469
Five- to Fifteen-Year Follow-up After
Fontan Operation
David J. Driscoll, MD; Kenneth P. Offord, MS; Robert H. Feldt, MD;
Hartzell V. Schaff, MD; Francisco J. Puga, MD; and Gordon K. Danielson, MD
Background. The purpose of this study was to estimate survival and quality of outcome and
factors associated with outcome for patients out 5 to 15 years from their Fontan
operation.
Methods and Results. We studied 352 patients who had the Fontan operation prior to 1985. The
overall 1-, 5-, and 10 -year survival was 77%, 70%1, and 60%1, respectively. The following factors
were significantly associated with lower survival: univentricular heart or complex congenital
anomalies other than tricuspid atresia, early calendar year of operation, heterotaxia syndromes, early age at operation, increased pulmonary artery pressure, atrioventricular valve
dysfunction, and higher (worse) New York Heart Association class. Reoperations were
necessary for 103 of the 352 patients. At least 20%, of the survivors have or have had cardiac
arrhythmias requiring antiarrhythmic medication or mechanical pacemaker insertion. Between 7% and 10%l of the patients have had or had protein-losing enteropathy/hypoproteinemia.
At 5 years postoperatively, 122 patients (34.7%) were alive with a better New York Heart
Association functional classification than preoperatively. Fifty-eight patients (16.5%) were alive
and in the same functional classification, but 126 (35.8%) died within the first 5 years or were
in a worse functional classification. Thirty-nine patients were doing excellently and 29 patients
poorly 5 years after the operation. Of the surviving patients, 43% can do as much exercise as
their peers, whereas 3% are incapable of exercise.
Conclusions. To assure good functional long-term outcome in addition to survival, clinicians
must exclude from selection for Fontan operation patients known to be at high risk for death
or poor outcome. (Circulation 1992;85:469-496)
assess
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
In 1971, Fontan and Baudet1 described a technique for successful definitive palliation for
patients with tricuspid atresia. Subsequently,
this technique has been applied to the treatment of
most forms of functional single ventricle. The operation results in separation of systemic and pulmonary
venous return, directing the former into the pulmonary artery without passing through a ventricle. Pulmonary venous return is directed to the single ventricle and subsequently pumped into the aorta.
Operative and early postoperative mortality and
morbidity have been addressed by numerous investigators.2-6 However, mid- to long-term follow-up
studies of these patients are few, have follow-up
times less than 5 years, include a relatively small
number of patients, or have follow-up information
for less than 100% of the cohort.2-1' Increasingly,
From the Section of Pediatric Cardiology, the Section of Biostatistics, and the Section of Cardiovascular Surgery, Mayo Clinic
and Mayo Foundation, Rochester, Minn.
Address for reprints: David J. Driscoll, MD, Mayo Clinic, 200
First Street SW, Rochester, MN 55905.
Received May 10, 1991; revision accepted September 24, 1991.
mid- and long-term problems have been recognized
after this operation, including arrhythmias, proteinlosing enteropathy, cirrhosis, subaortic obstruction,
atrioventricular valve insufficiency, cardiomyopathy,
and death.
The purpose of this study was to determine the 5- to
15-year survival and functional status of 352 patients
who had a modified Fontan operation at the Mayo
Clinic before December 31, 1984. We sought to identify
factors associated with patient survival and to determine which factors might further identify the survivors
with an excellent, intermediate, or poor outcome.
Methods
Original Cohort
Between October 31, 1973, and December 31, 1984,
352 patients had a modified Fontan operation at the
Mayo Clinic. These dates were chosen for this study to
allow for a potential of at least 5 postoperative years of
follow-up. There were 215 male patients and 137
female patients. The mean age at the time of the
Fontan operation was 11.1 years, the median was 10
years, and the range was from less than 1 year to 42
470
Circulation Vol 85, No 2 February 1992
TABLE 1. Number of Choussat* Risk Factors Present
Patients (n=352)
Risk factors (no.)
n
0
46
1
93
2
80
3
76
4
30
19
5
6
7
7
1
%
13
26
23
22
9
5
2
0.3
*Reference 12.
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
years. The primary cardiac malformation was tricuspid
atresia in 125 patients, univentricular heart (doubleinlet ventricle) in 114 patients, and other complex
forms of single ventricle in the remaining 113. Only 46
patients (13%) fulfilled all of the criteria for operability
for the Fontan operation as described by Choussat et
al.12 Indeed, more than 60% lacked two or more of
these criteria (Tables 1, 2, and 3).
Cohort of Suwvivors
Of 352 original patients, 122 are known to be dead,
230 are known to be alive at last follow-up, and all
except 19 of the 230 returned a follow-up medical
questionnaire for the present study. We obtained
follow-up information from next of kin for four patients who died shortly before receiving our follow-up
questionnaire. Hence, detailed follow-up information
was available for 215 patients. In addition, of the 19
known to be alive at the last known follow-up but who
did not return a questionnaire, current survival information was available from other sources for an additional three. The remaining 16 patients who were lost
to follow-up were last known to be alive from 174 days
to 10.9 years after operation.
For these 215 5-year survivors who returned a
medical questionnaire, the mean age at the time of
the Fontan operation was 11.3 years (range, 1-42
years). The present mean age of this subset is 19.1
years (range, 7-50 years). The mean follow-up interval for these surviving patients was 7.7 years (range,
0.5-15.5 years). For only 14 patients was the last
known follow-up information obtained less than 5
years postoperatively (Table 4).
Data Ascertainment
The Mayo Clinic medical records of these 352
patients were reviewed and abstracted for findings
before and after the Fontan operation. In addition, a
detailed health status questionnaire was sent to each
known survivor. Patients who did not return or did
not complete the questionnaire were sent a second
questionnaire. If the second questionnaire was not
returned or completed, an attempt was made to
contact the patients by telephone.
Statistical Analysis
Statistical analysis included Fisher's exact test, x2
test of association for comparing proportions,
Wilcoxon rank-sum test, Kaplan-Meier survival
curves, log-rank test, and Cox's proportional hazards
model for assessing multivariate associations between
survival and risk factors.13-18 For some survival analyses, the date of the Fontan operation was used as time
0, whereas for others, 30 days after the operation was
considered time 0. All deaths, regardless of cause, that
occurred after the initiation of the operation were
considered in the survival analysis irrespective of
whether they occurred intraoperatively, during the
postoperative hospitalization, or after hospitalization.
For determining which factors were predictors of
death or other end points including death, the predictor variables were analyzed as continuous or categorical variables as appropriate. In the multivariate
analysis, we used backward elimination of nonsignificant variables. For graphical presentation of the
significant variables found in multivariate analysis,
we used the concept of an integer-valued risk score to
summarize the significant predictor variables. The
levels of the significant predictor variables were
categorized and given integer weights ranging in
value from 0 to 5. The weights were chosen stepwise,
increasing the value until the underlying variable no
longer contributed to the prediction of the end point
after the risk score for that variable was included.
Logistic regression was used for assessing factors
associated with the 5-year survival status and functional outcome. The SAS statistical software system
was used throughout.19 Only two-tailed probability
values are reported; values <0.05 were considered
significant.
The variables evaluated for their association with
survival and functional status of survivors are listed in
Tables 2 and 3. With the exception of "normal right
atrial size," all of the criteria proposed by Choussat
et al12 for a Fontan operation were included in the
analysis. All variables except postoperative right
atrial pressure were preoperative observations.
Results
Mortality
Of the initial 352 patients, 230 were known to be
alive at last contact, and 216 of the 230 were known
to be alive at least 5 years postoperatively.
The overall 30-day, 1-, 5-, and 10-year survival from
time of the operation was 84%, 77%, 70%, and 60%,
respectively (Figure 1). The majority of the 122 deaths
were cardiac related, with 56 ascribed to ventricular
failure (Table 5). Other causes of death, such as renal
failure, may have resulted from ventricular failure.
Predictors of Mortality
Eighteen potential predictors of mortality were
examined (Tables 2 and 3). When appropriate, they
were considered in both a continuous and discrete
fashion. Long-term survivorship was assessed from
471
Driscoll et al Follow-up After Fontan Operation
TABLE 2. Univariate and Multivariate Assessment of Factors Associated With Survival for the End Point "Death Any Cause" With Time
Zero the Initiation of Fontan Operation
Pm (with Pm (w/o
Survival (%)
n
1
10
Years
Pc
6
Months
Year
Years
Pu
30 Days
5
RAP)
RAP)
Factor
76.7
59.8
...
78.7
70.0
352 83.8
Overall
Sex
57.5
NS
NS
NS
77.2
75.3
68.2
215 81.2
Male
81.0
78.8
63.7
72.9
137 86.9
Female
Lesion
125 90.4
84.0
84.0
79.9
70.1
<0.001
Tricuspid atresia
0.010
0.020
80.7
57.2 X 2=14.18
114 86.0
83.3
73.4
Univentricular heart
51.5
64.6
<0.001
0.033
113 74.3
68.1
55.6
Complex single ventricle
Calendar year of operation
37.5
8 62.5
37.5
37.5
25.0
0.019 0.011(-) <0.001(-) <0.001(-)
1970-1975
127 79.5
74.8
66.8
58.1 X2=7.95 X 2=6.47
75.6
1976-1980
217 87.1
79.2
73.1
62.6
82.0
1981-1984
Heterotaxia
62.7
...
79.7
73.7
<0.001
NS
0.006
81.5
325 86.8
Absent
27 48.1
40.7
25.9 X 2=36.37
44.4
25.9
Present
Asplenia
61.2
...
80.6
78.6
71.7
<0.001
341 85.6
Absent
18.2
18.2 X 2=30.01
11 27.3
18.2
18.2
Present
.
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
Polysplenia
Absent
Present
Age at operation (yr)
4-15
<4 or >16*
<4
4-15
>16
Preoperative sinus rhythm
Present
Absent*
<0.001
.
.
.
.
.
.
..
...
336
16
84.8
62.5
79.5
62.5
77.7
56.3
71.9
31.3
61.2
31.3
240
112
33
240
79
82.5
86.6
66.7
82.5
94.9
78.3
79.5
54.5
78.3
89.9
76.7
76.8
54.5
76.7
86.1
70.7
68.6
51.5
70.7
75.7
61.9
55.9
46.8
61.8
60.1
320
32
85.0
71.9
79.7
68.8
78.4
59.4
72.4
46.3
61.1
46.3
297
55
86.5
69.1
81.5
63.6
79.8
60.0
73.6
50.7
61.4
50.7
X 2=8.65
154
177
89.0
79.7
83.8
74.0
83.1
71.2
76.5
64.3
69.0
53.1
0.012 0.001(+) <0.001(+) <0.001(+)
X2=6.33 42=10.45
154
5
84.4
60.0
79.2
60.0
75.9
60.0
69.9
20.0
54.4
20.0
X2=12.32
NS
NS
NS
0.035(-)
0.011
X 2=6.45
...
NS
NS
0.003
...
NS
NS
0.025
X 2=7.40
Systemic venous drainage
Normal
Abnormal*
Mean PA pressure (mm Hg)
Normal (c15)
Abnormal (>15)*t
Pulmonary arteriolar resistance (U. mi2)
Normal (<4)
Abnormal (.4)*t
Mean PA size (% normal)
Normal (.60)
Abnormal (<60)*
LV ejection fraction (%)
Normal (.60)
Abnormal (<60)*t
AV valve dysfunction
Normal
Abnormal*t
...
NS
...
NS
0.022(-)
X 2=5.28
...
<0.001
...
<0.001
0.002
...
NS
NS
55
2
41
71
95.1
88.7
95.1
83.1
92.7
78.9
87.6
73.1
81.4
29.2
319
33
85.6
66.7
81.5
51.5
79.9
45.5
73.5
36.4
62.5
36.4
281
71
83.3
85.9
79.0
77.5
76.5
77.5
71.4
64.4
61.4
52.5
NS
X 2=21.02
PA architecture
Normal
Abnormal
NS
472
Circulation Vol 85, No 2 February 1992
TABLE 2. Continued.
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
Survival (%)
Pm (with Pm (w/o
Factor
n 30 Days 6 Months 1 Year 5 Years 10 Years
Pu
Pc
RAP)
RAP)
LV end-diastolic pressure (mm Hg)
Normal (.12)
121 89.3
83.5
81.8
67.9
77.5
0.025 0.013(+)
NS
NS
Abnormal (>12)*
187 80.2
75.9
74.3
65.1
54.4 X2=5.02 2=6.18
NYHA functional class
I
5 100.0
100.0
100.0 100.0
100.0
NS
0.025
NS
0.017
II
182 84.1
81.3
79.1
73.5
62.5
1=5.00
III
152 82.2
75.0
73.7
64.9
55.3
IV
7 85.7
71.4
57.1
57.1
57.1
Immediate postop RA pressure
(mm Hg)
Normal (.20)
271 91.1
87.1
85.2
79.2
70.1
<0.001 <0.001
<0.001
...
66 65.2
Abnormal (>20)
54.5
51.5
45.5
29.2
2=45.96 2=39.18
Risk group 9
All absent
46 89.1
87.0
87.0
87.0
70.7
0.051
...
...
...
>1 present
306 83.0
77.5
75.2
67.5
58.2
2=3.8
Risk group 4
All absent
143 88.8
84.6
84.6
77.5
68.1
0.006
...
.1 present
209 80.4
74.6
71.3
65.0
52.4
2=7.60
1~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
RAP, right atrial pressure; PA, pulmonary artery; LV, left ventricle; AV, atrioventricular; NYHA, New York Heart Association; RA, right
atrium.
n refers to number of patients with the factor. For some factors, a few patients are missing information; sum of the numbers in subgroups
will not equal overall total. Survival estimates are obtained from Kaplan-Meier survival curves for selected time points after time zero.
Pu refers to univariate probability value from log-rank test with NS (nonsignificant) if Pu >0.05. Below each significant finding is the
associated x2 statistic subscripted with its degrees of freedom. Comparison of the subgroups is an overall comparison of the survival
experience not restricted to the selected time points presented. Pc refers to univariate probability value from Cox regression model assessing
association between survival and factor considered as a continuous variable. Direction of association is indicated in parentheses. Minus (-)
sign indicates that survival is improving (the hazard is decreasing) as the level of the factor increases or if the condition is present. Pm (with
RAP) is the significance of that factor in a multivariate model arrived at after backward elimination of the nonsignificant (p>0.05, NS)
variables in a model that initially included the immediate postoperative right atrial pressure. Pm (w/o RAP) is like Pm (with RAP) except
that only preoperative factors were considered.
