Is Cardiac Catheterization a Prerequisite in AllPatients Undergoing Bidirectional

Cavopulmonary Anastomosis?Colin J. McMahon, MB BCh, MRCP, Benjamin W. Eidem, MD, Louis I. Bezold, MD,

Thomas Vargo, MD, Steven R. Neish, MD, J. Timothy Bricker, MD, John Kovalchin, MD,and Howaida El-Said, MD, Houston, Texas, and Cairo, Egypt

Traditionally, all patients undergo cardiac catheter-ization before bidirectional cavopulmonary anasto-mosis (BCPA). The purpose of this study was todetermine if preoperative catheterization is neces-sary when echocardiographic parameters appearfavorable. A retrospective review was performed ofall patients who underwent BCPA (n � 142) betweenFebruary 1996 and May 2001. Echocardiographiccriteria defining a favorable BCPA candidate in-cluded good ventricular function, moderate or lessatrioventricular and semilunar valve regurgitation,absence of ventricular outflow tract obstruction,normal proximal branch pulmonary artery (PA)size, and low PA pressures estimated by PA bandgradient or systemic PA shunt velocity. The medianage at operation was 7 months (range: 2-11) andweight was 6.2 kg (range: 2.7-7.1). There were 73unfavorable candidates. Patients with hypoplastic

left heart syndrome (n � 23) and pulmonary atresiawith intact septum (n � 15) predominated amongthe unfavorable group. All patients were catheter-ized. This provided additional information on PApressures in 3 patients in the favorable group butdid not defer operation or influence outcome (nomortality, prolonged pleural drainage, or longerintensive care department stay). All 3 patients arealive at a mean follow-up of 51 months. The 30-daymortality was 2% (4 of 151 patients), all in theunfavorable group. Overall, 20 patients (13%) re-quired arterioplasty of PAs at the time of BCPA.BCPA can be performed with a low risk of morbidityand mortality in a wide range of patients. By usingcommonly acquired echocardiographic parameters,a low-risk subgroup of patients can be identifiedwho can safely avoid preoperative cardiac catheter-ization. (J Am Soc Echocardiogr 2003;16:1068-72.)

In recent years there has been a significant reduc-tion in mortality associated with creation of cavopul-monary shunts in children.1 However, there aresignificant complications associated with palliateduniventricular hearts, irrespective of the ingeniousmethods used to augment efficiency and minimizeenergy expenditure in cavopulmonary circuits.2-5

The potential need for later cardiac transplantationsecondary to a failing single ventricle highlights theimportance of preserving venous access as patientsoften require multiple biopsies in the period aftertransplantation.6 Increasingly, cardiac catheteriza-tion is being avoided in the preoperative assessmentof patients with tetralogy of Fallot and transpositionof the great arteries. Hence, the question arises

whether hemodynamic cardiac catheterizations, tra-ditionally performed in all patients before creationof a bidirectional cavopulmonary anastomosis(BCPA), can be avoided if sufficient data are avail-able from transthoracic echocardiography.

METHODS

We retrospectively reviewed the medical records, echo-cardiograms, and cardiac catheterization data of all pa-tients who underwent creation of unilateral or bilateralBCPA at Texas Children’s Hospital, Houston, Texas, fromFebruary 1996 to May 2001. Inclusion criteria included allchildren with univentricular physiology who underwentBCPA, had cardiac catheterization before operation, andhad a complete echocardiogram performed within 2weeks of cardiac catheterization that was of sufficientquality to determine specific study variables. Exclusioncriteria included lack of clinical follow-up, adults whounderwent BCPA, and patients with 2 adequate ventricleswho underwent BCPA to volume unload the right ventri-cle such as Ebstein’s anomaly with severe tricuspid regur-gitation or pulmonary atresia with intact ventricular sep-tum. Echocardiograms were reviewed blindly by 1experienced echocardiographer to determine 6 specific

