Author + information
- Received June 17, 1996
- Revision received January 29, 1997
- Accepted February 21, 1997
- Published online June 1, 1997.
- Susan G MacLellan-Tobert, MD, FRCP(C)A,
- David J Driscoll, MDA,*,
- Carl D Mottram, RRTB,
- Douglas W Mahoney, MSABC,
- Peter C Wollan, PhDB and
- Gordon K Danielson, MDC
- ↵*Dr. David J. Driscoll, Section of Pediatric Cardiology, Department of Adolescent and Pediatric Medicine, Mayo Clinic, Rochester, Minnesota 55906.
Objectives. The purpose of this study was to identify the determinants of exercise tolerance in patients with Ebstein’s anomaly.
Background. Patients with Ebstein’s anomaly of the tricuspid valve may have exercise limitation that improves after surgical repair.
Methods. One hundred seventeen patients performed cycle ergometry for a total of 124 tests (preoperative test in 76 patients, postoperative test in 23, test but no operation in 18, preoperative and postoperative test in 7). Multiple linear regression analysis was used to identify predictors of maximal oxygen uptake, oxygen saturation and heart rate at peak exercise.
Results. Age at the time of exercise ranged from 6 to 60 years (median 15). An atrial septal defect was present in 67 patients (88%) preoperatively. Compared with the preoperative group, the postoperative group had significantly higher maximal oxygen uptake (mean [±SD] 20.5 ± 7.4 vs. 25.3 ± 7.0 ml/kg body weight per min, p = 0.006). Postoperative rest and exercise blood oxygen saturation was higher than that measured preoperatively (p = 0.0001). Six of seven patients tested before and after the operation showed improved exercise tolerance. Preoperatively, major predictors of maximal oxygen uptake were oxygen saturation at rest (p = 0.01) and age (p = 0.0001). Preoperatively, the major predictor of oxygen saturation at peak exercise was rest oxygen saturation (p = 0.0001), and major predictors of peak exercise heart rate were rest heart rate (p = 0.01) and rest oxygen saturation (p = 0.01). In the postoperative group, predictors of maximal oxygen uptake included age at exercise testing, male gender and heart size.
Conclusions. Definitive operation for Ebstein’s anomaly results in improved exercise tolerance. Before the operation, rest oxygen saturation is the major predictor of exercise tolerance, oxygen saturation at peak exercise and peak heart rate. Postoperatively, age, gender and heart size influenced maximal oxygen uptake.
(J Am Coll Cardiol 1997;29:1615–22)
Ebstein’s anomaly of the tricuspid valve is a relatively rare malformation that constitutes <1% of all congenital heart defects. The constellation of findings in patients with Ebstein’s anomaly includes downward displacement of the tricuspid valve into the right ventricle, an enlarged redundant anterior leaflet of the tricuspid valve and tethering of the septal and posterior leaflets to the underlying myocardium. A patent foramen ovale or atrial septal defect is present in the majority of patients with Ebstein’s anomaly (). In most cases significant tricuspid valve regurgitation is present, but in a minority of patients tricuspid stenosis can be present. Because of the combination of features present in Ebstein’s anomaly, patients often report significant dyspnea and exercise intolerance ([2, 3]). Using multivariate modeling, we examined a large group of patients with Ebstein’s anomaly to identify the determinants of aerobic capacity, blood oxygen saturation, maximal heart rate and cardiac output and to determine the effect of type of operation on the cardiorespiratory response to exercise.
1.1 Study Group.
One hundred seventeen patients with Ebstein’s anomaly underwent a total of 124 cycle exercise tests between September 1980 and November 1994 (Table 1). Seventy-six patients were tested before surgical repair, 14 patients after tricuspid valve repair and 9 patients after tricuspid valve replacement. The interval between operation and postoperative exercise testing ranged from 1 to 10 years. Eighteen patients who had exercise testing but did not subsequently have surgical repair were analyzed separately from the preoperative group who had subsequent operative repair. Three patients in the preoperative group were taking beta-adrenergic blocking agents at the time of exercise testing and had significantly lower rest and exercise heart rates and exercise tolerance. These three patients were excluded from the multivariate analysis because the beta-blockers may have negatively affected their performance. Seven patients were tested both preoperatively and postoperatively. Fifty-six of the 117 patients in this study were reported previously by Driscoll et al. ().
