Reflux is a bothersome outcome of LSG, and its aetiology is still not well understood. De novo reflux or worsening reflux is reported to be between 8 and 50% of all post sleeve gastrectomy patients [8, 19, 20].

Reflux can be related to defective lower oesophageal sphincter anatomy and function because of hiatus hernia, whilst some studies show good outcomes of hiatus hernia repair following LSG [20, 21]. Others suggest that reflux is not improved unless the hiatus hernia repaired is larger than 4 cm [12, 19, 21, 22]. In our study, reflux symptoms were present in both those who had hiatus hernia and those who did not. Further sub-analysis showed no correlation between symptomatology and hiatus hernia size; however, 6 (66.6%) of the patients undergoing hiatal repair alone reported a post-op Visick score of less than 3.

Weight regain is another risk factor for reflux; however, there was concordance for lower BMI at representation in patients presenting with reflux, regurgitation/dysphagia than those with weight regain in our study (p = 0.07). Other studies have also found that reflux was also independent of weight regain in LSG patients [23, 24].

There are increasing arguments that sleeve shape impacts the function of the sleeve and therefore might present a combined anatomical and functional cause of concomitant reflux [20, 25, 26]. Laplace’s law explains that the pressure inside an inflated and elastic container with a curved surface, e.g. a bubble or a blood vessel, is inversely proportional to the radius as long as the surface tension is stable, whilst Poiseuille’s law indicates that the flow rate is directly proportional to the pressure difference and to the fourth power of the radius of the pipe or vessel. This means that even small changes in the radius can have a significant impact on the flow rate [27] and that a dilated pipe, proximal to a stenosis, will be prone to more dilation.

Anatomic abnormalities in the LSG can be in the form of an internal indentation and/or a sharp angulation of the gastric lumen, which creates a flap valve producing a functional obstruction, typically at the incisura, hence angularis incisura stenosis [13, 24]. The narrowing or stenosis could be the result of over-tight stapling and/or twisting of the sleeve from misalignment of staple lines [19, 28]. An alternate or contributing cause could arise in the years following the procedure whereby the vertical part of the sleeve dilates and the sleeve tube lengthens, thereby creating a more acute angularis angle and relative stenosis. This will lead to increased proximal gastric pressures based on the above-mentioned laws [29]. The most prevalent surgical treatment for post-LSG reflux is conversion to gastric bypass which promotes improvement in reflux control via the putative mechanisms of reduced parietal cell mass, rapid transit time through to stomach and reduced intragastric pressures [22, 23]. Whilst proving the anatomic causes of reflux in LSG can be challenging [5, 14], very few surgeons would be happy to construct a gastro-enterostomy anastomosis below a gastric stenosis if their aim was to improve reflux symptoms.

Endoscopic evaluation is a valuable tool in the diagnosis of abnormalities such as reflux it can prove to be less sensitive in establishing narrowing of conduit, where passage of the scope could be interpreted as a normal examination. This comes in addition to the potential surgeons’ bias playing a role in under-interpreting a technical error which may contribute to patients’ symptoms [15, 30].

Our study identified 2 (3.1%) endoscopic evidence of stenosis, which is similar to reported incidence of up to 3.9% in other studies. However, there were other 7 (11.9%) patients in our cohort with endoscopic evidence of reflux without presence of other contributing factors on the endoscopy. 3DCT however was able, in these patients, to suggest a potential anatomic cause.

3D computed tomography scans can help determine gastric volume; the presence of hiatus hernia with the highest accuracy [31] can also provide detailed information of the shape and measurements of the sleeve [31, 32] including detection of relative or functional stenoses which may be missed on endoscopy or other studies.

If we follow the rules of physics of angulated non-reinforced tubes, we can deduct that the creation of a narrow distal sleeve might impact on its shape and increase the angulation at the incisura, therefore reducing the surface area at that point and creating a significant pressure gradient [33]. A higher proximal pressure may lead to subsequent proximal stretch and lengthening of the conduit overtime, similar to what is seen in vascular biomechanics [34]. The results of our study indicate a reduction of angularis surface area when sharper angulation of the sleeve occurs. Furthermore, the sharper the angle the wider the oesophageal diameter above the hiatus was (p = 0.008) and the longer the conduit length (p = 0.01). We also see a significant positive correlation between the sleeve lengthening and duration of primary surgery to 3DCT (p = 0.024). Similarly, patients with endoscopic evidence of stenosis and/or reflux presented long after those with normal or hiatus hernia only. All of this corresponds to previously reported timing of presentation of patients with stenosis that can take months or years to occur [5, 14, 35].

When we compare 3DCT findings to endoscopic findings, we find lower volumetry, angularis and distal surface area in those diagnosed with stenosis and/or reflux on endoscopy (Table 5). In specific, we found more patients with disparity in the shape of the sleeve in the form of (ASA/PMSA ≤ 0.5) in those diagnosed with stenosis and/or reflux on endoscopy (p = 0.02).

In those who underwent sleeve to bypass/or stricturoplasty for reflux, patients experienced better results if they had a disparity between angularis surface area and proximal conduit surface area, in particular more patients experienced better outcome if they had an ASA/PMSA ratio of 0.5 or less. Whilst bypass in general is thought to have a better therapeutic profile for management of sleeve reflux [36], our findings suggest that patients with a detectible stenosis did well if the surgery addressed the stenosis, whereas patients without an anatomic abnormality did less well.

Our study is limited by a retrospective design, a relatively small cohort of patients and lack of control of 3DCT imaging in asymptomatic patients post sleeve. Performing a prospective study with inclusion of asymptomatic patients would further help standardise measurements of the conduit and reduce bias. In addition, adding manometry might help establish proximal gastric pressures in patients with possible gastric obstruction and therefore would add another confirmatory test to compare with the CT findings.