AV valve abnormal if moderate or severe insufficiency preoperatively or necessity for valve repair or replacement prior to or concurrent
with Fontan procedure.
All factors except postop RA pressure are based on preoperative assessment except where indicated.
*Risk group 9 consists of the presence of any (.1) of the nine conditions (criteria of Choussaut's et al12).
tRisk group 4 consists of the presence of any (.1) of the four conditions.
the time of operation and from day 30 postoperatively; the latter analysis excluded patients who died
within 30 days of the operation. These analyses were
performed with and without the inclusion of postoperative right atrial pressure.
Analysis of Survival From Time of the Operation
As is apparent from Table 3, there are numerous
variables that are significantly associated with survival in a univariate fashion. In fact, of all the
variables, only sex, pulmonary arteriolar resistance,
and pulmonary artery architecture were not associated with survival. For the 46 patients who fulfilled
all of the criteria of Choussat et al, the survival at 5
and 10 years was 87% and 71%, respectively, whereas
the 306 patients who violated one or more of these
criteria had rates of 68% and 58%, respectively. We
elected to examine survival for patients who violated
one or more of the following four variables: mean
pulmonary pressure greater than 15 mm Hg, pulmonary arteriolar resistance equal to or greater than 4
U im2, left ventricular ejection fraction less than
60%, and atrioventricular (AV) valve dysfunction
(moderate or severe AV valve insufficiency or the
need to repair or replace the valve). The 5- and
10-year survival for 143 patients who did not violate
these criteria was significantly better (78% and 68%,
respectively) than for those who violated one or more
of these conditions (65% and 52%, respectively).
For each end point, two multivariate models were
constructed, one excluding and one including postoperative right atrial pressure. These predictor variables were weighted statistically by the model to
assess their relative effects on survival. The effect on
survival of one or more of these variables is illustrated in Figure 2. Excluding postoperative right
atrial pressure, the following variables were retained
in the model: univentricular heart, complex single
ventricle, early calendar year of operation, heterotaxia, young age at operation, increased pulmonary
artery pressure, AV valve dysfunction, and higher
New York Heart Association (NYHA) classification.
If postoperative right atrial pressure is allowed to
enter the model, age at operation, heterotaxia, and
Driscoll et al Follow-up After Fontan Operation
473
TABLE 3. Univariate and Multivariate Assessment of Factors Associated With Survival for End Point "Death Any Cause" With Time Zero
30 Days Post Fontan (Excludes Those Who Died Within 30 Days of Fontan Operation)
Survival (%)
Pm (with Pm (w/o
n 30 Days 6 Months 1 Year 5 Years 10 Years
Pu
Pc
Factor
RAP)
RAP)
295 96.3
93.2
91.5
83.2
71.3
...
...
Overall
...
...
Sex
176 97.2
93.2
92.0
82.7
70.2
NS
Male
NS
NS
119 95.0
93.3
90.8
83.9
73.3
Female
Lesion
113 96.5
92.9
92.9
88.4
77.5
Tricuspid atresia
NS
98 98.0
95.9
93.8
85.4
66.6
NS
Univentricular heart
NS
NS
84 94.0
86.9
90.5
73.5
69.2
NS
Complex single ventricle
Calendar year of operation
5 80.0
60.0
60.0
60.0
1970-1975
NS
NS
0.010(-) 0.005(-)
40.0
101 95.1
94.0
92.0
84.0
73.1
1976-1980
189 97.4
1981-1984
93.1
91.0
83.4
71.9
Heterotaxia
282 96.5
Absent
93.3
91.8
84.6
72.3
0.012
...
NS
0.043
13 92.3
84.6
69.2
53.8
53.8 X 2=6.27
Present
.
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
Asplenia
Absent
Present
Polysplenia
Absent
292
3
.
.
(1 died in <30 days, 1 alive out 5-6 years, 1 alive out 6-7 years)
285
10
96.1
100.0
93.0
100.0
91.6
90.0
84.4
50.0
72.1
50.0
0.019
XV=5.47
198
97
22
198
75
96.5
95.9
95.4
96.5
96.0
94.4
90.7
81.8
94.4
93.3
92.9
88.7
81.8
92.9
90.7
85.2
79.2
77.3
85.2
79.8
75.0
64.5
70.2
75.0
63.3
NS
...
...
NS
NS
NS
NS
272 96.3
23 95.7
Absent*
Systemic venous drainage
Normal
257 96.5
38 94.7
Abnormal*
Mean PA pressure (mm Hg)
137 96.4
Normal (<15)
141 95.7
Abnormal (>15)*t
Pulmonary arteriolar resistance (U m2)
Normal (<4)
130
3
Abnormal (.4)*t
Mean PA size (% normal)
49
Normal (.60)
2
Abnormal (<60)*
LV ejection fraction (%)
39 100.0
Normal (.60)
63 95.2
Abnormal (<60)*t
AV valve dysfunction
273 97.1
Normal
22 86.4
Abnormal*t
PA architecture
93.4
91.3
92.3
82.6
84.8
64.5
71.8
64.5
NS
...
NS
NS
93.4
92.1
92.2
86.8
84.7
73.4
71.0
73.4
NS
...
NS
NS
93.4
92.2
93.4
89.4
85.2
80.7
77.6
66.7
NS
0.013(+) 0.046(+)
X 2=6.15
...
0.045(+)
Present
Age at operation (yr)
4-15
<4 or .16*
<4
4-15
>16
Preoperative sinus rhythm
Present
Normal
Abnormal*
234
61
95.7
98.4
NS
...
x1=4.03
...
0.025(-)
...
X1=5.04
100.0
92.1
97.4
88.9
89.0
82.4
85.6
32.9
94.5
77.3
93.4
68.2
85.5
54.5
73.0
54.5
94.0
90.2
91.9
90.2
85.3
75.0
73.7
61.1
NS
0.020(-)
...
x2 =5.43
<0.001
...
0.001
<0.001
...
NS
0.024
X2= 13.32
0.050
X
2=3.85
474
Circulation Vol 85, No 2 February 1992
TABLE 3. Continued.
Survival (%)
n 30 Days 6 Months 1 Year 5 Years 10 Years
Pu
Pm (w/o
RAP)
Pm (with
RAP)
Pc
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
Factor
LV end-diastolic pressure (mm Hg)
NS
NS
108 96.3
92.6
91.6
86.8
76.0
NS
NS
Normal (s12)
81.2
94.0
92.7
67.8
150 96.0
Abnormal (>12)*
NYHA functional class
I
5 100.0
100.0
100.0
NS
0.016(+) 0.025(+) 0.011(+)
100.0 100.0
87.4
74.4
153 98.0
95.4
94.1
II
X 2=5.78
125 94.4
91.2
89.6
67.3
78.1
III
IV
66.7
66.7
6 83.3
83.3
66.7
Immediate postop RA pressure
(mm Hg)
0.012
...
93.5
86.9
76.9
<0.001 0.003(+)
247 97.2
95.1
Normal (.20)
67.4
44.8
43 90.7
81.4
79.1
l2=14.67 X2=9.04
Abnormal (>20)
Risk group 9
41 97.7
94.9
79.3
NS
...
...
...
97.6
97.6
All absent
81.4
70.1
92.5
90.6
254 96.1
.1 present
Risk group 4
127 96.8
78.7
95.3
95.3
86.4
NS
All absent
80.8
65.2
168 95.8
91.7
88.7
.1 present
RAP, right atrial pressure; PA, pulmonary artery; LV, left ventricle; AV, atrioventricular; NYHA, New York Heart Association; RA, right
atrium.
n refers to number of patients with the factor. For some factors, a few patients are missing information; sum of the numbers in subgroups
will not equal overall total. Survival estimates are obtained from Kaplan-Meier survival curves for selected time points after time zero.
Pu refers to univariate probability value from log-rank test with NS (nonsignificant) if Pu >0.05. Below each significant finding is the
associated x2 statistic subscripted with its degrees of freedom. Comparison of the subgroups is an overall comparison of the survival
experience not restricted to the selected time points presented. Pc refers to univariate probability value from Cox regression model assessing
association between survival and factor considered as a continuous variable. Direction of association is indicated in parentheses. Minus (-)
sign indicates that survival is improving (the hazard is decreasing) as the level of the factor increases or if the condition is present. Pm (with
RAP) is the significance of that factor in a multivariate model arrived at after backward elimination of the nonsignificant (p>0.05, NS)
variables in a model that initially included the immediate postoperative right atrial pressure. Pm (w/o RAP) is like Pm (with RAP) except
that only preoperative factors were considered.
AV valve abnormal if moderate or severe insufficiency preoperatively or necessity for valve repair or replacement prior to or concurrent
with Fontan procedure.
All factors except postop RA pressure are based on preoperative assessment except where indicated.
*Risk group 9 consists of the presence of any (>1) of the nine conditions (criteria of Choussaut's et all2).
tRisk group 4 consists of the presence of any (> 1) of the four conditions.
NYHA classification are eliminated from the model,
but univentricular heart, complex single ventricle, calendar year of operation, increased pulmonary artery
pressure, AV valve dysfunction, and increased postoperative right atrial pressure remain. The results of this
modeling effort are illustrated in Figure 3.
Analysis of Survival From 30 Days Postoperatively
Table 3 presents the candidate predictors of survival when time 0 for the analysis is the date of
Fontan operation plus 30 days (i.e., the cohort of
100
80
268
110
0.
2t
233
_
60 -
-
4
44
40 I20
n = 352
TABLE 4. Length of Follow-up for Patients Lost to Follow-up Less
Than 5 Years Postoperatively
Patients (No.)
Length of follow-up (yr)
2
<1
ito <2
0
3
2 to <3
4
3 to <4
4 to <5
5
n=14.
1
0
0
1
2
3
l
l
4
5
6
7
l
l
8
9
10
Years
Initiation of Fontan operation
FIGURE 1. Kaplan-Meier survival curve following initiation of
Fontan procedure, time 0 (n=352). Expected survival is based on
generalpopulation of West/North central region of United States
for persons of like age, ser, and year of bith and includes death
from all causes. Numbers of patients alive and in follow-up
cohort at 1, 5, and 10 years after Fontan are shown.
Driscoll et al Follow-up After Fontan Operation
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
TABLE 5. Mode of Death for 122 Patients*
Patients (No.)
Cause of death
91
Cardiovascular related
73
Ventricular failure
29
Renal failure or insufficiency
19
Respiratory failure or insufficiency
14
Sepsis
11
Sudden and unexpected
13
Arrhythmia
7
At cardiac reoperation
8
Bleeding
5
Hepatic failure or insufficiency
7
Brain death
3
Pulmonary embolus
Cardiac tamponade
1
1
Accidental
Unknown
8
*Per last attending physician or death certificate. More than one
cause may be listed for each patient.
early operative survivors). It is apparent that fewer
variables are statistically significant in a univariate
fashion for this analysis than for the analysis from the
date of Fontan operation.
Again, multivariate models were constructed with
and without postoperative right atrial pressure. If
postoperative right atrial pressure is excluded from
the model, early calendar year of operation, heterotaxia, AV valve dysfunction, abnormal pulmonary
artery architecture, and higher NYHA class are
significantly associated with mortality.
If postoperative right atrial pressure is included
in the model, early calendar year of operation,
Risk Patient,
score
.0
0
100
0
80
1-2
190
L
3-5
101
6-7
19
11
40
.................
20
" ~~~~~.. ... -
:
.60
0
1
2
3
4
5
2
CO
40
6
7
8
9
10
Years
t
Initation of Fontan operation
FIGURE 3. Kaplan-Meier survival curves displaying multivariate findings for the end point of death after initiation of
Fontan operation. The following risk factors were assigned a
risk score of 1: early year of operation (1972-1982); increased
pulmonary artery pressure (>15 mm Hg); univentricular heart
or complexforms ofsingle ventricle; and atrioventricular valve
insufficiency, repair, or replacement. Increased right atrial
pressure (>20 mm Hg) immediately postoperatively was assigned a risk score of 3.
increased pulmonary artery pressure, AV valve
dysfunction, higher NYHA class, and increased
right atrial pressure are significantly associated with
mortality (Figure 4).
Morbidity
Reoperation. Reoperation was necessary for 103 of
the 352 patients (29%). These 103 patients had a total
of 194 cardiac procedures during 158 operations (Table
6). The most common reoperations included pacemaker insertion or replacement, exploration for control
100
Risk Patient,
score no.
a
5
153
3-4
139
1
"---L
at 601
_
1
.,t
-
0
no.
1-2
6
'.
80
80
no.
100
60
Risk Patient,
score
475
2 40
0-1
95
2-3
150
4-6
22
-
C)
F
._.._.."_
...........
5-6
....... ..........
20
13
20
0
1
2
3
5
4
6
0
7
8
9
10
1
Initiation of Fontan operation
Years
FIGURE 2. Kaplan-Meier survival curves displaying multivariate findings for the end point of death after initiation of
Fontan operation. Postoperative right atrial pressure was not
considered in this modeL Each of the following risk factors
was assigned a risk score of 1: early year of operation
(1972-1982); increased pulmonary artery pressure (>15
mm Hg); univentricular heart or complex forms of single
ventricle; atrioventricular valve insufficiency, repair, or replacement; heterotaxia; preoperative New York Heart Association classification III or IV; and young age at operation (<4
years).
l
l
n
l-
n
1
1
2
3
5
4
6
7
8
9
10
Years
30 days after Fontan operation
FIGURE 4. Kaplan-Meier survival curves displaying multivariate findings for end point of death from 30 days after
Fontan operation. Postoperative right atrial pressure was
considered in this model but was not a significant variable.
Each of the following risk factors was assigned a risk score of
1: early year of operation (1972-1982), increased pulmonary
artery pressure (> 15 mm Hg), and preoperative New York
Heart Association classification III or IVand increased right
atrial pressure (21-22 mm Hg) immediately postoperatively.