From the Lillie Frank Abercrombie Section of Pediatric Cardiol-ogy, Texas Children’s Hospital and Baylor College of Medicine,Houston, Texas; and Department of Pediatric Cardiology, Univer-sity of Cairo (H.E-S.).Reprint requests: Colin J. McMahon, MB BCh, MRCP, LillieFrank Abercrombie Section of Pediatric Cardiology, Texas Chil-dren’s Hospital, 6621 Fannin, Houston, TX 77030 (E-mail:cmcmahon@bcm.tmc.edu).Copyright 2003 by the American Society of Echocardiography.0894-7317/2003/$30.00 � 0doi:10.1016/S0894-7317(03)00592-3

1068

variables. These included qualitative assessment of ven-tricular function, degree of atrioventricular (AV) valveregurgitation (AVVR) and semilunar valve regurgitation(SVR), the presence of adequate size proximal pulmonaryartery (PA), estimation of PA pressure (PAP)–systolic(EPAP) on the basis of either Doppler-derived PA bandgradient or systemic to PA shunt velocity, and presenceand degree of right or left ventricular outflow tract (LVOT)obstruction.

Patients were assigned a cumulative numeric score onthe basis of these preoperative echocardiographic data(Table 1). Normal ventricular function was qualitativelyscored as 0, mildly depressed as 1, and moderately toseverely depressed as 2. The absence or presence of mildAVVR/SVR was scored as 0 each, moderate AVVR/SVR as1, and severe AVVR/SVR as 2. Normal PA size was definedas a mean z score of left and right PAs between �0 to �1and scored as 0, between �1 to �2 as 1, and ��2 as 2.Estimated normal EPAP � 30 mm Hg was scored as 0,mildly elevated at 30 to 50 mm Hg was scored as 1, and�50 mm Hg was scored as 2. Absence of LVOT obstruc-tion, defined as peak velocity �1.5 m/s, was scored as 0,1.5 to 2.5 m/s was scored as 1, and �2.5 m/s was scoredas 2.

Patients were classified into 2 groups on the basis of theechocardiographic score. Group A was comprised ofpatients who had a total score � 2 with no score in anycategory �1. These patients were deemed favorable forBCPA without preceding cardiac catheterization. Group B,deemed unfavorable for BCPA without a preoperativecardiac catheterization, included those patients with atotal score � 2 or who had a score of 2 in any individualcategory, ie, estimated PAP � 50 mm Hg, PA size ��2 zscores, severe AVVR, moderate-severe ventricular dysfunc-tion, or significant LVOT. Patients were also included inthis group if there was echocardiographic evidence ofmajor aortopulmonary collaterals, a restrictive atrial sep-tum, or the presence of pulmonary venous anomalies.

Catheterization data were reviewed by 1 reviewer whowas blinded to the patients’ echocardiographic data.Specific attention was drawn to the PAP, pulmonaryvascular resistance, and PA anatomy, for which catheter-ization and angiography are held as the gold standard.Comparisons were made between noninvasive and inva-sive data to assess correlation between both modalitiesand specifically to determine if patients assigned to thefavorable echocardiographic group in fact had unfavor-able cardiac hemodynamics or anatomy by cardiac cathe-terization. Clinical end points, mortality and morbidity,postoperative duration of stay in the intensive care depart-ment, duration of chest tube drainage, duration of inotro-pic support, and length of mechanical ventilation werecompared between patients in group A and B.

Statistics

Comparison of clinical end points between groups A andB were made using Student t test. A P value � .05 wasregarded as statistically significant. Correlation was drawnbetween echocardiographic and catheterization assess-ment of specific study variables.

RESULTS

During the study period 157 patients underwentBCPA. Data were incomplete in 15 patients leaving142 patients available for analysis. Group A com-prised 69 patients whereas group B had 73 patients.Demographic data are presented in Table 2. Diag-noses for each echocardiographic group are pre-sented in Figure 1.

