Surgical right ventricular outflow tract (RVOT) reconstruction during repair of tetralogy of Fallot (TOF) often results in pulmonary valve (PV) insufficiency (PI) and, in the long-term, progressive right ventricular (RV) dysfunction due to chronic volume overload (1). Most clinical studies focusing on the sequelae of TOF repair have found that use of a transannular patch (TAP) is a major determinant of late RV dilation (2). Its detrimental role is even more pronounced by the additional development of extended RVOT akinesia or aneurysm, often occurring in patients who were operated on in former eras, when a large ventriculotomy was used for closure of the ventricular septal defect and relief of the infundibular obstruction (3). To overcome these late effects, recent surgical management of TOF involves pursuing a RVOT-sparing approach, including maximal preservation of PV function and/or minimal infundibular scarring (4- 6). However, the long-term results of these strategy preferences are unknown.

The chronic effects of pulmonary insufficiency–related volume overload on RV function have already been studied in animal models (7- 9). Kuehne et al. (8) demonstrated impaired biventricular systolic function and decreased RV contractility in growing swine, in relation to pulmonary insufficiency, through transcatheter stent implantation across the PV. Hence, in contrast with the usual clinical setting, the impact of the commonly associated surgically induced infundibular dysfunction (INF) on RV function has rarely been studied.

We developed an experimental model in growing pigs to investigate the differential contribution of dysfunction of each RVOT component on RV function. The goal was to mimic properly the acute and chronic physiological effects of surgical RVOT reconstruction. Hemodynamic assessment was performed by using conductance catheter technology for quantification of RV volumes and indices of systolic and diastolic RV performance.

The study protocol was performed according to the standards of The Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (publication 85-23, revised 1996) and approved by the local ethics committee of the University Hospital of Ghent (Ethical Committee Laboratory Animals 08/30).

The experimental model included 16 Landrace pigs (Rattlerow-Seghers, Lebbeke, Belgium). Three groups with 4 pigs each underwent a surgical RVOT dysfunction. In the PI group, an isolated PI was created by excision of 1 anterior PV leaflet through a transverse pulmonary arteriotomy. In the INF group, INF was obtained by infundibulotomy and closure with a polytetrafluorethylene patch of 30 by 20 mm (Gore-Tex, Gore & Associates, Newark, Delaware). The third group (TAP) combined a PI by excision of 1 leaflet and INF by using a transannular plasty with a 40-mm long polytetrafluorethylene patch. In both the INF and TAP groups, the length of infundibular incision was made equal at 25 to 30 mm. Four animals served as a control group and were sham-operated (SHAM). None of the animals had any RVOT obstruction. The conduct of the study protocol is depicted in (Figure 1).

After premedication with intramuscular tiletamine and zolazepam in a combined solution with xylazine 2% (0.2 ml/kg), anesthesia was induced with intravenous propofol 3 mg/kg, sufentanil 0.005 mg/kg, and rocuronium bromide 1 mg/kg. After endotracheal intubation, the animals were mechanically ventilated with FiO₂ 40% and tidal volume of 0.1 to 0.15 l/kg. Anesthesia was maintained with continuous end-tidal sevoflurane 2.5% administered by using the AnaConDa System (Sedana Medical A.B., Sundbyberg, Sweden) and eventual additional boluses of sufentanil 0.005 mg/kg. Basic monitoring included electrocardiogram, body temperature, and ventilatory CO₂ emission through capnography. Oxygenation was controlled by arterial blood gas sampling. Hemodynamic monitoring included continuous arterial pressure through an 8.5-F catheter into the left carotid artery and central venous pressure through a 7.5-F catheter into the right atrium via external jugular vein puncture.

Via a left thoracotomy, the different types of RVOT dysfunction (as described earlier) were performed with the use of right heart bypass and partial clamping of the RVOT. In SHAM animals, only right heart bypass was installed during an equal time period. At the end of the procedure, the animals were extubated and treated with intramuscular buprenorphine 0.03 ml/kg and an intercostal block with levobupivacaine 5 mg/ml.

A follow-up interval of 3 months was used to determine the chronic phase, according to the rapid organ maturation and growth of these animals (8,10). At that time, hemodynamic effects were assessed in a way similar to the acute phase, but cardiac exposure was obtained using median sternotomy.

