergometry may have limited the attainable pWe gratefully acknowledge the assistance of Mary Gould, Carol Page, and Brian Mist. Sanjay Sharma and Mark W Norman were supported by British Heart Foundation junior fellowships.