Postoperative pulmonary vascular growth in patients with congenital diaphragmatic hernia

The survival of patients with congenital diaphragmatic hernia (CDH) has recently improved due to modern treatment strategies, including gentle ventilation with permissive hypercapnia and delayed surgical repair after stabilization [1]. Conflicting reports exist on postoperative lung vascular growth and function after CDH repair [2‐6]. Arena et al. reported rapid improvement in lung perfusion after CDH repair, even in apparently hypoplastic and small lungs in infant, adolescent, and adult patients with CDH [4]. Another single-center study using multiple lung perfusion scintigraphy concluded that the majority of survivors showing significant vascular growth had symptom-free lives [5]. However, another study in a cohort of CDH survivors with more severe disease characteristics showed deterioration of ventilation/perfusion (V/Q) mismatch due to perfusion deficits in the ipsilateral lung [6]. Furthermore, factors affecting postoperative lung perfusion have not been thoroughly investigated previously in the literature.

This study aimed to determine the possible vascular growth in the ipsilateral lung and the factors affecting it in CDH survivors using serial lung perfusion scintigraphy.

Patients with left Bochdalek hernia, who were diagnosed prenatally or within 24 h after birth, underwent CDH repair during the neonatal period from January 1st, 2001 to December 31st, 2020, and who received at least 1 postoperative lung perfusion scintigraphy at our institution were included in the study. The following patient variables were retrospectively extracted from the medical records and used for analysis: sex, gestational age at prenatal diagnosis, gestational age at birth, birth weight, Apgar score (1 min, 5 min), alveolar–arterial oxygen difference (A-aDO₂) at birth and immediately before the operation, nitric oxide (NO) use during preoperative stabilization and at the time of surgery, liver up, defect size (congenital diaphragmatic hernia study group (CDHSG) staging A–D) [7], direct/patch closure, operative approach (thoracoscopy/open abdominal), and age at discharge. Age and ipsilateral lung perfusion rate at the first and serial lung perfusion scintigraphy were also obtained. The results from the first and last perfusion studies were used if three or more studies were performed.

For lung perfusion scintigraphy, macroaggregated albumin particles tagged with 99 technetium, at a dose of 50,000–100,000 particles were slowly injected. Planar images were obtained from the anterior, posterior, and lateral projections. The percentage function of the individual lungs was calculated.

The primary outcome of the study was the difference between the 1st and serial lung perfusion study in patients with CDH. As secondary outcomes, we also analyzed the result of 1st and serial study. Categorical and numerical variables were expressed as n (%) and median (interquartile range), respectively. The Wilcoxon test, Pearson’s correlation coefficient test, and paired t-test were used for categorical, numeric, and paired variables, respectively. We used JMP14, (SAS Institute Inc., NC, USA) for statistical analysis. The correlation coefficient was expressed as r. Statistical significance was set at p < 0.05. This study was approved by the Institutional Review Board of the Research Ethics Committee of the Faculty of Medicine, the University of Tokyo [No.2996-(10)].

Twenty-one patients with CDH who underwent surgery during the neonatal period and at least one postoperative lung perfusion scintigraphy were included in the study. Non-isolated patients with CDH and severe associated anomalies, including chromosomal abnormalities and major cardiac malformations, were not included in the study. All 21 patients were diagnosed prenatally, had a left-sided hernia, and were free from extracorporeal membrane oxygenation (ECMO) during the perioperative course. NO therapy was used in seven (33%) patients for preoperative stabilization. The number of patients with each CDHSG defect size were: A, 8 (38.1%); B, 7 (33.3%); C, 3 (14.3%); and D, 3 (14.3%). Of the total, 12 and 9 patients underwent open abdominal repair (O) and thoracoscopic repair (T), respectively. Among these patients, 11 underwent one lung perfusion study and 10 underwent multiple studies. The patient characteristics were similar. However, patients in multiple studies underwent surgery at older ages (4.5 days vs. 2 days, p = 0.03) and showed lower ipsilateral lung perfusion (25.45% vs. 38%, p = 0.049) in the 1st study, suggesting that patients in multiple studies had severe disease (Table 1).

Among the 21 patients who underwent one or multiple perfusion studies, the median age and the ipsilateral perfusion rate was 34 (22, 44.25) days and 32 (22, 38.5) % in the first perfusion study, respectively (Table 2). Gestational age at prenatal diagnosis (p = 0.02), A-aDO₂ at birth (p = 0.007), and preoperative NO use (p = 0.014, data not shown) significantly correlated with the 1st perfusion study (Fig. 1A, B). No other variables, including the operative approach (T 32.7 (20.7, 38.5) %; O 31 (21, 38.75) %, p = 0.943), were significantly associated with the 1st perfusion study (data not shown).

The median age at the time of the serial studies was 3.6 (1.1, 10.2) years among 10 patients with multiple studies (Table 2). The ipsilateral perfusion rate in the serial studies was 33 (28.8, 40.25) %. While there was a tendency for higher perfusion rates in patients with older gestational age at prenatal diagnosis (p = 0.10, Fig. 1C) or smaller A-aDO₂ at birth (p = 0.06, Fig. 1D), no perioperative parameters, including preoperative NO use (p = 0.24) and operative approach (p = 0.45), were significantly associated with the serial ipsilateral perfusion rate.