In the acute, postsurgical phase, a 7-F dual-field pressure-volume catheter (CD Leycom, Zoetermeer, the Netherlands) was introduced into the right ventricle from the RVOT directed toward the apex. Correct catheter positioning was confirmed by using radioscopy. The conductance catheter was connected to a Sigma M module and digitized at 250 Hz for online computer analysis with Conduct NT CFL-512 software (CD Leycom). A 16- or 20-mm perivascular flow probe (Transonic Systems Inc., Ithaca, New York) was placed around the pulmonary trunk for cardiac output measurement. Preload modulation was achieved by radioscopically guided placement of a pulmonary artery balloon catheter (PTS-303 NuMED, Heart Medical Europe B.V., Best, the Netherlands) into the inferior vena cava via puncture of the right jugular vein. In the chronic phase, conduction of the conductance catheter measurements was similar, but inferior vena cava occlusion was done surgically by using a tourniquet.

Acquisition of pressure and volume data were obtained at end-expiration. Volume calibration was performed by integration of slope factor α for cardiac output and by parallel conductance during injection of 0.02 ml/kg of hypertonic saline. Baseline measurements included end-systolic and end-diastolic RV volumes with subsequent calculation of stroke volume, ejection fraction, and cardiac output. To correct for growth variability between the animals over the groups, volume-dependent variables were indexed to body surface area (BSA), following the equation of Kelley et al. (11) for swine: BSA (cm²) = 734 × weight (kg)^0.656.

Based on the instantaneous pressure-volume relationship changes during transient occlusions of the inferior vena cava, RV contractile function was quantified by the slope (M_w) of the preload recruitable stroke work and the slope (E_max) of the end-systolic pressure-volume relationship (ESPVR). In addition, the volume intercept of the ESPVR was determined at the pressure level of 25 mm Hg. Only recordings with <10 % change of heart rate and a correlation coefficient of the linear regression line r² > 0.90 were considered eligible.

Evaluation of diastolic RV function was based solely on passive ventricular compliance, as active isovolumic relaxation (τ) was not reliable in the groups with PI. RV compliance was expressed as chamber stiffness constant β, derived from the exponential fit of the end-diastolic pressure-volume relationship.

Magnetic resonance imaging (MRI) was performed with a 1.5-T imager system (Siemens Avanto, Erlangen, Germany) with maximum gradient strength of 200 mT/m and a slew rate of 45 mT/m/ms. Under anesthesia with intravenous propofol 1% (3 mg/kg) and manual-hold ventilation via endotracheal intubation, the animals were positioned on the MRI table in right lateral position. The MRI protocol consisted of electrocardiography-triggered TrueFISP images in 3 orthogonal planes, electrocardiography-gated TrueFISP cine imaging of long and short cardiac axes, and phase-contrast acquisition through plane flow mapping perpendicular to the blood flow in the aorta and in the pulmonary artery. If the TrueFISP images included artifacts produced by vivid aortic flow, a spoiled gradient echo sequence was used instead. All images were taken during end-expiratory ventilation stop. Left ventricular and RV volume measurements on the short axis stack, as well as postprocessing of phase-contrast images, were performed by a single investigator with the use of Argus software package (Siemens). MRI allowed validation of the RV volumes obtained by the conductance technique as well as determination of the PI grade. Pulmonary regurgitation fraction was calculated as the ratio of pulmonary retrograde to antegrade flow volume (expressed as a percentage).

At the end of the study, the animals were euthanized with an intravenous solution of embutramide 200 mg, mebenzoniumiodide 200 mg, tetracaine hydrochloride 5 mg, and dimethylformamide 1 mg (T61) at a dose of 0.3 ml/kg. The heart was harvested, and tissue samples of the basal and free wall of the right and left ventricles and a papillary muscle of the tricuspid valve were excised for histological analysis. Hematoxylin-eosin and Masson's trichrome staining were used for delineation of interstitial collagen extent.

All data are expressed as mean ± SD. Data distribution was assessed for normality by using the Shapiro-Wilk test. Normal distributed data were compared between groups by using 1-way analysis of variance, with Tukey or Dunnett's T3 correction for multiple comparisons, depending on homogeneity of variance. The change of each variable per individual animal between the acute and chronic phase is expressed as the proportional change between both variables in percentages. This calculated variable was also analyzed by using 1-way analysis of variance with post-hoc Tukey correction. For nonnormally distributed data, group comparisons were made with Kruskal-Wallis analysis and subsequent Mann-Whitney testing, with adjustment for multiple comparisons for defining intergroup differences. Evolution of RV function variables for individual animals between the first and second phase was evaluated with a paired t test or Wilcoxon signed rank test.