Table 1 Echocardiographic scoring system for patients ineach category

Variable Score � 0 Score � 1 Score � 2

Ventricularfunction

Normal Mildly depressed �Moderatelydepressed

AV valveregurgtn

None-mild Moderate Severe

Semilunarvalveregurgtn

None-mild Moderate Severe

HypoplasticPAs

�0 to �1 Zs �1 to �2 Zs ��2 Zs

Outflowtractobstruction

0-1.5 m/s 1.5-2.5 m/s �2.5 m/s

EPAP �30 mm Hg 30-50 mm Hg �50 mm Hg

AV, Atrioventricular; EPAP, estimated pulmonary arterial pressure; PAs,pulmonary arterieis; regurgtn, regurgitation; Zs, Z-score.

Table 2 Demographics and clinical outcome of patientsundergoing bidirectional cavopulmonary anastomosis

No. patients 142

Female 67

Age at BCPA (mo) 7 (2-11)

Group Group A Group B

P

value

Duration ICU stay(d)

3.5 (1-6) 4.2 (1-14) NS

Duration hospitalstay (d)

7 (5-21) 10 (8-32) NS

Durationmechanicalventilation

1.1 h (0.5-4 d) 1.4 h (0.7-6 d) NS

Duration inotropicsupport

2.5 h (0.5-5 h) 3.1 h (2-6 h) NS

Duration chest tubedrainage (d)

2.1 (2-4.4) 3.8 (2.5-11) �.05

BCPA, Bidirectional cavopulmonary anastomosis; d, days; ICU, intensivecare unit; NS, not significant.Data expressed as median and range.

Journal of the American Society of EchocardiographyVolume 16 Number 10 McMahon et al 1069

Echocardiographic Findings

Qualitative ventricular function was normal in 59and mildly depressed in 10 patients in group Acompared with mildly depressed function in 18 andmoderately to severely depressed function in 6patients in group B. AVVR was quantified as mild for9 patients in group A compared with mild in 37patients and moderate-severe in 14 patients in groupB. In 21 patients in group B there was mild or lessAVVR. There was mild SVR in 7 patients in group Acompared with 9 patients in group B having mild-moderate SVR and 11 patients having severe SVR.The PA size was mildly hypoplastic in 21 patientsand moderate to severely hypoplastic in an addi-tional 8 patients in group B whereas all patients ingroup A had normal PA size (Figure 2). The esti-mated PAP was � 30 mm Hg in 45 patients and 30to 40 mm Hg in 24 patients in group A comparedwith 15 patients in group B, whereas 54 patients ingroup B had EPAPs � 40 mm Hg (Figure 2).

Catheterization Data

The PAP was obtained by directly entering the PA in108 of 142 patients. The remaining 34 patients hadan estimation of PAP from the pulmonary veinwedge pressure. We avoided, whenever possible,crossing the Blalock-Taussig shunt to obtain a PAPparticularly in children with hypoplastic left heartsyndrome, and relied on pulmonary venous wedgepressure if �20 mm Hg.

Correlation of Echocardiographic andCatheterization Data

There was 92% correlation between echocardio-graphically and angiographically determined qualita-tive ventricular function. Correlation for PA size was91% for both groups whereas the correlation forAVVR was 93% between both modalities. There wasno significant LVOT obstruction determined eitherby echocardiography or cardiac catheterization inany study patient. Three patients from group Ademonstrated higher PAP at cardiac catheterizationthan predicted by echocardiography. Additional ar-eas of PA branch stenosis and hypoplasia wereunexpectedly demonstrated in 8 patients by cardiaccatheterization. These 8 patients each had their PAsinspected and augmented at the time of BCPA.Predicted EPAP by echocardiography was minimallybut consistently higher than pressure measurementsobtained at catheterization: mean EPAP 35 mm Hgwith range of 29 to 85 mm Hg by catheterizationcompared with 41 mm Hg with range of 34 to 130mm Hg by echocardiography. This may have beeninfluenced by sedation used at catheterization ordifferences in volume status as a result of osmoticcontrast agents.

