Cardiac Magnetic Resonance Versus Routine Cardiac Catheterization Before Bidirectional Glenn Anastomosis in Infants With Functional Single Ventricle
A Prospective Randomized Trial
Background— Routine preoperative catheterization is standard practice in patients with single-ventricle physiology before bidirectional Glenn anastomosis. Because catheterization is invasive and exposes patients to ionizing radiation, cardiac magnetic resonance (CMR) may be a safe and effective alternative.
Methods and Results— We conducted a prospective, randomized, single-center clinical trial comparing catheterization with CMR in patients considered for bidirectional Glenn operation from February 2003 to June 2006. End points were frequency of adverse events of the preoperative evaluation and a composite score of clinically successful surgery. Of 92 eligible patients, 82 were enrolled on the basis of screening echocardiogram, fulfillment of inclusion criteria, and informed consent. Patients were randomized to catheterization (n=41) or CMR (n=41). There were no baseline differences between groups. Four treatment crossovers occurred, 3 to catheterization and 1 to CMR. Catheter interventions were performed in 17 patients (41%). Catheterization resulted in more minor adverse events (78% versus 5%; P<0.001), longer preoperative hospital stays (median, 2 versus 1 day; P<0.001), and higher hospital charges ($34 477 versus $14 921; P<0.001). There was 1 major adverse event in the CMR group (P=1.0). The operative course and frequency of postoperative complications were similar between the 2 groups. The proportion of patients who had a successful bidirectional Glenn operation was similar (71% versus 83%; P=0.3). At the 3-month follow-up, there were no differences in clinical status, oxygen saturation, or frequency of reinterventions.
Conclusions— CMR is a safe, effective, and less costly alternative to routine catheterization in the evaluation of selected patients before bidirectional Glenn operation. Further studies are necessary to determine whether there are long-term benefits from transcatheter interventions in these patients.
Received June 21, 2007; accepted September 7, 2007.
Patients with severe forms of congenital heart disease resulting in a functionally univentricular heart undergo a series of palliative operations that result in passive systemic venous flow through the pulmonary circulation and a functional single ventricle supporting the systemic circulation. Anastomosis of the superior vena cava to the pulmonary artery, or bidirectional Glenn operation, is a palliative procedure typically done in infancy as an intermediate step before completion of a Fontan operation. Routine cardiac catheterization before bidirectional Glenn operation is considered the standard of care in these patients.1 The main goals of the procedure include assessment of anatomic and hemodynamic suitability for surgery and catheter-based interventions such as balloon dilation of aortic coarctation.2 However, cardiac catheterization is associated with morbidity, especially in the high-risk group of infants with single-ventricle physiology,3–7 and exposure to ionizing radiation in childhood has been linked to increased risk of subsequent cancer.8 Moreover, 2 retrospective studies have demonstrated that patients are rarely excluded from bidirectional Glenn operation on the basis of findings at cardiac catheterization.9,10 Others have questioned the benefits of certain transcatheter interventions.11,12
Clinical Perspective p 2725
Although echocardiography is widely used in patients with single-ventricle physiology, its ability to comprehensively image all the relevant anatomy, particularly the thoracic vasculature, may be limited in some patients.9 Cardiac magnetic resonance (CMR) has been shown to be an accurate diagnostic tool in infants and children with a wide range of congenital heart disease, including functional single ventricle.13–16 Therefore, this noninvasive imaging modality can potentially provide an alternative to prebidirectional Glenn catheterization; indeed 1 retrospective cohort study demonstrated the feasibility of this approach.17 The goal of this prospective, randomized, single-center trial of CMR versus catheterization in infants presenting for bidirectional Glenn operation was to test the hypothesis that CMR can provide the necessary anatomic and functional information required to evaluate such patients safely with fewer adverse events, reduced costs, and equivalent short-term postoperative outcomes.
All patients between 2 months and 5 years of age presenting to Children’s Hospital Boston for consideration for bidirectional Glenn operation from February 2003 to June 2006 were eligible for enrollment. Patients were excluded if preoperative catheterization was performed at another institution or if they had contraindications to CMR (eg, pacemaker). Patients with pulmonary vein stenosis, pulmonary hypertension, severe ventricular dysfunction, severe atrioventricular valvar regurgitation, known large aortopulmonary or venous collateral vessels, or coarctation of the aorta (defined as >50% narrowing of the aortic isthmus relative to the descending aorta or ≥3 m/s peak Doppler flow velocity) were excluded because these conditions are generally considered to warrant catheterization for hemodynamic assessment or transcatheter intervention.
