Outcomes of surgical versus conservative treatment in pediatric median arcuate ligament syndrome.

This study was approved by the Institutional Review Board at Cincinnati Children's Hospital Medical Center (CCHMC). As this was a retrospective study, the requirement for informed consent was waived. A retrospective review of electronic medical records was conducted for symptomatic patients aged 7–21 years who were diagnosed with MALS by ultrasound (U/S) and/or CTA in the gastroenterology clinic at CCHMC between 2017 and 2024. All patients initially identified by ICD‐10 code I77.4 underwent detailed chart review to confirm they met these clinical and radiologic criteria prior to inclusion; one patient was excluded because he was asymptomatic. Specifically, Doppler ultrasound criteria included: peak systolic velocity >200 cm/s during expiration, end‐diastolic velocity >55 cm/s, and a decrease in peak systolic velocity >60 cm/s during inspiration or in the upright position. CTA criteria included: proximal celiac artery narrowing most evident during expiration, post‐stenotic dilation distal to the area of compression, a hooked appearance of the celiac artery due to ligament compression, and exclusion of other vascular abnormalities that could explain symptoms. At our institution, the disorders of gut–brain interaction (DGBI) clinic is a specialized multidisciplinary clinic focused on functional gastrointestinal (GI) disorders and brain–gut interactions. The team includes pediatric gastroenterologists with subspecialty expertise in DGBI, GI psychologists who provide cognitive behavioral therapy (CBT) and pain‐coping strategies, and registered dietitians experienced in functional GI nutrition and individualized dietary management. Children presenting with chronic abdominal pain are first evaluated in General Pediatric Gastroenterology, where a DGBI diagnosis is often considered. We included patients who underwent surgical decompression at CCHMC as well as those who received conservative management only. Patients who underwent surgery outside of CCHMC and asymptomatic patients were excluded from the analysis. We gathered information on demographics (age, gender, race, and ethnicity) medical history, surgical history (including endoscopies), medication history, diagnostic testing (U/S and CTA), and GI comorbidities. Outcomes were assessed based on the patients’ subjective responses documented in the charts. Outcomes were classified into worse, same, improved, and resolved based on the presenting symptoms change in intensity or frequency. Median arcuate ligament release and celiac plexus neurolysis are performed with a robotic‐assisted minimally invasive technique in which four ports are placed across the abdomen, and a liver retractor is used to aid with exposure. The left gastric vessel is followed to its take off of the celiac trunk and the tissue surrounding it is divided. Then the diaphragmatic crus is opened to expose the aorta, dissection is then carried down on this plane until the celiac artery take off is exposed. This helps identify the median arcuate ligament which is then divided along with the celiac plexus branches. Dissection is carried on either side of the artery surrounding the celiac trunk and cephalad 4 cm above its take off. The tissue and plexus branches on the celiac towards the trunk trifurcation is also divided until all arteries are exposed and no tissue is creating any compression. This allows for a full dissection and neurolysis. For the conservative group, two visits were analyzed: the initial visit to the GI clinic at CCHMC and the most recent follow‐up. In contrast, the surgical group had three visits reviewed: the initial GI visit, the post‐surgery visits, and the latest follow‐up. We examined responses collected for recent follow up (FU) visit—by each study group and between two the group. Then we examined co‐morbidities and interventions in relation to binary global scores (resolved + improved vs . same + worse) in the medical and surgical group separately. Finally, we examined global scores collected at post‐op and recent FU visits in surgical cohort ( n  = 24) longitudinally using proportional odds model as 4 and 2 levels ordinal response. Data were presented as median (interquartile range [IQR]) for continuous variables and as count (%) for categorical variables. Group differences examined using Wilcoxon rank‐sum test and Chi‐square or Fisher's exact test as applicable, respectively. Symptom response collected as 4‐level ordinal variable (resolved, improved, no‐change, and worse) was examined between the medical and surgical group using Chi‐square test. In addition, in the surgical group the ordinal symptom response was also examined between the recent follow‐up visit and earlier post‐operative visit using proportional odds model. The results were reported as cumulative odds and 95% confidence interval (CI) for symptom response to be resolved in comparison to other lower order responses when comparing recent follow‐up visit to earlier post‐operative visit. All analyses were conducted as two‐sided tests using SAS version 9.4 and p  ≤ 0.05 was considered to be statistically significant.