Bland-Altman processing was used for validation of agreement of volume measurement between MRI and conductance technique (12). Statistical analysis was conducted by using SPSS version 19.0 software (SPSS Inc., Chicago, Illinois). A p value <0.05 was considered significant.

At the start, all pigs had a comparable weight and BSA, according to their age of 8 to 9 weeks. After 3 months, some growth variability resulted in a relative BSA increase of 95 ± 28% in the SHAM group, 105 ± 12% in the INF group, 117 ± 6% in the PI group, and 121 ± 8% in the TAP group (p = 0.15). Although these differences were not significant, the use of indexed RV volumes seemed more appropriate to differentiate growth versus dilation. In addition, the extent of surgical infundibulotomy was identical in the INF and TAP groups, shown by an indexed incision length of 47.2 ± 1.6 mm/m² and 46.6 ± 2.3 mm/m² (p = 0.65), respectively.

Measurement of RV volumes showed a close reliability between the conductance method and MRI, with a correlation coefficient of 0.93 (p < 0.0001) and 0.85 (p < 0.0001) for end-diastolic volume and end-systolic volume, respectively. Bland-Altman analysis confirmed agreement between both methods, with a bias of 7.6 ml and limits of agreement (–18.9 to 58.3) for ESV, and a bias of 6.6 ml and agreement limits (–17.2 to 53.0) for EDV.

Postsurgical PI was only present in the PI and TAP groups (pulmonary regurgitation fraction [PRF] 29 ± 4% and 22 ± 9%; p = 0.14), respectively. After 3 months, PRF changed to 34 ± 5% and 36 ± 13%, resulting in significantly greater PRF progression in the presence of concomitant INF (14 ± 2% vs. 54 ± 24%; p = 0.03). Isolated infundibuloplasty did not induce PV dysfunction.

The PI in the PI and TAP groups induced a direct significant increase of mainly indexed end-diastolic volume and furthered RV dilation in the chronic phase, as evident from the significant increase of indexed end-systolic volume and indexed end-diastolic volume compared with the SHAM group and their postoperative volumes. Late indexed end-systolic volume was even significantly larger in the TAP group than in the PI group (p = 0.004). RV volumes of the INF group evolved similarly to the control group. Although stroke volume was increased in the TAP group versus the SHAM group (p = 0.03), the relative stroke volume index increase was not statistically different between groups. Conversely, RV ejection fraction decreased significantly (by 6 ± 3% and 9 ± 6%) compared with baseline values in the PI and TAP groups, respectively, whereas ejection fraction remained unchanged in both other groups. RV cardiac output decreased in both groups with PI, achieving nearly significant difference in the TAP group (16 ± 6%; p = 0.05), mainly due to physiological adoption of lower heart rates.

RV pressures decreased significantly between the acute and chronic phase in all groups and was more pronounced for diastolic pressures (end-systolic pressure p = 0.015; end-diastolic pressure p < 0.0001) (Table 1).

Surgical RVOT dysfunction induced acute impairment of RV contractile function as shown by the decreased E_max and M_w slope, with regard to the SHAM group. Compared with isolated PI, INF in the TAP and INF groups resulted in a lower E_max and M_w at the early stage, but the difference with the PI group was not significant. RV contractility decreased further after 3 months and was proportionally most significant in the TAP group as E_max and M_w decreased 64 ± 22% and 37 ± 4%, respectively, in relation to previous values (Table 2). According to the decrease in ESPVR slopes, V₂₅-intercepts shifted to the right, consistent with progressive deterioration of contractility in all treatment groups. In the chronic phase, the PI group showed less impaired contractility compared with the TAP group, as evidenced by a significantly steeper E_max slope (0.47 ± 0.12 vs. 0.20 ± 0.05; p = 0.02) and M_w slope (13.3 ± 1.6 vs. 8.9 ± 1.4; p = 0.01). The lower M_w and E_max slope of the INF and TAP groups illustrated the significant impact of surgical infundibular injury on RV contractile function compared with the reference group SHAM. Despite the absence of a secondary volume remodeling process, both contractility indices were still lower after 3 months compared with the PI group.

RV compliance improved in all groups after 3 months (Table 2). However, RV chamber stiffness decreased relatively more in the PI and TAP groups (84 ± 6% and 91 ± 5%, respectively) compared with the SHAM group, in relation to the extent of secondary RV dilation and increasing PRF in both groups.