Atrioventricular valve insufficiency, repair, or replacement and
increased right atrial pressure (>23 mm Hg) immediately
postoperatively were assigned a risk score of 2
476
Circulation Vol 85, No 2 February 1992
TABLE 6. Breakdown of 194 Cardiac Procedures Performed During 158 Reoperations in 103 of 352 Study Patients
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
Patients
Procedure
Procedures (No.)
% of 352
n
Pacemaker insertion or replacement/revision
40
29
8
Exploration for control of bleeding
31
28
8
Revise right atrium-pulmonary artery connection
18
17
5
Relieve subaortic obstruction
12
10
3
Close residual atrial septal defect
9
8
2
Repair dehisced atrioventricular valve patch
10
8
2
Replace atrioventricular valve
6
6
2
Take down Fontan
5
5
1
Right atrium conduit thrombectomy
6
6
2
Pericardiectomy or pericardiostomy
5
5
1
Pulmonary arterioplasty
3
3
<1
Repair chyle leak
2
2
<1
Explant inferior vena cava prosthetic valve
2
2
<1
Cardiac transplantation
1
1
<1
Aortic valve replacement
1
1
<1
Miscellaneous
43
29
8
n refers to the number of patients having the procedure on one or more occasion. A patient may have had a specific
procedure on more than one occasion.
of postoperative bleeding, and revision of a right atrialto-pulmonary artery connection. The Fontan operation
was taken down in five patients, and cardiac transplantation was performed in one patient.
Risk Factors for the Combined Event: Death
or Reoperation
The risk of reoperation or death was assessed from
the time of the Fontan operation and separately for
patients who were alive 30 days postoperatively,
considering day 30 as time 0. Reoperation for bleeding within 48 hours of the Fontan operation was
excluded from this analysis.
The set of potential predictor variables considered
is listed in Table 7. Excluding postoperative right
atrial pressure from the multivariate model, the
following variables were significant when considered
multivariately: univentricular heart, complex single
ventricle, early calendar year of operation, heterotaxia, young age at operation, increased pulmonary
artery pressure, and higher NYHA classification. The
multivariate findings for this combined death or
reoperation end point are illustrated in Figure 5.
When postoperative right atrial pressure is included in the model, heterotaxia and NYHA class
exit the model and increased postoperative right
atrial pressure and absence of sinus rhythm enter the
model (Figure 6).
Survival to death or initial reoperation for patients
who survived at least 30 days postoperatively is
displayed in Table 8. Excluding postoperative right
atrial pressure, the multivariate model includes the
following significant variables: univentricular heart,
complex forms of single ventricle, early calendar year
of operation, and AV valve dysfunction.
When postoperative right atrial pressure is included in the model, only calendar year of operation
and complex single ventricle in addition to increased
right atrial pressure are significant predictors of the
combined event variable of death or reoperation.
The effect on survival or reoperation of one or more
of these mathematically weighted predictor variables
is illustrated in Figure 7.
Hospitalization
For the 215 survivors who returned a questionnaire,
118 have been hospitalized for cardiac reasons subsequent to their Fontan operation (Tables 9 and 10). The
reasons for hospitalization are listed in Table 10. Except for cardiac reoperation, arrhythmias were the
most frequent single reason for hospitalization.
Arrhythmias
More than 20% of the surviving cohort had symptoms of tachycardia or palpitation (Table 11). Postoperatively, atrial flutter or fibrillation was reported
on the follow-up questionnaire by 12% and 19% of
the survivors at 5 years and currently, respectively.
Currently, at least one antiarrhythmic medication
other than digitalis was being taken by 40 (19%) of
the survivors, and eight (4%) took two. Twenty-two
patients (10%) had mechanical pacemakers. Fiftytwo patients had 74 hospitalizations for arrhythmias,
and 22 patients had 25 hospitalizations for pacemaker-related problems.
For the entire cohort of 352 patients, 73 (21%) had
or had had atrial flutter or fibrillation; 40 (11%) had
premature ventricular contractions, and 44 (13%)
required a pacemaker.
Protein-Losing Enteropathy or Hypoproteinemia
or Hypoalbuminemia
A patient was considered to have protein-losing
enteropathy or hypoproteinemia or hypoalbumin-
Driscoll et al Follow-up After Fontan Operation
477
TABLE 7. Univariate and Multivariate Assessment- of Factors Associated With Survival for End Point "Death Any Cause or First
Reoperation" With Time Zero the Initiation of Fontan Operation (Excluding Reoperations for Bleeding Within 2 Days of Fontan Operation)
Survival (%)
n 30 Days 6 Months 1 Year 5 Years 10 Years
352 78.4
69.9
67.0
55.7
46.4
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
Factor
Overall
Sex
Male
Female
Lesion
Tricuspid atresia
Univentricular heart
Complex single ventricle
Calendar year of operation
1970-1975
1976-1980
1981-1984
Heterotaxia
Absent
Present
Asplenia
Absent
Present
Polysplenia
Absent
Present
Age at operation (yr)
4-15
<4 or .16*
<4
4-15
.16
Preoperative sinus rhythm
Present
Absent*
Pm (with
Pm (w/o
RAP)
RAP)
...
...
...
NS
NS
NS
Pu
Pc
215
137
77.7
79.6
68.3
72.3
65.1
70.1
53.2
59.7
47.4
44.6
125
114
113
85.6
79.8
69.0
78.4
71.0
71.9
53.0
40.5
59.2
45.2
34.7
X 2=21.98
59.3
77.6
66.6
55.8
0.008
<0.001
8
127
217
50.0
73.2
82.5
37.5
66.1
73.3
37.5
63.0
70.5
12.5
53.3
58.7
12.5
43.1
53.4
0.010
0.011(-) <0.001(-) <0.001(-)
X 2=9.13
X 2=6.49
325
27
81.8
37.0
72.9
33.3
70.4
25.9
59.1
14.8
49.0
14.8
<0.001
...
NS
X2=36.16
341
11
80.4
18.2
71.8
9.1
68.9
9.1
57.2
9.1
47.7
9.1
...
...
X 2=30.54
336
16
79.8
50.0
70.8
50.0
68.4
37.5
57.5
18.7
47.7
18.7
X 2= 12.41
240
112
33
240
79
77.1
81.2
66.7
77.1
87.3
69.6
70.5
51.5
69.6
78.5
66.6
67.9
51.5
66.6
74.7
55.1
56.9
39.4
55.1
64.4
46.6
46.4
25.9
46.6
54.4
320
32
80.0
62.5
71.6
53.1
69.0
46.9
58.2
30.5
47.9
30.5
0.006
X 2=7.55
...
0.020
NS
297
55
81.5
61.8
73.1
52.7
70.7
47.3
60.0
32.2
50.0
27.1
<0.001
...
NS(-)
NS
154
177
83.1
74.0
75.3
63.8
74.0
61.0
58.0
52.4
49.4
45.3
NS
0.044(+)
0.014
0.006
154
5
77.3
40.0
69.5
40.0
66.8
40.0
54.7
20.0
47.6
0.0
NS
NS
NS
NS
<0.001
<0.001
<0.001
NS
...
.
..
.0..
0.007
0.004
0.002
...
...
...
0.017 0.022(-) 0.011(-)
X2=8.20 X 2=5.26
...
0.018(-)
Systemic venous drainage
Normal
Abnormal*
Mean PA pressure (mm Hg)
Normal (.15)
Abnormal (>15)*t
Pulmonary arteriolar resistance (U
Normal (<4)
Abnormal (.4)*t
Mean PA size (% normal)
Normal (.60)
Abnormal (<60)*
LV ejection fraction (%)
Normal (>60)
Abnormal (<60)*t
AV valve dysfunction
Normal
Abnormal*t
PA architecture
Normal
Abnormal*
2=16.92
X2=4.05
im2)
NS
-55
2
NS
41
71
87.8
87.3
82.9
77.5
80.5
71.8
70.5
58.9
64.5
55.4
NS
NS
319
33
79.6
66.7
72.1
48.5
69.9
39.4
58.0
33.3
48.0
33.3
0.005
x 1=8.02
...
281
71
78.3
78.9
70.8
66.2
67.6
64.8
56.4
53.2
47.3
42.5
NS
478
Circulation Vol 85, No 2 February 1992
TABLE 7. Continued.
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
Survival (%)
Pm (with Pm (w/o
n 30 Days 6 Months 1 Year 5 Years 10 Years
Pc
Pu
RAP)
RAP)
Factor
LV end-diastolic pressure (mm Hg)
121 81.0
72.7
71.0 60.7
49.9
NS
NS
NS
NS
Normal (.12)
187 75.4
67.9
65.2
53.3
45.2
Abnormal (>12)
NYHA functional class
I
5 100.0
0.045
0.010(+)
NS
0.026(+)
100.0 100.0 100.0 100.0
70.9
59.7
49.0 V 2=8.03 X 2=6.71
II
182 80.8
73.1
61.1
III
152 73.7
64.4
48.3
41.7
IV
7 85.7
57.1
57.1
28.6
71.4
Immediate postop RA pressure
(mm Hg)
74.9 63.1
53.4 <0.001(+) <0.001(+) <0.001(+)
...
Normal (.20)
271 86.7
78.2
43.9
66 56.1
45.4
34.8
25.4 X 2=27.95 X 2=28.67
Abnormal (>20)
Risk group 9
46 82.6
78.3
78.3
73.6
57.0
0.045
...
All absent
68.6
65.4
53.1
45.3 V2=4.01
306 77.8
.1 present
Risk group 4
143 83.2
76.2
All absent
74.8 61.2
49.5
NS
...
.1 present
209 75.1
65.6
61.7
45.3
52.0
RAP, right atrial pressure; PA, pulmonary artery; LV, left ventricle; AV, atrioventricular; NYHA, New York Heart Association; RA, right
atrium.
n refers to number of patients with the factor. For some factors, a few patients are missing information; sum of the numbers in subgroups
will not equal overall total. Survival estimates are obtained from Kaplan-Meier survival curves for selected time points after time zero.
Pu refers to univariate probability value from log-rank test with NS (nonsignificant) if Pu >0.05. Below each significant finding is the
associated x2 statistic subscripted with its degrees of freedom. Comparison of the subgroups is an overall comparison of the survival
experience not restricted to the selected time points presented. Pc refers to univariate probability value from Cox regression model assessing
association between survival and factor considered as a continuous variable. Direction of association is indicated in parentheses. Minus (-)
sign indicates that survival is improving (the hazard is decreasing) as the level of the factor increases or if the condition is present. Pm (with
RAP) is the significance of that factor in a multivariate model arrived at after backward elimination of the nonsignificant (p>0.05, NS)
variables in a model that initially included the immediate postoperative right atrial pressure. Pm (w/o RAP) is like Pm (with RAP) except
that only preoperative factors were considered.
AV valve abnormal if moderate or severe insufficiency preoperatively or necessity for valve repair or replacement prior to or concurrent
with Fontan procedure.
All factors except postop RA pressure are based on preoperative assessment except where indicated.
*Risk group 9 consists of the presence of any (.1) of the nine conditions (criteria of Choussaut's et al12).
tRisk group 4 consists of the presence of any (>1) of the four conditions.
Risk Patient,
emia when: 1) there was historical information in the
Mayo Clinic medical record substantiating that fact,
2) there was a postoperative serum protein level less
than 6.3 g/dl, 3) the postoperative serum albumin
level was less than 3.1 g/dl, or 4) a surviving patient
indicated on the medical questionnaire that he or she
had hypoproteinemia or protein-losing enteropathy 5
years postoperatively or currently.
Of the entire cohort of 352 patients, at some point in
the course of follow-up, 36 (10%) met the criteria for
hypoproteinemia or protein-losing enteropathy. Considering only the 215 patients who survived and returned a medical questionnaire for this study, 15 (7%)
had hypoproteinemia or protein-losing enteropathy.
The results of Kaplan-Meier analysis for hypoproteinemia or protein-losing enteropathy are displayed
in Tables 12 and 13. For Table 12, death was not
considered an event of interest. The variables significantly related, in univariate fashion, to the development of hypoproteinemia or protein-losing enteropathy include heterotaxia, polysplenia, anomalies of
systemic venous drainage, increased pulmonary arte-
score
0
100
no.
5
80
0-0
_: 60
1-2
160
Co
.it
3-4
123
25
22
2 40
20
0
0
t
1
2
3
5
4
6
7
8
9
10
Years
Initiaton of Fontan operation
FIGURE 5. Kaplan-Meier survival curves displaying multi-
vaniate findings for end point of death or reoperation from time
of Fontan operation. Postoperative right atrial pressure was
not considered in this model. Each of the following risk factors
was assigned a risk score of 1: early year of operation
(1972-1982), increased pulmonary artery pressure (>15
complex single ventricle,
preoperative New York Heart Association classification III or
IVt and young age at operation (<4 years). The presence of
heterotaxia was assigned a risk score of 5.
mm Hg), univentricular heart or
Driscoll et al Follow-up After Fontan Operation
Risk Patient,
score no.
6
0
100
1
80
-60
-i
2-5 265
.it
2
40
c)
29
SI
.
.
-i
20
0
0
1
1
2
3
l
l
4
5
6
7
8
9
10
6-7 16
Years
Initiation of Fontan operation
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
FIGURE 6. Kaplan-Meier survival curves disp Playing multivariate findings for endpoint ofdeath or reoperatiion from time
of initiation of Fontan operation. Each of the following sk
factors was assigned a risk score of 1: increase d right atrial
pressure (>20 mm Hg) immediately postoperati ively, univentricular heart or complex single ventricle, young iage at operation (<4 years), increased pulmonary artery pre. ssure (15-25
mm Hg), and preoperative absence of sinus rhyt thm. Each of
the following risk factors was assigned a risk sco re of 2: early
year of operation (1972-1982) and an increase d pulmonary
artery pressure (>25 mm Hg).
riolar resistance, and increased left venti ricular enddiastolic pressure. The highest occurrenc ,e rates for
hypoproteinemia or protein-losing entero pathy were
among the five patients with pulmonar' V arteriolar
resistance of 4 U. m2 or greater (67%) and the 27
patients with heterotaxia (36%). Less ti han a 30%
chance of survival free of hypoproteinuria or proteinlosing enteropathy at 5 years postoperatilvely (Table
13) was noted among the 27 patients withl heterotaxia,
the 11 patients with asplenia, the 16 pa itients with
polysplenia, the five patients with increa sed pulmonary arteriolar resistance, the eight patient s with early
calendar year of operation, and the 71 pc atients with
left ventricular ejection fraction less than 60%.
