Left ventricular (LV) systolic dysfunction has been observed in patients with repaired tetralogy of Fallot (TOF), although its clinical associations are unknown. Adults with repaired TOF were identified from 11 adult congenital heart disease centers. Clinical history was reviewed. Patients with pulmonary atresia were excluded. Echocardiograms were reanalyzed to estimate LV ejection fraction. LV function was defined as normal (LV ejection fraction ≥55%) or mildly (45% to 54%), moderately (35% to 44%), or severely (<35%) decreased. Right ventricular (RV) and LV dimensions and Doppler parameters were remeasured. Function of all valves was qualitatively scored. Of 511 patients studied, LV systolic dysfunction was present in 107 (20.9%, 95% confidence interval 17.4 to 24.5). Specifically, 74 (14.4%) had mildly decreased and 33 (6.3%) had moderately to severely decreased systolic function. Presence of moderate to severe LV dysfunction was associated with male gender, LV enlargement, duration of shunt before repair, history of arrhythmia, QRS duration, implanted cardioverter–defibrillator, and moderate to severe RV dysfunction. Severity or duration of pulmonary regurgitation was not different. In conclusion, LV systolic dysfunction was found in 21% of adult patients with TOF and was associated with shunt duration, RV dysfunction, and arrhythmia.
The primary aim of this study was to determine the prevalence of left ventricular (LV) systolic dysfunction in a large population of patients with repaired tetralogy of Fallot (TOF) seen at adult congenital heart disease centers throughout North America. The secondary aim was to explore demographic, clinical, and functional variables associated with LV systolic dysfunction in this cohort and set the stage for more focused future study.
Methods
We designed a multicenter, retrospective, cross-sectional study with standardized reassessment of echocardiographic data. Each participating center identified patients who had an echocardiogram and had been seen as outpatients within 2 years at the same center. Inclusion criteria were age >18 years and known TOF with successful intracardiac repair. Patients with pulmonary atresia with ventricular septal defect were excluded. We also excluded patients with complex coexisting congenital abnormalities or uninterpretable echocardiographic images as judged by the observer at each center. Investigators from each center obtained demographic, clinical, and echocardiographic data. De-identified data were then sent to a single center for consolidation (University of Washington, Seattle) and analysis. Institutional Review Board approval was obtained at each participating center.
Historical data obtained included year of birth, history of aortopulmonary shunt placement, age at shunt placement, age at intracardiac repair, and any subsequent cardiac surgeries or interventions. History of documented sustained atrial or ventricular arrhythmia, implantation of a pacemaker or cardioverter–defibrillator, and medication use at the time of clinic visit were also documented. We also recorded any known coronary artery disease (CAD), previous myocardial infarction, or coronary intervention. Information about atherosclerotic risk factors recorded included history of hypertension, diabetes, hyperlipidemia, and previous or current tobacco use. Height, weight, and blood pressure at the time of visit were recorded if available. Hypertension on the study day was defined as a recorded systolic blood pressure >140 mm Hg or diastolic blood pressure >85 mm Hg. Overweight was defined as a body mass index ≥25 kg/m 2 and obesity as a body mass index ≥30 kg/m 2 . QRS duration and current rhythm were also recorded. Number of previous pregnancies, if any, was recorded for women. Patients were not contacted to obtain any missing data.
Echocardiograms were reviewed and each observer was asked to provide a subjective estimate of LV ejection fraction (EF) based on any and all available echocardiographic data, as would be done for any standard clinical echocardiogram from each laboratory. Use of the Simpson biplane method was encouraged but not required. Instead, laboratories were asked to integrate all portions of the study for this interpretation. From the EF provided, we defined decreased systolic function as an estimated LVEF <55%. Categories for mildly, moderately, and severely decreased systolic function were defined as LVEFs 45% to 54%, 35% to 44%, and <35% respectively.