Macroscopic evaluation after 3 months showed a patent RVOT in all animals, with only scar fibrosis around the patched area in the INF and TAP groups. In animals with PI, the most anterior PV leaflet was lacking and was associated with effective PV annulus disruption in the TAP group. Both posterior leaflets were intact. The PV was anatomically and functionally normal in the SHAM and INF groups.

Microscopy revealed normal myocyte architecture and absence of interstitial fibrosis in the SHAM group. However, increased interstitial collagen proliferation and enhanced myocardial bundle disarray were seen in the INF and TAP groups, whereas these changes were only mildly present in the PI group. Tissue samples are illustrated in (Figure 2).

In this study in a growing swine model, RV function was assessed in relation to dysfunction of the different anatomical components of the RVOT. Effective surgically induced PI resulted in secondary RV remodeling, affecting mainly the end-diastolic volume in the acute phase. After 3 months, RV dilation was associated with regression of RV ejection fraction by the higher ratio of end-systolic to end-diastolic volume increase. The presence of INF concomitant to PI promoted the RV dilation process, merely by increasing the PRF. Because isolated INF had no effect on RV geometry, and because RV remodeling was less pronounced in isolated chronic PI, the role of the preserved structural integrity of the infundibulum in sustaining the PI-related volume overload is suggestive. These data are in accordance with those of Kuehne et al. (8), who found identical changes in RV size along with impaired pump function. In their model, isolated PI was created by PV stenting, which resulted in 33% PRF. This acute PI was higher compared with the one in our series, as well as with the usual early clinical observation. Moreover, PRF increased to 49% after 3 months, corresponding to 50% PRF progression in their experiment. This was only seen in our study by adding INF, which renders the comparison with the physiological sequelae of TOF after TAP repair more consistent.

Based on the analysis of indices such as preload recruitable stroke work and ESPVR, any RVOT dysfunction resulted in instant impairment of RV contractile performance. Contractile function seemed particularly more attenuated through infundibular dysfunction, certainly due to intrinsic myocardial damage of that part of the right ventricle. Even though further deterioration was enhanced by PI-related volume load, in correlation with the extent of the secondary RV remodeling process, contractility still seemed to be superior by preserving infundibular function. At last, systolic RV performance was worst after TAP reconstruction. This observation is currently in line with the clinical findings of Frigiola et al. (13), who demonstrated reduced RV contractility in relation to the degree of pulmonary regurgitation and subsequent RV dilation, occurring mainly after TAP repair of TOF patients. Here, they used isovolumic myocardial acceleration, which is a tissue Doppler–based parameter that has been previously validated as a load-insensitive marker of contractility, comparable to the conductance-derived slopes E_max and M_w (14).

In contrast with other animal models addressing the RV properties after chronic PI (8- 9), a decreased RV contractile state was noted after isolated PI at baseline conditions. Although contractility indices were lower at rest, contractile reserve only seemed to be impaired during dobutamine stress testing in these studies. This distinct finding might be attributed to differences in experiment conduction. First, the placement of the conductance catheter, which was aligned from the RV inlet to the apex in their studies, might have underestimated the real contribution of the infundibulum when challenged by the volume overload. Previous reports examined the effect of catheter position in the inhomogeneous right ventricle and demonstrated that a larger part of the ventricular cavity is assessed by placing the catheter in the RV apex from the RVOT (15- 16). From this point of view, the measurements in our study probably reflect more adequately the functional changes at the level of interest; namely, the RVOT. Second, the technical method of achieving PV dysfunction in our experiment might have been more detrimental for RV function. However, the use of mechanical right heart support could not be incriminated for its influence on RV contractility because the indices were significantly altered compared with the SHAM animals, which were also subjected to the myocardial effects of partial circulatory assistance.

Regarding diastolic performance, RV compliance increased proportionally the most in correlation with increasing PI-related volume overload and RV dilation. This finding is in accordance with Pasipoularides et al. (17), who demonstrated improved passive myocardial compliance in a canine model of volume overload, due to significant tricuspid regurgitation. In addition, the experiment of Kuehne et al. (8) confirmed the transition of a noncompliant restrictive right ventricle to a compliant nonrestrictive right ventricle as a natural adaptation mechanism to volume overload. Nonetheless, the impact of cardiopulmonary bypass on diastolic RV properties must be assumed in our model, as the decreases in the stiffness coefficient in the control group were equal to the decreased end-diastolic RV pressures at 3 months.