Chart review showed that ~90% of patients in the medical (nonsurgical) cohort were evaluated by pediatric surgery and had the opportunity to discuss the risks and benefits of operative intervention. Among patients who did not undergo surgery, the reasons included insurance denial of a planned surgery ( n  = 2), patient or family preference ( n  = 6), including significant anxiety about surgery ( n  = 1), families declining after reviewing operative risks ( n  = 4), or clinical improvement on medical therapy while awaiting surgery ( n  = 1), external specialist recommendation ( n  = 1), where a family reported being advised against surgery by an outside MALS specialist andnot being referred to surgery ( n  = 1), a patient improving clinically in the DGBI clinic for presumed irritable bowel syndrome (IBS) who later had imaging suggestive of MALS but opted to continue conservative management. Initially, ICD‐10 code screening yielded 40 patients. Five patients were excluded because they underwent surgery outside CCHMC, and one patient was excluded because he was completely asymptomatic. Our cohort includes 34 patients with median age of 15 years (13,17), 27 (79%) females, and 32 (94%) Caucasians. Ten patients with median age 15 years (7,16), 9 females (90%), and 9 (90%) Caucasians received conservative treatment with pharmacotherapy only. Twenty‐four patients with median age 15 years (13,17), 18 (75%) females, and 23 (96%) Caucasians underwent surgical release (Table  1 ). The conservative group had the most recent follow‐up duration of median of 5 months (range: 1–22 months). In the surgical group follow up included the post‐surgery visit (median: 20 days post‐surgery, range: 7–39 days), and the most recent follow‐up duration of median: 9.8 months, range: 7 days to 4.2 years. Demographic data. Of the total cohort, 33 had both ultrasound and CTA, while 1 had only ultrasound. Among the 33 who had both studies, 21 had abnormalities on both modalities and 12 had only one abnormal result. The single patient who underwent only ultrasound also had an abnormal finding. In comparing conservative and surgical treatment groups, abdominal pain and nausea were the most prevalent symptoms in both. In the conservative group 90% had abdominal pain and nausea each while in surgical group 100% had abdominal pain and 63% had nausea. The two groups, however, were not significantly different for these two symptoms ( p  = 0.3 and p  = 0.2153 respectively). Surgical patients had a significantly higher prevalence of periumbilical pain (88% vs. 70%, p  = 0.0154) and a lower prevalence of upper abdominal pain (4% vs. 30%, p  = 0.01). Other presenting symptoms reported include vomiting (10% vs. 33%, p  = 0.225) and regurgitation (0% vs. 8%, p  = 1, Table  2 ). Presenting symptom. All patients in the conservative group met criteria for DGBI, with diagnoses including IBS (50%), functional dyspepsia/gastroparesis (40%), and functional nausea (10%). In contrast, 46% of the surgical group visited the DGBI clinic, and only 21% met IBS criteria ( p  = 0.0021). Pharmacotherapy use was similar between groups (90% vs. 83%, p  = 1). Furthermore, there were no qualitative differences in the types of medications used, as both groups were treated with similar major medication classes and comparable specific agents, suggesting that the cohorts were similarly medically managed. Though behavioral interventions were more common in the conservative group (80% vs. 50%) it was not significantly different ( p  = 0.14). Other treatments, such as pyloric botulinum injection (30% vs. 29%) and percutaneous electrical nerve stimulation (20% vs. 17%), were used at similar rates. More patients within the conservative group adhered to dietary restrictions (30% vs. 9%, p  = 0.1493, Table  3 ). Comorbidities (chronic pancreatitis, anxiety, postural orthostatic tachycardia syndrome (POTS), hypermobility syndrome, endometriosis, and eosinophilic esophagitis) were largely comparable, except for biliary dyskinesia, which was more common in the conservative group (10% vs. 0%, p  = 0.022). Hypermobility syndrome was also more frequent in the conservative group (40% vs. 17%, p  = 0.2054), but not statistically significantly different (Supporting Information S1: Table  S1 ). Comparison of therapeutic interventions between conservative and surgical groups. Note : This table summarizes the management strategies received by patients in the conservative ( n  = 10) and surgical ( n  = 24) cohorts. Values are presented as number (percentage). p ‐Values were calculated using Fisher's exact test to compare proportions between groups. Outcomes were classified into four‐levels as resolved, improved, no‐change, and worse based on the change in intensity and frequency of presenting symptoms (Figure  1 ). Examining symptom response between the conservative and surgical group, no significant difference was noticed ( p  = 0.21). Postoperative and follow‐up outcomes are presented as the percentage of patients who experienced symptom resolution, improvement, no change, or worsening after surgical or conservative treatment. Recent follow‐up data show comparable rates of resolution and improvement between the two groups, with no statistically significant difference ( p  = 0.21). Notably, worsening symptoms were reported only in the surgical group. F/U, follow up; post op, post operative. Neither the comorbidities nor the interventions received in each study group, showed an association with binary symptom responses (resolved/improved vs. same/worse), results not presented. Examining surgical group only with 4‐level ordinal response using proportional odds model, the cumulative odds ratio (OR) (95% CI) for symptom resolution in comparison to other lower order responses at the most recent follow‐up visit compared to immediate post‐operative visit was 0.11 (0.02–0.56), p  = 0.009. Similarly, when examining symptom resolution and improvement combined versus no‐change and worsening combined, the odds of symptom being resolved/improved were still lower (OR [95% CI] = 0.14 [0.02–0.97], p  = 0.047). To address variability in follow‐up timing, we conducted a sub‐analysis restricted to patients with outcomes available at approximately 5 months. All medical patients ( n  = 10) met this criterion, while five surgical patients were excluded due to lack of follow‐up within this window. At 5 months, there remained no statistically significant differences in either binary improvement (60% vs. 68.4%; p  = 0.698) or 4‐level global scores ( p  = 0.434) between medical and surgical management. Within the surgical cohort, proportional‐odds modeling showed that, compared with the immediate postoperative visit, the odds of full symptom resolution at recent follow‐up were significantly lower (OR 0.21; 95% CI 0.06–0.79; p  = 0.02). In contrast, the binary collapsed model (resolved/improved vs no change/worse) was not significant (OR 0.40; 95% CI 0.08–1.94; p  = 0.26). Overall, the 5‐month sub‐analysis was fully consistent with the primary analysis and did not change the study's conclusions. In the surgical group, the median post‐operative hospital stay was 3 days (IQR 2–5 days). Complications were reported by five patients (21%). Hematochezia and rash were noted in two patients (8%) each, while one patient (4%) experienced a splenic artery injury during surgery which required repair.

Our cohort included 34 pediatric patients, the majority of whom were females and Caucasians. This is consistent with previous studies highlighting that MALS is more common in females and Caucasians with some studies reporting it to be four times more in females in both adults and pediatric populations. 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 , 11 , 12 The clinical presentation of MALS in this pediatric population was consistent with previous reports, with abdominal pain and nausea being the most reported symptoms. 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 Patients in the surgical group were more likely to report periumbilical pain, while those in the conservative group had a higher prevalence of upper abdominal pain. One theory about MALS pathophysiology suggests it arises from ischemia, where blockage of the celiac artery causes intestinal angina. 9 , 13 , 14 This difference along with this vascular theory may reflect the anatomical variability in MALS by which symptoms are dependent on the areas supplied by the narrowed celiac artery. In the conservative group, all patients were diagnosed with concomitant DGBI. This might be explained by the second theory of MALS pathophysiology: the steal phenomena. 9 , 13 It results in a water‐shed effect on the celiac plexus leading pain; and/or disruption of neuro‐enteric pain pathways affecting visceral hypersensitivity mediated through the celiac ganglion. 9 , 13 Visceral hypersensitivity plays a role in the pathogenesis of DGBI, 9 , 15 , 16 , 17 and there might be a common upstream cause for both DGBI and MALS. Our institution provides a multidisciplinary approach to MALS patients, which is not widely available in all hospitals. Both patient groups were evaluated by various specialties (GI, DGBI, surgery, obstetrics and gynecology, and behavioral clinics) to rule out other conditions and address comorbidities. While both groups underwent this approach with no significant differences, significantly more patients in the conservative group received behavioral therapy, including pain coping techniques and CBT. Developing effective coping strategies before surgery may be essential for managing chronic pain, and although not yet studied in pediatric MALS, CBT has shown benefits in other surgical and chronic pain contexts, making it a promising intervention. 