Hemodynamic outcomes in patients undergoing bidirectional cavopulmonary connection with additional or antegrade pulmonary blood flow: a single-centre retrospective study

Introduction

In recent decades, a growing number of adults worldwide have survived childhood undergone a Fontan or total cavopulmonary connection (TCPC) procedure (Reddy et al., 2020). Cardiac surgical techniques have been evolving to some extent but are still largely unsupported by high-quality evidence of optimal outcomes. Interstage palliative surgeries, such as bidirectional cavopulmonary connection (BCPC), have become a pivotal staged procedure that prepares single ventricle patients for TCPC operation (Glenn & Patino, 1954; Glenn, 1958; Konstantinov & Alexi-Meskishvili, 2000). Candidacy for TCPC operation depends on low pulmonary vascular resistance, good ventricular function, proper pulmonary artery (PA) size, and a well-functioning atrioventricular valve (AVV) (Venna, Cetta Jr & d’Udekem, 2021).

The question then arises: how can these conditions for TCPC candidacy be attained? The prospect of a failed TCPC is a nightmare scenario accompanied by multisystem complications. Therefore, BCPC operation is the ultimate long-term palliation for some patients (Calvaruso et al., 2008; Miyaji et al., 1996). What if BCPC with antegrade or additional pulmonary blood flow (AAPBF) could be used to better prepare patients in terms of promoting PA growth, improving oxygen saturation and preventing pulmonary arteriovenous malformation (AVM) (Miyaji et al., 1996; Demirtürk et al., 2013; Ferns et al., 2013; Berdat et al., 2005; Baek et al., 2021)? A counterargument to performing BCPC with AAPBF is the potential increase in adverse outcomes, including a higher incidence of pulmonary hypertension, deterioration of ventricular function, and exacerbation of AVV regurgitation (Ferns et al., 2013; Berdat et al., 2005; Baek et al., 2021). One study reported an increased incidence of AVV repair and/or replacement associated with maintaining AAPBF (Davidson et al., 2023). The current study hypothesizes that performing BCPC with AAPBF can promote adequate PA growth, thereby optimizing the opportunity for patients to become suitable candidates for TCPC procedure. Therefore, this study aimed to evaluate the effect of AAPBF on PA growth and hemodynamics, with the hope of improving TCPC-related outcomes.

Patients and Methods

Patient selection

The present study was a single-center, retrospective cohort analysis using a hospital database from a large referral cardiac center in Thailand. The study had been approved by the Siriraj IRB, Faculty of Medicine Siriraj Hospital, Mahidol University (SIRB protocol No. 905/2565 (IRB1), COA no. Si 045/2023) since January 17, 2023. The requirement of informed consent was waived due to the minimal risk posed to the participants. Between 2006 and 2022, a total of 181 patients underwent BCPC and provided complete pre-BCPC and pre-TCPC catheterization data. After excluding patients with one-and-a-half ventricle physiology who underwent a Kawashima operation (n = 14), 167 patients were eligible for the analysis (Fig. 1).

Results

Baseline characteristics

The cohort of 167 patients was divided into two categories based on preservation of AAPBF: group 1 underwent BCPC with AAPBF (n = 44), and group 2 underwent BCPC without AAPBF (n = 123; Fig. 1). Baseline characteristics were not significantly different (Table 1), with a comparable proportion of presenting physiology: pulmonary obstruction (86.4% in group 1, 81.3% in group 2), systemic obstruction (2.3% in group 1, 6.5% in group 2), and excessive pulmonary blood flow (11.4% in group 1, 12.2% in group 2; p = 0.45, p = 0.45, p = 0.89, respectively). The cardiac morphologies are demonstrated in Fig. 2. The initial ventricular function and AVV regurgitation were not statistically different between group 1 and 2.

Table 1:

Patients’ characteristics (n = 167).

