J THoRAc CARDIOVASC SURG 92:1021-1028, 1986 Extending the limits for modified Fontan procedures During the early development of atriopulmonary anastomotic operations (Fontan-Kreutzer), a number of physiologic and anatomical limits were proposed by the Fontan group as selection criteria. Among 167 consecutive patients undergoing modified Fontan procedures from 1973 through 1985, 109 (65%) patients exceeded one or more of the original selection criteria in areas of age, anomalies of systemic or pulmonary venous connection, pulmonary artery distortion, and pulmonary artery pressure. Twenty-six patients had a mean pulmonary artery pressure greater than 15 mm Hg, with 16 operative survivors (62%). Nineteen patients had anomalies of systemic and/or pulmonary venous connection, and 16 survived (84 %). There were 44 patients under the age of 4 years, and 26 survived (59 %). Twenty-five patients were older than 15 years, and 23 (92 %) survived the Fontan procedure. Pulmonary artery distortion, relating to prior palliative operations, was found in 34 patients. Seventeen of these 34 surviveda modified Fontan procedure (SO %). Twenty-six patients had a pulmonary arteriolar resistance more than 2 Wood units times square meter, and 14 survived(54%), whereas 81 of 93 with a pulmonary arteriolar resistance of less than 2 U . m' survived(87 % ). Multivariate analysis showed that pulmonary arteriolar resistance and pulmonary artery distortion had a significant, negative impact on survival, but age and anomalies of systemic and/or pulmonary venous connection did not. Pulmonary artery pressure was not an independent predictor of outcome. The results show that the original criteria may be exceeded in the areas of age and anomalies of pulmonary or systemic venousconnection. Pulmonary artery pressure alone should not contraindicate a Fontan procedure if pulmonary arteriolar resistance is low. Pulmonary artery distortion from a prior palliative operation and elevated pulmonary arteriolar resistance increase the risk of a Fontan procedure. John E. Mayer, Jr., M.D. (by invitation), Hrodmar Helgason, M.D. (by invitation), Richard A. Jonas, M.D. (by invitation), Peter Lang, M.D. (by invitation), Florentino J. Vargas, M.D. (by invitation), Nancy Cook, D.Sc. (by invitation), and AIdo R. Castaneda, M.D., Boston, Mass. During the early development of atriopulmonary anastomotic operations, a number of criteria were suggested by Choussat, Fontan, and co-workers' to aid in the selection of patients for these procedures. However, the extent to which these criteria can be extended has not been evaluated. In a review of all 167 consecutive patients undergoing modified Fontan procedures at From the Departments of Cardiology and Cardiovascular Surgery, The Children's Hospital, and the Departments of Pediatrics and Surgery, Harvard Medical School, Boston, Mass. Read at the Sixty-sixth Annual Meeting of The American Association for Thoracic Surgery, New York, N. Y., April 28-30, 1986. Address for reprints: John E. Mayer, Jr., M.D., Cardiovascular Surgery, The Children's Hospital, 300 Longwood Ave., Boston, Mass. 02115. Children's Hospital, Boston, from 1973 through 1985, a total of 109 (65%) patients were identified who exceeded one or more of the original ten criteria in the areas of age, anomalies of systemic and/or pulmonary venous connection, pulmonary artery deformity, and pulmonary vascular hemodynamics. This report describes the surgical results in this group of patients and explores the relative importance of these four criteria in predicting surgical outcome. Methods The hospital records and catheterization data of all 167 patients undergoing modified Fontan procedures at Children's Hospital between 1973 and 1985 were reviewed. For this report, the following variables were recorded: 1021 1 0 2 2 Mayer et al. Age Year of operation Diagnosis Anatomy of systemic atrioventricular (AV) valve (left or right sided; mitral, tricuspid, or AV canal) Spatial relationship of atrial baffle and systemic AV valve Prior shunt (yes/no) Prior pulmonary artery band (yes/no) Anomaly of systemic venous return (yes/no) Anomaly of pulmonary venous return (yes/no) Hemoglobin concentration Arterial saturation Mean pulmonary artery pressure (PAP) Ventricular end-diastolic pressure Pulmonary arteriolar resistance (indexed) (PARI) Pulmonary-systemic flow ratio Pulmonary artery distortion (yes/no) Type of atriopulmonary connection Presence of valve in atriopulmonary connection or in right side of circulation (yes/no) History of heart failure History of pulmonary hypertension A diagnosis was assigned to each patient on the basis of a synthesis of preoperative echocardiography and cardiac catheterization, operative findings, and autopsy results (when available). Tricuspid atresia was the diagnosis assigned in patients with situs solitus, an atretic right AV valve, and dextro-loop of the ventricle, with or without transposition of the great arteries. The diagnosis of single ventricle was assigned to patients with both atria emptying into a single ventricular chamber (double-inlet left or right ventricle). One or both AV valves could be patent, and there could be an "infundibular" chamber connected to the dominant ventricle via a bulboventricular foramen. Either dextroloop or levo-loop of the ventricle could be present. Heterotaxy syndrome was identified when typical systemic and/or pulmonary venous connection anomalies, atrial septal defect, and AV valve anomalies (generally complete AV canal) were present. Pulmonary atresia with intact ventricular septum was diagnosed when there were two ventricles with dextro-loop, each fed by a patent AV valve, and there was atresia of the outlet valve of the right ventricle. Hypoplastic left heart syndrome was defined as a prohibitively small left ventricle (end-diastolic volume less than 20 ml/m') in association with mitral and/or aortic atresia or stenosis and an intact ventricular septum. A diagnostic category of hypoplastic right ventricle and ventricular septal defect was used for patients with a right AV valve that straddled the ventricular septum and an underdeveloped The Journal of Thoracic and Cardiovascular Surgery right ventricle. Included in this group were patients with dextro-transposition and straddling tricuspid valve and patients with superoinferior ventricles. A final cetegory of "other" single ventricle was used for patients with a functionally single ventricular mass who did not fit into one of the other categories. Hemodynamic variables were derived from cardiac catheterization data acquired by standard methods. Pulmonary and systemic blood flows were determined by the Fick method. Pulmonary arteriolar resistance (indexed) (PARI) was determined from mean PAP, pulmonary venous atrial pressure, and calculated pulmonary blood flow (indexed) and expressed as Wood units times square meter. No calculation of pulmonary blood flow or resistance was made when the pulmonary arteries were not entered unless the only source of pulmonary flow was from a single source (single ventricle without systemic-pulmonary shunt or single systemic-pulmonary shunt with pulmonary atresia). Pulmonary venous wedge pressures were used to approximate mean PAP if the pulmonary arteries were not entered. The presence of pulmonary artery distortion was recorded when a review of the radiologic and catheterization reports combined with the recorded operative findings were found to show distorted or markedly hypoplastic peripheral or central pulmonary arteries. In doubtful cases the original angiograms were reviewed and a qualitative assessment was made by two of the authors. Anomalous systemic venous drainage was identified when there existed systemic venous connections to the heart other than a right-sided superior and inferior vena cava (or a left-sided superior and inferior vena cava in situs inversus). Included among these were connections of a left superior vena cava to the left atrium or coronary sinus. Anomalous pulmonary venous connection was defined when the pulmonary veins made no direct connection to the left atrium but drained to a systemic vein or to the same side of the atrium as the venae cavae. The technique of operation has been described previously.' All operations were performed with cardiopulmonary bypass and moderate hypothermia (25 0 to 28 0 C), and in 31 patients, weighing less than 10 kg, a period of less than I hour of profound hypothermia and circulatory arrest was used. Atriopulmonary connections were fashioned by a variety of techniques including right atrial-right "ventricular" outflow chamber connection (direct anastomosis or conduit) and right atrial-pulmonary artery connections (direct anastomosis or conduit). A "modified Glenn procedure" was defined as a