Cerebral and Flank Oxygenation Trends in Neonatal Surgery for Abdominal Wall Defects: The Role of Near‐Infrared Spectroscopy

Neonates undergoing abdominal wall defect repair are at risk of bowel ischemia due to increased intra-abdominal pressure during surgical manipulation [1]. This pressure can compromise perfusion, particularly in patients undergoing primary closure [2, 3]. Detecting early signs of compromised oxygenation remains difficult, as traditional monitoring tools may not accurately reflect regional perfusion [4]. Near-infrared spectroscopy (NIRS) offers continuous, non-invasive monitoring of regional tissue oxygen saturation (rSO2), including bowel perfusion [4]. Its use in neonatal surgery is expanding [1, 5], though standardized approaches to interpretation remain lacking [6]. We aimed to evaluate intraoperative changes in cerebral and flank oxygenation using NIRS in neonates undergoing abdominal wall defect repair, and to assess its role in identifying hypoperfusion in giant omphalocele and complicated gastroschisis. This retrospective observational study was conducted at Necker Children's Hospital and included neonates with gastroschisis or omphalocele undergoing surgical repair from January 2015 to December 2024. Institutional approval was obtained from the Assistance Publique–Hôpitaux de Paris (IRB20250124140845). Explicit refusal by parents or legal guardians was the only exclusion criterion. Data were extracted from anesthetic charts and surgical records. NIRS monitoring became standard practice in 2020, although documentation remained inconsistent and sensor placement was not routinely noted. Patients were categorized by diagnosis and complexity: giant omphalocele (≥ 5 cm or liver involvement) and complicated gastroschisis (atresia, volvulus, perforation). Anesthesia included ketamine, midazolam, or propofol for induction, atracurium or rocuronium for relaxation, and sevoflurane with sufentanil for maintenance. Norepinephrine was used in six cases, none NIRS-monitored. Monitoring was performed with the Medtronic INVOS 5100C system, with cerebral sensors on the forehead and flank sensors along the posterior axillary line. Values were recorded every 10 min at three timepoints: T1 (surgery start), T2 (5–10 min post-reintegration), and T3 (final value before ICU transfer). Changes above 20% from baseline were considered significant. Fluid boluses (NaCl 0.9% or Albumin 5%) at 10–20 mL/kg were administered in these cases. Changes below 20% from baseline were considered within the range. Delta NIRS (ΔNIRS) was calculated as changes between T2–T1 and T3–T1. Analyses were performed using R (v4.4.2). A total of 152 neonates were included: 77 with gastroschisis and 75 with omphalocele. Cerebral NIRS monitoring was performed in 19.7% overall, more in gastroschisis (23.4%) than omphalocele (16%). Most (86.7%) exhibited values within the recommended range, with a trend toward greater stability in gastroschisis (94.4%) compared to omphalocele (75%). Flank NIRS monitoring was performed in 15.8% overall, with similar rates in gastroschisis (16.9%) and omphalocele (14.7%). Of those monitored, 75% maintained values within the recommended range, with a non-significant trend toward better stability in gastroschisis (84.6%) compared to omphalocele (63.6%) (Figure 1). Within the gastroschisis cohort, 28.6% had complicated forms. No significant differences were observed between complicated and non-complicated cases in gestational age, sex, birth weight, or preoperative intubation. NIRS monitoring was used in 24.7% of cases. Cerebral and flank NIRS values at T1–T3 were comparable between non-complicated and complicated gastroschisis. Flank ΔNIRS.F1 and ΔNIRS.F2 also showed similar distributions between the two groups (median −5 [IQR −13 to −3] vs. −5.5 [IQR −10 to −3]; −3 [IQR −7 to 0] vs. −2.5 [IQR −8.5 to 4.5]). In the omphalocele cohort, 46.7% had giant defects. These infants had lower birth weights (median 2700 g vs. 3135 g) and were more frequently intubated preoperatively (74.3% vs. 43.6%). NIRS monitoring was performed in 18.7% of cases. Cerebral NIRS at T1 was higher in giant (median 93.5 [IQR 90.5–95]) compared with non-giant omphaloceles (75 [IQR 69.75–81.25]). Flank ΔNIRS.F1 was −24.5 [IQR −32.75 to −19.75] vs. −2 [IQR −7.5 to 0], and ΔNIRS.F2 was −23.5 [IQR −29.75 to −18.25] vs. −2 [IQR −6.5 to 5] in giant vs. non-giant omphaloceles (Table 1). Our findings suggest that cerebral and flank NIRS behave independently. In gastroschisis, they fluctuated but generally remained within acceptable limits (< 20% change from baseline value). Baseline cerebral saturations were higher in omphalocele patients, likely reflecting the higher rate of preoperative intubation, though missing documentation on other data like oxygenation and hemoglobin limited interpretation. In giant omphaloceles, flank desaturation was significantly greater at T2 and T3, suggesting vulnerability to perfusion compromise, possibly due to increased intra-abdominal pressure. These observations align with prior studies [5, 7] showing somatic NIRS sensitivity in detecting ischemia. Limitations of this study include its retrospective design, incomplete NIRS documentation, and inconsistent recording of flank sensor placement. NIRS monitoring only became routine in 2020, and early cases often lacked flank data. Nonetheless, our results support dual-site NIRS monitoring to detect regional perfusion changes during surgery. Improved documentation and adherence to protocols are needed to optimize perioperative care. In conclusion, neonates undergoing abdominal wall defect repair are at risk of regional perfusion compromise not always reflected in cerebral NIRS alone. This highlights the region-specific nature of tissue oxygenation and supports dual-site NIRS monitoring in neonatal surgery. The authors have nothing to report. The study was approved and registered in the Assistance Publique–Hôpitaux de Paris (APHP) research registry (IRB20250124140845). Written non-opposition was obtained from all parents or legal guardians, and explicit refusal was the only reason for exclusion. Written notification of non-opposition was obtained from the parents or legal guardians of all participants. Parental opposition was the only exclusion criterion. The authors declare no conflicts of interest. The datasets generated and analyzed during the current study are not publicly available due to institutional restrictions but are available from the corresponding author on reasonable request.