1.2 Exercise Protocol.
This study was performed as a retrospective review of exercise tests. Institutional Review Board approval was granted. Exercise testing was performed using cycle exercise as described previously ([2, 4]). Seven patients were tested before 1982 using individualized cycle protocols. Since 1982, testing has been performed on an electronically braked cycle ergometer using a protocol of 3-min incremental steps of increasing work load, as described by James et al. ([5, 6]). Oxygen saturation by pulse oximetry was monitored before and during exercise. Patients were coached to exercise to exhaustion, and only tests that represented a peak exercise cardiorespiratory effort, as determined by the examiner, were included in the study. Cardiac output, or effective pulmonary blood flow, was measured using an acetylene-helium rebreathing technique (). Baseline spirometry, including forced expiratory volume in 1 s (FEV1), forced vital capacity (FVC) and maximal voluntary ventilation, was performed in all patients before exercise testing. Measurement of maximal voluntary ventilation was the most effort-dependent variable measured, and in the youngest patients a trend toward suboptimal maximal voluntary ventilation results was noted. Therefore, for purposes of analysis, FEV1× 40 was used to approximate maximal voluntary ventilation for each patient in the study (). Cardiac output determination was attempted in all patients who did not have an atrial septal defect or whose shunt was determined to be small (blood oxygen saturation by pulse oximetry >92% before and during exercise). Fifteen of the 76 patients in the preoperative group and 20 of the 23 patients in the postoperative group had cardiac output determinations at peak exercise. In all patients, the cardiothoracic ratio was measured from the chest X-ray film taken within 1 day of the exercise test.
1.3 Statistical Analysis.
The results of the baseline spirometry and exercise tests were compared with normal values published previously ([6, 9–11]). Continuous variables were reported as mean value ± SD, and categoric variables were presented as percentage of the total. The Wilcoxon rank sum test was used to compare continuous variables between patient groups (e.g., preoperative and postoperative), and the Wilcoxon signed-rank test was used for those patients with both preoperative and postoperative exercise tests. The primary dependent variables of interest were maximal oxygen consumption (V̇o2max, ml/kg body weight per min), blood oxygen saturation as measured by pulse oximetry, heart rate and cardiac output at peak exercise. The independent variables that were investigated for their association with the dependent variables of interest included age at exercise, body surface area, gender, medication usage, FVC, FEV1, respiratory exchange ratio (V̇co2/V̇o2), cardiothoracic ratio, rest heart rate, systolic blood pressure, diastolic blood pressure and log of V̇o2. For the patients with a preoperative test, whether or not an atrial septal defect was present at the time of exercise was also considered as an independent predictor, and the degree of tricuspid regurgitation was considered as an independent predictor in those patients with a postoperative exercise test. Univariate associations were investigated by simple linear regression. Multiple linear regression with a backward variable selection technique was used to identify independent predictors. For each of the dependent variables, the final model was reported along with the associated univariate results. Logarithmic transformation was used to stabilize the variance in the V̇o2data. A p value ≤0.05 was considered statistically significant.
Age at the time of exercise ranged from 6 to 60 years (median 15). An atrial septal defect was present in 67 (88%) of the 76 preoperative patients and in 5 (28%) of the 18 patients who did not have subsequent surgical repair. None of the patients in the postoperative group had an atrial septal defect or a patent foramen ovale.
No significant differences were found between the preoperative and postoperative groups for age, height, weight, body surface area, FEV1, FVC, oxygen uptake at rest or systolic or diastolic blood pressure at rest or with exercise. All patients tolerated exercise testing, and no complications occurred during any of the exercise studies.