Other Medical Events
Of the 215 5-year survivors who returr ied a questionnaire, two patients reported the p resence of
cirrhosis, four had strokes, one had a bra tin abscess,
and seven had infective endocarditis.
Functional Status of Survivors
General status. At 5 years postoperativ ely, 89 and
83 of the survivors who returned a qu estionnaire
were in NYHA class I and II, respectively (Table 14).
Both in absolute numbers and as a percen tage of the
specific cohort, this represents a greater number of
patients in NYHA class I postoperativel3y than preoperatively. However, if patients in NY]HA class I
and II are combined, 187 of the original cc hort of 352
patients were in one of these two claLssifications
preoperatively, and 172 of the 215 survrivors were
similarly classified 5 years postoperatively z. Although
the relative percent was greater post operatively
(80%) than preoperatively (53%), the abs olute num-
479
ber of patients in class I and II was slightly less
(172) than preoperatively (187).
postoperatively
The information in Table 14 illustrates that 122
patients (35%) were alive postoperatively with a
better functional classification than preoperatively.
Fifty-eight patients (17%) were alive and in the same
functional classification 5 years postoperatively, but
126 patients (36%) had died within the first 5 years or
were in a worse functional classification 5 years
postoperatively. Of the remaining, 46 patients (13%)
were known to be alive at 5 years but lacked NYHA
functional class information either preoperatively,
postoperatively, or both.
We arbitrarily defined an "excellent" 5-year postoperative status for survivors as a patient fulfilling all
the following criteria: alive in NYHA class I, no
cardiac symptoms, no cardiac medications (other
than digitalis), and the ability to do as much or more
physical activity as one's peers (or the ability to do
heavy housework). Using these criteria, 37 patients
were defined as excellent 5 years postoperatively.
Similarly, "poor" status was defined as any one of the
following: NYHA class IV, the presence of proteinlosing enteropathy, hypoproteinemia, cirrhosis, the
inability to climb one flight of stairs or to do any
exercise, unemployment that resulted from the Fontan operation, or incapacity for any type of employment. By these criteria, 29 patients 5 years postoperatively can be defined as doing poorly.
By exclusion from an excellent or poor classification, 149 patients were considered to be in an intermediate category. In addition to these 149 patients,
we placed into the intermediate group 29 patients
who were alive 5 years after the Fontan operation but
for whom no follow-up questionnaire was available.
Tables 15, 16, and 17 contain the data and results of
univariate and multivariate analyses of the predictors
of excellent, poor, and intermediate outcome. In
Table 15, all 352 patients are analyzed, and death as
well as excellent, intermediate, and poor outcomes at
5 years postoperatively are considered as individual
outcome categories. Of these 352 patients, 10.5%
had an excellent outcome; 50.6%, intermediate;
8.2%, poor; and 30.7% were dead. The preoperative
factors significantly associated with outcome in a
multivariate fashion included complex forms of single
ventricle, early calendar year of operation, heterotaxia, increased pulmonary pressure, AV valve dysfunction, and higher NYHA classification.
In Table 16, only 5-year survivors are considered for
analysis. The end points are for 5 years postoperatively.
For these 244 patients, the factors significantly associated with outcome in a multivariate logistic model for
predicting outcome other than death included complex
single ventricle, early calendar year of operation, heterotaxia, abnormal systemic venous drainage, and abnormalities of pulmonary artery architecture.
In Table 17, only current survivors are analyzed,
and the end points are for current status. For these
analyses, the factors significantly associated with outcome in a multivariate model included early calendar
480
Circulation Vol 85, No 2 February 1992
TABLE 8. Univariate and Multivariate Assessment of Factors Associated With Survival for End Point "Death Any Cause or First
Reoperation" With Time Zero 30 Days Post-Fontan Operation (Excludes Those Who Died and Reoperations Within 30 Days of
Fontan Operation)
Survival (%)
Pm (with Pm (w/o
Factor
n 30 Days 6 Months 1 Year 5 Years 10 Years
Pu
Pc
RAP)
RAP)
Overall
295 94.9
88.5
84.4
69.2
57.7
...
...
...
Sex
176 95.5
Male
NS
88.1
82.9
66.0
58.4
NS
NS
119 94.1
Female
89.1
86.6
73.8
56.2
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
Lesion
Tricuspid atresia
Univentricular heart
Complex single ventricle
Calendar year of operation
1970-1975
1976-1980
1981-1984
Heterotaxia
Absent
Present
Asplenia
Absent
Present
Polysplenia
Absent
Present
Age at operation (yr)
4-15
<4 or .16*
<4
4-15
.16
113
98
84
96.5
95.9
91.7
90.3
89.8
84.5
89.4
83.5
78.6
81.3
65.7
56.9
65.5
57.1
50.2
0.006
X 2= 10.42
5
101
189
80.0
94.1
95.8
60.0
90.1
88.4
60.0
85.1
84.6
20.0
70.0
70.0
0.0
57.1
64.4
0.002
X2=12.69
282
13
95.0
92.3
88.3
92.3
84.7
76.9
70.2
46.2
58.3
46.2
NS
285
10
94.7
100.0
88.1
100.0
84.5
80.0
70.2
40.0
58.3
40.0
NS
...
...
198
97
22
198
75
94.9
94.8
95.5
94.9
94.7
90.4
84.5
77.3
90.4
86.7
86.3
80.4
77.3
86.3
81.3
70.9
65.6
59.1
70.9
67.7
61.0
51.3
38.8
61.0
54.5
NS
...
...
NS
NS
NS
NS
272
23
94.9
95.7
89.0
82.6
85.3
73.9
71.0
46.6
58.5
46.6
NS
...
NS
NS
257
38
94.9
94.7
88.7
86.8
85.2
78.9
71.3
54.6
59.5
46.1
NS
NS
NS
137
141
95.6
93.6
89.1
86.5
86.8
82.3
65.8
70.8
57.2
59.6
NS
NS
NS
NS
130
3
...
NS
...
...
49
2
...
NS
.
NS
.
...
.
.
.
NS
0.006
0.026
0.002
0.001(-)
<0.001
NS
NS
292
3
Preoperative sinus rhythm
Present
Absent*
Systemic venous drainage
Normal
Abnormal*
Mean PA pressure (mm Hg)
Normal (c15)
Abnormal (>15)*t
Pulmonary arteriolar resistance (U. mi2)
Normal (<4)
Abnormal (.4)*t
Mean PA size (% normal)
Normal (.60)
Abnormal (<60)*
LV ejection fraction (%)
Normal (.60)
Abnormal (<60)*t
AV valve dysfunction
Normal
39
63
100.0
90.5
94.9
87.3
89.7
81.0
70.9
66.4
67.7
62.5
NS
NS
273
22
95.6
86.4
89.7
72.7
86.4
59.1
70.7
50.0
58.6
50.0
0.044
X 2=4.06
...
NS
0.041
Abnormal*t
PA architecture
Normal
Abnormal*
234
61
94.9
95.1
90.2
82.0
85.4
80.3
70.2
65.2
59.0
52.1
NS
...
NS
NS
Driscoll et al Follow-up After Fontan Operation
481
TABLE 8. Continued.
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
Survival (%)
Pm (with Pm (w/o
n 30 Days 6 Months 1 Year 5 Years 10 Years
Pu
Pc
RAP)
RAP)
Factor
LV end-diastolic pressure (mm Hg)
NS
108 95.4
89.8
87.9
73.5
59.3
NS
NS
NS
Normal (.12)
88.7
84.7
68.5
58.4
Abnormal (>12)*
150 94.7
NYHA functional class
I
5 100.0
100.0
100.0 100.0
100.0
NS
0.017
NS
NS
II
153 96.1
90.2
86.9
73.6
59.6
X f=5.65
III
125 93.6
86.4
61.8
54.6
81.5
IV
6 83.3
83.3
66.7
66.7
33.3
Immediate postop RA pressure
(mm Hg)
247 95.5
0.010
0.018
Normal (<20)
89.9
85.4
72.1
60.5
0.041
43 90.7
Abnormal (>20)
79.1
76.7
55.8
44.0 X 2=4.19 X2=6.57
Risk group 9
41 97.6
95.1
84.7
...
All absent
95.1
69.4
0.031
...
...
.1 present
254 94.5
87.4
82.7
66.7
56.4 X12=4.64
Risk group 4
127 96.1
90.6
All absent
88.1
69.7
57.4
NS
86.9
.1 present
168 94.0
81.5
68.9
58.6
RAP, right atrial pressure; PA, pulmonary artery; LV, left ventricle; AV, atrioventricular; NYHA, New York Heart Association; RA, right
atrium.
n refers to number of patients with the factor. For some factors, a few patients are missing information; sum of the numbers in subgroups
will not equal overall total. Survival estimates are obtained from Kaplan-Meier survival curves for selected time points after time zero.
Pu refers to univariate probability value from log-rank test with NS (nonsignificant) if Pu >0.05. Below each significant finding is the
associated x2 statistic subscripted with its degrees of freedom. Comparison of the subgroups is an overall comparison of the survival
experience not restricted to the selected time points presented. Pc refers to univariate probability value from Cox regression model assessing
association between survival and factor considered as a continuous variable. Direction of association is indicated in parentheses. Minus (-)
sign indicates that survival is improving (the hazard is decreasing) as the level of the factor increases or if the condition is present. Pm (with
RAP) is the significance of that factor in a multivariate model arrived at after backward elimination of the nonsignificant (p>0.05, NS)
variables in a model that initially included the immediate postoperative right atrial pressure. Pm (w/o RAP) is like Pm (with RAP) except
that only preoperative factors were considered.
AV valve abnormal if moderate or severe insufficiency preoperatively or necessity for valve repair or replacement prior to or concurrent
with Fontan procedure.
All factors except postop RA pressure are based on preoperative assessment except where indicated.
*Risk group 9 consists of the presence of any (.1) of the nine conditions (criteria of Choussaut's et a112).
tRisk group 4 consists of the presence of any (>1) of the four conditions.
of operation, age <4 years or >16 years at the
time of operation, and abnormalities of systemic
venous drainage.
no.
120
144
Exercise Capability
Most of the surviving patients can perform light
(dust, wash dishes, wipe counters, empty wastebaskets) or moderate (vacuum, strip and make a bed,
clean windows) housework (Table 18). Half of the
patients can do heavy housework (i.e., move heavy
furniture).
Similarly, most patients can walk 400 yards or
climb one flight of stairs but only half can run 100
9
TABLE 9. Number of Hospitalizations Since Fontan Operation for
215 Surviving Patients Who Returned a Questionnaire
year
Risk
score
0
1-2
Patient,
n
3
Patients
0
t
1
2
3
5
4
6
7
8
9
10
Years
30 days after Fontan operation
FIGURE 7. Kaplan-Meier survival curves displaying multivariate findings for the end point of death or reoperation from
30 days after Fontan operation. Each of the following factors
was assigned a risk factor of 1: increased postoperative right
atrialpressure (>20 mm Hg), later calendaryear of operation
(1983-1984), and complex forms of single ventricle.
Hospitalizations (No.)
0
1
2
3
4
5
.6
n
97
46
18
12
16
3
23
S
45
21
8
6
7
1
11
482
Circulation Vol 85, No 2 February 1992
TABLE 10. Indications for Hospitalizations After Fontan Procedure for 215 Surviving Patients Who Returned a Questionnaire
Patients hospitalized (No.)*
Indication
62
Cardiac operation
57
Other
52
Arrhythmia
22
Pacemaker insertion or replacement
14
Heart failure
10
Abdominal swelling
9
Leg edema
7
Endocarditis
6
Protein-losing enteropathy
4
Hypoproteinemia
4
Stroke
1
Liver problems
1
Brain abscess
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
*A patient may have provided multiple indications for
hospitalization.
one
yards. Nausea when exercising was reported by 18
(8%) and 21 (10%) of surviving patients 5 years
postoperatively and currently, respectively.
Slightly fewer than half of the surviving patients
consider themselves capable of doing as much or
more physical activity than their peers (Table 19).
Five years postoperatively, 18 patients (8%) and
currently 13 patients (6%) describe themselves as
able to do less than one quarter as much exercise as
their peers.
For 140 students, 30 (21%) participate fully, 46
(33%) less than fully, and 36 (26%) do not participate in regular school physical education. Only seven
of 28 patients who have access to special physical
education classes participate in them.
TABLE 11. Arrhythmias in 215 Survivors
Results from follow-up questionnaire
Syncope
Rapid heart rate (tachycardia)
Slow heart rate (bradycardia)
Palpitations
Atrial flutter or fibrillation
Premature ventricular contractions
Ventricular tachycardia
Pacemaker
Number of antiarrhythmic medications*
0
1
2
5 Years
postop
Patients
No. %
18
8
44 20
17
8
51 24
26 12
13
6
4
9
t
t
179
31
5
83
14
2
Currently
Patients
No. %
17
8
45 21
15
60
41
15
13
22
7
28
19
7
6
10
167
40
8
78
19
4
*Excluding digitalis.
tPresence of a pacemaker was only asked for patient's current
status.
Education and Employment
Education. School is attended by 140 of the 215
survivors. Classes are attended full-time by 125 of the
140 and part-time by 15. Six of the 15 part-time
students are part-time because of their heart problem. During a 6-month school period, 92 students
(66%) missed no school days, 23 (16%) missed 1-5
days, and six (4%) missed more than 30 days because
of cardiac problems (Table 20).
For 73 people who have completed formal education, 10 stopped their effort to obtain formal education because of cardiac problems. At least some
post-high school education was obtained by 44% of
these 73 subjects (Table 21).
Employment. Eighty patients were gainfully employed: 52 full-time and 28 part-time. Because of
cardiac problems, four of the 28 part-time employees
were not employed full-time. There were 50 unemployed patients, 11 of whom were unemployed because
of their cardiac problem and two because of their
Fontan operation. There were very few days lost from
work among the 80 employed individuals (Table 22).
The patient's own perception of employment capability was obtained from 149 patients (Table 23). Fulltime employment was thought to be possible by 72%
and no employment was thought to be possible by 3%.