18 , 19 Although referral to the DGBI clinic versus direct surgical evaluation was based on initial GI assessment of clinical phenotype rather than a standardized algorithm, this variability may have influenced exposure to conservative therapies such as behavioral interventions. Standardizing the diagnostic and referral pathway (e.g., consistent symptom phenotyping, imaging criteria, and multidisciplinary review) may reduce practice variation and ensure more uniform access to conservative management prior to surgery. Prospective studies are needed to determine whether such standardization improves patient selection and long‐term outcomes in pediatric MALS. Pediatric patients with MALS show high rates of anxiety and mood disorders, similar to those with other diseases with chronic abdominal pain. 20 Another study showed that more than half of the pediatric patients assessed for MALS had at least one psychological condition prior to surgery. 20 Prior studies have noted that a subset of patients with POTS also carry a diagnosis of MALS, and that the gastrointestinal symptoms seen in both conditions can overlap, making clinical distinction challenging. 9 A study done in 2021 recommended screening for celiac artery stenosis after POTS diagnosis is confirmed. 21 It also found an increased prevalence of Ehler‐Danlos Syndrome patients in patients with MALS. 22 , 23 , 24 In our cohort, one surgical patient with vascular Ehlers–Danlos syndrome experienced an intraoperative splenic artery injury. In line with previous research, 83% of our operated patients noted improvement or resolution of symptoms after undergoing surgery. According to earlier studies, 67%–98% of their study population had also noted similar results. 5 , 13 , 18 , 20 , 25 , 26 , 27 , 28 , 29 A recent systematic review, which includes 880 adult and 195 pediatric MALS patients and is thus the largest patient population described to date, showed that 77%–98% of patients had noted improvement or resolution of symptoms postoperatively, consistent with our findings. 28 No other systematic reviews about MALS treatments and outcomes have been published in the past decade. In contrast to the previously published reports, only 54.1% of our surgical patients sustained those improvements at a median follow‐up of 9.8 months. Prior studies had shown sustained outcomes in 70% of patients, but these studies have significantly longer follow‐up, between 6 and 228 months. The newest and largest‐scale review suggests a 70% symptom relief rate in three of six pediatric studies, with follow‐ups of 6–62 months after laparoscopic MALS release. 28 The shorter duration of follow‐up in our study offers a plausible explanation for the differences in results. While our study did not find any predictor for surgery success, a study conducted in 1985 did find that the predictors contributing to better surgical outcomes included female gender, younger age, postprandial abdominal pain, marked weight loss, and absence of a psychiatric history. 30 A more recent study conducted in 2017 determined that only post‐exertional abdominal pain was a predictor of a good surgical outcome. 31 It should also be taken into consideration that the shorter median follow‐up time of our study may have limited the detection of similar predictors. In accordance with the prior findings, no predictors of successful outcomes after conservative management were identified in our study. MALS surgery remains a reasonably safe procedure with a reported complication rate of about 0%–10%. 5 , 26 , 28 , 29 , 32 The only major complication observed in our study was an intraoperative splenic artery injury in a patient with EDS, which may or may not be related to the underlying connective tissue disorder. Areas where patients exhibited postoperative symptoms included pain rash and hematochezia, resolved within 1 day. This study has several limitations that should be acknowledged. The sample size is relatively small, and the retrospective nature of the study means that outcomes are based on subjective clinical documentation, which may introduce bias. Additionally, while we attempted to capture all relevant demographic, clinical, and treatment‐related data, there may be other confounding factors that we were unable to control for. Furthermore, the lack of universally accepted diagnostic criteria for MALS remains a challenge. Moreover, there are inherent limitations to the use of ICD‐10 codes that could introduce bias; for instance, some patients in the conservative group may not have been assigned an ICD‐10 diagnosis code despite having abnormal imaging findings. Finally, there was a variable follow‐up duration between both groups (5 vs. 9.8 months). Our study relied on imaging findings (ultrasound and CT) to confirm the diagnosis, but future research should explore the development of more definitive diagnostic tools and criteria, including biomarkers or functional tests, to aid in the identification of MALS.