2.1 Exercise Tolerance.
Compared with the preoperative group, the postoperative group had significantly greater exercise tolerance, defined by V̇o2at peak exercise (p = 0.006) (Table 2, Fig. 1). Furthermore, preoperatively, oxygen uptake at rest was positively correlated with the logarithm of V̇o2at peak exercise (r = 0.46). In the group of 18 patients who had exercise testing but did not have subsequent surgical repair, the V̇o2was 31.1 ± 9.0 ml/kg per min (95 ± 44% of predicted). Five of the seven patients tested both before and after the operation showed significant improvement in V̇o2(Fig. 2). One patient showed slight improvement and another had significant weight gain (25 kg) in the 2.5 years between exercise tests, resulting in her V̇o2being lower postoperatively than preoperatively.
On multivariate analysis, using logarithmic transformation, independent predictors of V̇o2in the patients in the preoperative group were younger age at exercise testing and blood oxygen saturation at rest (Table 3). There was no significant relation between heart size and V̇o2preoperatively. In the patients in the postoperative group, major predictors of V̇o2were younger age at exercise testing, male gender and smaller heart size on chest x-ray film (Table 3). Tricuspid valve replacement or repair and degree of residual tricuspid regurgitation after repair did not significantly influence exercise tolerance in patients in the postoperative group.
2.2 Arterial Blood Oxygen Saturation.
Postoperative blood oxygen saturation measured at rest was significantly higher than that measured preoperatively (Table 2). Furthermore, blood oxygen saturation at peak exercise was higher in the postoperative group than in the preoperative group (Table 2, Fig. 3). In the group of 18 patients who did not subsequently have an operation, blood oxygen saturation at rest was 95.2 ± 2.6% and at peak exercise it was 92.4 ± 3.6%. Only 5 of those 18 patients had an atrial septal defect. On multivariate analysis, the major predictor of blood oxygen saturation at peak exercise in the preoperative group was blood oxygen saturation at rest (Table 4). Significant correlates for oxygen saturation at peak exercise in the postoperative group were use of medication and minute ventilation (V̇e) at rest (Table 4).
2.3 Presence or Absence of Atrial Septal Defect.
In the preoperative group, comparisons were made between 67 patients with an atrial septal defect and 9 patients without an atrial septal defect (Table 5). No differences were noted between the two groups with regard to age, but patients without an atrial septal defect weighed more and had a larger body surface area. No significant difference for V̇o2existed between preoperative patients with and without an atrial septal defect. However, preoperative patients with an atrial septal defect had significantly lower blood oxygen saturation compared with those without an atrial septal defect. Tidal volume (Vt) at rest and during exercise was significantly higher in preoperative patients without an atrial septal defect.
No significant differences existed between the preoperative group without an atrial septal defect and the postoperative group with regard to V̇o2, blood oxygen saturation or cardiac output. However, FVC was significantly lower in the postoperative group than in the preoperative group of patients without an atrial septal defect (2.9 ± 1.2 vs. 3.8 ± 1.0 liters, p = 0.05). Significant differences were also noted for V̇eat rest (postoperative: 9.5 ± 2.4 liters/min; preoperative without atrial septal defect: 12.9 ± 3.0 liters/min, p = 0.01) and Vtat rest (postoperative: 0.49 ± 0.16 liter; preoperative without atrial septal defect: 0.70 ± 0.19 liter, p = 0.007).
2.4 Heart Rate and Cardiac Output.
Peak heart rate was below that expected in both the preoperative and postoperative groups (Table 2). However, there was no significant difference between these groups with regard to heart rate either at rest or at peak exercise. Rest heart rate and rest blood oxygen saturation were positive predictors of peak heart rate in the preoperative group (Table 6). Predictors of peak heart rate in the postoperative group, using multivariate analysis, included systolic blood pressure at rest, respiratory exchange ratio and smaller heart size on chest X-ray film (Table 6).