The study was approved by the Scientific Review Committee of the Department of Cardiology and by the Children’s Hospital Boston Committee on Clinical Investigation. Written informed consent was obtained from all parents or legal guardians. A Data and Safety Monitoring Board monitored the study.
Eligible patients underwent a screening echocardiogram, which was independently interpreted by a staff pediatric cardiologist not affiliated with the study and by a study investigator. Standard echocardiographic techniques were used to assess cardiovascular anatomy and function as previously described by our center in this patient population.9 Briefly, echocardiograms included assessments of the systemic and pulmonary veins, atrial septum, atrioventricular valve morphology and function, ventricular anatomy and function, outflow tract and semilunar valve function, aorta, and branch pulmonary arteries. Atrioventricular valve regurgitation was graded as mild, moderate, or severe on the basis of the width of the vena contracta by color Doppler relative to valve annulus diameter visualized from 2 orthogonal views. Systemic ventricular function was graded on the basis of visual assessment of ejection fraction. Mild dysfunction was an ejection fraction ≈50%; moderate dysfunction, ≈40%; and severe dysfunction, ≈30% or less. A consensus between 2 echocardiographers was required for the diagnosis of severe atrioventricular valve regurgitation or severe systemic ventricular dysfunction.
In the absence of exclusion criteria, patients were enrolled and randomized in a 1:1 fashion to undergo either cardiac catheterization or CMR. Randomization was performed with a computerized random-number generator in permuted blocks of sizes 2 and 4. Modality crossover after randomization was allowed and analyzed on an intention-to-treat basis. Data collection began at the time of enrollment and continued prospectively through the preoperative evaluation, operative and postoperative hospitalization, and clinical assessment 3 months after surgery.
Patients randomized to cardiac catheterization underwent the procedure according to clinical practice at our laboratory.18 Accordingly, patients were admitted to the hospital the evening before catheterization for intravenous hydration. Catheterizations were performed under intravenous sedation or general anesthesia as deemed appropriate by the managing team.
CMR examinations were performed under general anesthesia according to our published clinical practice.16 To facilitate potential crossover from CMR to catheterization, the study protocol required preadmission and intravenous hydration in all patients enrolled in the CMR arm. CMR studies were performed on a 1.5-T scanner, and vital signs (cutaneous oxygen saturation, end-tidal CO2, and ECG) were continuously monitored. The imaging protocol included ECG-gated, breath-hold steady-state free-precession cine sequences in the axial, vertical, and horizontal long-axis, ventricular short-axis, and oblique planes parallel to the aortic arch; ECG-gated, breath-hold fast spin echo with double inversion recovery parallel to the aortic arch and the pulmonary arteries; velocity-encoded cine phase-contrast sequences for measurements of systemic and pulmonary blood flow and, when appropriate, in the branch pulmonary arteries and atrioventricular valve(s); and gadolinium-enhanced (0.2 mmol/kg) 3-dimensional magnetic resonance angiography.
Before anesthesia was discontinued, the image data were reviewed with the managing cardiologist, and if consensus was reached that catheterization was not indicated, the patient was extubated in the magnetic resonance imaging suite and transferred to the recovery room for monitoring. Patients could then be discharged to home at the discretion of the managing team.
The following outcomes were determined before the bidirectional Glenn operation: frequency of adverse events associated with the preoperative test as measured by a set of predefined minor and major criteria (Appendix), hospital length of stay, and total hospital charges associated with the preoperative evaluation. Outcomes during and after the bidirectional Glenn surgery included intraoperative and postoperative complications; a composite score of a successful bidirectional Glenn procedure comprising no unplanned postoperative interventions or procedures, oxygen saturation ≥75% at the time of hospital discharge, eligibility for discharge from the cardiac intensive care unit by postoperative day 4, and eligibility for hospital discharge by postoperative day 7; and clinical assessment at 3 months after surgery, including cutaneous oxygen saturation, clinical well-being, and frequency of new diagnoses or reintervention.
On the basis of data from our previous study,9 we calculated that a sample size of 41 patients in each arm would provide 80% power to detect a 19% difference in the incidence of adverse events associated with the preoperative evaluation using a 2-sample test of proportions conducted at the α=0.05 level of significance. All analyses were performed on an intention-to-treat basis. Patient characteristics and outcomes were compared for those randomized to CMR versus catheterization using the 2-sample t test or Wilcoxon rank-sum test for continuous variables and Fisher’s exact test for categorical variables. Linear and logistic regression analyses compared outcomes after adjustment for baseline patient characteristics not equally balanced between the 2 groups.
