Comparison of Cardiopulmonary Adaptation During Exercise in Children After the Atriopulmonary and Total Cavopulmonary Connection Fontan Procedures
Background There are several potential physiological differences between the atriopulmonary (AP) and the total cavopulmonary connection (TCPC) Fontan circulations. Studies suggest that the TCPC reduces energy loss due to turbulence and may have more dependence on respiratory movement for pulmonary blood flow. We compared cardiopulmonary physiology during rest and exercise in patients who had undergone the AP Fontan procedure with those who had undergone the TCPC Fontan procedure.
Methods and Results Forty-three children were studied more than 6 months after undergoing a Fontan procedure (23 AP and 20 TCPC); 106 healthy children were also studied as a control group. Measurements of effective pulmonary blood flow, stroke volume, arteriovenous oxygen difference, minute ventilation, heart rate, and oxygen and carbon dioxide consumption were made with an Innovision quadrupole mass spectrometer. Data from the control group allowed calculation of z scores for the Fontan groups matched for age, sex, pubertal stage, and body surface area. Maximal exercise performance was equal in the two Fontan groups, but it was below normal. However, adaptation to exercise was different in the Fontan groups. After 9 minutes of exercise, pulmonary blood flow rose less in the AP group than in the TCPC group (P<.01), and the stroke volume in the AP group also tended to be lower (P=.057) and their arteriovenous oxygen difference was significantly greater (P<.01). Although minute ventilation per unit of carbon dioxide production was similar in the Fontan groups at this level of exercise, children in the TCPC group breathed faster by approximately 10 breaths per minute (P<.005).
Conclusions At submaximal exercise, children who had undergone the TCPC Fontan procedure had pulmonary hemodynamics superior to those of children who had undergone the AP procedure, largely because of respiratory adaptation that permitted blood to be “sucked” into the lungs. To achieve the same maximal exercise performance, children who had undergone the AP procedure had a superior metabolic adaptation to exercise stress.
Since the Fontan procedure was first described in 1971,1 many modifications have been introduced. Currently there are two main surgical techniques for the procedure. In the atriopulmonary (AP) procedure, a direct connection is formed between the right atrium and main pulmonary artery, providing the potential advantage of the pulsatile action of the right atrium. The total cavopulmonary connection (TCPC), originally proposed by Kawashima et al,2 showed that it was possible to bypass the atrium. Modifications of this technique have been described to direct inferior vena caval flow to the pulmonary artery using an intra-atrial baffle combined with a bidirectional Glenn procedure. In vitro studies indicated that this circulation results in less turbulence,3 4 which may be advantageous in vivo, especially in the presence of a dilated and poorly contracting atrium. Exercise performance is usually reduced compared with that of normal control subjects after the Fontan operation,5 6 7 8 9 10 but findings after AP and TCPC procedures in patients who are otherwise similar have not been compared. Our previous data suggest that the determinants of pulmonary blood flow in the two groups may be significantly different; both groups demonstrate augmentation of blood flow, the augmentation of pulmonary blood flow with inspiration being markedly more important in the TCPC patients.11 12 In this study we therefore compared resting and exercise cardiopulmonary hemodynamics in patients after AP or TCPC Fontan procedures using respiratory mass spectrometry.
Forty-three patients from a cohort of 129 who had undergone a Fontan procedure at the Royal Brompton Hospital or the Hospital for Sick Children between 1979 and 1992 were studied (Table⇓). An AP procedure had been performed in 23 patients and a TCPC procedure in 20. No patient had evidence of pathway obstruction or residual shunt. Selection criteria were: Each patient was (1) a resident of England or Wales, (2) at least 6 months post–Fontan procedure, (3) >7.5 years old and >125 cm tall (for use of exercise bicycle), (4) free of complicating acute or chronic respiratory or other conditions that might affect exercise performance, and (5) free of evidence of pathway obstruction or residual shunt. No other selection criteria were used, and 80% of eligible children were studied. As a control group, 106 healthy children (55 boys) from three local schools were also studied. They were matched for age, sex, height, and weight to the Fontan patients.
