The posterior gastric artery: A meta‐analysis and systematic review

The aim of this study was to review the literature on the posterior gastric artery, estimate its prevalence and summarize its reported origins. The databases Pubmed, Scopus, Web of Science and Google Scholar were searched to find all studies describing the prevalence and origin of the posterior gastric artery. Pooled prevalences were estimated using a random effects model. Thirty‐eight studies with a total of 3366 subjects were included in the analysis. The overall prevalence of the posterior gastric artery was 57.4% (95% CI = 49.1%–65.7%). The prevalence of the posterior gastric artery was significantly higher in surgical studies than in cadaveric and angiographic studies. There were no differences in prevalence between multi‐detector computed tomography studies and cadaveric studies, nor were there differences when comparing geographical location or study size. Origin data were extracted from 34 studies, with a total of 1533 cases. The posterior gastric artery arose as a single vessel from the splenic artery in 1160 cases (pooled prevalence 86.5% [95% CI = 78.5%–94.7%]), from the superior polar splenic artery in 339 cases (pooled prevalence 11.8% [95% CI = 3.7%–19.9%]) and from other origins in 50 cases (pooled prevalence 0.27% [95% CI = 0.00–0.71%]). The posterior gastric artery is present in 57.4% of cases and most commonly arises from the splenic artery. It should be identified before gastric resections as it may be an important source of blood to the gastric stump. Multi‐detector computed tomography has sufficient sensitivity to detect it before surgery.


| INTRODUCTION
First described by Walther (1740) and named by Haller (1745), the posterior gastric artery (PGA; arteria gastrica posterior) has been often neglected in the literature, with recent anatomical textbooks providing only short descriptions or not mentioning it at all.Nonetheless, this artery has important clinical implications in gastric surgery.It may be an important source of blood to the remnant stomach after resection and if accidentally severed may result in gastric ischemia or bleeding (DiDio et al., 1980;Suzuki et al., 1978;Yu & Whang, 1990).
No uniformly accepted definition of the posterior gastric artery currently exists.In most cases, it is described as a branch of the splenic artery traveling to the upper portion of the posterior surface of the stomach.It has also been described as arising from or giving off the superior polar splenic artery.Other authors consider it only to be a proximal branch of the splenic artery, disregarding its origins from other arteries (Helm, 1915;Ishii et al., 2018).Loukas et al. define the artery not by its origin but instead by the area of the stomach it supplies, which they report as the posterior wall of the upper part of the body of the stomach near the cardia and fundus (Loukas et al., 2007).
The posterior gastric artery may be encountered during several surgical procedures involving the posterior aspect of the stomach.The posterior gastric artery may, under certain circumstances, be a significant, if not the only, source of arterial blood supply to the gastric remnant after subtotal gastrectomy.Injury or interruption of this artery can therefore lead to necrosis of the gastric remnant or impaired healing of the anastomosis (Suzuki et al., 1978;Yu & Whang, 1990).Additionally, it can be a source of complications during mobilization of the gastric fundus during fundoplication or sleeve gastrectomy.Thus, preoperative knowledge of its anatomical features and variations has clinical importance for safe and effective practice of surgery.
The reported prevalence of the posterior gastric artery varies considerably in the literature, from as low as 4% up to 100% (Laude et al., 1972;Okabayashi et al., 2006).We hypothesize that the various different reported prevalences of the posterior gastric artery are expected to be due to the differences in study modalities, sample sizes as well as lack in standardization of reporting anatomical data.
The aim of this study was thus to summarize the current data on the posterior gastric artery in the literature, to calculate its pooled prevalence by means of meta-analytic techniques, to investigate sources of heterogeneity and summarize its clinical importance.database PubMed the MeSH terms "gastric artery" and "splenic artery" were used.Studies that reported complete prevalence data on the origin of the posterior gastric artery were selected for inclusion into the meta-analysis.Studies that did not contain quantitative data on the PGA or had insufficient data were excluded.For studies that had incomplete data we always attempted to contact the authors, whenever possible.In accordance with the guidelines for writing anatomical meta-analyses, case reports, conference abstracts, and letters to the editor were excluded.However, these sources were reviewed for anatomical or clinical interest and were used for the discussion whenever appropriate.No date or language restrictions were applied.
Data extracted from the studies included: country, date of publication, study type (adult cadavers, fetal and neonatal cadavers, angiography, computed tomography, surgical), sample size and the number of posterior gastric arteries reported, their origin and when available diameter.
We used Google sheets to collect the extracted data.The estimated pooled prevalence of the origins of the posterior gastric arteries were calculated in a random effects meta-analysis using the program R version 3.4.3(R Foundation for Statistical Computing) with the software package meta.All estimated pooled prevalences were made using a random effects model unless stated.For testing heterogeneity between the studies, we used the Cochran's Q test and Higgins I 2 statistic.Cochran's Q p-value of <0.10 was used to indicate significant heterogeneity.Higgins I 2 values range from 0% to 100% with higher values indicating higher heterogeneity.Subgroup analysis was performed according to sample size, study type and geographical location.