Medications
At 5 years postoperatively and currently, 87 (40%)
and 94 (44%) of the surviving cohort took no cardiac
medication, respectively (Table 24). For those taking
medication, digoxin and furosemide were the most
common cardiac medications. Slightly more than
10% of the patients were taking one or more of the
following antiarrhythmic agents: quinidine, procainamide, verapamil, propranolol, and amiodarone.
Symptoms
For the surviving patients, 56 (26%) and 60 (28%)
were free of significant cardiac symptoms at 5 years
postoperatively and currently, respectively (Table 25).
For those with symptoms, the most commonly reported
symptoms were easy fatigue (36% and 37%), shortness
of breath (28% and 30%), palpitations (24% and 28%),
and rapid heart rate (23% and 27%). Nausea or
vomiting with exercise was reported by 8-10% of the
patients at 5 years postoperatively and currently.
Fertility
One woman had two and another woman had one
pregnancy prior to the Fontan operation. All three
terminated in abortion, one spontaneously and two
therapeutically.
There were seven pregnancies among six women
after the Fontan operation. Of the seven pregnancies, there were three spontaneous abortions, three
therapeutic abortions, and one liveborn child. Although the numbers are small, the spontaneous
abortion rate before the Fontan operation was 33%
and after the Fontan operation was 43%.
Driscoll et al Follow-up After Fontan Operation
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
TABLE 12. Univariate Assessment of Factors Associated With End Point of Protein-Losing Enteropathy/Hypoproteinemia or
Hypoalbuminemia With Time Zero the Initiation of Fontan Operation
Absence of PLE (%)
n
30 Days
6 Months
1 Year 5 Years
Factor
10 Years
Pu
Overall
352
100.0
97.9
97.1
89.5
85.3
Sex
Male
215
100.0
99.4
98.8
91.0
85.1
NS
Female
137
100.0
95.6
94.6
87.1
85.3
Lesion
125
Tricuspid atresia
100.0
99.0
99.0
94.2
89.3
NS
114
97.9
Univentricular heart
100.0
97.9
86.6
83.0
100.0
96.2
113
93.5
86.2
82.4
Complex single ventricle
Calendar year of operation
1970-1975
8
100.0
100.0
100.0
100.0
66.7
NS
1976-1980
127
100.0
97.9
97.9
87.0
82.4
1981-1984
217
100.0
97.8
96.7
90.6
88.3
Heterotaxia
Absent
98.2
325
100.0
97.8
86.2
0.006
90.5
Present
27
100.0
91.7
82.5
64.2
64.2
X l=7.55
Asplenia
Absent
Present
Polysplenia
Absent
Present
Age at operation (yr)
4-15
<4 or 216*
<4
4-15
.16
Preoperative sinus rhythm
Present
Absent*
Systemic venous drainage
Normal
Abnormal*
Mean PA pressure (mm Hg)
Normal (c15)
Abnormal (>15)*t
Pulmonary arteriolar resistance (U mi2)
Normal (<4)
Abnormal (>4)*t
Mean PA size (% normal)
Normal (.60)
Abnormal (<60)*
LV ejection fraction (%)
Normal (.60)
Abnormal (<60)*t
AV valve dysfunction
Normal
Abnormal*t
PA architecture
Normal
Abnormal
483
Pc
...
NS
341
11
100.0
100.0
97.9
100.0
97.1
100.0
89.4
100.0
85.2
100.0
336
16
100.0
100.0
98.2
90.0
97.8
78.8
90.6
56.3
86.3
56.3
240
112
33
240
79
100.0
100.0
100.0
100.0
100.0
99.0
95.7
95.2
99.0
95.8
97.9
95.7
95.2
97.9
95.8
90.6
87.0
89.6
90.6
86.4
84.5
87.0
89.6
84.5
86.4
320
32
100.0
100.0
97.7
100.0
96.9
100.0
89.5
88.2
85.6
80.9
NS
297
55
100.0
100.0
98.8
91.7
98.4
88.6
91.4
76.2
87.7
68.3
0.001
X 2 10.32
154
177
100.0
100.0
98.5
97.0
98.5
96.2
91.9
86.4
89.6
81.2
NS
NS
154
5
100.0
100.0
96.8
100.0
96.8
100.0
89.5
33.3
85.8
33.3
<0.001
NS
55
2
NS
<0.001
xv= 11.75
NS
NS
...
NS
Xv2=11.03
...
NS
NS
41
68
100.0
100.0
97.4
96.7
97.4
94.9
88.9
87.3
82.1
72.8
NS
319
33
100.0
100.0
98.1
94.7
97.3
94.7
90.0
81.6
85.6
81.6
NS
281
71
100.0
100.0
97.3
96.8
98.2
90.6
85.1
86.7
79.3
NS
100.0
484
Circulation Vol 85, No 2 February 1992
TABLE 12. Continued.
Factor
LV end-diastolic pressure (mm Hg)
Normal (.12)
Abnormal (> 12)*
NYHA functional class
I
II
III
IV
Immediate postop RA pressure
Absence of PLE (%)
6 Months
1 Year 5 Years
n
30 Days
121
187
100.0
100.0
97.1
97.9
97.1
96.5
5
182
152
7
100.0
100.0
100.0
100.0
100.0
96.7
99.2
100.0
100.0
95.3
99.2
100.0
10 Years
Pu
Pc
92.9
85.3
90.9
81.6
0.043
X 2=4.09
NS
100.0
88.2
89.7
100.0
100.0
84.6
84.4
100.0
NS
NS
(mm Hg)
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
271
Normal (<20)
100.0
97.5
96.6
90.8
86.6
NS
NS
66
100.0
100.0
100.0
84.9
81.4
Abnormal (>20)
Risk group 9
94.7
All absent
46
100.0
100.0
100.0
87.2
NS
...
96.7
88.6
85.0
>1 present
306
100.0
97.5
Risk group 4
89.4
143
100.0
98.4
97.5
91.5
NS
...
All absent
96.8
87.8
81.6
.1 present
209
100.0
97.5
Deaths not preceded by evidence of protein-losing enteropathy (PLE) are treated in the analysis as censored observations. RAP, right
atrial pressure; PA, pulmonary artery; LV, left ventricle; AV, atrioventricular; NYHA, New York Heart Association; RA, right atrium.
n refers to number of patients with the factor. For some factors, a few patients are missing information; sum of the numbers in subgroups
will not equal overall total. Survival estimates are obtained from Kaplan-Meier survival curves for selected time points after time zero.
Pu refers to univariate probability value from log-rank test with NS (nonsignificant) if Pu >0.05. Below each significant finding is the
associated x2 statistic subscripted with its degrees of freedom. Comparison of the subgroups is an overall comparison of the survival
experience not restricted to the selected time points presented. Pc refers to univariate probability value from Cox regression model assessing
association between survival and factor considered as a continuous variable. Direction of association is indicated in parentheses. Minus (-)
sign indicates that survival is improving (the hazard is decreasing) as the level of the factor increases or if the condition is present. Pm (with
RAP) is the significance of that factor in a multivariate model arrived at after backward elimination of the nonsignificant (p>0.05, NS)
variables in a model that initially included the immediate postoperative right atrial pressure. Pm (w/o RAP) is like Pm (with RAP) except
that only preoperative factors were considered.
AV valve abnormal if moderate or severe insufficiency preoperatively or necessity for valve repair or replacement prior to or concurrent
with Fontan procedure.
All factors except postop RA pressure are based on preoperative assessment except where indicated.
*Risk group 9 consists of the presence of any (.1) of the nine conditions (criteria of Choussaut's et a112).
tRisk group 4 consists of the presence of any (.1) of the four conditions.
Postoperative Cardiac Cathetenzation
Of the 352 patients who had a modified Fontan
operation, 133 had at least one postoperative cardiac
catheterization. These cardiac catheterizations were
performed from 2 days to 9 years (mean, 1.8 years)
postoperatively and were performed at various medical
institutions. The pertinent hemodynamic data for this
very select group of patients is summarized in Table 26.
Discussion
The operative mortality for the modified Fontan
operation has been well defined for a large number of
patients by numerous authors.2-1' As experience with
this procedure has increased, the operative risk has
decreased. This is well illustrated by our own experience. Prior to 1975, the 30-day mortality for all
patients having the modified Fontan operation at the
Mayo Clinic was 38%; between 1976 and 1980 it was
20%; between 1981 and 1985 it was 13%; and since
1985 it has been 8%. The reduction of 30-day mortality occurred despite increasing complexity of the
defects being repaired.
It is essential, at this point, that the mid- to
long-term results of this procedure be measured and
the determinants of the quality of outcome be identified. To do this, we selected the earliest cohort of
patients to have had the modified Fontan operation
at the Mayo Clinic. This cohort was chosen so that all
patients were at least 5 years postoperative when the
study began. In contrast to previous studies, this
design allows collection of uniform information at
specified times postoperatively. An obvious consequence of this study design, however, is that patients
operated on most recently (when the 30-day mortality is lowest) are excluded from the study. One could
argue that the results may be different when this
group is analyzed some time in the future. In an
attempt to obviate this problem, we performed parallel analyses of our data, both for all patients from
the time of Fontan operation and for only those
patients who were alive 30 days after the Fontan
operation.
A second more difficult problem in interpreting
the results of this study is the lack of an unoperated
Driscoll et al Follow-up After Fontan Operation
485
TABLE 13. Univariate Assessment of Factors Associated With End Point Protein-Losing Enteropathy/Hypoproteinemia or
Hypoalbuminemia or Death With Time Zero the Initiation of Fontan Operation
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
Factor
Overall
Sex
Male
Female
Lesion
Tricuspid atresia
Univentricular heart
Complex single ventricle
Calendar year of operation
1970-1975
1976-1980
1981-1984
Heterotaxia
Absent
Present
Asplenia
Absent
Present
Polysplenia
Absent
Present
Age at operation (yr)
4-15
<4 or >16*
<4
4-15
>16
Preoperative sinus rhythm
Present
Absent*
Systemic venous drainage
Normal
Abnormal*
Mean PA pressure (mm Hg)
Normal (.15)
Abnormal (>15)*t
Pulmonary arteriolar resistance (U. m2)
Normal (<4)
Abnormal (24)*t
Mean PA size (% normal)
Normal (>60)
Abnormal (<60)*
LV ejection fraction (%)
Normal (260)
Abnormal (<60)*t
AV valve dysfunction
Normal
Abnormal*t
PA architecture
Normal
Abnormal
n
352
215
30 Days
83.8
Survival free of PLE (%)
6 Months
1 Year
5 Years
77.0
74.7
65.7
10 Years
57.8
Pu
Pc
...
...
NS
137
81.9
86.9
76.7
77.4
74.9
74.5
65.3
66.4
57.6
58.1
125
114
113
90.4
86.0
74.3
83.2
81.6
65.5
83.2
79.8
60.2
76.7
68.9
50.3
68.1
58.8
45.1
X2=18.20
8
127
217
62.5
79.5
87.1
37.5
74.0
80.2
37.5
73.2
76.9
37.5
62.0
68.9
25.0
55.9
58.2
0.029
0.008(-)
x2=7.08
X2=7.09
325
27
86.8
48.1
80.0
40.7
78.1
33.3
69.3
22.2
60.8
22.2
X 2=37.55
341
11
85.6
27.3
78.9
18.2
76.5
18.2
67.2
18.2
59.2
18.2
X 2=26.56
336
16
84.8
62.5
78.0
156.3
76.2
43.8
67.7
25.0
59.2
25.0
X 2= 14.62
240
112
33
240
79
82.5
86.6
66.7
82.5
94.9
77.5
75.9
51.5
77.5
86.1
75.4
73.2
51.5
75.4
82.3
67.3
62.3
48.5
67.3
68.1
58.5
56.5
43.6
58.5
61.4
320
32
85.0
71.9
77.8
68.8
76.2
59.4
68.0
43.0
59.6
39.4
X 2=7.13
297
55
86.5
69.1
80.5
58.2
78.8
52.7
69.9
43.4
61.3
38.9
X 2= 17.20
154
177
89.0
79.7
82.5
71.8
81.1
68.9
72.5
59.2
68.1
51.3
0.004
X2=8.39
0.002(+)
154
5
84.4
60.0
76.6
60.0
74.0
60.0
64.5
20.0
56.0
20.0
0.027
NS
X 2=4.90
<0.001
<0.001
<0.001
<0.001
NS
0.040
0.008
<0.001
55
2
41
71
95.1
88.7
92.7
80.3
92.7
74.6
82.0
66.0
72.7
26.1
319
33
85.6
66.7
79.9
48.5
77.7
45.5
68.8
36.4
60.2
36.4
281
71
83.3
85.9
76.9
77.5
74.4
76.1
66.7
61.7
59.8
48.3
NS
X 2=6.43
x2=9.40
...
NS
NS
0.046(-)
xl=4.00
X
<0.001
= 16.18
NS
486
Circulation Vol 85, No 2 February 1992
TABLE 13. Continued.
Factor
LV end-diastolic pressure (mm Hg)
Normal ( 12)
Abnormal (>12)*
NYHA functional class
I
II
III
IV
Immediate postop RA pressure
Survival free of PLE (%)
6 Months
1 Year
5 Years
n
30 Days
121
187
89.3
80.2
81.0
74.3
80.2
71.1
5
182
152
7
100.0
84.1
82.2
85.7
100.0
78.6
74.3
71.4
100.0
75.8
73.0
57.1
10 Years
Pu
75.0
59.2
65.3
53.2
0.007
0.006
X2=7.22
X2=745
100.0
68.6
60.9
57.1
100.0
60.2
53.2
57.1
NS
0.037
X 1=4.34
Pc
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
(mm Hg)
Normal (<20)
271
91.1
84.9
66.4
82.6
74.3
<0.001
<0.001
66
65.2
Abnormal (>20)
54.5
51.5
42.4
32.0
Xl=34.83 X2=28.77
Risk group 9
46
All absent
89.1
87.0
87.0
82.3
68.9
...