Each child had his or her height and weight measured with a Harpenden stadiometer (Holtain Ltd) and electronic scales (SECA), respectively. In addition, the stage of puberty was assessed,13 two-site skinfold measurements (triceps and subscapular) were made with a Holtain skin caliper (Holtain Ltd), and three flow-volume loops were measured (Compact Vitallograph), the best being documented according to American Thoracic Society standards.14
After local ethical committee approval was given and written informed consent from each subject’s parents was obtained, a standard protocol was used for all children. They arrived at the laboratory having fasted for 1 hour. After anthropometric and flow-volume loop measurements were made, the children practiced with the equipment. Once they were confident with the technique involved, they rested for 10 minutes and then performed, from functional residual capacity, five 20-second rebreathing maneuvers every 3 minutes for 15 minutes; a metronome was used to control the respiratory rate to 40 breaths per minute. The rebreathing bag (with a volume of 40% of predicted vital capacity17 ) contained 35% oxygen, 5% sulfur hexafluoride, 0.3% acetylene, and 0.3% carbon monoxide labeled with 18O. The balance was nitrogen. At 0.3%, acetylene gives repeatable results15 and is both odorless and nonexplosive. Functional residual capacity was determined with a calibrated pneumotachograph linked to a pair of pneumatic valves that automatically switched breathing from room air to the rebreathing bag. During the maneuvers, the continuous pulse rate and arterial saturation were measured with a surface oximeter (Nellcor) placed over the right supraorbital artery.
After the resting measurements were made, the subject exercised on an electromagnetic bicycle (Seca 100), electrically calibrated before every study, that produced a constant workload independent of pedal speeds over the range of 6 to 150 rpm. After a 3-minute rest, the subject performed a 12-second rebreathing maneuver (the shortened time ensured that carbon dioxide did not build up sufficiently during rebreathing to cause distress during exercise) and then began cycling, initially backward at zero load to loosen up and then forward at 25 W/m2, increasing in 15-W/m2 increments every 3 minutes until exhaustion. During the last 20 seconds of each 3-minute stage, the subject performed a 12-second rebreathing maneuver while continuing to pedal. For consistency, the child was encouraged to pedal at a rate between 50 and 70 rpm during the entire duration of exercise. At exhaustion, the child stopped pedaling but remained on the bicycle for 9 more minutes, performing three more 12-second rebreathing maneuvers. During the initial rest, exercise, and recovery phases, continuous mixed– expired gas analysis was undertaken when the subject was not performing rebreathing maneuvers. A typical study lasted 75 minutes, during which time the child performed 12 to 16 rebreathing maneuvers.
Respiratory Mass Spectrometry
Mass spectrometry is used to measure the fractional concentrations of the components of a gaseous mixture on the basis of their mass-to-charge ratio alone.16 An Innovision 2000 quadrupole mass spectrometer was used to continuously sample each subject’s ventilated gas at a rate of 5 mL/min down a 0.2-mm-ID Teflon tubing. The individual components of the gas were analyzed semicontinuously, with each component gas requiring a minimum of 5 milliseconds for analysis, the data from the first 2 milliseconds being discarded to eliminate errors due to machine equilibration. During the study, each gas was analyzed for at least 10 milliseconds; acetylene and carbon monoxide were analyzed for 15 and 20 milliseconds, respectively. All gases in this study were therefore measured at least 9.7 times per second. The delay from sampling to measurement was 450 milliseconds. The stable isotope of carbon monoxide, CO labeled with 18O, was present to measure transfer factor, but those data are not part of the present study. During the exercise study of each subject, the device underwent a two-point calibration three times, by exposure to both a calibration gas and zero gas (vacuum), and a one-point calibration at least eight times. The mass spectrometer was controlled with an IBM clone PC 386DX with a coprocesssor to ensure rapid data handling during the rebreathing maneuvers, and the results were downloaded and analyzed by customized software.