direct Volume 92 Modified Fontan procedures Number 6 December 1986 Table I. Causes of death after Fontan operation (N = 167) No. Cause of death Low cardiac output Elevated pulmonary vascular resistance Failure of single ventricle Systemic venous obstruction Pulmonary venous obstruction Uncertain origin of low cardiac output Tamponade Noncardiac Total 12 9 2 4 7 I 3 38 (23%) connection between a superior vena cava and a pulmonary artery, either end of cava to side of pulmonary artery or end of pulmonary artery to side of vena cava (without ligation of superior vena cava-atrial junction in the latter). Techniques for dealing with anomalous systemic or pulmonary venous connection have been previously described. 3 1023 Table II. Fontan procedures Diagnosis Success Total 55 43 9 65 54 12 19 8 6 3 Tricuspid atresia Single ventricle Heterotaxy Hypoplastic RV + VSD PAjlVS HLHS Other 13 4 2 I ~ 85 80 75 68 50 33 33 Legend: RV, Right ventricle. VSD, Ventricular septal defect. PA, Pulmonary atresia. IVS, Intact ventricular septum. HLHS, Hypoplastic left heart syndrome. Table m. PARI: Effect on survival after Fontan operation PARI (V. m 2) o -0.99 I -1.49 1.5-1.99 ~2.0 Success Total % 40 18 22 14 42 22 29 26 95 82 76 54 Statistical methods The outcome variable of interest in this analysis was the failure or success of the Fontan procedure. An operation was considered a failure if the patient died of any cause within 30 days of the operation with or without the Fontan connection in place or if the Fontan connection was taken down after the initial operation. The relationship of this outcome with each of the potential predictor variables was examined with a general chi square test of association. When appropriate, categories were consolidated because of small numbers or similarity in meaning. Continuous predictor variables were categorized roughly into quartiles. When categories were ordered, a chi square test for trend in proportions was used. All predictor variables were then tested in a stepwise logistic regression model with the Statistical Analysis System on an IBM mainframe computer.' Indicator variables were created for levels of categorical variables, and trends across quartiles were tested for continuous variables. For variables with several missing values, a missing value indicator was created and included in the test for that variable. A significance level of 0.05 was required for entry into the logistic regression model. Results In the total group of 167 patients, 129 survived the operative period (77%). In two survivors the atriopulmonary connection was taken down in the operating room with reestablishment of a systemic-pulmonary shunt. Thus 127 patients had a successful modified Fontan operation (76%). The causes of death are shown in Table I. The results by diagnostic group are shown in Table II. The best survival rate occurred in patients with tricuspid atresia (85% operative success). The results in the single ventricle and heterotaxy groups were nearly equal to those of the tricuspid atresia group. PARI had an important impact on outcome after the modified Fontan procedure. Among the 119 patients for whom PARI was available, the success rate declined from 95% for PARI of less than 1 U . m 2 to 54% for PARI of 2 U . m' or more (Table III). In Table IV, the relationship of mean PAP to PARI and to successful outcome with a Fontan procedure is shown. When PAP was 15 mm Hg or less, the success rate was 80%, and this declined to 62% in the group with a PAP of more than 15 mm Hg, The greater importance of PARI is shown by the success rate of 80% in patients with a PAP of more than 15 mm Hg but with a PARI of less than 2 U . m', compared to a success rate of only 65% in the group with a PAP of less than 15 mm Hg but a PARI of 2 U . m' or more. These data are also displayed in Fig. 1, which shows the distribution of PAP and PARI. The results of modified Fontan procedures in patients with anomalies of systemic and/or pulmonary venous connection are shown in Table V. Eight patients had left 1024 The Journal of Thoracic and Cardiovascular Surgery Mayer et al. Table IV. PAP: Operative success after Fontan operation PARI PAP (mm Hg) SIT ~15 >15 69/78 12/15 Totals 80/93 I % SIT 87 80 11/17 3/9 86 14/26 I Total No PARI data ;;;2 U. m2 <2 U. m' % SIT 65 33 27/35 1/2 54 28/37 I % SIT 80 50 106/130 16/26 76 122/156 I % 82 62 78 Legend: S, Success. T, Total. Table V. Anomalies of venous connection-Fontan procedures Survivors LSVC to CS Left Glenn Atrial baffle Ligation LSVC to LA Left Glenn Atrial baffle/conduit Occlusion (cath, lab.) TAPVC Total Total 3 3 3 4 I I 3* 3* 3 5 I I 2 2 16 (84%) Table VI. Pulmonary artery distortion-i-Foman procedures 19 Legend: LSVC, Left superior vena cava. CS, Coronary sinus. LA, Left atrium. Cath. lab., Catheterization laboratory. TAPVC, Total anomalous pulmonary venous connection. 