0.033
.1 present
306
83.0
72.9
63.3
75.5
56.5
x2=4.54
Risk group 4
All absent
143
88.8
83.2
82.5
73.9
67.2
0.002
.1 present
80.4
209
72.7
69.4
60.1
51.0
X 2=9.68
For this analysis, evidence of protein-losing enteropathy (PLE) or death was considered as the end point. RAP, right atrial pressure; PA,
pulmonary artery; LV, left ventricle; AV, atrioventricular; NYHA, New York Heart Association; RA, right atrium.
n refers to number of patients with the factor. For some factors, a few patients are missing information; sum of the numbers in subgroups
will not equal overall total. Survival estimates are obtained from Kaplan-Meier survival curves for selected time points after time zero.
Pu refers to univariate probability value from log-rank test with NS (nonsignificant) if Pu >0.05. Below each significant finding is the
associated x2 statistic subscripted with its degrees of freedom. Comparison of the subgroups is an overall comparison of the survival
experience not restricted to the selected time points presented. Pc refers to univariate probability value from Cox regression model assessing
association between survival and factor considered as a continuous variable. Direction of association is indicated in parentheses. Minus (-)
sign indicates that survival is improving (the hazard is decreasing) as the level of the factor increases or if the condition is present. Pm (with
RAP) is the significance of that factor in a multivariate model arrived at after backward elimination of the nonsignificant (p>0.05, NS)
variables in a model that initially included the immediate postoperative right atrial pressure. Pm (w/o RAP) is like Pm (with RAP) except
that only preoperative factors were considered.
AV valve abnormal if moderate or severe insufficiency preoperatively or necessity for valve repair or replacement prior to or concurrent
with Fontan procedure.
All factors except postop RA pressure are based on preoperative assessment except where indicated.
*Risk group 9 consists of the presence of any (.1) of the nine conditions (criteria of Choussaut's et al12).
tRisk group 4 consists of the presence of any (>1) of the four conditions.
or
conventionally palliated (systemic-to-pulmonary
artery shunt or pulmonary artery band) control
group. Thus, although one can describe the outcome
events, it is impossible to know if the outcome after
the Fontan operation is better or worse than it would
have been without the Fontan operation.
Survival
Fontan et a17 reported the outcome of 334 patients
operated on in Bordeaux, France, or Birmingham,
Ala. They were operated on from 1975 to 1988, and
the follow-up period ranged from 1 month to 20
years. The survival was 79% (30 days), 74% (6
months), 73% (1 year), 69% (5 years), and 63% (10
years). This is remarkably similar though slightly
lower than overall survival in our study (84%, 30
days; 79%, 6 months; 77%, 1 year; 70%, 5 years; and
60%, 10 years). Our data substantiate the observation by Fontan et al that there is a continued
greater-than-expected risk of death even after the
Fontan operation.
It is difficult to know if survival after the Fontan
operation is better or worse than if the Fontan
operation had not been performed. Taussig20 reported eight of 24 patients (33%) with univentricular
heart who survived 20 years after a Blalock-Taussig
shunt. Moodie et al21,22 reported a 50% survival 14
years after diagnosis in patients with an unoperated
type A univentricular heart and a 50% 4-year survival
for patients with unoperated type C univentricular
heart. For patients who received conventional palliation, the survivorship for 5 years after diagnosis was
72%; for patients with type A and type C univentricular heart it was 68%. Unfortunately, it is unclear
from Moodie's two studies the ages of the patients at
the time of diagnosis and, hence, impossible to
reconstruct survival curves from those data.
Determinants of Survival
In 1977, Choussat et al'2 listed 10 criteria for
operability for a low-risk Fontan operation. It has
been well established that operative survival is pos-
Driscoll et al Follow-up After Fontan Operation
TABLE 14. New York Heart Association Classification
At 5 years post-Fontan, functional class of
questionnaire responders
Subtotal
Preop
NYHA
Questionnaire
I
II
III
IV
(D+)
Missing
class
n
not returned
5 (0)
0
2
0
0
3
Missing
6
0
5
1
0
I
5
1
3
0
(0)
0
4
II
8
3
44
12
118 (1)
55
182
11
5
85 (3)
III
9
4
27
38
152
IV
0
7
2
1
1
0
2(0)
0
19
89
26
9
215 (4)
Total
8
83
352
487
Known to be deceased, not sent
questionnaire
Died
<5 years
1
0
48
52
3
104
>5 years
0
0
8
6
0
14
Alive at 5 years
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
Subtotal:
worse at 5 yr
Questionnaire
Same
Worse
Died <5
or died <5
not returned/
Better
Preop
years
class
class
years
NYHA class
n
class
missing
6
1
1
5*
0
0
0
Missing
1
1
I
5
NP
3
1
0
64
11
II
44
48
16
182
55
III
52
57
13
11
5
152
65
IV
3
3
2
7
2
0
NP
122
22
104
126
32
58
Total
352
NYHA, New York Heart Association; D+, responded but now known to be deceased; NP, not possible.
*Missing baseline NYHA class (no basis for judging change); all alive at 5-yr post-Fontan.
Subtotal
1
0
56
58
3
118
sible, even in the absence of one or more of these
criteria. It also is clear that operative mortality
increases significantly as more of these risk factors
are violated and that some of these risk factors are
more important than others. Additionally, other risk
factors have been identified, including lesions other
than tricuspid atresia, presence of ventricular hypertrophy, early calendar year of operation, subaortic
obstruction, long operative ischemic time, and use of
prosthetic valves in the systemic venous-to-pulmonary artery connections.3,10,23
The level of right atrial pressure postoperatively is
strongly associated with survival. This is because
postoperative right atrial pressure is reflective of and
determined by several other risk factors, including
systemic ventricular function, AV valve competency
and function, pulmonary arteriolar resistance, and
pulmonary artery size. Obviously, because it is a
postoperative measurement, it is useless as a guide
for selecting patients for operation.
One must be cautious in interpreting the results of
analysis of risk factors. For example, pulmonary artery
size was not significantly associated with survival in our
study. However, there were few patients with small
pulmonary arteries sent for operation. Also, several risk
factors that were important in the analysis of survivorship from date of operation lost their importance when
survivorship was assessed only for patients who were
alive 30 days after operation. One must not conclude
that these risk factors are unimportant. Indeed, they
may have accounted for early death.
Reoperation
The need for reoperation after the Fontan procedure was relatively frequent. Reoperations related to
Died >5
years
0
0
8
6
0
14
mechanical pacemakers were most frequent and
were necessary for 8% of the total cohort of 352
patients. Excluding reoperation to control bleeding,
revision of the right atrial-to-pulmonary artery connection was the next most common reason for reoperation (5% of the cohort). Revision of right atrialto-pulmonary artery connections has become much
less frequent since we abandoned the use of prosthetic conduits and tissue valves. Current techniques
for establishing patulent autologous tissue connections have reduced the need for reoperation for
pathway stenosis. It remains to be seen if pathway
stenosis will again become a problem because of
growth as the modified Fontan operation is done in
younger patients.
In this series, five Fontan operations had to be taken
down. It is hoped that with more stringent selection
criteria, these patients could be identified prior to
establishment of the Fontan type of connection.
Arrhythmias
Atrial arrhythmias, particularly atrial flutter, occur
in patients with single ventricle both before and after
the Fontan operation. Presumably, stretch of the
right atrium, decreased ventricular function, and AV
valve insufficiency contribute to this problem. The
occurrence of arrhythmias in our study population is
based on evidence for such an arrhythmia in the
patient's medical history, evidence of an arrhythmia
on an electrocardiographic or 24-hour electrocardiographic recording, or patient self-reporting. It is
possible that patients' self-reporting of an arrhythmia
lacks precision. However, taking this caveat into
account, between 12% and 22% of the patients have
postoperative arrhythmias. More than 12% required
488
Circulation Vol 85, No 2 February 1992
TABLE 15. Health Status at 5 Years Post-Fontan in All 352 Patients
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
Factor
Overall
Sex
Male
Female
Lesion
Tricuspid atresia
Univentricular heart
Complex single ventricle
Calendar year of operation
1970-1980
1981-1984
Heterotaxia
Absent
Present
Asplenia
Absent
Present
Polysplenia
Absent
Present
Age at operation (yr)
4-15
<4 or >16
<4
4-15
.16
Preoperative sinus rhythm
Present
Absent
Systemic venous drainage
Normal
Abnormal
Mean PA pressure (mm Hg)
Normal (c15)
Abnormal (>15)
Pulmonary arteriolar resistance (U. m2)
Normal (<4)
Abnormal (>4)
Mean PA size (% normal)
Normal (.60)
Abnormal (<60)
LV ejection fraction (%)
Normal (.60)
Abnormal (<60)
AV valve dysfunction
Normal
Abnormal
PA architecture
Normal
Abnormal
352
Excellent Intermediate
No. % No. %
37 10.5 178 50.6
Died
before 5
Poor
years
No. % No. %
29
8.2 108 30.7
215
137
22
15
10.2
11.0
110
68
51.2
49.6
13
16
6.0
11.7
70
38
125
114
113
17
13
7
13.6
11.4
6.2
72
61
45
57.6
53.5
39.8
8
9
12
6.4
7.9
10.6
135
217
8
29
5.9
13.4
70
108
51.8
49.8
9
20
325
27
35
2
10.8
7.4
174
4
53.5
14.8
341
11
36
1
10.6
9.1
177
1
336
16
36
1
10.7
6.3
240
112
33
240
79
30
7
1
30
6
320
32
n
X2
Univariate
df
p
Multivariate
p
...
...
...
32.6
27.7
3.9
3
NS
28
31
49
22.4
27.2
43.4
17.7
6
0.007
NS
NS
0.009
6.7
9.2
48
60
35.6
27.6
6.8
3
NS
0.001
28
1
8.6
3.7
88
20
27.1
74.1
26.3
3
<0.001
0.021
51.9
9.1
29
0
8.5
0.0
99
9
29.0
81.8
14.5
3
0.002
175
3
52.1
18.7
28
1
8.3
6.3
97
11
28.9
68.7
11.6
3
0.009
12.5
6.3
3.0
12.5
7.6
120
58
14
120
44
50.0
51.8
42.4
50.0
55.7
18
11
2
18
9
7.5
9.8
6.1
7.5
11.4
72
36
16
72
20
30.0
3.5
3
NS
NS
32.1
48.5
30.0
25.3
9.6
6
NS
NS
36
1
11.3
3.1
167
11
52.2
34.4
26
3
8.1
9.4
91
17
28.4
53.1
9.5
3
0.023
NS
297
55
34
3
11.4
5.5
160
18
53.9
32.7
22
7
7.4
12.7
81
27
27.3
49.1
14.5
3
0.002
NS
154
177
17
18
11.0
10.2
85
80
55.2
45.2
15
14
9.7
7.9
37
65
24.0
36.7
6.3
3
NS
154
5
19
0
12.3
0.0
79
0
51.3
0.0
9
1
5.8
20.0
47
4
30.5
80.0
8.4
3
0.038
55
2
12
0
21.8
0.0
25 45.4
2 100.0
3
0
5.5
0.0
15
0
27.3
0.0
41
71
4
11
9.8
15.5
28
34
68.3
47.9
3
6
7.3
8.4
6
20
14.6
28.2
4.7
3
NS
319
33
36
1
11.3
3.0
167
11
52.3
33.3
29
0
9.1
0.0
87
21
27.3
63.6
20.0
3
<0.001
0.032
281
71
31
6
11.0
8.5
148
30
52.7
42.2
19
6.8
14.1
83
25
29.5
35.2
5.9
3
NS
NS
10
NS
0.021
Driscoll et al Follow-up After Fontan Operation
489
TABLE 15. Continued.
Died
before 5
years
No. %
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
Poor
Univariate
Multivariate
Excellent Intermediate
n No. % No. % No. %
Factor
p
p
X2 df
LV end-diastolic pressure (mm Hg)
0.047
8
6.6 27 22.3
7.9 3
...
Normal (<12)
121 16 13.2 70 57.8
9.6 66 35.3
187 19 10.2 84 44.9 18
Abnormal (>12)
NYHA functional class
4 80.0
0
0
0.031
5
1 20.0
0.0
0.0 12.5 9
NS
I
8.2 50 27.5
II
182 26 14.3 91 50.0 15
III
9
5.9 75 49.4 14
9.2 54 35.5
152
IV
4 57.1
7
0
0.0
0 0.0
3 42.9
Immediate postop RA pressure
(mm Hg)
Normal (.20)
271 29 10.7 157 57.9 26 9.6 59 21.8 28.8 3 <0.001
Abnormal (>20)
66
6
9.1
21 31.8
3
4.5 36 54.6
See text for definitions of excellent, intermediate, and poor.
Univariate associations between factors and health status were made using x2 statistics for a two-dimensional contingency table.
Multivariate associations between factors and health status were made using logistic regression with backward elimination of nonsignificant
(NS) variables treating the dependent health status variable on an ordinal scale with excellent at one extreme and death at the other. Some
of the subgroups did not have a sufficient number of patients to allow a meaningful analysis but are presented to indicate frequency of
patients in the subgroups.
RAP, right atrial pressure; PA, pulmonary artery; LV, left ventricle; AV, atrioventricular; NYHA, New York Heart Association; RA, right
atrium.
n refers to number of patients with the factor. For some factors, a few patients are missing information; sum of the numbers in subgroups
will not equal overall total. Survival estimates are obtained from Kaplan-Meier survival curves for selected time points after time zero.
Pu refers to univariate probability value from log-rank test with NS (nonsignificant) if Pu >0.05. Below each significant finding is the
associated x2 statistic subscripted with its degrees of freedom. Comparison of the subgroups is an overall comparison of the survival
experience not restricted to the selected time points presented. Pc refers to univariate probability value from Cox regression model assessing
association between survival and factor considered as a continuous variable. Direction of association is indicated in parentheses. Minus (-)
sign indicates that survival is improving (the hazard is decreasing) as the level of the factor increases or if the condition is present. Pm (with
RAP) is the significance of that factor in a multivariate model arrived at after backward elimination of the nonsignificant (p>0.05, NS)
variables in a model that initially included the immediate postoperative right atrial pressure. Pm (w/o RAP) is like Pm (with RAP) except
that only preoperative factors were considered.
AV valve abnormal if moderate or severe insufficiency preoperatively or necessity for valve repair or replacement prior to or concurrent
with Fontan procedure.