Echocardiographic measurements were also made expressly for study purposes in each patient according to a standardized protocol. From parasternal 2 dimensional long-axis views, measurement of LV internal dimension (diastole and systole) and LV wall thickness were made (leading edge to leading edge). Fractional shortening of the left ventricle was quantified from these measurements. Right ventricular (RV) outflow diameter (diastole) was obtained from parasternal short-axis views at the level of the aortic valve. RV long-axis length (diastole and systole) and inlet diameter (diastole) were measured from the apical 4-chamber view. Categorical assessment of RV systolic function (normal, mildly decreased, or moderately to severely decreased) was made subjectively by each observer. Pulmonary valve regurgitation severity was similarly categorized. Duration of pulmonary regurgitation was calculated as age at time of valve replacement (when relevant) minus age at repair. Any other valve dysfunction, particularly aortic regurgitation, was also categorized according to established criteria for valve assessment. Peak tricuspid valve regurgitation velocity was recorded. Atrioventricular valve closure time and ejection time were measured for the right and left ventricles for quantification of the myocardial performance index for the 2 ventricles. Right or left atrial enlargement was indicated as present or absent.
All data were collected on preprinted case-report forms and mailed without patient identifiers to a single center for consolidation. The dataset was reviewed for internal consistency and validity. Data quality checks consisted of identifying and tracking missing, incomplete, or inconsistent data. Illegible data, invalid formats, and invalid codes were flagged. Data queries were issued to each center to clarify and resolve discrepancies on a per-patient basis.
Statistical analysis was performed using SPSS 11.0 for Windows (SPSS, Inc., Chicago, Illinois). Three groups were defined by LVEF, namely normal, mildly decreased, or moderately to severely decreased. Groups were compared using analysis of variance with Tukey post hoc testing, chi-square testing, or nonparametric tests as appropriate. Univariate and multivariate binary logistic regression models were performed with moderately to severely decreased EF as the dependent variable. Variables associated with p values <0.2 in univariate analyses were considered in automated stepwise multivariate selection models, with p values <0.05 for entry and >0.1 for removal. Results are expressed as mean ± SD, median (interquartile range), or number (percentage). A p value <0.05 was considered statistically significant.
Results
Data on 511 patients from 11 centers were included for analysis, all with acceptable echocardiograms for measurement. Of the total cohort, 54% of patients were women and 43% had a previous palliative shunt. Most patients had received a transannular patch repair (81%). Mean age at the time of study was 37.2 ± 12.1 years. Median age at TOF repair for the entire cohort was 6.0 years (interquartile range 6.6).
A histogram of estimated LVEF is shown in Figure 1 . For the entire cohort, mean LVEF was 58 ± 9%. LV systolic dysfunction was found in 107 patients (20.9%, 95% confidence interval [CI] 17.4 to 24.5). Mild dysfunction was present in 74 (14.4%, 95% CI 11.4 to 17.5). Moderately decreased function was found in 27 (5.2%, 95% CI 3.3 to 7.2) and was severely decreased in 6 (1.1%, 95% CI 0.2 to 2.1).
Variables compared between groups based on LV systolic function are presented in Table 1 . There were more men than women with LV dysfunction. Patients with LV dysfunction were older at the time of study. History of previous shunt placement was not different, but shunt duration before repair was longer in patients with moderate to severe LV dysfunction. Age at time of repair was not significantly different, although patients who were ≥18 years old at the time of intracardiac repair were more likely to have a decreased LVEF. Patients with ≥4 previous cardiac surgeries were more likely to have mild LV dysfunction (p = 0.001), whereas patients with moderate to severe LV dysfunction did not differ. Patients with LV dysfunction were more likely to be taking angiotensin-converting enzyme inhibitors, β blockers, or diuretics. Patients with LV dysfunction had a wider QRS duration. These patients were also more likely to have a history of arrhythmia, pacemaker, or implanted cardioverter–defibrillator. Arrhythmia data from this cohort have been recently published.
