Retrospective Review of Perioperative Neonatal Echocardiography for Non‐Cardiac Surgery at a Quaternary Level Academic Hospital in Johannesburg, South Africa

Congenital heart disease (CHD) has a global incidence of 9 in 1 000 live births [1]. There is a paucity of data on the incidence of concurrent CHD in neonates undergoing surgery worldwide, and even less information is available on this patient population in our resource-limited setting. Neonates may present for non-cardiac surgery for conditions commonly associated with CHD: esophageal atresia (EA) with or without tracheoesophageal fistula (TEF), omphalocele, congenital diaphragmatic hernia (CDH), anorectal malformation (ARM), duodenal atresia, jejunal and other intestinal atresias. Given the time-sensitive nature of these surgeries, more information regarding which clinical features are most suggestive of CHD would aid in guiding the safe administration of neonatal anesthesia in the absence of a preoperative echocardiogram. Neonates with CHD have a two-fold increase in mortality from non-cardiac surgery [2], and while transthoracic echocardiography is the gold standard for diagnosing CHD, it is not always possible for this to be performed preoperatively in resource-limited settings. Our study set out to determine the incidence of CHD in neonates undergoing non-cardiac surgery. We additionally conducted an exploratory analysis of the association between specific clinical features and special investigations with CHD. A single-center retrospective cohort study was conducted on 198 neonates during their first 28 days of life (corrected gestational age to 37 weeks for premature neonates) who underwent non-cardiac surgery for general surgical procedures between October 2021 and December 2023. The neonates who received perioperative echocardiography (103) were further investigated, and their echocardiographic findings, clinical features, chest radiographs, and electrocardiograms (ECG) were analyzed. Those neonates who were not referred for echocardiography did not have any concerning clinical features. Ethical approval was granted, and findings were reported according to the STROBE recommendations for observational studies. The majority of the echocardiographic findings (82/103, 79.6%) were isolated physiological lesions and were considered normal echocardiograms in this study. The most common physiological lesion was a patent foramen ovale (PFO), present in 76.7% of neonates (Table 1). CHD was identified in 16 neonates (15.5%). Our study found only one neonate with a cyanotic lesion: namely, tricuspid atresia with ductal-dependent circulation, a critical CHD. Fifteen neonates had acyanotic lesions, with the most common being a ventricular septal defect (VSD), followed by a pathological patent ductus arteriosus (PDA). Echocardiography was performed preoperatively in 54.4% (56/103) of neonates. The most frequent reason for emergency surgery (35.4%) was necrotizing enterocolitis (NEC) or spontaneous intestinal perforation (SIP). The most common diagnosis amongst neonates referred for echocardiography was anorectal malformation (ARM), and this group also had the highest yield of CHD (Table 2). 1.4 (0 to 7.7) 5.6 (0.1 to 27.3) 3.6 (0.1 to 18.3) 7.1 (0.2 to 33.9) 29.6 (13.8 to 50.2) 30.8 (14.3 to 51.8) 4.2 (0.1 to 21.1) 20 (0.5 to 71.6) 11.1 (1.4 to 34.7) 11.1 (1.4 to 34.7) 29.4 (10.3 to 56) 29.4 (10.3 to 56) Our rates of CHD per surgical condition, which were similar to international figures, included ARM 30.8% (vs. 20%–40%) [3], TEF/EA 29.4% (vs. 25%–30%) [3], and other intestinal atresia 7.1% (vs. 6.6%–21%) [4]. Our rate of CHD in CDH of 33% was much higher than international figures (10.6%) [5], however, due to the small number of neonates in this group, the result should be interpreted with caution. Surgical conditions that we identified to have lower rates included DA 11.1% (vs. 39%–44%) [6], and omphalocele 0% (vs. 35%) [4] (Table 2). We identified that the following clinical features had significant associations with CHD: Dysmorphic features (p-value: 0.002, OR: 6.12, 95% CI: 2 to 21.23), suspected or confirmed congenital abnormalities (p-value: 0.004, OR: 20.27, 95% CI: 3.84 to 374.7), the presence of an abnormal CXR (p-value: 0.009, OR: 6.03, 95% CI: 1.73 to 28.22) or an abnormal ECG (p-value: 0.002, OR: 39.43, 95% CI: 5.02 to 832.2). CHD may be present in the absence of a murmur, and our study revealed that seven out of the sixteen neonates with CHD (43.8%) had a murmur. Dysmorphic features included any clinical abnormality detected on physical examination suggestive of genetic disorders, such as low-set ears, microcephaly, midface hypoplasia, epicanthic folds, syndactyly, extra digits, clubbed feet, vertebral abnormalities, or genital malformations. Congenital abnormalities included VACTERL associations (suspected in 37/103, 35.9%), and genetic syndromes of Trisomy 21 (8/103, 7.8%), Trisomy 18 (4/103, 3.9%), Trisomy 13 (2/103, 1.9%), and other (4/103, 3.9%). An abnormal CXR included any abnormality detected on cardiac or respiratory evaluation, such as pulmonary infiltrates, dextrocardia, abnormal heart shapes, or features suggestive of cardiomegaly. Five of the eighty ECGs performed were found to be abnormal. Three of these displayed a left axis with paucity of left ventricular forces, one had an abnormal p-wave axis, and the final ECG displayed a north-west axis. The limitations of this study include the conduct at a single center, limiting the generalizability of the findings. The postoperative outcomes, morbidities, and mortalities associated with CHD were not explored. Sample sizes in certain surgical pathologies were small, and while the findings remain clinically relevant, further larger-scale studies are recommended to expand on these results. In conclusion, the majority of neonates in this study had physiological lesions, with the minority (15.5%) having CHD. The most common surgical pathology associated with CHD was ARM, and the most frequently encountered lesion was VSD. Almost half of the neonates with CHD had a murmur. We found the presence of dysmorphic features, congenital abnormalities, an abnormal chest radiograph, or an abnormal electrocardiogram to be associated with CHD. In the absence of these features, the likelihood of CHD is low, and most neonates could safely undergo anesthesia for general surgical procedures in resource-limited settings without a preoperative echocardiogram, considering the time-sensitive nature of surgery. The data that support the findings of this study are available from the corresponding author upon reasonable request.