All factors except postop RA pressure are based on preoperative assessment except where indicated.
implantation of a mechanical pacemaker, and at least
20% of the surviving cohort require at least one
antiarrhythmic medication other than digitalis 5
years after operation.
Because the insertion and management of cardiac
pacemakers in this study were performed at a variety
of institutions, the indications for pacemaker insertion, repair, or replacement were not uniform. However, among the indications for initial insertion were
the presence of preoperative complete atrioventricular block, late postoperative "tachy-brady" syndrome, and the perceived need for antitachycardia
pacemakers. Surgically induced complete atrioventricular block occurred but was uncommon.
It is important to recognize the presence of atrial
flutter in patients who have had the Fontan operation
because it may result in congestive heart failure and
fluid retention. We have cared for several patients in
whom this arrhythmia went unrecognized for several
weeks because the patient was thought to have sinus
tachycardia with first-degree atrioventricular block
instead of atrial flutter with 2:1 conduction. One
should suspect the presence of atrial flutter in patients who have had a Fontan operation who have a
persistent and relatively constant tachycardia (100150 beats per minute) and a prolonged PR interval. It
will be important to find modifications of the surgical
technique that will lessen the frequency of postoperative arrhythmias.
Protein-Losing Enteropathy or Hypoproteinemia
Protein-losing enteropathy is a serious problem
after the Fontan operation.24 Its cause is unknown
but probably is related to systemic venous hypertension and increased thoracic duct pressure. There also
may be a local autoimmune or allergic component in
the gut wall. Technically, one cannot be certain that
a patient with hypoproteinemia or hypoalbuminemia
after the Fontan operation has protein-losing enteropathy unless excess loss of protein from the gastrointestinal tract can be demonstrated. For this reason,
we assessed the presence and determinants of protein-losing enteropathy, hypoproteinemia, or hypoalbuminemia in this study. We found an estimated
occurrence rate of protein-losing enteropathy, hypoproteinemia, or hypoalbuminemia of 10.5% and
14.7% at 5 and 10 years postoperatively, respectively.
There was a relatively strong relation between pro-
490
Circulation Vol 85, No 2 February 1992
TABLE 16. Health Status at 5 Years Post-Fontan in 244 Surviving Patients
Excellent Intermediate
n
No.
No.
%
Factor
%
244
37
Overall
15.2
178
72.9
Sex
22
145
Male
15.2
110 75.9
15.1
68
68.7
15
Female
99
Lesion
Tricuspid atresia
Univentricular heart
Complex single ventricle
Calendar year of operation
1970-1980
1981-1984
Heterotaxia
Absent
Univariate
Poor
Multivariate
p
No.
29
%
x2
11.9
...
...
...
...
13
16
9.0
16.2
3.0
2
NS
NS
df
p
97
83
64
17
13
7
17.5
15.7
10.9
72
61
45
74.2
73.5
70.3
8
9
12
8.3
10.8
18.8
4.9
4
NS
NS
NS
0.032
87
157
8
29
9.2
18.5
70
108
80.5
68.8
9
20
10.3
12.7
4.5
2
NS
0.038
237
7
35
2
14.8
28.6
174
4
73.4
57.1
28
1
11.8
14.3
1.1
2
NS
0.033
242
2
36
1
14.9
50.0
177
1
73.1
50.0
29
0
12.0
0.0
...
239
5
36
1
15.1
20.0
175
3
73.2
60.0
28
11.7
20.0
0.5
2
NS
1
168
76
17
168
59
30
7
1
30
6
17.9
9.2
5.9
17.9
10.2
120
58
14
120
44
71.4
76.3
82.3
71.4
74.6
18
11
2
18
9
10.7
14.5
11.8
10.7
15.2
3.4
2
NS
NS
3.8
4
NS
NS
229
15
36
1
15.7
6.7
167
11
72.9
73.3
26
3
11.4
20.0
1.6
2
NS
NS
216
28
34
3
15.7
10.7
160
18
74.1
64.3
22
7
10.2
25.0
5.3
2
NS
0.048
117
112
17
18
14.5
16.1
85
80
72.7
71.4
15
14
12.8
12.5
0.1
2
NS
NS
107
1
19
0
17.8
0.0
79
0
73.8
0.0
9
1
8.4
100.0
...
40
2
12
0
30.0
0.0
25
2
62.5
100.0
3
0
7.5
0.0
...
35
51
4
11
11.4
21.6
28
34
80.0
66.7
3
6
8.6
11.8
1.9
2
NS
...
232
12
36
1
15.5
8.3
167
11
72.0
91.7
29
0
12.5
0.0
2.5
2
NS
NS
Abnormal*t
PA architecture
Normal
Abnormal*
198
46
31
6
15.7
13.0
148
30
74.7
65.2
19
10
9.6
21.7
5.3
2
NS
0.025
Present
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
Asplenia
Absent
Present
Polysplenia
Absent
Present
Age at operation (yr)
4-15
<4 or 216*
<4
4-15
>16
Preoperative sinus rhythm
Present
Absent*
Systemic venous drainage
Normal
Abnormal
Mean PA pressure (mm Hg)
Normal (<15)
Abnormal (>15)*t
Pulmonary arteriolar resistance (U mi2)
Normal (<4)
Abnormal (.4)*t
Mean PA size (% normal)
Normal (.60)
Abnormal (<60)*
LV ejection fraction (%)
Normal (.60)
Abnormal (<60)*t
AV valve dysfunction
Normal
Driscoll et al Follow-up After Fontan Operation
491
TABLE 16. Continued.
Excellent
No.
%
Intermediate
No.
%
Poor
Univariate
df
p
X2
Multivariate
p
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
n
No.
%
Factor
LV end-diastolic pressure (mm Hg)
94
70
8
2
16
17.0
74.5
8.5
2.0
NS
...
Normal (.12)
84
121
19
69.4
18
14.9
Abnormal (>12)*
15.7
NYHA functional class
5
1
4
0
6
I
20.0
80.0
0.0
7.2
NS
NS
II
132
26
91
15
11.4
19.7
68.9
14
III
98
9
9.2
75
76.5
14.3
4
IV
4 100.0
0
0
0.0
0.0
Immediate postop RA pressure
(mm Hg)
212
Normal (.20)
29
13.7
157 74.1
26
12.3
0.9
2
NS
...
30
6
3
Abnormal (>20)
20.0
21
70.0
10.0
See text for definitions of excellent, intermediate, and poor.
Univariate associations between factors and health status were made using x2 statistics for a two-dimensional contingency table.
Multivariate associations between factors and health status were made using logistic regression with backward elimination of nonsignificant
(NS) variables treating the dependent health status variable on an ordinal scale with excellent at one extreme and death at the other. Some
of the subgroups did not have a sufficient number of patients to allow a meaningful analysis but are presented to indicate frequency of
patients in the subgroups.
RAP, right atrial pressure; PA, pulmonary artery; LV, left ventricle; AV, atrioventricular; NYHA, New York Heart Association; RA, right
atrium.
n refers to number of patients with the factor. For some factors, a few patients are missing information; sum of the numbers in subgroups
will not equal overall total. Survival estimates are obtained from Kaplan-Meier survival curves for selected time points after time zero.
Pu refers to univariate probability value from log-rank test with NS (nonsignificant) if Pu >0.05. Below each significant finding is the
associated x2 statistic subscripted with its degrees of freedom. Comparison of the subgroups is an overall comparison of the survival
experience not restricted to the selected time points presented. Pc refers to univariate probability value from Cox regression model assessing
association between survival and factor considered as a continuous variable. Direction of association is indicated in parentheses. Minus (-)
sign indicates that survival is improving (the hazard is decreasing) as the level of the factor increases or if the condition is present. Pm (with
RAP) is the significance of that factor in a multivariate model arrived at after backward elimination of the nonsignificant (p>0.05, NS)
variables in a model that initially included the immediate postoperative right atrial pressure. Pm (w/o RAP) is like Pm (with RAP) except
that only preoperative factors were considered.
AV valve abnormal if moderate or severe insufficiency preoperatively or necessity for valve repair or replacement prior to or concurrent
with Fontan procedure.
All factors except postop RA pressure are based on preoperative assessment except where indicated.
*Risk group 9 consists of the presence of any (.1) of the nine conditions (criteria of Choussaut's et all2).
tRisk group 4 consists of the presence of any (.1) of the four conditions.
tein-losing enteropathy, hypoproteinemia, or hypoalbuminemia and the presence of heterotaxia syndromes, anomalies of systemic venous drainage, and
increased pulmonary arteriolar resistance.
Functional Status
There are, of course, many ways to assess functional status. Most commonly, the NYHA classification is used. In the follow-up study by Fontan et al,7
155 (48%) patients were in NYHA class I, 51 (16%)
were in class II, eight (2%) were in class III, and 105
(33%) were dead. These investigators reported that
the greater the length of follow-up, the fewer patients
there were in NYHA classes I and II. Indeed, in our
study, which had a longer follow-up interval than the
study by Fontan et al, only 49% of the survivors were
in NYHA class I and II. Girod et al,9 however,
reported 26 patients (80% of their cohort) to be in
class I or II at the mean follow-up period of 8.9 years.
DeVivie et all' studied a relatively small group of
patients and found 88% of the survivors in NYHA
class I or II.
When assessing functional status with the NYHA
classification, it is important to compare patient
status after operation with that before operation. In
our study, 172 patients were in NYHA classes I and
II postoperatively, and 187 were in NYHA classes I
and II preoperatively. Considering class I alone,
there were only five patients so classified preoperatively but 89 so classified postoperatively. We found
that 180 patients were alive and in the same or better
NYHA classification 5 years postoperatively than
preoperatively. In contrast, 126 were dead or in a
worse functional classification postoperatively than
preoperatively.
Another index of functional status is exercise tolerance. We found that approximately 50% of the
surviving cohort could do heavy housework or run
100 yards 5 years postoperatively. In contrast, 15
patients were incapable of climbing one flight of
stairs 5 years postoperatively.
We attempted to define the factors associated with
excellent outcome and poor outcome after the Fontan operation. The criteria we used to define an
excellent outcome were strict. Also, it is difficult to
know how to handle death in this type of analysis.
Surely, death is a poor outcome. If death is included
in the analysis, then there were numerous factors
492
Circulation Vol 85, No 2 February 1992
TABLE 17. Current Health Status in 211 Surviving Patients
Excellent
n
No.
Factor
%
211
39
18.5
Overall
Sex
123
25
20.3
Male
14
15.9
88
Female
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
Lesion
Tricuspid atresia
Univentricular heart
Complex single ventricle
Calendar year of operation
1970-1980
1981-1984
Heterotaxia
Absent
Present
Asplenia
Absent
Present
Polysplenia
Absent
Present
Age at operation (yr)
4-15
<4 or >16*
<4
4-15
.16
Preoperative sinus rhythm
Present
Absent*
Systemic venous drainage
Normal
Abnormal
Mean PA pressure (mm Hg)
Normal (. 15)
Abnormal (>15)*t
Pulmonary arteriolar resistance (U . m2)
Normal (<4)
Abnormal (>4)*t
Mean PA size (% normal)
Normal (.60)
Abnormal (<60)*
LV ejection fraction (%)
Normal (.60)
Abnormal (<60)*t
AV valve dysfunction
Normal
Abnormal*t
PA architecture
Normal
Abnormal*
Intermediate
No.
%
143
67.8
Univariate
Poor
No.
29
%
13.7
x2
df
p
...
84
59
68.3
67.0
14
15
11.4
17.0
1.7
2
NS
84
68
59
17
14
8
20.2
20.6
13.6
56
47
40
66.7
69.1
67.8
11
7
11
13.1
10.3
18.6
2.8
4
NS
69
142
8
31
11.6
21.8
47
96
68.1
67.6
14
20.3
10.6
5.8
2
0.054
15
204
7
39
0
19.1
0.0
137
6
67.2
85.7
28
1
13.7
14.3
1.7
2
NS
209
2
39
0
18.7
0.0
141
2
67.5
100.0
29
0
13.9
0.0
1.0
2
NS
206
5
39
0
18.9
0.0
139
4
67.5
80.0
28
1
13.6
20.0
1.2
2
NS
149
62
15
149
47
32
7
1
32
6
21.5
11.3
6.7
21.5
12.8
102
41
11
102
30
68.5
66.1
73.3
68.5
63.8
15
14
3
15
7.5
2
0.024
8.0
4
NS
11
10.1
22.6
20.0
10.1
23.4
198
13
39
0
19.7
0.0
132
66.7
84.6
27
2
13.6
15.4
3.2
2
NS
11
185
26
39
0
21.1
0.0
124
19
67.0
73.1
22
7
11.9
26.9
9.4
2
0.009
100
99
19
18
19.0
18.2
71
63
71.0
63.6
10
18
10.0
18.2
2.8
2
NS
91
1
17
0
18.7
0.0
64
0
70.3
0.0
10
1
11.0
100.0
...
37
2
10
0
27.0
0.0
24
2
64.9
100.0
3
0
8.1
...
0.0
30
43
3
12
10.0
27.9
24
25
80.0
58.1
3
6
10.0
14.0
4.2
2
NS
201
10
38
1
18.9
10.0
134
9
66.7
90.0
29
14.4
0.0
2.6
2
NS
172
39
32
7
18.6
18.0
119
24
69.2
61.5
21
12.2
20.5
1.9
2
NS
0
8
Driscoll et al Follow-up After Fontan Operation
493
TABLE 17. Continued.
Excellent
No.
%
Intermediate
No.
%
Poor
Univariate
df
p
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
n
No.
%
Factor
x2
LV end-diastolic pressure (mm Hg)
Normal (<12)
82
17
20.7
70.7
7
3.0
2
58
8.5
NS
Abnormal (>12)*
105
19
18.1
68
64.8
18
17.1
NYHA functional class
I
1
20.0
4
5
80.0
0
0.0
6.9
6
NS
II
117
26
22.2
64.1
16
75
13.7
III
82
10
12.2
60
12
73.2
14.6
IV
2
0
0.0
1
1
50.0
50.0
Immediate postop RA pressure
(mm Hg)
185
32
27
Normal (<20)
17.3
126
68.1
14.6
0.8
2
NS
24
5
Abnormal (>20)
17
70.8
2
8.3
20.8
See text for definitions of excellent, intermediate, and poor.