Impact of Discrepant Data on PostoperativeOutcome

Despite any discrepant findings, all patients under-went successful BCPA creation. There were 14patients, 8 of whom were from group A, with PAstenoses who concomitantly underwent pulmonaryarterioplasty. Clinical outcomes were not signifi-cantly different in patients with unexpected PAanomalies or elevated PAP. The duration of intensivecare stay, need for and duration of inotropic sup-port, and length of mechanical ventilation werecomparable between both groups and also in those

Figure 1 Cardiac diagnoses (x axis) in favorable (group A,gray bars) and unfavorable (group B, black bars) echocar-diographic patient groups (y axis) undergoing bidirectionalcavopulmonary anastomosis. DILV, Double inlet left ven-tricle; DORV, double outlet right ventricle; HLHS, hypo-plastic left heart syndrome; HTX, heterotaxy; IVS, intactventricular septum; PAT, pulmonary atresia; TAT, tricus-pid atresia.

Figure 2 Comparison of specific echocardiographic param-eters of patients in group B (unfavorable anatomy) (x axis)and number of patients (y axis). EP, Estimated peak systolicpulmonary artery (PA) pressure; Hypo, hypoplastic; MAP-CAs, major aortopulmonary collaterals; Score, patients withcombined echocardiographic score � 2 (gray bar) or � 2(black bar).

Journal of the American Society of Echocardiography1070 McMahon et al October 2003

patients with unexpected catheterization findings(Table 2). The only outcome variable that differedsignificantly was a longer duration of chest tubedrainage in the group with unexpected findings(mean 3.8 vs 2.1 days, P � .05). All patients fromgroup A remained alive at a mean follow-up of 51months. There were 4 deaths (2%) in the 30-daypostoperative period in group B. Causes of deathincluded sepsis in 3 patients with heterotaxy/asple-nia and sepsis with cardiac arrest in 1 patient withhypoplastic left heart syndrome.

DISCUSSION

The benefits derived from creation of BCPAs inchildren with univentricular hearts results from areduction in volume loading conditions on the ven-tricle and reduction in AVVR.7-9 Traditionally inmost institutions, routine assessment of childrenundergoing BCPA includes preoperative cardiaccatheterization. Catheterization and angiography re-main the gold standard to assess PA anatomy and tocalculate pulmonary vascular resistance and PAP.Increasingly, noninvasive modalities including echo-cardiography and magnetic resonance imaging de-termine data routinely obtained by catheterizationincluding ventricular function, severity of AVVR orSVR, presence and degree of outflow tract obstruc-tion, and ratio of pulmonary to systemic bloodflow.10-12

Our study has demonstrated that by using specificechocardiographic criteria, a candidate for low-riskBCPA can be predicted. A significant number ofpatients can, therefore, avoid undergoing routinepreoperative cardiac catheterization. Good correla-tion was demonstrated between assessment of ven-tricular function and AVVR and SVR determined byechocardiography and cardiac catheterization.There were only 3 patients (4%) who had unexpect-edly elevated PAPs by catheterization comparedwith echocardiography. Conversely, many patientsdemonstrated higher echocardiographic-estimatedpeak EPAPs than directly measured by catheteriza-tion. In 8 patients (6%), proximal PA stenoses werefound on catheterization, which were not antici-pated from echocardiography. Irrespectively all pa-tients had their PA anatomy inspected at operationand were repaired when necessary.

Interestingly clinical end points did not differsignificantly between groups with the exception ofa longer duration of chest tube drainage in theunfavorable group by echocardiography. The 30-daymortality was only 2% (4 patients) overall with alldeaths occurring in the unfavorable echocardio-graphic group.