When the 1st and the serial perfusion study were compared, perfusion rate increased with age in all patients (difference:  + 7 (4.3, 13.25) %, p = 0.001) (Fig. 2A). The difference was significantly larger in patients who underwent thoracoscopic repair than in those who underwent the open abdominal approach (T: 10.6 (5.0, 17.1) %; O: 4.25 (1.2, 7.9) %, p = 0.042) (Fig. 2B). This difference was not significantly associated with other variables, including gestational age at prenatal diagnosis (p = 0.16, Fig. 2C), A-aDO2 (p = 0.96, Fig. 2D), NO use (p = 0.28), liver up (p = 0.90), or patch closure (p = 0.08).

In this study, we investigated the postoperative status and natural history of ipsilateral lung perfusion in patients with CDH. Our cohort consisted of patients with relatively mild CDH, which was suggested by the smaller CDHSG defect size and the lack of need for ECMO. Gestational age at prenatal diagnosis, A-aDO2 at birth, and preoperative NO use significantly correlated with the 1st lung perfusion study at a median age of 34 days. All patients with serial perfusion studies showed significant improvement in ipsilateral lung perfusion (Fig. 2A). The thoracoscopic approach was significantly associated with a larger difference between the 1st and serial studies (Fig. 2B).

The natural history of the ipsilateral lung function is a matter of concern in patients undergoing CDH repair. Significant lung growth occurs at the alveolar level with postnatal vascular remodeling, resulting in larger and fewer muscular arteries after CDH repair [8], suggesting that the anatomical changes caused by CDH are, at least in part, reversible [8‐10]. While some studies have shown persistent postoperative reduction of ipsilateral lung perfusion and impaired ventilation/perfusion (V/Q) mismatch [2, 3, 6, 11], other studies have suggested that lung perfusion improves postoperatively with age [4, 5]. Spirometry studies after CDH repair have also shown conflicting results, ranging from normal to obstructive, restrictive, and mixed dysfunction [12‐14]. This inconsistency could be due to the differences in the populations and designs of each study. Recent progress in the management of CDH may have led to the differences between studies. However, even with the possible impairment of ventilation and perfusion functions in the ipsilateral lung, the majority of these studies demonstrated that cardiopulmonary symptoms were rarely observed in patients with CDH, especially long after surgery [12‐14]. Similarly, all the patients in the present study were free of respiratory symptoms.

It is still controversial whether the postoperative ipsilateral lung perfusion improves, worsens or remains unchanged in patients with CDH. Previous studies with small cohorts demonstrated poor lung perfusion with limited follow-up periods [13, 15]. Other studies have suggested that factors related to severe diseases, such as liver up, patch closure, and ECMO usage [16, 17] are associated with poor perfusion. The results of our 1st perfusion study are consistent with those of previous studies. Gestational age at prenatal diagnosis, A-aDO2 at birth, and preoperative NO, all of which are indicators of CDH severity, were associated with the 1st perfusion study. A relatively large cohort from an American group showed that the V/Q mismatch worsened gradually in survivors of CDH, especially in those with severe disease characteristics, as a result of a decrease in ipsilateral lung perfusion [6]. Their cohort consisted of patients with severe disease, with most having a CDHSG defect size of C/D. In contrast, patients in studies showing improvement in ipsilateral lung perfusion with age were less severe, such as those without ECMO or patch closure [4, 5]. In our study, in which the majority of patients had a CDHSG defect size A/B with none requiring ECMO, all patients demonstrated improvement in ipsilateral lung perfusion with age, and the thoracoscopic approach was associated with a larger increase in perfusion. Considering the results of these studies, postoperative improvement in ipsilateral perfusion can occur in patients with mild CDH while it worsens with age in patients with severe CDH. This is consistent with a previous study that showed a continuous improvement in lung perfusion in a patient with modest CDH, but not in patients with severe CDH [5].

Our study implied that a minimally invasive approach may enhance improvement in patients with mild CDH. Due to the small number of patients and potential bias for the retrospective study, this finding might be a coincidence. However, our study showed no significant association between postoperative improvement in lung perfusion and factors associated with severity of CDH such as A-aDO₂, NO use, liver up, or patch closure, thoracoscopic approach rather than severity of CDH would affect postoperative improvement in lung perfusion. A possible explanation is that limited damage to the respiratory system during thoracoscopic surgery may have contributed to this improvement. Previous studies have described possible increases in respiratory morbidity or functional derangement due to iatrogenic damage during the neonatal perioperative period [13, 18].

The limitations of our study include its retrospective design and small number of patients. This study is subject to selection bias resulting from a tendency for serial studies to be performed in patients with severe CDH indicated by lower ipsilateral lung perfusion. Although modern treatment strategies, including gentle ventilation and surgery after stabilization, were applied for CDH over the study period, a possible change in treatment for CDH at our institution during the study period might have affected the outcome. Due to the retrospective nature of the study, prenatal imaging and postoperative echocardiogram were not systematically performed in all cases, and data obtained from these analyses, including the observed to expected lung area to head circumference ratio (O/E LHR) and pulmonary artery diameter, were not available for the study. The adoption of the thoracoscopic approach during the study period is another potential source of bias. However, the open abdominal approach for patients with mild disease was included in this study with this study period. Most patients with mild disease have recently undergone thoracoscopic repair at our institution.

In conclusion, our study using postoperative lung perfusion scintigraphy demonstrated that vascular growth in the ipsilateral lung was impaired, and that the severity of impairment was associated with the severity of the disease in patients with CDH. Postoperative vascular growth in the ipsilateral lung occurs with age, at least in patients with mild disease, as observed in our cohort. The thoracoscopic approach might contribute to better vascular growth in the ipsilateral lungs. Further large-scale clinical and experimental studies are necessary to understand postoperative lung development in patients with CDH.