'One patient each with associated TAPVC. superior vena cava draining to the coronary sinus. This was managed by placing the atrial baffle which separated systemic from pulmonary venous return to the left of the coronary sinus in four. In one patient in whom the right AV valve was an important part of the pathway between the pulmonary veins and the ventricle, death occurred when the baffle partially obstructed the right AV valve orifice. In three patients an end-to-side connection between the left superior vena cava and left pulmonary artery (modified Glenn) was created to prevent obstruction of the right AV valve, and all three patients survived. In one patient with an innominate vein, the left superior vena cava was ligated. Nine patients had a left superior vena cava connecting directly to the left atrium. In five patients, a complex atrial baffle was sewn to segregate all of the systemic venous return on one side of the baffle and connect the pulmonary venous drainage to the AV valve(s). Three survived (60%). In three others a modified left Glenn anastomosis was created, and all three survived. In one patient the diagnosis of left superior vena cava was not made preoperatively or intraoperatively. The cava was Success Total % No distortion Distortion 110 17 133 34 83 50 Total 127 167 76 successfully occluded in the catheterization laboratory postoperatively. Two patients with left superior vena cava-left atrial connections had associated total anomalous pulmonary venous connection (TAPVC) , which was successfully managed by a modified left Glenn and intra-atrial baffle in one (TAPVC to left superior vena cava) and by an atrial baffle in the other (TAPVC to right atrium). Two others had isolated TAPVC managed by pulmonary vein-atrial anastomosis in one (supracardiac TAPVC) and by atrial baffle placement in the other (TAPVC to right atrium). In the total group of 19 patients with venous anomalies, there were 16 survivors (84%). Pulmonary artery distortion, related to previous palliative operations, was identified in 34 patients. Pulmonary artery reconstruction, in conjunction with the modified Fontan procedure, was attempted in 27. The results are shown in Table VI. Only 50% of patients with pulmonary artery distortion had a successful outcome, whereas 83% of those without pulmonary artery distortion were treated successfully. There were no survivors among the seven patients in whom a pulmonary artery reconstruction was not possible. The success rate was 63% (17/27) among those undergoing satisfactory pulmonary artery reconstruction in conjunction with the Fontan procedure. In six patients the stenosis was enlarged at the site of the atriopulmonary connection, and all survived. All but one of the remainder had patches placed across the stenotic area. One patient with a proximal right pulmonary artery stenosis had a right-sided Glenn anastomosis (without caval ligation) Volume 92 Number 6 December 1986 Modified Fontan procedures 1025 Table VII. Age-operative success after Fontan operation 36 Success Total % 0-3.99 4-15.99 26* 78 23 44 98 25 59 80 92 Total 127 167 76 "Two additional patients survived but with the Fontan shunt taken down. -:-.. 28 ~ 24 ~20 and a right atrium-main pulmonary artery conduit placed successfully. The impact of age on operative success is shown in Table VII. A total of 69 patients were outside the 4 to 15 year age range and 74% survived, although two patients in the under 4 year group had early takedown of their atriopulmonary connection and their operations were considered failures. Age under 4 years did not exert an independent influence on survival in the multivariate analysis. A partial explanation for this is apparent from an examination of Table VIII. Fourteen of the 44 patients under 4 years of age had a PARI of 2 Wood units or more, and the survival rate in this group was 43% (6/14). Of those under 4 years of age with a low