The physiological measurements and equations are detailed in the “Appendix.” During rebreathing, the absorption of acetylene was used to calculate the effective pulmonary blood flow in contact with ventilated alveoli,17 18 19 20 21 and the absorption of oxygen was used to measure oxygen consumption. With the Fick equation, the ratio of oxygen consumption to pulmonary blood flow can be used to calculate the arteriovenous oxygen content difference. Helium-dilution mixed–expired gas analysis22 allows calculation of oxygen consumption, carbon dioxide production, minute and alveolar ventilation, and the anaerobic threshold.23 The maximum work performed by the patients was expressed as a percentage of the median maximum work performed for each sex and age group.
All traces were visually checked to ensure that the software had both determined the point of complete mixing of the rebreathing and lung gases and excluded pulmonary blood recirculation. This was determined to have occurred correctly in 2933 of 2943 measurements. The first 40% of each expired breath was excluded from data analysis because it is composed largely of dead-space gas. The last 20% was also excluded because it represents underventilated lung elements.
Data from the healthy control subjects were used to derive means, corrected for age, surface area, and sex, for all the parameters at rest and during each exercise stage. Resting values were determined from the average of the last three rebreathing measurements at rest.17 Children were divided into four age groups (8 to 10.5, 10.6 to 12.5, 12.6 to 14.5, and more than 14.5 years) and three pubertal groups (pre-, early, and late puberty based on Tanner stages 1, 2 to 3, and 4 to 5, respectively). From these normal results, z scores were calculated so that for the control children, evaluated as a group, the mean z score for any parameter during rest or any stage of exercise was 0 with an SD of 1. Thus, during exercise, even though all raw parameter values may have increased in the control group, the mean z score remained 0. A deviation from a mean of 0 in a patient group during exercise was interpreted as the patient group’s failing to match the changes expected rather than as an absolute fall in that parameter.
z score results are presented as means with 95% CIs.24 Other results are expressed as medians with 95% CIs.25 z score differences between patient groups and control subjects at rest and at each exercise stage were analyzed by a one-way ANOVA with Duncan’s correction for multiple contrasts; z score differences within each patient group were analyzed by a two-way ANOVA with the subject as a blocking variable and Duncan’s correction for multiple contrasts; and z score differences between patient groups at each rest and exercise stage were analyzed using the unpaired Student’s t test. The null hypothesis was rejected when P<.05. This methodology was used to strike a balance between the effects of multiple contrasts and the maximum reduction in the error. Other group differences were analyzed by the Mann-Whitney U and χ2 tests.
Characteristics of the study groups are summarized in the Table⇑. The AP group was slightly but not significantly older (median, 14.1 years; 95% CI, 11.0 to 16.1) than the TCPC group (median, 10.9 years; 95% CI, 10.0 to 12.4; P=.10) with a nonsignificant excess of males (P=.23). These differences were eliminated when age- and sex-specific z scores were determined. Although age at Fontan surgery was similar, follow-up was significantly longer in the AP group (P<.005). The TCPC group had a preponderance of complex heart disease rather than classic tricuspid atresia compared with the AP group (P<.01).
At rest, effective pulmonary flow was markedly lower in subjects who had undergone a Fontan procedure than in control subjects (P<.001), as expected. There was no significant difference between the two Fontan groups (AP mean, 2.2 L · min−1 · m−2; TCPC mean, 2.3 L · min−1 · m−2). At 40 W/m2 of exercise (9 minutes of exercise in total), the increase in effective pulmonary blood flow seen in AP subjects, which was significantly less than the increase in control subjects (P=.05), was less than that in the TCPC patients (P<.01; AP mean, 3.7 L · min−1 · m−2; TCPC mean, 4.8 L · min−1 · m−2) (Fig 1A⇓). This trend was maintained at 55 W/m2, but the difference was not significant because of the smaller number of subjects reaching this workload. Resting heart rate was significantly higher in both Fontan groups at rest than in control subjects, but the rise during exercise was less than that in control subjects (P<.05 for AP, P<.01 for TCPC) (Fig 1B⇓). At rest, stroke volume in both Fontan groups was lower than in the control subjects (P<.0001). The increase in stroke volume during exercise was slightly greater in the TCPC than in the AP patients, but the difference just failed to reach significance (P=.057) at 40 W/m2 (Fig 1C⇓). The difference was approximately 5 mL per beat per meter squared and was maintained at the next exercise stage.