| Study selection
Searching the medical databases using the search criteria identified 2808 articles.An additional 31 were found from other sources.After removing duplicates 900 articles underwent abstract and title screening, which resulted in exclusion of 821 articles.A total of 79 full texts were then assessed for inclusion; 41 were removed and 38 met the inclusion criteria.The study selection process is shown in Figure 1.

| Study characteristics
In total 38 studies were included with a combined total of 3366 subjects.The studies originated from 14 countries, from five continents.
The most frequent country of origin was Japan (n = 8), followed by the United States of America (n = 6) and India (n = 4).Twenty-six studies were based on anatomical specimens, five on multi-detector computed tomography (MDCT), four on angiography and three on surgical procedures.Concerning the studies based on anatomical specimen 19 used adult cadavers, three used both adult and fetal cadavers, two used only fetal cadavers, one used full-term stillborns and one used neonatal cadavers.

| Prevalence of the posterior gastric artery
Using a random effect model the overall prevalence of the posterior gastric artery was 57.4% (95% CI = 49.1%-65.7%).Using a common effect model the prevalence was 87.2% (95% CI = 80.1-82.2%).The prevalence of the posterior gastric artery in each included study is shown in Table 1.A subanalysis showing the prevalence of the posterior gastric artery according to study modality, geographical location and study size is shown in Table 2.When analyzing the prevalence by study modality, the prevalence was the highest in surgical studies (79.8% [95% CI = 67.2-92.3]),followed by MDCT-based studies (60.5% [95% CI = 36.2%-84.8%])and cadaveric studies (56.5 [95% CI =46.0%-67.1%])and the lowest prevalence was in angiographic studies (41.8% [95% CI = 28.8%-54.9%]).Prevalences were significantly different between surgical and angiographic studies and between cadaveric and surgical studies.There were no differences when comparing geographical location or study size.

| Origin of the posterior gastric artery from the superior polar splenic artery
This has been described as a gastric branch of the superior polar splenic artery or a splenic branch of the posterior gastric artery.Some authors describe the posterior gastric artery arising from the superior polar splenic artery (Ishikawa et al., 2018;Okabayashi et al., 2006;Pity nski et al., 1996;Skawina et al., 1999).Other authors describe the superior polar splenic artery arising from the posterior gastric artery (Rossi & Cova, 1904;Sahni et al., 2003;Trubel et al., 1985).A few studies have adopted a compromise, describing all such cases as gastrosplenic arteries (Tanigawa, 1963;Trubel et al., 1988).Further, Trubel et al. describes three subtypes depending on if the diameter of one branch is larger, smaller or the same size as the other (Trubel et al., 1988).Tanigawa describes three subtypes depending on whether both branches are present or if one of the branches is missing (Tanigawa, 1963).As the term gastrosplenic artery also refers to a variant of the coeliac trunk where the common hepatic artery has an aberrant origin, we decided not to use this term and to instead refer to all these cases as posterior gastric arteries arising from the superior polar splenic artery for the sake of this meta-analysis.In total there were 339 reported cases of this origin from a total of 708 posterior gastric arteries from 10 studies.

| Origin of the posterior gastric artery from the splenic artery
Seven studies with a total of 209 cases specified which part of the splenic artery the posterior gastric artery originated from (proximal, middle, or distal third).In 74 cases it arose from the proximal third (pooled prevalence 48.9% [95% CI 15.5-82.4%]); in 96 cases from the middle third (pooled prevalence 37.0% [95% CI 11.1%-62.8%]),and in 43 cases from the distal third (pooled prevalence 15.1% [95% CI 2.4%-27.9%]).The origin of the posterior gastric artery from the proximal, middle and distal thirds of the splenic artery as well as from the superior polar splenic artery is shown in Figure 2.