PARI (1 Wood unit or less), all five survived. The relationship between the number of ChoussatFontan criteria exceeded (maximum of four in this study) and the outcome is shown in Table IX. Clearly, one criterion can be exceeded without adversely affecting results. The relative importance of the four factors (age, pulmonary artery distortion, pulmonary hemodynamics, and venous anomalies) is addressed by the multiple logistic regression analysis (Table X). In this analysis, both pulmonary artery distortion and elevated PARI were associated with increased operative risk, which confirms the impression from the univariate assessment. The presence of a left-sided AV valve with mitral valve morphology was associated with a reduced risk. Included in this group of patients were those with classical tricuspid atresia (63) as well as eight patients with pulmonary atresia and intact ventricular septum, 21 patients with single ventricle, and 14 patients with hypoplastic right ventricle and ventricular septal defect. This factor, which was associated with reduced risk, was of greater significance than any of the standard diagnostic groups, explaining why the diagnosis of tricuspid atresia did not appear as an independent risk factor (see Table II). Neither age nor the presence of anomalies of systemic or pulmonary venous connection was associated with increased risk. The values for these variables • ~ ~ ~ 16fo12 8 0 Ieo Success I Failure 32 Age (yr) ~16 0 0 • • •• ••• 0 • 0 • • • •t •• ••• 0 ~···U e.l.. • I .0 0 ~ • ': • • •• • •• o• •• •;: •• t •••••• • • b .. ~ 00 • • 0 ~ 2 PARI(IJ·rrIJ >15112/15 (80%) II 0 • 0 0 0 •• 3 3/9 (33%) 4 I <15169/78(87%) 1111/17(65%) 1 L - <2------..J l - - . - >2----...l Fig. 1. Relationship between mean pulmonary artery pressure (PjiA), pulmonary arteriolar resistance indexed (PARI), and outcome after Fontan procedure. shown in the table were obtained only by forcing these variables into the model. The decreased risk associated with the later years of operation is attributed to the acquisition of experience in the operative and postoperative management of these patients. Discussion The dismal natural history of tricuspid atresia' and univentricular heart"!' combined with the decreasing mortality rates for modified Fontan procedures and the symptomatic improvement which most patients experience after this procedure have led to attempts to extend this procedure to patients who do not meet one or more of the ten criteria originally proposed by Choussat, Fontan, and their associates. I The results of the current review help to define the areas in which these criteria can and cannot be safely extended in applying the Fontan principle. The results in patients outside the 4 to 15 year age group suggest that modified Fontan procedures can be offered to patients outside this age range. The very favorable results in the over 15 year age group suggest that this limit alone should not contraindicate a Fontan procedure. Others have reported good results in this age group.' The group under 4 years of age requires closer examination. In this younger age group, choices frequently must be made between a palliative shunt procedure (which may be the second or third palliative procedure) and a modified Fontan procedure. In the 1026 The Journal of Thoracic and Cardiovascular Surgery Mayer et al. Table VIII. Age and PARI-Operative success after Fontan operation PARI (Wood units indexed) ND 0-0.99 Age (yr) (SfT) (S/T) o - 3.99 6/10* 22/32 5/6 5/5 25/27 10/10 3/5 9/11 6/6 34/48 40/42 18/22 4.0-15.99 ~16 Total 1.0-1.49 I I (S/T) 1.5-1.99 (S/T) 6/10* 13/16 3/3 ~2.0 I 22/29 (SfT) 6/14 8/11 0/1 14/26 Legend: ND, No PARI data. S, Success. T, Total. 'One patient survived takedown in the operating room (counted as failed Fontan). Table IX Survival rate No. of criteria exceeded o I 2 3 Table X. Risk factors for poor outcome of Fontan procedure (N = 167) (%) 49 58 22 o 58 72 33 4 84 81 67 o Variable PA distortion PARI (vs. 0.99) o -0.99 I -1.49 1.5-1.99 ~2.0 current group under 4 years, 12 of 14 (86%) patients with low PARI and nondistorted pulmonary arteries underwent a successful Fontan procedure. Age alone was not an independent risk factor when pulmonary artery distortion and PARI were taken into account. These results are in contrast to those of Fontan," Cleveland," Mair,' and their colleagues, all of whom found young age alone to be an