	Group 1 with APBF (n = 44)	Group 2 Without APBF (n = 123)	p-value
Gestational age (<37 weeks)	5 (11.4%)	10 (8.1%)	0.532
Low birth weight (<2,500 g)	8 (18.2%)	15 (12.2%)	0.324
SpO2 (%)	84 (72–89)	80 (70–86)	0.084
Heterotaxy syndrome	17 (38.6%)	51 (41.5%)	0.743
Right isomerism	11 (25%)	28 (22.8%)	0.764
Initial good ventricular function	44 (100%)	123 (100%)	–
Initial moderate to severe AVVR	4 (9.1%)	17 (13.8%)	0.417
Atrial situs			0.532
Normal Situs inversus Common atrium	24 (54.5%) 6 (13.6%) 14 (31.8%)	75 (61%) 10 (8.1%) 38 (30.9%)
Confluence pulmonary artery	44 (100%)	121 (98.4%)	>0.99
Operation before BCPC
Systemic to pulmonary artery shunt	13 (29%)	65 (53%)
Ductal stenting	–	3 (2%)
Pulmonary artery banding	7 (16%)	14 (11%)
RV to pulmonary artery shunt	–	5 (4%)
Norwood operation	–	1 (0.8%)
TAPVR repair	2 (4%)	5 (4%)
Aortic arch repair	–	2 (1.6%)
No previous surgery	21 (47%)	33 (26%)

DOI: 10.7717/peerj.20021/table-1

Notes:

Data represented as mean ± SD, median (IQR; p25-p75), and n (% within column).

*Statistical significance at p-value <0.05.

AAPBF

, additional or antegrade pulmonary blood flow

AVVR

atrioventricular valve regurgitation

SpO₂

oxygen saturation

MAPCAs

major aortopulmonary collateral arteries

Change in pulmonary size, EDP, and mPAP

Figure 3 and Table 3 present significant changes in absolute pulmonary artery size after applying Bonferroni correction with the significance threshold set at p = 0.017. The time intervals between the two assessments were 54.77 ± 25.63 months for group 1 and 69.39 ± 33.60 months for group 2 (p = 0.011), therefore further analysis using ANCOVA was used to adjusted the baseline different. Relative to their original sizes, the RPA increased by 35% in group 1 (with AAPBF) compared to 18% in group 2 (without AAPBF; p = 0.001), while the LPA increased by 26% in group 1 versus 10% in group 2 (p = 0.011). Although the initial p-value for RPA change (p = 0.019) did not initially meet this Bonferroni-adjusted threshold, further analysis using ANCOVA revealed a statistically significant difference (p < 0.001). The z-scores for RPA and LPA also showed significant increases in size and rates of changes (Table S1). Furthermore, the z-scores clearly indicated a proportional deterioration in pulmonary artery growth relative to somatic growth in patients without AAPBF. There was a marked decline in the Nakata index in group 2 (−29% of the original index; p < 0.05) compared to a 9.4% increase in group 1 (p < 0.05). Although the McGoon ratio was not statistically significant between the groups, follow-up analysis using ANCOVA revealed statistically significant differences across all pulmonary arteries growth parameters, as presented in Table S2.

Table 3:

Changes in pulmonary size, EDP, and mPAP (n = 167).