Oxygen Consumption and Arteriovenous Oxygen Difference
Resting oxygen consumption was normal in the AP patients but significantly lower (approximately 0.01 L · min−1 · m−2) in the TCPC group (P<.01) (Fig 2A⇓). During exercise, oxygen consumption per unit of work was lower in both Fontan groups compared with control subjects. This difference was most marked for the AP patients (P<.01). At rest and throughout exercise, the arteriovenous oxygen difference was significantly higher in both Fontan groups compared with controls (P<.001) (Fig 2B⇓). Furthermore, the increase in the arteriovenous oxygen difference was consistently greater in the AP group than in the TCPC group (P<.01).
At rest, both Fontan groups had a minute ventilation higher than controls’ (P<.01) (Fig 3A⇓). At low levels of exercise, minute ventilation in both Fontan groups remained significantly higher than in the control group, but this difference was lost at higher workloads. However, the ventilatory pattern was very different between the two Fontan groups. In the TCPC group, there was a significant increase in respiratory rate (Fig 3B⇓) early in exercise (P<.03) with a lower tidal volume (Fig 3C⇓), so that at 40 W/m2 there was a difference of approximately 10 breaths per minute. In contrast, the AP patients followed a normal pattern with no difference in respiratory rate and tidal volume compared with control subjects.
The change in respiratory rate in TCPC patients could not be explained by increased carbon dioxide production, because both Fontan groups had the same minute ventilation per unit carbon dioxide production (Fig 3D⇑). The AP group’s anaerobic threshold z score was significantly lower than the control group’s, but the TCPC group’s was not. In the control group the mean was −0.03 (95% CI, 0.17 to −0.23); in the AP group it was −0.77 (95% CI, −0.27 to −1.27; P<.05); and in the TCPC group it was −0.35 (95% CI, 0.09 to −0.78; NS).
Maximum Exercise Performance
There was no difference in maximum workload between the two Fontan groups (Fig 4⇓); it was subnormal in both compared with the control group (P<.001). There was no effect of age at surgery, time from surgery, age at testing, or surgical diagnosis, although for the Fontan group as a whole, females tended to perform better than males (P<.07).
At maximum exercise, stroke volume (Fig 5A⇓) and effective pulmonary blood flow (Fig 5B⇓) were both significantly higher (P=.05 and P<.001, respectively) in the TCPC group than in the AP group, but in the TCPC group the arteriovenous oxygen difference was lower (P<.05) (Fig 5C⇓). These differences persisted 6 minutes into recovery. Respiratory rate was significantly higher in the TCPC group until 3 minutes after exercise (P<.05) (Fig 5D⇓).
Comparison of cardiopulmonary function in patients who had undergone AP or TCPC Fontan procedures indicated that the two groups of patients had similar maximum exercise performance but different physiological adaptations to exercise. Maximum work, aerobic capacity, stroke volume, and heart rate response after both Fontan procedures were markedly less, when these patients were evaluated as a group, than those of a control group of healthy children, although many of the Fontan patients’ individual z scores were in the normal range.
In the AP group, there was metabolic adaptation even at low workloads. Their cardiac output at rest was lower than in the control subjects, and this difference increased with increasing exercise. Predictably, their anaerobic threshold was lower than in the control subjects, but tissue oxygen extraction was maintained. Consequently, during exercise the AP patients developed a markedly abnormal arteriovenous oxygen difference that was significantly greater than that of TCPC patients or control subjects at all workloads. The mechanism for their relatively high tissue oxygen uptake in the face of reduced oxygen delivery is unclear but is important to their ability to sustain exercise despite a cardiac output significantly lower than TCPC patients’.