independent risk factor. However, for an equivalent level of PARI, results are less favorable in the younger group although the number of patients is relatively small. Because of these data, we believe that patients under 4 years of age can be offered a Fontan procedure when low PARI (less than 2 U . m') and nondistorted pulmonary arteries are present. The encouraging results in patients with anomalies of pulmonary and systemic venous connection suggest that their presence should not necessarily contraindicate a Fontan procedure. These results are in contrast with those of King," DjCarlo.!' and their co-workers. Several technical features are important in this group of patients and have been detailed in a prior publication.' An end-to-side left superior vena cava-left pulmonary artery anastomosis (modified left Glenn) has been very useful. Complex inta-atrial baffles should be used cautiously because of the risk of distortion by the right atrial-left atrial pressure difference that will occur postoperatively. This distortion can obstruct the pulmo- Left-sided mitral valve (vs. any other systemic AV valve) Year of operation (vs. 84-85) 73-79 80-81 82-83 84-85 Age (vs. 4-15) o -3.99 4 -15.99 ~16 ASVC APVC p Value Odds ratio 0.0164 3.50 0.0724 0.0221 0.0047 6.14 9.40 15.82 0.22 0.0537 0.001 0.6552 3.69 8.55 1.36 0.2326 1.87 0.4472 0.7773 0.9046 0.57 0.87 0.82 om Legend: PA, Pulmonary artery. PARI, Pulmonary arteriolar resistance (indexed). AV, Atrioventricular. ASVC, Anomalous systemic venous connection. APYC, Anomalous pulmonary venous connection. nary venous pathway, particularly when a right-sided or common AV valve will serve as the systemic AV valve. The anatomy of the venous connections should be clearly delineated by preoperative echocardiography and catheterization, but with careful preoperative study and an individualized operative approach, favorable results can be achieved. The dominant effects of the pulmonary vascular hemodynamics on outcome after Fontan procedures are confirmed by this group of patients. PAP as an isolated measurement is less useful than calculated PARI, because an 80% survival rate was achieved in the group with a PAP of more than 15 mm Hg but a PARI of less than 2 U . m-, In addition, PAP did not appear as an Volume 92 Number 6 December 1986 independent risk factor in the multivariate analysis. Similar results have been obtained by the Mayo Clinic group.' The current group of patients contains only two patients with a PARI of 4 U . m' or more, and both died. The level of PARI of 2 U . m 2 or more leading to increased operative risk in our patients is lower than the 4 U . m 2 level suggested by Choussat, Fontan, and their colleagues', In the group with a PARI of 2 U . m' or more, the anatomic diagnosis seems to be related to outcome, as eight of 10 with tricuspid atresia survived operation but only six of 16 non-tricuspid atresia defects survived. This difference was possibly related to the type of atriopulmonary connection, as five of seven with right atrial-right "ventricular" connections survived and nine of 19 with right atrial-pulmonary arterial connections survived. However, in the total group of 167 patients, the type of atriopulmonary connection (to "ventricle" or pulmonary artery) or the presence or absence of a valve anywhere in the right side of the circulation did not have an independent effect on outcome. Intermittent abdominal compression to support the circulation in the early postoperative period (when pulmonary arteriolar resistance may be temporarily elevated) may be useful in managing patients with a high PARI, but we have not found this technique to be consistently useful. The calculation of PARI is not always possible, particularly when there are multiple sources of pulmonary blood flow, and the fact that PARI is a derived variable must also be kept in mind. Finally, the importance of pulmonary artery distortion was emphasized by this review. The disastrous impact of residual pulmonary artery distortion at the completion of the Fontan procedure is shown by the 100% mortality in the group in whom pulmonary artery reconstruction was not possible. However, when a good pulmonary artery reconstruction can be performed, reasonable results can be achieved (63% survival). Some authors have emphasized the importance of the size of the pulmonary arteries in the outcome of Fontan procedures,14 