| Variable | LVEF | p Value | ||
|---|---|---|---|---|
| ≥55% | 45–54% | <45% | ||
| Men | 169/404 (42%) | 45/74 (61%) ⁎ | 21/64 (64%) † | <0.001 |
| Age at study visit | 35.5 ± 10.9 | 40.5 ± 13.8 | 39.7 ± 16.8 † | 0.002 |
| Any shunt placemen | 168/404 (44%) | 38/74 (51%) | 16/33 (48%) | 0.153 |
| Age at first shunt (years) | 2.0 (2.5) | 2.0 (3.5) | 2.0 (5.2) | 0.22 |
| Age at repair (years) | 6.0 (5.0) | 6.0 (6.5) | 11.0 (14.0) † | 0.09 |
| Age ≥18 years at time of repair | 25/365 (6%) | 7/64 (10%) † | 8/21 (28%) ⁎ | <0.001 |
| Shunt duration (years) | 4.0 (4.0) | 4.0 (5.5) | 7.0 (10.2) † | 0.005 |
| Transannular patch | 294/346 (85%) | 49/57 (86%) | 21/25 (84%) | 0.97 |
| Conduit from right ventricle to pulmonary artery | 14/329 (4%) | 4/55 (7%) | 0/28 (11%) | 0.81 |
| Previous pulmonary valve replacement | 153/404 (38%) | 36/74 (49%) | 15/33 (46%) | 0.11 |
| Age at pulmonary valve replacement (if relevant) | 28.9 ± 13.1 | 34.6 ± 18.4 | 38.6 ± 17.1 | 0.07 |
| Cardiac surgeries | 3.3 ± 1.6 | 3.3 ± 0.7 | 3.3 ± 1.6 | 0.8 |
| ≥4 cardiac surgeries | 60/404 (15%) | 23/74 (31%) ⁎ | 8/38 (24%) | 0.005 |
| Angiotensin-converting enzyme inhibitor use currently | 38/397 (10%) | 21/72 (29%) ⁎ | 15/33 (46%) ⁎ | <0.001 |
| β-Blocker use currently | 93/397 (23%) | 29/37 (40%) † | 19/32 (59%) ⁎ | <0.001 |
| Aspirin use currently | 88/398 (22%) | 31/74 (42%) ⁎ | 10/32 (31%) | 0.005 |
| Diuretic use currently | 40/397 (10%) | 19/73 (26%) ⁎ | 19/33 (58%) ⁎ | <0.001 |
| Digoxin use currently | 31/395 (8%) | 10/72 (14%) | 7/32 (22%) † | 0.004 |
| QRS duration (ms) | 148 ± 26 | 150 ± 35 | 167 ± 36 ⁎ | 0.001 |
| Previous arrhythmia | 147/404 (36%) | 35/74 (47%) | 25/33 (76%) ⁎ | <0.001 |
| Implanted cardioverter/defibrillator | 32/393 (8%) | 7/69 (10%) | 11/30 (37%) ⁎ | <0.001 |
| Pacemaker | 40/391 (10%) | 11/70 (16%) | 14/33 (42%) ⁎ | <0.001 |
Comparison of echocardiographic variables between groups is presented in Table 2 . Mild and/or moderate to severe LV systolic dysfunction was associated with larger LV diastolic and systolic internal diameters, lower fractional shortening, higher myocardial performance index, and left atrial enlargement. Although mild aortic regurgitation was common, only 17 patients had moderate to severe aortic regurgitation, which was more prevalent in patients with low LVEF.