Cardiac output was measured at rest in 19 patients, at peak exercise in 15 patients preoperatively and 20 patients postoperatively and in 15 of the 18 patients who did not have a subsequent operation. At rest the cardiac output for all patients was less than the mean for our laboratory but was within the limits of normal (Fig. 4). At peak exercise the majority of patients had slightly lower cardiac output for oxygen uptake. There was depression of the slope of the cardiac output curve in the postoperative group and in the “no operation” group. There were no differences between the preoperative and postoperative groups for cardiac output at rest or with exercise. Cardiac output at peak exercise was significantly higher in the patients who did not have a subsequent operation than in the preoperative patients who had a subsequent operation (cardiac index 6.1 ± 1.8 liters/min per m2in the “no operation” group vs. 4.7 ± 1.6 liters/min per m2in the preoperative group, p = 0.02). However, the patients who did not have a subsequent operation also had a higher V̇o2. No significant predictors of cardiac output were found in either the preoperative or postoperative patients.
2.5 Electrocardiographic Findings.
For the preoperative patients without right bundle branch block, ST segment changes were not present either at rest or with exercise. No significant arrhythmias developed during exercise in either the preoperative or postoperative patients. Preoperatively, 24 patients at rest and 23 patients during exercise had first-degree atrioventricular block. In the postoperative group, seven patients had first-degree atrioventricular block at rest and during exercise. Junctional rhythm was noted in three patients at rest preoperatively and in one patient at rest postoperatively. No patients had junctional rhythm during exercise. Second- and third-degree atrioventricular blocks were not present in any patient. An accessory pathway was noted in 18 patients preoperatively, with eight patients having a manifest pathway noted at rest and during exercise. No accessory pathway was apparent in any of the postoperative patients. Premature atrial contractions were found in two postoperative patients at rest, one postoperative patient during exercise and four postoperative patients during postexercise recovery. Two preoperative patients developed premature atrial contractions during the recovery period. Preoperatively, 10 patients had isolated premature ventricular contractions at rest, 13 with exercise and 18 during the recovery period. Postoperatively, one patient had isolated premature ventricular contractions at rest, four with exercise and three during the recovery period. No additional ventricular arrhythmias were observed in any patient at rest or during exercise.
2.6 Pulmonary Function Variables.
There were no differences between the preoperative and postoperative groups for the baseline spirometric measurements of FEV1, FVC or maximal voluntary ventilation. The mean FEV1for all patients was 2.5 ± 1.0 liters/s (100% of predicted), and the mean FVC for all patients was 2.9 ± 1.3 liters (90% of predicted). Minute ventilation at rest was significantly higher in the preoperative group than in the postoperative group (p = 0.002) (Table 2), and a trend toward significance was noted in Vtat rest. No difference in respiratory frequency was present between the two groups either at rest or with exercise. The ventilatory equivalent for oxygen (V̇e/V̇o2) was higher in the preoperative group both at rest and with exercise than in the postoperative group (Fig. 5). No difference in V̇e/V̇o2existed between the groups with and without an atrial septal defect at rest or with exercise. The 18 patients who did not subsequently have an operation demonstrated a normal response of V̇e/V̇o2during exercise (37.3 ± 6.3 at rest, decreasing to 35.1 ± 6.3 with exercise).
Ebstein’s anomaly of the tricuspid valve is characterized by a redundant anterior leaflet with tethering and displacement of the valve, usually associated with significant regurgitation. An atrial septal defect is found in as many as 79% of patients (). A number of factors may contribute to impaired exercise performance in this group of patients. These factors may include the presence of significant tricuspid regurgitation, poor right ventricular function and decreased cardiac output, right to left shunt and associated abnormal ventilation, cardiomegaly and use of medication. Previous investigators have shown that hypoxia plays a primary role in exercise intolerance, and definitive repair of Ebstein’s anomaly significantly improves exercise performance. In 1985, Barber et al. () described the cardiorespiratory responses to exercise in 14 patients with Ebstein’s anomaly and demonstrated V̇o2was reduced to 43 ± 17% of the predicted normal rate. Subsequently, Driscoll et al. () reported on the exercise response for 38 patients before surgical repair of the tricuspid valve and for 11 patients after repair of the tricuspid valve. Exercise tolerance improved in the postoperative patients; however, the preoperative patients without an atrial septal defect did not differ significantly from postoperative patients. Because of the small number of patients in these previous studies, statistical analysis was limited to paired analysis and simple linear regression. Using multivariate analysis, our data continue to support the previous findings of Barber et al. () and Driscoll et al. ().
