Author + information
- Received August 28, 2002
- Revision received December 16, 2002
- Accepted January 16, 2003
- Published online June 18, 2003.
- Chetan Varma, MBBS, MDa,
- Matthew R Warr, BSca,
- Aaron L Hendler, MDa,
- Narinder S Paul, MDa,
- Gary D Webb, MD, FACCa and
- Judith Therrien, MDa,* ()
- ↵*Reprint request and correspondence:
Dr. Judith Therrien, Toronto Congenital Cardiac Center for Adults, Toronto General Hospital, University of Toronto, Eaton Building, 200 Elizabeth St., Toronto, 12-215EN, ON M5G 2C4, Canada.
Objectives The study was done to determine the prevalence of pulmonary emboli (PE) in asymptomatic adult Fontan patients and to identify the risk factors associated with PE.
Background Right atrial thrombi and systemic thromboembolic complications have been reported after the Fontan procedure. However, the frequency of silent PE in this patient population is not known.
Methods All consecutive adult Fontan patients attending the adult congenital clinic over a six-month period underwent ventilation-perfusion (VQ) scanning and blood testing for thrombophilia tendency. If the VQ scan showed an intermediate or high probability for PE, a computerized tomography (CT) pulmonary angiogram was performed to confirm the presence of PE.
Results Thirty patients (mean age 26 ± 7 years, 57% men) were included in this study. Five (17%) adult Fontan patients had an intermediate or high probability for PE on VQ scan, all of which were confirmed on CT pulmonary angiography. No patient had a thrombophilia tendency. Pulmonary emboli were not present in any patients (30%) taking warfarin. Late age at time of Fontan operation (19 ± 6 years vs. 11 ± 6 years, p = 0.012) and type of Fontan anatomy (p = 0.001) were associated with increased risk of silent PE.
Conclusions Seventeen percent of adult patients with Fontan procedure have clinically silent PE. The long-term hemodynamic implications of this with respect to Fontan attrition over time are unknown. Large randomized prospective studies looking at anticoagulation therapy in all Fontan patients are urgently needed.
The Fontan procedure was initially described by Fontan and Baudet in 1971 for the surgical management of tricuspid atresia (1). The principle of the Fontan operation is the diversion of systemic venous return directly to the pulmonary arteries (PA), bypassing the pulmonic ventricle when the latter is nonexistent, too small, or dysfunctional (2). Numerous variations exist in the surgical approach. The most likely types of Fontan procedure to be encountered today are: 1) the direct right atrium (RA)–PA connection; 2) the total cavopulmonary connection (superior vena cava to PA and inferior vena cava to PA through an intra-atrial tunnel); and 3) the extracardiac conduit (superior vena cava to PA and inferior vena cava to PA through an external conduit). Thrombi and thromboembolic complications have been reported after the Fontan operation, and modifications of the original atriopulmonary procedure have been developed to prevent this and other longer-term complications (3–5).
Studies have determined the point prevalence of atrial thrombus using echocardiography (6–10). However, to date the frequency of silent pulmonary emboli (PE) has not been investigated. Over time some recipients of the Fontan operation will have failure of the previously functional Fontan circuit. The underlying causes have been attributed to disturbed flow dynamics or ventricular dysfunction. Occult PE could cause increasing pulmonary vascular resistance and inefficiency of the Fontan circulation, but the prevalence of silent PE in the Fontan population is unknown. This study aimed to determine the prevalence of silent PE in the adult Fontan population and to identify the predisposing risk factors.
The study included all consecutive adult patients with a modified Fontan procedure who were seen for routine follow-up between November 2001 and April 2002 at the University of Toronto Congenital Cardiac Center for Adults. Informed consent was obtained from all participants. Patients with acute symptomatic PE were excluded.
Patient age, gender, operative details of the Fontan procedure, New York Heart Association functional class, current prescription medications, oxygen saturation monitoring in room air (after 5 min of rest), and electrocardiography were recorded. Complete two-dimensional transthoracic echocardiography (TTE) and Doppler study were performed in all patients. The TTE was performed using a Hewlett-Packard model 5500 with standard views. The presence of right atrial clot was recorded. Ventricular function was assessed qualitatively and graded 1 (normal), 2 (mild dysfunction), 3 (moderate impairment), and 4 (severe impairment).