	Group 1 With AAPBF (n = 44)			Group 2 Without AAPBF (n = 123)			p-value
	Pre-BCPC	Pre-TCPC	p-value	Pre-BCPC	Pre-TCPC	p-value
RPA size (mm)	9.3 (7.31–11.98)	12.38 (10.35–15.68)	0.001***	8.9 (7.28–11.0)	11 (9.33–12.90)	<0.001***
ΔRPA	3.27 (0.52–5.07)			1.60 (−0.05–3.53)			0.019***0.001**
RPA growth rate (mm/month)	0.06 (0.01–0.11)			0.03 (−0.01–0.51)			<0.001***
LPA size (mm)	9.15 (7.07–10.15)	11.2 (8.75–12.75)	0.011***	8.5 (7.07–10.80)	9.7 (8.25–11.20)	<0.001***
ΔLPA	2.38 (0.96–4.48)			0.88 (−0.63–2.93)			0.004***0.002**
LPA growth rate (mm/month)	0.05 (0.02–0.06)			0.01 (−0.01–0.39)			<0.001***
McGoon ratio	2.21 (1.80–2.60)	2.23 (1.83–2.76)	<0.001***	2 (1.75–2.30)	1.96 (1.75–2.20)	<0.001***
ΔMcGoon ratio	0.035 (−0.37–0.50)			−0.04 (−0.40–0.20)			0.253 0.009***
Nakata index (mm2/m2)	283.93 (176.07–421.61)	317.43 (215.65–400.09)	0.034***	291.83 (229.06–451.56)	217.85 (170.81–268.64)	<0.001***
ΔNakata index	26.7 (−97.69–123.63)			−84.67 (−189.75–14.28)			<0.001*** <0.001**
Nakata index change rate (mm2/m2 per month)	0.64 (−2.18–2.02)			−1.38 (−2.88–0.19)			<0.001***
mPAP (mmHg)	18 (15–20.50)	16 (13–18)	0.006*	18 (15–20.75)	15 (13–17)	0.182
ΔmPAP	−2 (−4.5–1)			−3 (−7–1)			0.118
Ventricular EDP (mmHg)	12 (10–13)	12 (10–14.75)	0.046*	12 (10–14)	12 (10–13)	0.164
ΔEDP	1 (−2–4)			−1 (−3–3)			0.219
PVRi (WU m2)	2.03 (1.28–2.6)	1.77 (1.03–2.33)	0.567	2.03 (1.21–2.88)	1.65 (0.99–2.50)	0.140
ΔPVRi	−0.27 (−1.18–0.83)			−0.41 (−1.35–0.47)			0.190

DOI: 10.7717/peerj.20021/table-3

Notes:

Data represented as mean ± SD, median (interquartile range; p25-p75), and n (% within column).

*Statistical significance at p-value < 0.05.

**Statistical significance at p-value < 0.05 by ANCOVA.

***Statistical significance at p-value < 0.017 (Bonferroni correction).

****Statistical significance at p-value < 0.025 (Bonferroni correction).

AAPBF

additional or antegrade pulmonary blood flow

SpO₂

oxygen saturation

BCPC

bidirectional cavopulmonary connection

TCPC

total cavopulmonary connection

RPA

right pulmonary artery

LPA

left pulmonary artery

PVRi

pulmonary vascular resistance index

EDP

end-diastolic pressure

mPAP

mean pulmonary arterial pressure

An insignificant increase in ventricular EDP when compared between groups was noted in group 1 (with AABPF) even though a decrease was observed in group 2 (without AABPF; p = 0.219). A decrease in mPAP was observed within groups for both groups, with a more prominent decrease in group 1 (−11%, p < 0.05) compared to group 2 (−17%, p = 0.182). However, the different did not reach significant when compare between groups (p = 0.12). A total of 106 patients in the study (63.5%) had an elevated mPAP of >16 mmHg at the pre-BCPC assessment, with 26 patients from group 1 (59%) and 80 patients from group 2 (65%). Subsequent pre-TCPC data showed that 23 patients in group 1 (52.3%) and 54 patients in group 2 (43.9%) had retained a high mPAP without statistically significant between groups (p = 0.339). Of these 23 patients in group 1, 13 patients underwent TCPC with seven fenestration (53.8%). Likewise, 29 patients from group 2 underwent TCPC with 13 fenestration (44.8%). No statistical difference between groups in the proportion of TCPC with fenestration was observed (p = 0.588).

Notably, four patients exhibited unfavorable hemodynamic data in pre-TCPC catheterization and were thus identified as high-risk patients not suitable for TCPC. Of these four patients, one patient from group 1 exhibited a high mPAP of >20 mmHg and elevated EDP. The remaining three patients from group 2 experienced pulmonary hypertension and exhibited small sized PA branches.

Discussion

The unilateral Glenn operation was first successfully performed in 1958 to alleviate cyanosis and unload the ventricles (Glenn & Patino, 1954; Glenn, 1958). In 1961, the first BCPC procedure was completed, resulting in better systemic arterial oxygen saturation (Konstantinov & Alexi-Meskishvili, 2000). BCPC, however, still results in a much lower pulmonary blood flow (PBF) compared to baseline or post-TCPC, and this decreased PBF jeopardizes PA growth (Uemura et al., 1995; Mendelsohn et al., 1994; Slavik, Franklin & Radley-Smith, 1999; Kansy et al., 2013). Therefore, AAPBF was introduced to help promote PA growth (Ferns et al., 2013; Berdat et al., 2005; Baek et al., 2021; Mendelsohn et al., 1994; Van de Wal et al., 1999).