The authors had full access to and take full responsibility for the integrity of the data. All authors have read and agree to the manuscript as written.
Of the 92 patients eligible for inclusion, 10 had exclusion criteria by screening pre-enrollment echocardiogram (aortic coarctation in 7 patients and pulmonary vein stenosis, restrictive atrial septum, and suspected pulmonary hypertension in 1 patient each). All remaining 82 patients had informed consent, were enrolled in the study, and were randomized (the Figure). No patient withdrew from the study. Table 1 summarizes the baseline characteristics of the study patients. The groups were similar in age, gender, race, ventricular anatomy, prior operations, oxygen saturation, and frequency of treatment with an angiotensin-converting enzyme inhibitor.
The preoperative echocardiographic findings were similar in the 2 study groups. Most patients had normal or mildly depressed ventricular function and trivial or mild atrioventricular valvar regurgitation (Table 1). Mild distal aortic arch obstruction was common but occurred with similar frequency in the 2 groups. At least mild branch pulmonary artery stenosis was present in many patients in both groups. In 6 patients, either 1 or both branch pulmonary arteries could not be imaged adequately because of poor acoustic windows.
Of the 41 patients in the catheterization group, 18 (44%) received general anesthesia, and 23 (56%) received intravenous sedation. One patient crossed over from catheterization to CMR for further delineation of complex systemic venous anatomy. Median cardiac index was 3.3 L · min−1 · m−2 (range, 2.1 to 6.1 L · min−1 · m−2); median pulmonary artery pressure was 16 mm Hg (range, 8 to 40 mm Hg); and median pulmonary vascular resistance was 2.0 Woods units/m2 (range, 0.9 to 4.7 Woods units/m2). Catheter interventions, performed in 17 patients (41%), are listed in Table 2. Median fluoroscopy time was 58 minutes (range, 19 to 149 minutes).
Cardiac Magnetic Resonance
Of the 41 patients enrolled in the CMR arm, 3 crossed over to catheterization. In 1 patient, CMR detected a previously unsuspected virtual atresia of the left pulmonary artery, and catheterization was requested for possible balloon dilation. In a second patient, catheterization was requested for further evaluation of surgical options (Fontan completion as opposed to a bidirectional Glenn shunt). Catheterization was performed in a third patient to manage shunt thrombosis (see Adverse Events). Median single-ventricle ejection fraction was 61% (range, 35% to 77%), and at least moderate atrioventricular valvar regurgitation was found in 2 patients (5%). Median CMR scan time was 53 minutes (range, 45 to 70 minutes), and median length of poststudy observation was 4 hours (range, 3 to 90 hours).
Table 3 summarizes adverse events associated with preoperative evaluation. There were no major adverse events in the catheterization group, but minor events were frequent (34 patients, 83%). One major adverse event occurred in the CMR group. A patient with pulmonary atresia palliated by a Blalock-Taussig shunt developed shunt thrombosis during the course of the CMR scan. The patient was immediately resuscitated and moved to the catheterization laboratory, and the shunt was successfully stented. The patient was placed on extracorporeal membrane oxygenator support and 4 days later underwent an uneventful bidirectional Glenn operation. The patient was discharged from the hospital 31 days after surgery and was in good clinical status at the 3-month follow-up. There also were 2 minor adverse event in the CMR group: A patient with mild respiratory distress after extubation was observed overnight in the hospital (unplanned admission), and 1 patient who crossed over to cardiac catheterization had acidosis and unplanned intubation during the procedure.
Compared with the CMR group, the incidence of minor adverse events was significantly higher in the catheterization group (36 versus 3 patients; P<0.001). Only 1 major adverse event occurred in the trial, as described above (P=1.0).
Hospital length of stay for the preoperative evaluation was longer (2 versus 1 days; P<0.001) and hospital charges for the preoperative evaluation were higher ($34 477 versus $14 921; P<0.001; Table 2) in the catheterization group compared with the CMR group. The incidence of minor adverse events, hospital length of stay, and hospital charges remained higher in the catheterization group after adjustment for baseline patient characteristics.
Patients in the catheterization group who had catheter-based interventions compared with those who did not had similar rates of minor adverse events (94% versus 83%; P=0.38), similar use of general anesthesia (47% versus 42%; P=0.76), similar hospital length of stay (2 versus 2.5 days; P=0.49), and a trend toward higher hospital charges (median, $40 068 versus $33 143; P=0.06).