Pediatric MALS presents with variable clinical features and heterogeneous responses to both surgical and conservative management. In our cohort, both treatment approaches produced comparable long‐term improvements, despite a stronger early postoperative response among surgical patients. The decline in sustained improvement over time, combined with the absence of identifiable predictors of treatment success, highlights the need for prospective, standardized studies. Future research should focus on refining patient selection, incorporating consistent diagnostic criteria, and evaluating the long‐term durability of conservative and surgical strategies in pediatric MALS.

Median arcuate ligament syndrome (MALS), also known as celiac artery compression syndrome or Dunbar syndrome, refers to a group of clinical signs and symptoms resulting from the compression of the celiac artery by the median arcuate ligament (MAL). 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 First described in 1917, MALS remains controversial due to its unclear pathophysiology, variable presentation, and lack of universally accepted diagnostic criteria. 1 It primarily affects females (4:1 ratio) with a median age of 30–50 years, though it can also occur in children. 2 , 3 Common symptoms include epigastric pain, nausea, vomiting, weight loss, and postprandial or exercise‐induced abdominal pain. 2 , 3 Psychiatric comorbidities, such as anxiety and depression, are frequently observed. 5 , 8 , 9 Diagnosis of MALS typically involves excluding other causes of abdominal pain. 2 , 3 , 4 , 6 , 7 , 8 , 9 Duplex ultrasonography is a useful initial screening tool for detecting changes in blood flow. 2 , 4 , 6 Other imaging modalities, such as computed tomography angiogram (CTA) with inspiratory and expiratory phases, magnetic resonance imaging (MRI), or angiogram, further confirm the diagnosis by demonstrating the classic “J‐hook” shape at the origin of the celiac artery, which normalizes during inspiration. 9 In the early 1960s, Harjola and later Dunbar et al. showcased some alleviation of symptoms following surgical decompression of the celiac artery in individual case reports and a series of cases, respectively. 2 , 3 , 4 Interventions for MALS aim to relieve constriction of the celiac artery by surgical decompression of the MAL, often combined with celiac ganglion neurolysis to address neuropathic pain. 2 , 3 , 4 Historically, the MAL release was performed via laparotomy where the diaphragmatic crura are divided and separated from the celiac axis. Advances in technology have led to minimally invasive techniques for release becoming the standard approach, offering smaller incisions, fewer complications, decreased length of stay and opioid use, and in the case of robotic surgery, better surgical views and instrument range of motion, though it may have drawbacks like incomplete artery release or a risk of aorta injury. 3 , 6 Unsurprisingly, there has been considerable debate not only over whether to offer surgical treatment but also over which surgical intervention is optimal. 9 The existing data on the efficacy of surgery is quite limited, especially in pediatric patients, with relatively short follow‐up periods. Most of the literature consists of retrospective reviews with relatively small case numbers. 9 The aim of this study is to assess the outcomes after surgical and conservative treatment in children with MALS. The primary aim of this study was to evaluate and compare clinical outcomes in pediatric patients diagnosed with MALS who underwent either surgical decompression or conservative medical management. Additionally, we aimed to examine potential predictors of favorable or unfavorable outcomes and identify characteristics of patients most likely to benefit from surgery or conservative therapy. We also evaluated the safety profile and complication rates associated with robotic‐assisted surgical decompression.

Dr. Neha R. Santucci is a key opinion leader for Kate Farms, Gi Health Foundation, and Ironwood Pharmaceuticals. The remaining authors declare no conflicts of interest.

Supporting file 1.