The advantages of AAPBF, including enhanced PA growth and improved oxygen saturation, have been previously reported (Ferns et al., 2013; Baek et al., 2021; Kowatari et al., 2020; Sughimoto et al., 2015; Ferns et al., 2021). The current study corroborates these findings, revealing a trend toward slightly higher oxygen saturation prior to discharge from the BCPC operation, prior to the TCPC operation, and at the subsequent follow-up when compared to patients without AAPBF. However, the differences observed in this investigation were not statistically significant, contrary to what has been reported in previous studies (Berdat et al., 2005; Chen et al., 2015). Patients with AAPBF in the current study had slightly larger PA branches before undergoing the BCPC operation, indicating that the underlying reasons for AAPBF were beyond solely the augmentation of PAs. At the pre-BCPC assessment, there were no statistically significant differences in RPA and LPA sizes between the two groups (p = 0.412 and p = 0.971, respectively). These variables were therefore adjusted for in the subsequent analysis since the interval between two catheterizations were difference. In this study, the modified Blalock-Thomas-Taussig shunt constituted a minority of the cases with AAPBF. This may be attributed to the limitations of the surgical technique, which could potentially increase the risk of PA distortion due to somatic outgrowth relative to the fixed shunt length (Godart et al., 1998). The cardiopulmonary bypass time and total operative time were significantly lower in group 1, likely due to the omission of time-consuming procedures required for pulmonary blood flow elimination (p < 0.05 for both comparisons). In the study cohort, there were no significant differences in ICU or hospital length of stay among patients with AAPBF when compare to patients without AAPBF. Similarly, there were no significant differences in the incidence of reoperations and reinterventions during the admission period. These findings were consistent with those of previous studies (Chen et al., 2015; Gray et al., 2007).

During the interstage period, the cardiac catheterization data showed a greater change in size and rate of growth of both PAs in group 1 (with AAPBF; p < 0.05 for all). The analysis of the z-scores of RPA and LPA in group 1 also showed significant increases in size and rates of change based on the absolute values and mean differences (Table S1). The findings clearly highlight a significant deterioration in PA growth in group 2 (without AAPBF) relative to overall somatic growth. The Nakata index demonstrated a marked increase in group 1, whereas a significant decline was observed in group 2 (p < 0.05 for both). This statistically significant difference persisted across all assessed parameters including RPA and LPA size, the McGoon ratio, and the Nakata index after adjusting for baseline values (Table 3, Table S2, and Fig. 3). Importantly, this statistical significance remained robust after the interval between the pre-BCPC and pre-TCPC assessment, pre-BCPC parameters, age at assessment, and body weight were accounted for using ANCOVA. These findings further underscore the substantial PA growth observed in group 1. Collectively, these results illustrate a preserved capacity for PA growth in proportion to body surface area throughout the evaluation period. This observation is particularly relevant since both parameters would generally be expected to decline over time in univentricular patients (Kansy et al., 2013).

Notably, the growth of the RPA size was greater than that of the LPA, even among patients who did not undergo BCPC with AAPBF. These findings align with those of previous studies. For example, Slavik et al. (1995) observed a trend towards a reduced postoperative left pulmonary artery size in patients who underwent BCPC without AAPBF. Additionally, this study assessed the association between LPA regression and several operation variables including the type of BCPC (bilateral vs. unilateral), pulmonary artery plasty during BCPC, and left pulmonary plasty during BCPC. However, none of these factors showed a statistically significant association with LPA regression. Additionally, a case series by Javadi et al. (2022) demonstrated that mean blood flow through the RPA was higher than that through the LPA, as assessed by cardiac magnetic resonance imaging (MRI). These observations led to the hypothesis that factors beyond anatomical structure, particularly those related to fluid dynamics and energy loss associated with various surgical approaches, may contribute to differential pulmonary artery growth.