Patients were screened for hypercoagulable abnormalities including deficiencies of protein C, protein S, antithrombin 3, and for antiphospholipid syndrome. Protein C and protein S levels were not included in the analysis if the patient was taking warfarin. Packed cell volume and albumin as a percentage of total protein were measured.
Ventilation-perfusion (VQ) scan
The VQ scan protocol was based upon the published reports for Fontan patients (11–15). Planar lung imaging in six projections was performed for both ventilation and perfusion scans. Images were acquired using an ADAC Forte large field-of-view gamma camera interfaced to a PEGASYS processing station (Phillips Medical, Malpedia, California). For the ventilation images, the patient breathed through a CADEMA aerosol unit (Tri-Bor Medical, New York, New York) containing Tc-99m DTPA, for 3 min. Following the ventilation images, the patient was injected with Tc-99m macroaggregates into the right arm and any foot vein so as to prevent false positive diagnoses due to preferential blood flow from the superior vena cava to the right PA and inferior vena cava to the left PA. Perfusion images were then performed in the same projections as the ventilation scan. The VQ scan was interpreted by an experienced nuclear medicine physician (A.L.H.) and classified as normal, very low probability, low probability, intermediate probability, and high probability for PE using modified Prospective Investigation of Pulmonary Embolism Diagnosis criteria (16). If the VQ scan was classified as either intermediate or high probability, the patient was requested to undergo a computerized tomography (CT) pulmonary angiogram to confirm the presence of PE using a PE protocol.
A CT pulmonary angiogram was performed on a multidetector spiral scanner (Lightspeed, General Electric, Milwaukee, Wisconsin). The patient was placed supine with the arms extended behind the head. Bilateral 18-g cannulas were inserted into antecubital veins and connected to a power injector (Medrad, Pittsburgh, Pennsylvania). Frontal and lateral scout views (120 kV, 10 mA) were used to plan the limits of the CT pulmonary angiogram from the dome of the diaphragm to the sternoclavicular joints.
A timing bolus was performed during free breathing using 15 cc of iohexol 65% (Omnipaque 300 mg/ml, Nycomed, Oslo, Norway) injected at 5 cc/l in order to determine the contrast peak in the area of interest. In a patient with normal cardiac anatomy, peak contrast would be at the level of the PA bifurcation. Because of the altered cardiac anatomy in the Fontan patients, this timing bolus was performed in the left atrium at the insertion of the inferior pulmonary veins. The time delay between the injection of contrast agent and image acquisition was then determined and the patient given instructions to perform a breath hold in inspiration during image acquisition. The patient was then injected with 60 to 80 cc of iohexol 65% at 5 cc/s, and image acquisition was commenced in a caudal-cranial direction.
The acquisition parameters were a 512 × 512 matrix, 2.5 × 1.5 mm slice width, a table speed of 7.5 to 15 mm/s in HS mode, 120 kV and 300 to 420 mA. The volume of injected contrast agent, table speed, and exposure factors varied with the patient body habitus. Soft-copy images were interrogated using scroll mode on a stand-alone PACS workstation by an experienced chest radiologist (N.S.P.). The diagnosis of acute or chronic thromboembolic disease was made in accordance with published criteria (17–19).
The primary end point was presence or absence of PE. Comparison between those with PE and without PE was made with independent ttest (equal variances not assumed) for continuous variables and the Fisher exact test for categorical variables. Comparison between multiple groups was performed with a Kruskal-Wallis test. Values are expressed as mean ± 1 SD. Logistic regression was performed to test whether variables were associated with the presence of a PE. A p value < 0.05 was considered statistically significant (two-sided).