The functional RV properties of the different groups correspond globally to the histological changes after 3 months. The impaired contractility observed in the 3 groups of RVOT dysfunction correlated with the increased presence of myocardial bundle disarray and interstitial fibrosis, suggesting some secondary remodeling at the cellular level. Similarly, modest increase in collagen concentration, with preserved diastolic compliance, was noticed in a pig model of volume overload through an arteriovenous fistula (18). Our results further support the clinical findings of Babu-Narayan et al. (19) that the extent of late cellular RV remodeling, indicated by fibrosis, is adversely related to the systolic RV function after TOF repair. Moreover, this process seemed to be advanced by the magnitude of the surgical RVOT insult as by progressive ventricular dilation.

In clinical practice, the type of applied RVOT reconstruction is mainly determined by the RVOT morphological features, including the degree of RV hypertrophy. Hence, a TAP, with its subsequent deleterious long-term expectations, is still used clinically in many patients (20).

The current study confirmed that a TAP affects the RV performance the most by combining the adverse early effect on myocardial contractility by INF and, on top of this, the gradual RV volume remodeling due to chronic and progressive PI. Our data support therefore the adoption of a RVOT-sparing approach, although it remains open whether this should be achieved by pursuing a PV-sparing versus infundibulum-sparing technique. Protagonists of the PV-sparing repair admit that the major advantage is merely related to preservation of the pulmonary annulus, as more than one-third of patients experience PI at midterm follow-up. In addition, in >25% of cases, an infundibular patch is associated for adequate relief of the infundibular hypertrophic stenosis (4).

The role of the infundibulum after TOF repair has currently been studied in patient cohorts undergoing procedures in former eras, with the use of extensive RV incisions closed with liberal patches (21). Two MRI studies on the functional analysis of the RV components in TOF patients demonstrated that the decreased ejection performance of the RV outlet portion was related to the extent of postsurgical scarring. RV dysfunction was only evident once the adaptive response of the larger trabecular part of the right ventricle to PI became insufficient (22- 23). Our experimental findings indirectly underscore the importance to maintain the infundibular function because: 1) large infundibular scarring negatively affects the RV contractile function; and 2) the preserved infundibular integrity resists progression of associated pulmonary regurgitation and delays the secondary dilation process. Meanwhile, early clinical data on the use of an infundibulum-sparing RVOT repair are promising in those infants with severe pulmonary valve hypoplasia, in whom relief of the PV annulus is deemed necessary at the cost of a minimal infundibular incision (5- 6).

Even though the induced RVOT dysfunctions were clinically relevant and the hemodynamic measurements were homogeneous, extrapolation of these animal data to the clinical setting of patients with TOF needs careful consideration. Confounding effects on RV function by factors such as cyanosis and RV hypertrophy were not taken into account. The investigation of INF was solely based on the complete surgical incision of the infundibulum. Consequently, the effect of a more conservative infundibular surgery, appropriate to actual techniques, was not fully ascertained. In addition, this model did not include the contribution of some residual low-gradient RVOT stenosis, with its beneficial effect on the PI-related RV dysfunction (24).

The chronic phase in this porcine model was determined by a 3-month interval, corresponding approximately to midterm follow-up in humans. However, it remains debatable whether longer follow-up in these animals is desired for clearer discrimination of the effects of RVOT component dysfunction.

Due to several periprocedural incidents, the sample size of each group was limited, which might undermine the power of the intentional study sample. Nonetheless, we feel confident with the conclusions on our data because the differences between the hemodynamic effects among the groups were indicative within an acceptable range of variance.

In a juvenile chronic pig model, RV function was assessed in relation to dysfunction of the main components of the RVOT based on analysis of pressure-volume loops. Although RV performance was significantly affected by any surgically induced RVOT dysfunction involving the infundibulum and/or the PV, the most detrimental effect was observed by combined dysfunction of both components through TAP repair. RV contractility was immediately impaired by infundibular distortion due to intrinsic myocardial damage, whereas it was chronically furthered by PI-related RV dilation. Preservation of the structural and functional integrity of the infundibulum seems to delay the extent of the volume load–dependent secondary RV remodeling by impeding pulmonary regurgitation progression. Based on these findings, the adoption of a RVOT-sparing strategy for repair of TOF seems justified.

The authors thank Dr. Ingrid Van Overbeke and Deborah Croes for their dedicated care of the animals, and particularly Maria Olieslagers, for the many efforts to smooth the organization of the experiments by technical assistance and caretaking of the well-being of both animals and study participants.