There is debate as to whether AAPBF poses a risk for pulmonary hypertension (McElhinney, Marianeschi & Reddy, 1998; Chen et al., 2020). Pulmonary vascular resistance index (PVRi) was increased in group 1 and decreased in group 2 (p = 0.190). Moreover, the data from the current study showed a significant reduction in mPAP within group 1 (p < 0.05), in accordance with previous reports (Berdat et al., 2005; Caspi et al., 2003). However, the decline in mPAP was not statistically significant between the groups (p = 0.118), suggesting an inconsequential impact on ventricles. Therefore, the current study did not reach a conclusion regarding whether AAPBF significantly affects mPAP. Currently, there is no universal guideline for defining elevated mean pulmonary artery pressure in single ventricle patients. Prior research indicated that an mPAP ≥16 mmHg was associated with adverse events following TCPC (Kido et al., 2021). Using this cutoff point, patients with an mPAP ≥16 mmHg following pre-BCPC catheterization decreased after BCPC in both groups. Analysis of the incidence of fenestrated TCPC revealed no significantly higher proportion in group 1 compared to group 2 (p = 0.223). This finding may suggest that patients who underwent BCPC with AAPBF potentially exhibited a physiologic status prior to TCPC that is not detrimental when compare to BCPC without AAPBF.

Concerns have been raised regarding the potential impact of AAPBF on ventricular volume loading (Miyaji et al., 1996; Uemura et al., 1995). Although the current study revealed a significant increase in ventricular EDP between two cardiac catheterizations in group 1 (p < 0.05), the change was not statistically significant when the intergroup comparison to group 2 (p = 0.219). Further analysis revealed that a LVEDP ≥13 mmHg was significantly associated with mortality in group 2 patients (without AAPBF; p = 0.026). In contrast, no significant association was observed between elevated LVEDP and mortality in group 1 (with AAPBF). These findings suggest the presence of additional volume loading in patients with AAPBF; however, this volume did not appear to negatively impact clinical outcomes. These findings are also consistent with previous studies (Berdat et al., 2005; Kowatari et al., 2020). A higher incidence of AVV failure necessitating surgical intervention in patients with AAPBF has been reported (Davidson et al., 2023). At the initial assessment for the current study, moderate to severe atrioventricular valve regurgitation (AVVR) was observed in four patients (9.1%) with AAPBF and 17 patients (13.8%) without AAPBF (p = 0.417). A subgroup analysis was performed on patients who exhibited moderate to severe AVVR at both baseline and at the most recent follow-up. There were no statistically significant differences found between the groups regarding AV valve morphology, ventricular morphology, heterotaxy status, or history of prior surgical interventions (Supplementary S3). Interestingly, nine patients (6.3%) without AAPBF required AVV repair following the BCPC operation, whereas none of the patients in the AAPBF group underwent such intervention. Other previously published studies have reported findings consistent with those from the current study (Ferns et al., 2013; Baek et al., 2021).

Recent research has highlighted the potential benefits of AAPBF, demonstrating its role in mitigating the development of pulmonary AVM and collateral vessel (Gray et al., 2007; Henaine et al., 2013; Yoshida et al., 2005). Abnormal vascular connections between arteries and veins could precipitate systemic desaturation and eventual ventricular dysfunction (Heinemann et al., 2001; Mohammad Nijres et al., 2020; Ichikawa et al., 1995). However, the current study was unable to demonstrate prevention of the development of AVM in AAPBF. Based on our findings, maintaining AAPBF might be beneficial in single ventricle patients with borderline hypoplastic PA to optimize PA growth without jeopardizing the candidacy of TCPC.

Conclusions

BCPC with AAPBF effectively maintains the indexed pulmonary artery growth without a significant impact on ventricular volume loading, mean pulmonary arterial pressure, or the severity of atrioventricular valve regurgitation. This approach may offer potential benefits for patients who are initially considered suboptimal candidates for TCPC completion, or, alternatively, may support long-term palliation by optimizing BCPC physiology in those who are unsuitable for TCPC procedure.

Supplemental Information