| Variable | LVEF | p Value | ||
|---|---|---|---|---|
| ≥55% | 45–54% | <45% | ||
| Left ventricular end-diastolic dimension (mm) | 45 ± 5 | 50 ± 9 ⁎ | 57 ± 5 ⁎ | <0.001 |
| Left ventricular end-systolic dimension (mm) | 30 ± 7 | 38 ± 10 ⁎ | 40 ± 6 ⁎ | <0.001 |
| Fractional shortening (%) | 34 ± 11 | 26 ± 12 ⁎ | 29 ± 8 ⁎ | <0.001 |
| Left ventricular myocardial performance index | 0.39 ± 0.28 | 0.57 ± 0.18 † | 0.55 ± 0.28 † | 0.001 |
| Left atrial enlargement (yes/no) | 65/395 (17%) | 18/72 (25%) | 15/32 (47%) ⁎ | <0.001 |
| Moderate to severe aortic regurgitation | 10/401 (2.5%) | 4/73 (5.5%) | 3/33 (9.1%) † | 0.022 |
| Right ventricular long-axis length diastole (cm) | 8.5 ± 1.7 | 10.1 ± 2.7 † | 9.6 ± 2.3 ⁎ | 0.001 |
| Right ventricular inflow diameter (cm) | 4.2 ± 0.9 | 4.7 ± 1.5 | 5.2 ± 1.3 | 0.05 |
| Right ventricular outflow diameter (cm) | 3.0 ± 0.9 | 3.7 ± 1.5 | 4.2 ± 1.8 † | 0.012 |
| Tricuspid annular excursion (cm) | 1.5 ± 0.9 | 1.1 ± 0.6 † | 1.0 ± 0.4 | 0.013 |
| Right ventricular myocardial performance index | 0.20 ± 0.10 | 0.22 ± 0.08 | 0.33 ± 0.12 | 0.22 |
| Right atrial enlargement (yes/no) | 194/389 (50%) | 46/70 (66%) † | 28/33 (85%) ⁎ | <0.001 |
| Moderate to severe tricuspid regurgitation | 55/397 (14%) | 12/73 (15%) | 13/32 (41%) | <0.001 |
† p <0.05 versus group with normal left ventricular ejection fraction.
Moderate to severe RV systolic dysfunction (n = 85) was more prevalent in patients with LV dysfunction ( Figure 2 ). Categorically, most patients with normal LV function had normal RV function (67%). In patients with moderately to severely decreased LV function, only 28% had a normal right ventricle, whereas 44% had moderate to severe RV dysfunction (p <0.001). RV long-axis length in diastole and outflow diameter was increased in patients with LV dysfunction. Tricuspid annular plane systolic excursion was shorter in patients with mild LV dysfunction but not in those with moderate to severe LV dysfunction. Myocardial performance index of the right ventricle was not different. Right atrial enlargement was more common in patients with LV dysfunction (p <0.001). Tricuspid valve regurgitation was trace/none in 104 (21%), mild in 318 (63%), moderate in 73 (15%), and severe in 7 (1%). Moderate to severe tricuspid regurgitation was more common in those with moderate to severe RV dysfunction (p <0.001) and in those with moderate to severe LV dysfunction ( Table 2 ). We found no relation between tricuspid regurgitation peak velocity and RV or LV systolic dysfunction.
Two hundred sixty-one patients (48%) had moderate to severe pulmonary regurgitation and 237 (39%) had undergone previous pulmonary valve replacement. Pulmonary regurgitation severity was not different in those with or without LVEF <55% (p = 0.17) even after excluding patients with previous pulmonary valve replacement (p = 0.93). There was also no difference in length of time between intracardiac repair and pulmonary valve replacement (mean 23 ± 12 years, p = 0.36).
Known CAD was documented in 19 patients (3.7%), which was not different between groups. Myocardial infarction was documented in only 3 patients. Overall, 21% of patients had ≥1 risk factor for CAD ( Table 3 ), which was disproportionately higher in patients with moderately to severely decreased EF. Patients with any CAD risk factor were older (49.2 ± 11.8 vs 44.1 ± 10.1 years, p <0.001), mostly attributable to diabetic patients who were also older (51.4 ± 12.5 vs 36.9 ± 11.8 years, p <0.001). Diabetes was uncommon overall (17 patients, 3.3% of entire cohort) but more prevalent in patients with LV dysfunction. Hyperlipidemia (11% of patients) was slightly more prevalent in the moderate to severe dysfunction cohort. History of hypertension (13% of total cohort), increased systemic blood pressure on the study day (4%), current tobacco use (11%), and history of tobacco use (16%) were not statistically different between groups. Subjects meeting criteria as overweight were more common in the low EF groups. Patients with previous cerebrovascular events (5.7%) were not more prevalent in either LV dysfunction cohort. After exclusion of any patient with atherosclerosis risk factors, there was no difference in age between LVEF categories. Apart from age at time of visit, analysis was not changed after exclusion of patients with CAD risk factors.