3.1 Exercise Tolerance and Oxygen Saturation.
It is apparent that patients with Ebstein’s anomaly have limited aerobic capacity when significant cyanosis and decreased blood oxygen saturation are present. Indeed, exercise tolerance positively correlates with the level of blood oxygen saturation. Driscoll et al. () described a strong positive correlation between V̇o2and rest oxygen saturation >90%, but no correlation between exercise tolerance and rest blood oxygen saturation <90%. Our findings demonstrate little correlation between V̇o2and any oxygen saturation level at rest in the simple linear model (r = 0.22). However, there was a strong correlation between exercise tolerance and rest blood oxygen saturation in the multivariate model.
The group of 18 patients who had exercise testing but did not subsequently have an operation were unique in that they appeared to have milder disease. This was apparent in the higher V̇o2when compared with the preoperative patients who subsequently had surgical repair. Additional supporting evidence for milder disease in the group that did not subsequently have an operation included a higher mean oxygen saturation at peak exercise than that in the preoperative group. Recall that only 5 of the 18 patients who did not have a subsequent operation had an atrial septal defect. The higher oxygen saturations at rest and with exercise reflect the insignificant or absent right to left shunting at the atrial level.
Exercise tolerance after surgical repair was significantly greater than that found preoperatively. Patients with Ebstein’s anomaly who have had surgical repair appear to have better exercise tolerance postoperatively than patients who have had the Fontan operation (V̇o2max, 50% to 60% predicted) (), but less aerobic capacity than patients who have had tetralogy of Fallot repair (V̇o2max, 77% predicted) (). Heart size appeared to negatively influence postoperative exercise tolerance, although tricuspid valve repair, replacement and degree of residual tricuspid regurgitation did not. Heart size in this instance likely reflects right ventricular dilation and dysfunction. Thus, postoperative patients with significant right ventricular dilation and dysfunction will have impaired exercise performance. The finding that blood oxygen saturation at peak exercise postoperatively was negatively influenced by the use of medications may also be a reflection of residual cardiac dysfunction.
3.2 Presence or Absence of Atrial Septal Defect.
Exercise tolerance did not differ between the preoperative patients with or without an atrial septal defect. This finding may seem surprising, but is likely a reflection of the small number of patients in the group without an atrial septal defect. However, differences between oxygen saturation at rest and with exercise were significant, reflecting changes in the volume of right to left shunting at the atrial level. Patients without an atrial septal defect preoperatively were similar to the patients in the postoperative group. This finding suggests that the major factor contributing to exercise intolerance in patients with unrepaired Ebstein’s anomaly is the presence and size of the intracardiac right to left shunt.
3.3 Heart Rate and Cardiac Output.
Peak heart rate was less than that predicted in both the preoperative and postoperative groups. Oxygen saturation had a major influence on heart rate preoperatively but not postoperatively. One may question whether the exercise tests reflected maximal effort. Clinical assessment of exercise effort or a V̇co2/V̇o2ratio >1 (Table 2) supported that the tests likely represented either a peak exercise measurement in patients with significant hypoxia or maximal effort in the others. Furthermore, three patients in whom beta-blockade was thought to negatively influence heart rate were not included in the final analysis. Reduced heart rate at peak exercise has been observed in patients with a variety of cyanotic congenital heart defects both preoperatively and postoperatively and in our patients ([14–16]). The cause of this blunted heart rate response is unclear and probably is multifactorial. It is interesting that perfusion of the carotid body with hypoxic blood can cause bradycardia (). Also, Åstrand and Åstrand () described relative bradycardia in subjects after acclimatization to higher altitude. The lower than expected peak heart rate in the postoperative group may reflect sinus node dysfunction after heart surgery.