A total of 44 adult patients with modified Fontan procedure were seen in our adult congenital clinic during this period. Three patients had a diagnosis of symptomatic PE during this study period, and they were excluded from the study (two patients were known for atrial flutter on Coumadin and presented with acute shorteness of breath, and one patient not anticoagulated presented with hemoptysis and desaturation). Of the remaining 41 patients requested to participate in this study, 31 consented and 10 declined. One of the VQ studies for a patient was incomplete and was excluded from the analysis. Therefore, 30 patients in total were studied.
The mean baseline characteristics are described in Table 1. The mean age of the study group was 26 ± 7 years (range 18 to 45 years). The mean age at the time of the last Fontan operation was 12 ± 7 years. All patients were in either New York Heart Association functional class 1 (23 patients) or class 2 (7 patients). Ventricular function was graded as 1 or 2 in 24 (80%) patients on TTE.
Individual anatomy, surgery, and anticoagulation indications are shown in Table 2. Atrial flutter or fibrillation, either chronic or paroxsymal, was present in nine patients (30%). Three patients had a past history of right atrial clot, and one other patient had residual thrombus detectable in the right atrium on TTE. In total, nine subjects were taking warfarin and two were on aspirin.
Mean packed cell volume was 44 ± 5% (normal range 40% to 54%). The mean albumin as a percentage of total protein was 59 ± 4% (normal 50% to 70%). In those patients not on warfarin (n = 21), no patient was found to have deficiency of protein C or protein S. No patient had abnormal antithrombin levels. No patient was positive for antiphospholipid antibodies or anticardiolipin antibodies.
The VQ scan
The VQ scan was normal in 18 patients, low probability in 7 patients, intermediate probability in 3 patients, and high probability in 2 patients. In all five patients (17%) with an intermediate or high probability study, CT pulmonary angiogram was undertaken, and PE was confirmed in all these patients. Distribution of the PE and the patient characteristics for those with PE are shown in Table 3. Figure 1depicts an example of a VQ scan and CT pulmonary angiogram from one of the patients.
Older age at the time of the Fontan operation was associated with the presence of PE (age 19 ± 6 vs. 11 ± 6 years, p = 0.012) as well as the type of Fontan correction (atriopulmonary 0/18, RA–RV 2/6, and lateral tunnel 3/4, p = 0.001). No patient on warfarin had a PE. Not associated with PE were atrial arrhythmias, right atrial thrombus, or previous systemic thrombotic events.
Prevalence of PE
The prevalence of silent PE in adult patients with the Fontan operation was unknown until now. The findings of this study reveal a one in six chance of a silent PE being present in asymptomatic adult Fontan patients. It is of interest to note that no patients on coumadin (n = 9) were diagnosed with an asymptomatic PE, raising the prevalence of silent PE in nonanticoagulated Fontan patients to 24% (5/21). The lifetime prevalence of asymptomatic PE in these patients may well be even greater as this study only looked at the prevalence at one point in time (20). Furthermore, each CT pulmonary angiogram performed confirmed the VQ scan result, implying the study threshold for CT scanning may have been too high, and changing this to include those with a lower probability for PE on VQ scan, which might have yielded more patients with asymptomatic PE. Finally, the total occurrence of PE—both symptomatic and asymptomatic—is actually still higher than one in six in the Fontan population, as this study excluded those with symptomatic PE.
The technique to determine the presence of PE is of great consequence. Ventilation perfusion scanning was chosen as the screening test in our study because of its availability, simplicity, safety, and comfort for the patient, and for its relatively good sensitivity and specificity as a screening test in non-Fontan patients. The imaging methodology was designed to prevent false positive diagnoses. Misinformation regarding pulmonary flow patterns can occur because of preferential blood flow from the superior vena cava to the right PA and from the inferior vena cava to the left PA. In our center, Fontan patients have been referred with a diagnosis of acute PE on imaging, which has subsequently been proven to be an artefact due to isotope injection from the arm only. This problem may be resolved by injection of macroaggregates for perfusion scanning into the arm and any foot vein. The timing of image acquisition with contrast during CT pulmonary angiography is also critical to prevent false positive diagnoses.