but others have not reproduced these results." In our group of patients, all of the distortions in the pulmonary artery were related to prior palliative procedures. This finding has provided additional impetus toward minimizing or eliminating palliative procedures before a modified Fontan procedure. When deformities more than a simple, centrally located, and readily accessible stenosis are identified, we now prefer to reconstruct the pulmonary arteries in preliminary operations. The technique reported by Uretzky, Puga, and Danielson," in which a Glenn shunt is combined with a connection between the right atrium and right Modified Fontan procedures 1 027 ventricle or main pulmonary artery, is an alternative to patching a stenotic area and was used successfully once in this series. In conclusion, the results of this review suggest that one or more of at least four of the original ten criteria of Choussat and Fontan may be exceeded, except in the area of PARI. PAP should not be considered alone as a contraindication to a Fontan procedure as long as PARI is low. Anomalies of venous connection may be successfully dealt with, and easily accessible localized distortions of the pulmonary may be repaired, but should be viewed with caution. Age alone does not contraindicate a Fontan procedure although criteria for younger patients should be more strict than those for the older age groups. 2 3 4 5 6 7 8 9 10 11 REFERENCES Choussat A, Fontan F, Besse P, Vallot F, Chauve A, Bricaud H: Selection criteria for Fontan's procedure, Paediatric Cardiology, RH Anderson, EA Shinebourne, eds., Edinburgh, 1978, Churchill Livingstone, pp 559566 Mayer J, Jonas R, Castaneda A: Modified Fontan procedure. Modern techniques in surgery. Cardiac/Thorac Surg. 78:1-11, 1985 Vargas F, Mayer J, Jonas R, Castaneda A: Anomalous systemic and pulmonary venous connections in conjunction with atriopulmonary anastomosis (Fontan-Kreutzer). Technical considerations. J THORAC CARDIOVASC SURG (in press) SAS User's Guide, SAS Institute, Inc., Cary, N.C., 1982 Dick M, Fyler D, Nadas A: Tricuspid atresia. Clinical course in 101 patients. Am J Cardiol 36:327-337, 1975 Moodie D, Ritter D, Tajik A, O'Fallon W: Long term follow-up in the unoperated univentricular heart. Am J Cardiol 53: 1124-1128, 1984 Mair D, Rice M, Hagler D, Puga F, McGoon D, Danielson G: Outcome of the Fontan procedure in patients with tricuspid atresia. Circulation 72:Suppl 2:88-92, 1985 Fontan F, Devill C, Quaegebeur J, Ottenkamp J, Sourdille N, Choussat A, Brom G: Repair of tricuspid atresia in 100 patients. J THORAC CARDIOVASC SUG 85:647-660, 1983 Cleveland D, Kirklin J, Naftel D, Kirklin J, Blackstone E, Pacifico A, Bargeron L: Surgical treatment of tricuspid atresia. Ann Thorac Surg 38:447-457, 1984 Stefanelli G, Kirklin J, Naftel D, Blackstone E, Pacifico A, Kirklin J, Soto B, Bargeron L: Early and intermediateterm (10 year) results of surgery for univentricular atrio-ventricular connection ("single ventricle"). Am J Cardiol 54:811-821, 1984 Sales H, Milliken J, Perloff J, Hellenbrand W, George B, Chin A, DiSessa T, Williams R: Experience with the The Journal of Thoracic and Cardiovascular Surgery 1028 Mayer et af. Fontan procedure. J THORAC CARDIOVASC SURG 88:939951,1984 12 King R, Puga F, Danielson G, Julsrud P: Extended indications for the modified Fontan procedure in patients with anomalous systemic and pulmonary venous return (abstr). Second World Congress of Pediatric Cardiology, New York, 1985, Springer-Verlag, p 7 13 DiCarlo D, Marcelletti C, Nijveld A, Lubbers L, Becker A: The Fontan procedure in the absence of the interatrial septum. Failure of its principle? J THORAC CARDIOVASC SURG 85:923-927, 1983 14 Nakata S, Jasuhara I, Yoshinori T, Hiromi K, Kisuhiro T, Mahoto N, Masahiko A, Atsuyoshi T: A new method for the quantitative standardization of cross sectional areas of the pulmonary arteries in congenital heart diseases with decreased pulmonary blood flow. J THORAC CARDIOVASC SURG. 88:610-619, 1984 15 Girod D, Rice M, Mair D, Julsrud P, Puga F, Danielson G: Relationship of pulmonary artery size to mortality in patients undergoing the Fontan operation. Circulation 72:Suppl 2:93-96, 1985 16 Uretzky G, Puga F, Danielson G: Modified Fontan procedure in patients with previous ascending aortapulmonary artery anastomosis. J THORAC CARDIOVASC SURG 85:447-450, 1983 (For Discussion see page 1044.)