Limitations

The echocardiographic score we chose to define afavorable operative candidate may have been toostringent. However, the goal of our study was toidentify a patient group with excellent preopera-tive data that did not need routine cardiac cathe-terization. Despite few patients in the favorablegroup having unexpected findings at the time ofcardiac catheterization, there was an excellentoverall outcome. The identification of stenoseswithin the proximal and distal PA branches re-mains difficult by echocardiography and cardiaccatheterization remains the gold standard. Thisstudy is subject to well-recognized inherent limi-tations of a retrospective study. A prospectivestudy is required to further test the validity of ourproposed scoring system.

Conclusions

Using commonly obtained echocardiographic pa-rameters a subgroup of patients can be identifiedwho can safely avoid catheterization before creationof a BCPA. However, echocardiographic suggestionof a high-risk parameter, particularly pulmonaryhypertension or stenosis within the PA tree, shouldprompt further investigation with cardiac catheter-ization or other imaging modalities before surgicalintervention.

REFERENCES

1. Mott AR, Spray TL, Gaynor JW, Godinez RI, Nicolson SC,Steven JM, et al. Improved early results with cavopulmonaryconnections. Cardiol Young 2001;11:3-11.

2. de Leval MR, Kilner P, Gewillig M, Bull C. Total cavopulmo-nary connection: a logical alternative to atriopulmonary con-nection for complex Fontan operations; experimental studiesand early clinical experience. J Thorac Cardiovasc Surg 1988;96:682-95.

3. Migliavacca F, de Leval MR, Dubini G, Pietrabissa R. Acomputational pulsatile model of the bidirectional cavopul-monary anastomosis: the influence of pulmonary blood flow.J Biomech Eng 1996;118:520-8.

4. de Leval MR, Dubini G, Migliavacca F, Jalali H, Camporini G,Redington AN, et al. Use of computational fluid dynamics inthe design of surgical procedures: application to the study ofcompetitive flows in cavo-pulmonary connections. J ThoracCardiovasc Surg 1996;111:502-13.

5. Pizarro C, de Leval MR. Surgical variations and flow dynamicsin cavopulmonary connections: a historical review. SeminThorac Cardiovasc Surg Pediatr Card Surg Annu 1998;1:53-60.

6. Carey JA, Hamilton JR, Hilton CJ, Dark JH, Forty J, Parry G,et al. Orthotopic cardiac transplantation for the failing Fontancirculation. Eur J Cardiothorac Surg 1998;14:7-13.

7. Trusler GA, Williams WG, Cohen AJ, Rabinovitch M, MoesCA, Smallhorn JF, et al. William Glenn lecture: the cavopul-monary shunt; evolution of a concept. Circulation 1990;82:IV131-8.

Journal of the American Society of EchocardiographyVolume 16 Number 10 McMahon et al 1071

8. Mahle WT, Cohen MS, Spray TL, Rychik J. Atrioventricularvalve regurgitation in patients with single ventricle: impact ofthe bidirectional cavopulmonary anastomosis. Ann ThoracSurg 2001;72:831-5.

9. Uemura H, Yagihara T, Kawashima Y, Okada K, Kamiya T,Anderson RH. Use of the bidirectional Glenn procedure inthe presence of forward flow from the ventricles to the pulmo-nary arteries. Circulation 1995;92:II228-32.

10. Be’eri E, Maier SE, Landzberg MJ, Chung T, Geva T. In vivoevaluation of Fontan pathway flow dynamics by multidimen-

sional phase-velocity magnetic resonance imaging. Circula-tion 1998;98:2873-82.

11. Sharma S, Ensley AE, Hopkins K, Chatzimavroudis GP, HealyTM, Tam VK, et al. In vivo flow dynamics of the total cavopul-monary connection from three-dimensional multislice magneticresonance imaging. Ann Thorac Surg 2001;71:889-98.

12. Fratz S, Hess J, Sachwaiger M, Martinoff S, Stern HC. Moreaccurate quantification of pulmonary blood flow by magneticresonance imaging than by lung perfusion scintigraphy in pa-tients with Fontan circulation. Circulation 2002;106:1510-3.

Journal of the American Society of Echocardiography1072 McMahon et al October 2003