Univariate associations between factors and health status were made using x2 statistics for a two-dimensional contingency table.
Multivariate associations between factors and health status were made using logistic regression with backward elimination of nonsignificant
(NS) variables treating the dependent health status variable on an ordinal scale with excellent at one extreme and death at the other. Some
of the subgroups did not have a sufficient number of patients to allow a meaningful analysis but are presented to indicate frequency of
patients in the subgroups.
RAP, right atrial pressure; PA, pulmonary artery; LV, left ventricle; AV, atrioventricular; NYHA, New York Heart Association; RA, right
atrium.
n refers to number of patients with the factor. For some factors, a few patients are missing information; sum of the numbers in subgroups
will not equal overall total. Survival estimates are obtained from Kaplan-Meier survival curves for selected time points after time zero.
Pu refers to univariate probability value from log-rank test with NS (nonsignificant) if Pu >0.05. Below each significant finding is the
associated x2 statistic subscripted with its degrees of freedom. Comparison of the subgroups is an overall comparison of the survival
experience not restricted to the selected time points presented. Pc refers to univariate probability value from Cox regression model assessing
association between survival and factor considered as a continuous variable. Direction of association is indicated in parentheses. Minus (-)
sign indicates that survival is improving (the hazard is decreasing) as the level of the factor increases or if the condition is present. Pm (with
RAP) is the significance of that factor in a multivariate model arrived at after backward elimination of the nonsignificant (p>0.05, NS)
variables in a model that initially included the immediate postoperative right atrial pressure. Pm (w/o RAP) is like Pm (with RAP) except
that only preoperative factors were considered.
AV valve abnormal if moderate or severe insufficiency preoperatively or necessity for valve repair or replacement prior to or concurrent
with Fontan procedure.
All factors except postop RA pressure are based on preoperative assessment except where indicated.
*Risk group 9 consists of the presence of any (.1) of the nine conditions (criteria of Choussaut's et al2).
tRisk group 4 consists of the presence of any (.1) of the four conditions.
significantly associated with the type of functional
outcome. If death is excluded as an outcome event,
early calendar year of operation, complex forms of
single ventricle, abnormalities of systemic venous
drainage, and abnormalities of pulmonary artery
architecture were significantly associated with health
status 5 years postoperatively. Current health status
was only influenced by calendar year of operation,
abnormalities of systemic venous drainage, and rela-
tively young (less than 4 years) or relatively old
(greater than or equal to 16 years) age at the time of
operation. Considerable caution is needed in interpreting these results. For example, abnormalities of
AV valve function were not significantly associated
with health status if death was not considered as an
outcome event. However, abnormalities of AV valve
function were strongly associated with death. Thus,
patients with AV valve abnormality were more likely
TABLE 18. Activity Capacity
Patients (n=215)
5 Years postop
Capable
Activity
Light housework
Moderate housework
Heavy housework
Climb one flight stairs
Walk 400 yards
Run 100 yards
No.
193
185
114
191
182
103
%
Incapable
No.
%
90
86
53
89
85
48
4
13
73
15
27
94
2
6
34
7
13
44
Unknown
No.
%
18
8
17
8
28
13
4
9
3
6
8
18
Capable
No.
192
185
115
192
182
109
%
89
86
53
89
85
51
Currently
Incapable
%
No.
1
1
4
9
65
30
10
5
24
87
11
40
Unknown
No.
%
10
22
21
10
35
16
13
6
4
9
19
9
494
Circulation Vol 85, No 2 February 1992
TABLE 19. Exercise Ability Relative to Peers
Patients (n=215)
5 Years
postop
Currently
No.
%
No.
%
1
0
0
0.5
More than peers
92
43
97
45
As much as peers
77
36
82
38
50% As much as peers
16
7
19
9
25% As much as peers
4
12
8
6
<25% As much as peers
TABLE 23. Patient Perception of Employment Capability Among
149 Patients Currently Older Than 16 Years
Patients
Perception
No.
S
Capable of no employment
3
5
23
15
Part-time, not physically demanding
9
13
Part-time, physically demanding
44
66
Full-time, not physically demanding
42
28
Full-time, physically demanding
i-f eAlris
lit-nraeh1o oUL
rric
inlap4aUi
Unknown
1
results be
5
deficiencies in the study design. The data from these
352 patients represent the initial half of the experience at the Mayo Clinic with the Fontan operation.
We already know that the operative mortality since
1985 is considerably less than that during the period of
A
U
J
4
2
J
11
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
to die and, hence, to be excluded from the analysis of
health status at 5 years after operation.
This study is perhaps the most detailed analysis of
the largest cohort of patients followed for the longest
time after the Fontan operation. How should these
interpreted? First, one must recognize the
TABLE 24. Medication Use Among Surviving Cohort
Patients (n=215)
5 Years
TABLE 20. School Absenteeism
Days absent in past
6 school months
0
1-5
6-10
11-20
21-30
>30
Unknown
Patients (n= 140)
No.
92
23
4
5
1
6
9
6
TABLE 21. Educational Level Achieved by Patients Who Have
Completed Formal Education
n
Highest level achieved
2
3
No formal education
Grade
<8
3
8
3
14
10
9-11
12
23
32
4
6
Some technical or vocational
6
8
Completed technical or vocational
14
10
Some college
8
11
Four years of college
4
5
Beyond college
n=73 patients.
TABLE 22. Absenteeism Among 80 Employed Individuals
Days absent in
6-month period
None
1-5
6-10
11-20
21-30
Unknown
Patients
n
65
6
0
2
1
6
Currently
postop
%
66
16
%
81
8
0
3
1
8
Medication
Digoxin
Furosemide
Spironolactone
Chlorothiazide
Diuretic (unspecified type)
Captopril
Hydralazine
Quinidine
Procainamide
Verapamil
Propranolol
Amiodarone
Warfarin
Other
None of above
No.
83
36
25
1
3
14
4
8
4
7
3
3
3
12
87
%
39
17
12
0.5
1
7
2
4
2
3
1
1
1
6
40
No.
92
41
25
4
3
16
1
11
5
7
5
3
6
28
94
%
43
19
12
2
1
7
0.5
5
2
3
2
1
3
8
44
TABLE 25. Symptoms Reported by Surviving Cohort
Symptom
Easy fatigue
Shortness of breath
Palpitations
Rapid heart beats
Swelling of abdomen
Chest pain
Swelling of ankles or legs
Syncope
Slow heart beats
Nausea or vomiting with
exercise
None of above
Patients (n=215)
5 Years
postop
Currently
No.
No.
%
%
79
37
78
36
61
28
30
65
51
24
28
60
49
27
23
58
26
12
28
13
24
11
34
16
19
9
20
9
17
8
18
8
17
17
8
8
18
56
8
26
21
60
10
28
Driscoll et al Follow-up After Fontan Operation
TABLE 26. Postoperative Cardiac Catheterization Findings
n
Mean
Finding
Operation to
674
146
catheterization (days)
Right atrial pressure
16.5
(mm Hg)
117
Mean pulmonary artery
81
15.7
pressure (mm Hg)
Ventricular end-diastolic
96
9.8
pressure (mm Hg)
Ventricular systolic
92
106
pressure (mm Hg)
Femoral artery systolic
101
89.6
pressure (mm Hg)
91.4
Blood 02 sat aorta (%)
41
Blood 02 sat femoral
64
88.6
artery (%)
89
13.9
Hemoglobin (g/dl)
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
this study (i.e., prior to 1985). Whether or not the
determinants of survival and outcome that were described in this study in a retrospective fashion can be
used prospectively in a meaningful manner remains to
be seen. Second, Choussat et al12 were correct. The
best results after the Fontan operation will be for
patients who fulfill the criteria that they outlined.
Third, Fontan et all17,8 were correct that the operation
that bears Fontan's name is palliative in nature and is
associated with ongoing mortality and morbidity.
Fourth, we have reached a point where more insightful selection of patients for this operation is warranted. Historically, cardiologists and surgeons have
been deviating further and further from the original
criteria of Choussat et al and have demonstrated that
patients lacking many of these criteria can survive
operation. Perhaps to ensure good functional longterm outcome in addition to survival, we must be
diligent in excluding from the Fontan operation patients known to be at high risk for death or poor
outcome. This is especially important because there
now are alternative treatment options. These options
include techniques to improve pulmonary blood flow
without concomitant ventricular volume overload
(variations of the Glenn anastomosis) and cardiac
transplantation. Fifth, it may be true (but not necessarily true, and certainly unproved at this time) that
offering the Fontan operation at an earlier age before
ventricular dysfunction and AV valve dysfunction occur may result in improved long-term results.
References
1. Fontan F, Mounicot F, Baudet E, Simonneau J, Gordo J,
Gouffrant J: "Correction" de L'Atresie Tricuspidienne: Rapport de deux cas "corriqes" par 1 'utilisation d'une technique
chirurqicale nouvelle. Ann Chir Thorac Cardio-vasc 1971;10:
39-47
2. Russo P, Danielson G, Puga F, McGoon D, Humes R:
Modified Fontan procedure for biventricular hearts with complex forms of double-outlet right ventricles. Circulation 1988;
78(suppl III):III-20-III-25
Median
Minimum
Maximum
483
2
3,277
16
6
33
15
3
33
9
0
24
105
74
178
92
93
58
65
100
99
91
13.9
58
9.1
100
18.5
495
3. Mayer J, Helgason H, Jonas R, Lang P, Vargas F, Cook N,
Castaneda A: Extending the limits for modified Fontan procedures. J Thorac Cardiovasc Surg 1986;92:1021-1028
4. Humes R, Feldt R, Porter C, Julsrud P, Puga F, Danielson G:
The modified Fontan operation for asplenia and polysplenia
syndromes. J Thorac Cardiovasc Surg 1988;96:212-218
5. Humes R, Porter C, Mair D, Rice M, Offord K, Puga F, Schaff
H, Danielson G: Intermediate follow-up and predicted survival after the modified Fontan procedure for tricuspid atresia
and double-inlet ventricle. Circulation 1987;76 (suppl III):
III-67-III-71
6. Bartmus D, Driscoll D, Offord K, Humes R, Mair D, Schaff H,
Puga F, Danielson G: The modified Fontan operation for
children less than 4 years old. J Am Coll Cardiol 1990;15:
429-435
7. Fontan F, Kirklin J, Fernandez G, Costa F, Naftel D, Tritto F,
Blackstone E: Outcome after a perfect Fontan operation.
Circulation 1990;81:1520-1536
8. Fontan F, Fernandez G, Costa F, Naftel D, Tritto F, Blackstone E, Kirklin J: The size of the pulmonary arteries and the
results of the Fontan operation. J Thorac Cardiovasc Surg
1989;98:711-724
9. Girod D, Fontan F, Deville C, Ottenkamp J, Choussat A:
Long-term results after the Fontan operation for tricuspid
atresia. Circulation 1987;75:605-610
10. Kirklin J, Blackstone E, Kirklin J, Pacifico A, Bargeron L: The
Fontan operation: Ventricular hypertrophy, age, and date of
operation as risk factors. J Thorac Cardiovasc Surg 1986;92:
1049-1064
11. DeVivie E, Rupprath G: Long-term results after Fontan
procedure and its modifications. J Thorac Cardiovasc Surg
1986;91:690-697
12. Choussat P, Fontan F, Besso P, Vallot F, Chauve A, Bricaud H:
Selection criteria for Fontan's procedure, in Anderson R, Shineborne E (eds): Paediatric Cardiology. New York, ChurchillLivingstone, 1977, pp 559-560
13. Kaplan EL, Meier P: Non-parametric estimation from incomplete observations. JAm Stat Assoc 1958;53:457-481
14. Peto R, Peto J: Asymptotically efficient rank invariant procedures (with discussion). J R Stat Soc 1972;135(A):185-207
15. Mantel N: Evaluation of survival data and two new rank order
statistics arising in its consideration. Cancer Chemother Rep
1966;50:163-170
16. Cox DR: Regression models and life-tables (with discussion).
JR Stat Soc 1972;34(B):187-220
17. Bergstralh EJ, Offord KP: Conditional Probabilities Used in
Calculating Cohort Expected Survival. Section of Biostatistics,
Mayo Clinic, Technical Report Series No 37, January 1988
496
Circulation Vol 85, No 2 February 1992
18. Walker SH, Duncan DB: Estimation of the probability of an
event as a function of several independent variables. Biometnka 1967;54:167-179
19. SAS Institute Inc: SAS® User's Guide: Basics. Version 5 ed.
Cary, NC, SAS Institute Inc, 1985
20. Taussig H: Long-term observations on the Blalock-Taussig
operation. The Johns Hopkins Medical Joumal 1976;139:69-76
21. Moodie D, Ritter D, Tajik A, McGoon D, Danielson G,
O'Fallon W: Long-term follow-up after palliative operation
for univentricular heart. Am J Cardiol 1984;53:1648-1651
22. Moodie D, Ritter D, Tajik A, O'Fallon W: Long-term follow-up in the unoperative univentricular heart. Am J Cardiol
1984;53:1124-1128
23. Barber G, Hagler D, Edwards W, Puga F, Danielson G,
McGoon D, Driscoll D: Surgical repair of univentricular heart
(double inlet left ventricle) with obstructed anterior outlet
chamber. JAm Coll Cardiol 1984;4:771-778
24. Hess J, Kruizinga K, Bijleveld C, Hardjowijuvo R, Eygelsar A:
Protein-losing enteropathy after Fontan operation. J Thorac
Cardiovasc Surg 1984;88:606-609
KEY WORDS a single ventricle * univentricular heart a
modified Fontan operation * congenital heart disease * surgery
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
Five- to fifteen-year follow-up after Fontan operation.
D J Driscoll, K P Offord, R H Feldt, H V Schaff, F J Puga and G K Danielson
Downloaded from http://circ.ahajournals.org/ by guest on June 15, 2017
Circulation. 1992;85:469-496
doi: 10.1161/01.CIR.85.2.469
Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 1992 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7322. Online ISSN: 1524-4539
The online version of this article, along with updated information and services, is located on
the World Wide Web at:
http://circ.ahajournals.org/content/85/2/469
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally
published in Circulation can be obtained via RightsLink, a service of the Copyright Clearance Center, not the
Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further
information about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation is online at:
http://circ.ahajournals.org//subscriptions/