The physiological response to exercise in the TCPC group was quite different. These patients had a higher effective pulmonary blood flow with a lower arteriovenous oxygen difference and a respiratory pattern significantly different from that of AP patients. TCPC patients take smaller but more frequent breaths. Although the difference is small (0.01 L · min−1 · m−2), their lower resting oxygen consumption at rest compared with AP patients’ is an unexpected finding. Because both Fontan groups had slightly higher heart and respiratory rates at rest compared with controls, a lower resting metabolic rate is an unlikely explanation, although catecholamine levels and thyroid status were not investigated. Methodological bias due to investigator training effects or sequence bias is also unlikely because the children were tested in random order.
We have previously shown that postoperative pulmonary blood flow after a TCPC procedure is clearly dependent on respiratory motion.12 Pulmonary blood flow increases during normal inspiratory effort and is augmented during both the Muller maneuver (inspiration against a closed glottis)12 and negative pressure ventilation.26 Although after the AP procedure, resting pulmonary blood flow is slightly augmented by respiration,11 the work of breathing appears to be a much more important energy source for pulmonary blood flow after the TCPC procedure. The present study indicates that this respiratory mechanism is also important during exercise. Starting with the onset of exercise, TCPC patients become relatively tachypneic compared with both normal subjects and AP patients, and this finding holds at all workloads. This is not a physiological response to excess carbon dioxide production (Fig 3D⇑); it may be an adaptive response to harness the energy of the “respiratory pump” to generate pulmonary blood flow. At maximal workload, the respiratory rates of the Fontan groups converge, indicating that there may be an optimal respiratory rate beyond which no further increase in cardiac output is possible. Experimental studies have indeed shown that the energy efficiency of the TCPC circulation, by avoiding energy loss due to turbulence, may be superior to that of the AP connection.3 4
Circulatory adaptation to submaximal workloads is more relevant to ordinary daily life than maximal exercise performance, so even though the two Fontan procedures provide similar maximum performance, the TCPC procedure results in a more efficient circulation, at least in the short term. Nevertheless, this advantage may be lost with even minor pathway obstruction (there was none in our study subjects) because of the lower mechanical reserve in the TCPC group.
If maximal exercise performance is achieved in AP patients by generation of a large arteriovenous oxygen difference, it is difficult for this to be enhanced by exercise training. Exercise performance in some patients who have undergone the TCPC procedure may be limited by an inability to enhance cardiac output by increased respiratory movement. However, they may still be able to increase oxygen extraction and increase their arteriovenous oxygen difference by an exercise training program.
There are unavoidable differences between the study groups. A randomized study comparing the two treatment approaches would be ideal but impractical. Selection criteria and institutional differences lead to an excess of patients with tricuspid atresia in the AP group, whereas the TCPC group includes subjects with more complex intracardiac anatomy. However, a relationship between intracardiac morphology and exercise performance after AP connections6 has not been demonstrated. The interval between Fontan procedure and study participation was slightly greater in the AP group, and there were small differences in length of follow-up between the groups, although length of follow-up was unrelated to performance in this study, confirming the findings of Nir et al.27 The differences in exercise performance are thus likely to be due to the type of surgical connection.
We conclude that, although children who have undergone either the AP or the TCPC Fontan procedure have the same maximum exercise performance, the TCPC procedure, though dependent on respiratory movement, results in a more efficient circulation at exercise levels relevant to ordinary daily life.
Effective Pulmonary Blood Flow
The absorption of acetylene from a closed rebreathing system is monoexponential and proportional to the pulmonary blood flow perfusing ventilated alveoli.17 18 19 20 21 This effective pulmonary flow, Qeff, in liters per minute, can be calculated with the equation
where βAc is the slope of the natural logarithm of the disappearance of acetylene with time corrected for changes in the total volume of gas within the lung/rebreathing bag system, which are calculated from changes in the fractional concentration of a nonsoluble inert gas (in this study, sulfur hexafluoride). Fi0 is the volume of the inert gas at the start of an experiment, and Fieq is the volume of the gas at equilibrium of the test gas mixture with the subject’s lung gases. VRB is the total volume (in liters) of the test gas mixture at the start of the experiment; intAc is the intercept of the disappearance curve of acetylene with time extrapolated back to time 0 once mixing of the test and native gases has occurred; PB is ambient atmospheric pressure, measured in mm Hg; and αb is the solubility constant of acetylene in blood.