| Diameter
The mean diameter of the posterior gastric artery was reported in 7 studies (477 cases).The combined mean was 1.67 mm.

| DISCUSSION
The posterior gastric artery (PGA) is variably present with a total pooled prevalence of 57.4% (95% CI 49.1-65.7%).Our analysis indicates that T A B L E 1 Prevalence of the posterior gastric artery.there is a large amount of heterogeneity both between studies and within the subgroups.Consequently, we used a random effects model to deal with this, but we cannot be certain that all confounds are appropriately controlled so our findings must be treated with caution.
They high may be due to differences in methodology and lack of the use of guidelines for standardized reporting of anatomical data.
In almost all studies the posterior gastric artery was defined as arising from the splenic artery or the superior polar splenic artery.
T A B L E 2 Prevalence of the posterior gastric artery subanalysis.F I G U R E 2 Origin of the posterior gastric artery.PGA, posterior gastric artery; SA, splenic artery; SSPA, superior splenic polar artery.
Other origins were very rare and included the two cases of the coeliac trunk (Okabayashi et al., 2006;Okabayashi, Kobayashi, Morishita, et al., 2005), sixteen cases of the left inferior phrenic artery (Tang et al., 2010) and nineteen cases of the left gastric artery (Gupta et al., 2020).Loukas et al. used a different definition to the other studies.They defined the posterior gastric artery by the area of the stomach that it supplied and not by its origin.In doing so they found a substantial amount of PGAs arising from the left gastric artery and the coeliac trunk.These cases may be considered branches of the left gastric arteries by other authors.For this reason, we did not include them in the origin analysis (Loukas et al., 2007).
Since we did not find a difference between the prevalence of PGA in cadaveric studies compared to studies based on MDCT, it can be concluded that MDCT is a sufficiently sensitive method for its detection.This fact allows surgeons to make the decision on the management of PGA and gastric stump already before the procedure.
The clinical significance of PGA relates mainly to gastric cancer surgery.In subtotal gastrectomy with D2 lymphadenectomy, the PGA can potentially be the only artery supplying the gastric stump (Yu & Whang, 1990).Knowledge of its existence should lead to greater caution with the goal of preserving it during lymphadenectomy around the splenic artery.On the contrary, its absence may lead to the decision to choose total gastrectomy instead of subtotal gastrectomy for fear of necrosis of the gastric stump or a leak from the gastro-enteroanastomosis-already within the preoperative planning process.
Ishii et al. demonstrated the existence of lymph nodes surrounding the PGA (Ishii et al., 2018).PGA-preserving lymphadenectomy of this area is challenging and is only possible with knowledge of its existence and detailed anatomy.
The PGA has clinical implications in pancreas transplantation.
Overlooking the PGA during pancreas harvesting can lead to significant bleeding and its treatment to damage the vascular supply and loss of the graft.Troppman et al. reported a case where overlooking the PGA lead to a splenic artery injury during procurement (Troppmann et al., 2004).
Aneurysms of gastric arteries account for 4% of visceral artery aneurysms, with the majority affecting the left and right gastric arteries.
Less commonly the gastro-omental arteries are involved.An aneurysm of the PGA has been reported in the literature.Gomes et al. reported an incidental finding of a PGA that was treated by coil embolisation (Gomes et al., 2018).It is recommended to treat all gastric artery aneurysms as they present a high risk of bleeding (Chaer et al., 2020).
Gupta et al. reported on the importance of recognition of the PGA during bariatric surgery.They reported that it is necessary to divide the PGA in patients undergoing laparoscopic sleeve gastrectomy to mobilize the fundus and in patients undergoing and laparoscopic Roux-en-Y gastric bypass it is necessary to dissect or divide the PGA to facilitate firing of the longitudinal stapler (Gupta et al., 2020).

| CONCLUSION
The posterior gastric artery is present in 57.4% of cases and most commonly arises from the splenic artery.Multi-detector computed tomography is a sufficient examination technique for the preoperative identification of the PGA.It is important to identify the PGA when performing gastric resections as it may be an important source of blood to the gastric stump and when mobilizing the fundus and during pancreatic procurements it may be a source of bleeding.

F
I G U R E 1 Study selection flowchart.