The acetylene-helium technique used to measure cardiac output underestimates cardiac output with exercise measured by dye dilution by 7.5% to 12% (). However, because our patient data were compared with normal data collected by using the same technique, the problem of underestimation is obviated. Cardiac output at rest was similar for the preoperative patients, for those who did not subsequently have an operation and for the postoperative patients. However, for all patients the cardiac output at peak exercise was less than the mean for our laboratory, but fell within the limits of normal (Fig. 4). The depression of the slope of the cardiac output curve and the lower cardiac output relative to oxygen uptake may be a reflection of right ventricular dysfunction. Recall that the patients who had cardiac outputs measured had blood oxygen saturations during exercise that remained >90%. Therefore, significant hypoxia was not a factor influencing exercise tolerance and cardiac output. Right ventricular function has been shown to be a major determinant of maximal oxygen consumption in patients with left ventricular dysfunction (). Furthermore, decreased cardiac output response to exercise occurs in patients with chronic right heart failure and correlates with severity of congestive heart failure ().
3.4 Ventilatory Responses.
Abnormal ventilatory patterns previously have been described in cyanotic patients ([2–4]). Typically these patterns are characterized by excessive ventilation at rest and during exercise. Respiratory frequency was increased at rest in both the preoperative and postoperative patients, and this may reflect mild hyperventilation secondary to anxiety in patients about to perform an exercise test. An additional important cause of increased respiratory frequency in the preoperative group may be the presence of a shunt at the atrial level. A right to left shunt can lead to overventilation, as evidenced by the increased respiratory frequency in the patients with an atrial septal defect. In the presence of a right to left shunt, only a portion of the systemic venous return reaches the lung for ventilatory exchange of carbon dioxide. To maintain normal systemic arterial partial pressure of carbon dioxide the patient must compensate by increasing V̇eto allow for removal of carbon dioxide. Although the end-tidal carbon dioxide may be low, the arterial carbon dioxide level is maintained near normal ([2, 21]). Elevation of the V̇e/V̇o2and V̇ealso occurs in the presence of hypoxemia and excessive ventilation or hyperventilation (). A higher V̇e/V̇o2was observed in the preoperative patients compared with the postoperative patients (Fig. 5).
In patients with Ebstein’s anomaly, exercise testing may assist in determining the severity of disease. One of the major contributing factors to preoperative exercise intolerance is hypoxemia. Before the operation, rest blood oxygen saturation is the major predictor of exercise tolerance, blood oxygen saturation at peak exercise and peak heart rate. As expected, the presence of an atrial septal defect significantly influences blood oxygen saturation at rest and at peak exercise. However, in patients with an insignificant shunt, right ventricular dysfunction may be a contributing factor to lower cardiac output response relative to oxygen uptake.
Finally, definitive operation for Ebstein’s anomaly results in improved exercise tolerance. Although the surgical elimination of a right to left shunt plays a significant role in improving exercise performance, the type of operation (tricuspid valve repair or replacement) does not appear to influence postoperative exercise tolerance.
☆ This study was supported by Grant 0138-94 from the Mayo Foundation.
- forced expiratory volume in 1 s
- forced vital capacity
- respiratory exchange ratio
- minute ventilation
- ventilatory equivalent for oxygen
- maximal oxygen consumption
- tidal volume
- Received June 17, 1996.
- Revision received January 29, 1997.
- Accepted February 21, 1997.
- The American College of Cardiology
- Driscoll DJ,
- Fuster V,
- McGoon DC
- Driscoll DJ,
- Staats BA,
- Heise CT,
- Rice MJ,
- Puga FJ,
- Danielson GK
- James FW,
- Kaplan S,
- Glueck CJ,
- Tsay J-Y,
- Knoght MJS,
- Sarwar CJ
- Wasserman K,
- Hansen JE,
- Sue DY,
- Whipp BJ,
- Casaburi R
- Knudson RJ,
- Lebowitz MD,
- Holberg CJ,
- Burrows B
- Driscoll DJ,
- Danielson GK,
- Puga FJ,
- Schaff HV,
- Heise CT,
- Staats BA
- Gold WM,
- Mattioli LF,
- Price AC
- Åstrand P-O,
- Åstrand I
- Weber KT,
- Janiki JS
- ↵Wasserman K, Hansen JE, Sue DY, Whipp BJ, Casaburi R. Principles of Exercise Testing and Interpretation. 2nd ed. Philadelphia: Lea & Febiger, 1994:63–64, 125.