Patients in the CMR and catheterization arms did not differ with regard to age at bidirectional Glenn, time from preoperative test to surgery, or associated surgical procedures (Table 4). There was 1 intraoperative complication in each group: 1 patient with narrowing of the bidirectional Glenn anastomosis that required revision and another with cyanosis on weaning from bypass that required pulmonary arterioplasty. Total cardiopulmonary bypass times, postoperative superior vena cava pressure, and mean atrial pressure did not differ between groups.
There were no postoperative deaths. Major complications occurred in 4 patients: 3 in the catheterization group (mediastinitis in 2 and persistent low cardiac output in 1) and 1 in the CMR group (stroke) (P=0.62). Minor postoperative complications occurred with similar frequency in both groups (Table 4). Although the duration of thoracostomy tube drainage (median, 2 days; P=0.10) and oxygen saturation at hospital discharge (83% versus 84%; P=0.13) were similar, the duration of mechanical ventilation (17 versus 11 hours; P<0.001), length of stay in the intensive care unit (3 versus 2 days; P=0.003), and length of hospital stay (6 versus 5 days; P=0.03) all were slightly longer in the catheterization group.
Patients in both groups met the 4-element definition of a clinically successful bidirectional Glenn operation with similar frequencies (71% in the catheterization group versus 83% in the CMR group; P=0.30; Table 5). When analyzed by whether patients received catheter-based interventions, those who had a preoperative transcatheter intervention had the same rate of achieving the clinical definition compared with those who did not (71% versus 71%; P=1.0).
The diagnostic accuracy of the preoperative test was assessed on the basis of intraoperative observations and findings at postoperative diagnostic tests such as catheterization. Four patients in the catheterization group (10%) had findings that were not appreciated at the initial procedure, including a large systemic venovenous collateral vessel, a diffusely small left pulmonary artery requiring intraoperative arterioplasty, a right ventricular aneurysm at a shunt insertion site, and an abscess around a Blalock-Taussig shunt. There were no new intraoperative or postoperative findings in the CMR group.
Nearly all patients were clinically well at the 3-month follow-up evaluation (Table 5). Oxygen saturation was similar in both groups (84% versus 85%; P=0.84). Two patients in the catheterization arm had new diagnoses: 1 with recurrent coarctation and 1 with deteriorating ventricular function. No patients required additional interventions.
This randomized clinical trial showed that in selected patients with single-ventricle physiology considered for a bidirectional Glenn operation, routine cardiac catheterization is associated with higher rates of minor adverse events, longer hospital stay, and higher hospital charges than CMR. The study also found no detectable differences in immediate and short-term postoperative outcomes. These findings follow a trend over the years in which routine preoperative catheterization has been supplanted by noninvasive strategies for a variety of congenital heart lesions.17,19–24 Although echocardiography provides sufficient diagnostic imaging in many patients with single-ventricle physiology,9,10 the results of this trial illustrate the additional utility of CMR, especially in imaging extracardiac anatomy in patients with suboptimal acoustic windows or other technical limitations. Although not evaluated in this study, the clinical utility of detecting scar tissue by myocardial delayed enhancement deserves further investigation in these patients.25 Furthermore, this imaging strategy was applicable for the majority of patients presenting for Glenn operation; only 10% of eligible patients were excluded on the basis of screening echocardiogram.
One of the theoretical advantages of routine catheterization before bidirectional Glenn operation is that anomalies such as systemic-to-pulmonary or venous collateral vessels and aortic coarctation can be palliated using transcatheter techniques. However, the precise indications and clinical benefits of these procedures are controversial.11,12 In 10% of our patients randomized to catheterization, findings that might prompt such palliative interventions were found only intraoperatively. Still, although the entry criteria of this trial were designed to exclude patients who might benefit from catheter-based therapeutic procedures, intervention was performed in 41% of the patients randomized to the catheterization group. Most of these interventions were coil occlusion of aortopulmonary collateral vessels, a procedure that some centers do not perform routinely.11,12 Furthermore, the long-term effect of systemic-to-pulmonary flow through these vessels and its effect on the single-ventricle circulation are largely unknown. Retrospective studies have found no clinical impact of aortopulmonary collaterals on resource use or immediate postoperative outcomes in similar patients undergoing Fontan operation.11,12 The results of this study showed no evidence of demonstrable clinical benefit from these interventions in the perioperative and early postoperative periods. Recognizing that the potential benefits of these interventions can manifest months or years later, it is clear that this study cohort requires long-term follow-up to assess whether any beneficial effect exists either at Fontan operation or over the longer term. Moreover, the question of whether occlusion of aortopulmonary collateral vessels before bidirectional Glenn surgery is clinically beneficial would ideally be addressed in a prospective randomized trial designed specifically to answer this question.