Previous studies of thrombus in the Fontan population
Thrombosis is a major cause of morbidity and mortality after Fontan procedures. Prior cross-sectional studies looking at the prevalence of thrombus within the right atrium after Fontan procedures have reported a prevalence of thrombi varying anywhere between 9% and 33% (7–10). The highest prevalence of thrombi in the right atrium was observed when transesophageal echocardiography was used to detect thrombus (9), and the lower prevalence occurred using TTE (10).
Studies of venous thromboembolic events as a primary outcome in Fontan patients have quoted incidences ranging from 5% to 16% (21–25). The higher incidences are in more recent studies, reflecting a longer duration of follow-up, and/or increased awareness of the potential for thrombosis with improved diagnostic testing. Presentation may be embolus to the lung, obstruction to flow in the right atrium, or peripheral venous thrombosis. The finding from this study of a 17% prevalence of silent PE would be consistent with the previously published data.
Risk factors for PE occurrence
Following the Fontan operation there may be a low flow state, stasis in venous pathways, right to left shunts, blind cul de sacs, prosthetic materials, and/or atrial arrhythmia as risk factors for thrombophilia (26). In addition, some studies have demonstrated a prothrombotic tendency in Fontan patients (27–30). We were surprised to find that a prothrombotic tendency was not observed in our adult patient population, either in those with or without PE, although those on warfarin did not have a complete screen. In some of the studies showing a prothrombotic tendency in a pediatric Fontan population, a normal adult range was used as a reference instead of normal pediatric values, and it has been stated that no accurate data implicate abnormalities of the coagulation system in thromboembolism after Fontan procedures (6). In one study (29), the extent of prothrombotic tendency was observed to be less in patients with longer follow-up, which may support our thrombophilia data. The lack of a prothrombotic tendency within our population implies that chronic PE formation may be a flow-related phenomena as opposed to an abnormal hematological tendency. Further in vitro and in vivo studies of flow hemodynamics in Fontan models and patients are needed to clarify the contribution of the so-called low flow state physiology to the prothrombotic tendency seen in these patients.
Our findings of the association of PE with an older age at the time of Fontan operation and lateral tunnel surgery need further confirmatory studies. Other studies with a high incidence of atrial thrombus did not identify predictors of thrombus formation (8–10). Higher left ventricular, end-diastolic pressure from prolonged volume overload condition or chronic cyanosis in patients operated on at a later age might predispose these patients to higher pulmonary vascular resistance and greater right atrial stasis. Likewise, the use of Dacron material in the lateral tunnel may act as a nidus for clot formation. Modifications to the original atriopulmonary Fontan procedure (lateral tunnel and external conduit) have been developed to prevent in part the prothrombotic complications of the RA to PA Fontan connection. It is of interest to note that none of our patients with an RA to PA connection were found to have an asymptomatic PE.
Anticoagulation and the Fontan patient
Despite our results and previous reports on the incidence of thrombi in this patient population, the role of long-term prophylactic anticoagulation in the Fontan patients remains controversial. The evidence to prescribe warfarin to all patients with the Fontan operation will continue to be poor without proper randomized controlled studies (6). Results of cohort studies on the effect of aspirin and/or warfarin treatment are contradictory. In a study of 72 postoperative Fontan patients, with a median age of 25 months and treated with aspirin only, no thrombi were detected with TTE at a mean follow-up of 40 months (31). Conversely, the follow-up study from Coon et al. (10)in 552 patients using TTE showed that almost 9% of patients could develop atrial thrombus while on aspirin or warfarin at a median follow-up of 22 months. In our study, silent PE were diagnosed in one patient while taking aspirin but none while taking Coumadin.
The morbidity related to the risk of bleeding with warfarin in this young adult population is not inconsequential, with serious bleeding in at least 0.5% per patient year (32). The optimal international normalized ratio one should aim for in the Fontan population is unknown. Fontan patients required 25% lower dosage as compared with other congenital heart disease patients in an analysis of warfarin therapy in pediatric patients (32). Compliance with regular medication in a young population may be inconsistent (5), and this was reflected in the baseline data presented, in which some patients with previous thrombi detected still declined to take warfarin. In our study over 50% of the cohort had at least one conventional indication (atrial arrhythmia, right atrial clot or PE) for warfarin. Large prospective studies comparing alternative strategies of warfarin, aspirin, and no therapy to prevent thrombi and PE in this patient population are urgently needed to determine whether all Fontan patients should be anticoagulated or whether a strategy of targeting selected patients should be followed.