The absorption of oxygen from the same closed system is also monoexponential and can be used in the following equation to calculate consumption in L/min16 :
Mixed–Expired Gas Analysis
For helium-dilution mixed–expired gas analysis,22 a 12-L baffled mixing box is used. The principle is that the addition of a tracer gas at a known flow rate into the stream of expired gas from a subject, followed by its remeasurement after perfect mixing, allows calculation of the flow rate of the expired gas and its components. The faster the flow of expired gas, the more dilute the tracer gas becomes. For example, minute ventilation (VE) is calculated as
where VTr is the flow rate of the tracer gas and FTreq is the fractional concentration of tracer gas after complete mixing with the expired gas. The term 1−FTreq is the correction for the addition of tracer gas to the system, and further corrections need to be added to the equation for ambient temperature, pressure, and humidity. In this study, helium was the tracer gas, used at a typical flow rate of 300 mL/min, and the equipment was calibrated for every study with a 3-L syringe (Hans Rudolph Inc). The coefficient of flow variability was always less than 2.5%. In a similar manner, carbon dioxide production can be calculated as well as oxygen consumption.
This study was funded in part by grants from Innovision (Denmark) and the British Heart Foundation (PG/91088). We are grateful to all parents and children who cooperated so admirably; to Cnud Pedersen for software modifications; and to Messrs Lincoln, Shore, Stark, Elliot, and DeLeval for allowing us to study their patients.
- Received May 2, 1994.
- Accepted August 15, 1994.
- Copyright © 1995 by American Heart Association
Fontan F, Baudet E. Surgical repair of tricuspid atresia. Thorax. 1971;26:240-248.
Cumming GR. Maximal exercise haemodynamics after the Fontan procedure. In: Doyle EF, Engle MA, Gersony WM, Rashkind WJ, Talner NS, eds. Pediatric Cardiology. New York, NY: Springer-Verlag; 1986:257-266.
Shacher BG, Fuhrman BP, Wang Y, Lucas RV, Lock JE. Rest and exercise hemodynamics after the Fontan procedure. Circulation. 1982;65:1043-1048.
Penny DJ, Redington AN. Doppler echocardiographic evaluation of pulmonary blood flow after the Fontan operation: the role of the lungs. Br Heart J. 1991;65:372-374.
Redington AN, Penny D, Shinebourne EA. Pulmonary blood flow after total cavopulmonary connection. Br Heart J. 1991;65:213-217.
Tanner JM. Growth at Adolescence. 2nd ed. Oxford, UK: Blackwell Scientific Publications, Inc; 1962.
Schied P. Respiratory mass spectroscopy. In: Laszlo G, Sudlow MF, eds. Measurement in Clinical Respiratory Physiology. London, UK: Academic Press, Inc; 1983:163-190.
Bush A, Busst C, Johnson S, Denison DM. Rebreathing method for the simultaneous measurement of oxygen consumption and effective pulmonary blood flow during exercise. Thorax. 1988;43:268-275.
Cander L, Forster RE. Determination of pulmonary parenchymal tissue volume and pulmonary capillary flow in man. J Appl Physiol. 1959;14:541-551.
Petrini MF, Peterson BT, Hyde RW. Lung tissue volume and blood flow by rebreathing: theory. J Appl Physiol. 1978;44:795-802.
Bowyer JJ, Warner JO, Denison DM. Effective pulmonary blood flow in normal children at rest. Thorax. 1988;43:972-977.
Beaver WL, Wasserman BJ, Whipp A. A new method for detecting anaerobic threshold by gas exchange. J Appl Physiol. 1986;60:2020-2027.
Gardner MJ, Altman DG. Confidence intervals rather than P values: estimation rather than hypothesis testing. Br Med J. 1986;292:746-750.
Campbell MJ, Gardner MJ. Calculating confidence intervals for some non-parametric analyses. Br Med J. 1988;296:1454-1456.