Another implication of our study is that in the absence of evidence of pulmonary hypertension, routine measurement of pulmonary vascular resistance is not necessary before bidirectional Glenn operation. Indeed, the range of resistance values found in the catheterization group was within acceptable limits, and patients in the CMR group did not fare differently after bidirectional Glenn operation as a result of the absence of these data. This is consistent with previous studies in which few patients were excluded from bidirectional Glenn on the basis of findings at cardiac catheterization.9,10
It is also worth noting that the incidence of minor adverse events recorded in the catheterization group is higher in this study than in our previous retrospective study of a similar cohort.9 This may be due to the meticulous prospective recording of all adverse events using predefined criteria in this trial versus retrospective data collection.
The only major adverse event in the study was that a patient in the CMR group developed modified Blalock-Taussig shunt thrombosis during the CMR scan. This is a well-known and life-threatening complication of such shunts,26,27 but an occurrence during CMR has not been previously reported. A thorough review of all similar single-ventricle patients (n=99) who have undergone CMR at our institution failed to show any similar events. Although the exact mechanism that led to shunt thrombosis remains unclear, we have considered modifying our CMR protocol in such patients to include preadmission for intravenous hydration and heparin infusion.
Finally, given that this was a single-center trial, confirmation of the results from a multicenter study is desirable. In addition, the study was powered to measure a modest difference in the incidence of adverse events but not to demonstrate the equivalency of postoperative outcome. It is conceivable that with a larger sample size some differences would be more apparent.
The results of this trial indicated that CMR is a safe and cost-effective alternative to catheterization in properly selected patients before bidirectional Glenn. Further studies are necessary to determine whether there are long-term benefits from certain transcatheter interventions in these patients and the generalizability of our results to other centers.
We thank the members of the Data and Safety Monitoring Board for this study: Kathy J. Jenkins, MD, MPH, Peter Laussen, MBBS, Gus J. Vlahakes, MD, and David Wypij, PhD. This study was supported by the Higgins Family Noninvasive Cardiac Imaging Research Fund. We are indebted to the parents and infants participating in this trial.
Keane J, Fyler D. Single ventricle. In: Keane J, Lock J, Fyler D, eds. Nadas’ Pediatric Cardiology. 2nd ed. Philadelphia, Pa: Saunders Elsevier; 2006: 743–751.
Booth P, Redington AN, Shinebourne EA, Rigby ML. Early complications of interventional balloon catheterisation in infants and children. Br Heart J. 1991; 65: 109–112.
Geva T, Greil GF, Marshall AC, Landzberg M, Powell AJ. Gadolinium-enhanced 3-dimensional magnetic resonance angiography of pulmonary blood supply in patients with complex pulmonary stenosis or atresia: comparison with x-ray angiography. Circulation. 2002; 106: 473–478.
Muthurangu V, Taylor AM, Hegde SR, Johnson R, Tulloh R, Simpson JM, Qureshi S, Rosenthal E, Baker E, Anderson D, Razavi R. Cardiac magnetic resonance imaging after stage I Norwood operation for hypoplastic left heart syndrome. Circulation. 2005; 112: 3256–3263.
Lock JE, Keane JF, Perry SB. Diagnostic and Interventional Catheterization in Congenital Heart Disease. 2nd ed. Norwell, Mass: Kluwer Academic Publishers; 2000.
This prospective, randomized clinical trial compared noninvasive (cardiac magnetic resonance imagine) and invasive (cardiac catheterization) preoperative diagnostic strategies in infants with severe congenital heart disease (functional single ventricle) before bidirectional Glenn operation. Forty-one infants were randomized to each imaging strategy and were followed up prospectively from the preoperative evaluation to clinical assessment 3 months after bidirectional Glenn operation. Minor adverse events were more frequent in the catheterization group (78% versus 5%), with longer preoperative hospital stays and higher hospital charges. Both groups did well with surgery, with similar rates of postoperative complications and length of stay and no clinical differences at 3-month follow-up. We found that cardiac magnetic resonance imaging was a safe, effective, and less costly alternative to cardiac catheterization in the routine evaluation of such patients. These results support the use of a noninvasive strategy in carefully selected patients before bidirectional Glenn surgery.
Clinical trial registration information—URL: http://www.clinicaltrials.gov. Unique identifier: NCT00112424.
The online-only Data Supplement, which contains an Appendix, is available at http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.107.723213/DC1.
Guest Editor for this article was Robyn J. Barst.