Potential implication of silent PE
The asymptomatic PE may carry significant clinical consequences. The hemodynamic consequence of silent emboli to the lungs is presently unknown in this patient group. It was recently demonstrated that pulmonary resistance rather than geometry decides flow distribution in the modified Fontan circulation (33). Computer simulations of the circulation in patients with cavopulmonary connection show a marked interrelation of cardiac and vascular pressure, flow, resistance, and capacitance (34). It is suggested that higher shear stress (from PE, for example) on the wall of the pulmonary vasculature may alter endothelial function (that is, increase vasoconstriction) and affect the longevity of the functional Fontan (35). If silent PE indeed increases pulmonary vascular resistance in these patients, any strategy to prevent PE formation could potentially improve long-term function of the Fontan circuit and improve survival of these patients. Further studies into the hemodynamic consequences of PE on the Fontan attrition rate are needed.
In quantifying a prevalence, bias can occur because of selection into or out of this population. Numbers in this study are small and conclusions regarding predisposing factors to PE formation should be viewed with caution because of a potential type 2 statistical error. Additionally, we did not perform a transesophageal echocardiogram concurrently to determine thrombus within the Fontan circuit at the time of VQ scan or CT in our patients. Although more information regarding concomitant right atrial thrombus would have been interesting, the discomfort engendered by transesophageal echocardiography would have been substantial without changing the result of our primary end point. Lastly, the relatively high incidence of pulmonary segmental flow disturbances in these patients from intrinsic PA stenosis or surgically created PA distortion ought to make us cautious in interpreting the results of the VQ scan, although a congenitally underdeveloped PA or branch would most likely be associated with a proportional decrease in ventilation and not give rise to a mismatch defect.
Seventeen percent of adult patients with the Fontan procedure have clinically silent PE. The hemodynamic long-term implications of this with respect to Fontan attrition over time are unknown. Large randomized prospective studies looking at anticoagulation therapy in all Fontan patients are urgently needed.
☆ Dr. Therrien is supported by the Heart and Stroke Foundation of Canada.
- computerized tomography
- pulmonary artery
- pulmonary embolus/emboli
- right atrium
- right ventricle
- transthoracic echocardiography
- ventilation perfusion
- Received August 28, 2002.
- Revision received December 16, 2002.
- Accepted January 16, 2003.
- American College of Cardiology Foundation
- Fontan F.,
- Baudet E.
- Bull K.
- Stumper O.,
- Sutherland G.,
- Geukens R.,
- Roelandt J.,
- Bos E.,
- Hess J.
- Fyfe D.A.,
- Kline C.H.,
- Sade R.M.,
- Gillette P.C.
- Pruckmayer M.,
- Zacherl S.,
- Salzer-Mhar U.,
- Schlemmer M.,
- Leitha T.
- Matsushita T.,
- Matsuda H.,
- Ogawa M.,
- et al.
- Gottschalk A.,
- Sostman H.D.,
- Juni J.E.,
- et al.
- Greaves S.M.,
- Hart E.M.,
- Brown K.,
- et al.
- ↵Coggon D, Rose G, Barker D. Epidemiology for the Uninitiated. 4th edition. BMJ Publishing Group, 1997; Ch. 2
- Rosenthal DN, Friedman MD, Kleinman CS, Kopf GS, Rosenfeld LE, Hellenbrand WE. Thromboembolic complications after Fontan operations. Circulation 195;92:287–93.
- Therrien J.,
- Webb G.
- Van Nieuwenhuizen R.C.,
- Peters M.,
- Lubbers L.J.,
- et al.
- Streif W.,
- Andrew M.,
- Marzinotto V.,
- et al.
- Pedersen E.M.,
- Stenbog E.V.,
- Frund T.,
- et al.