Anatomy of the greater palatine foramen and canal and their clinical significance in relation to the greater palatine artery: a systematic review and meta-analysis

Purpose Accurate knowledge of greater palatine foramen (GPF) and greater palatine canal (GPC) anatomy is necessary to avoid injury to the greater palatine artery (GPA) when performing a variety of anesthesiologic, dental or surgical procedures. The aim of this paper was to perform a systematic review and meta-analysis of literature on the anatomy and localization of bony structures associated with the GPA, namely the GPF and GPC. Methods A systematic literature search was performed using PubMed, Embase, ScienceDirect, and Web of Science databases. Seventy-five studies were included in the meta-analysis (n = 22,202 subjects). Results The meta-analysis showed that the GPF is positioned 17.21 mm (95% CI = 16.34–18.09 mm) from the posterior nasal spine, 2.56 mm (95% CI = 1.90–3.22 mm) from the posterior border of the hard palate, 46.24 mm (95% CI = 44.30–48.18 mm) from the anterior nasal spine, 15.22 mm (95% CI = 15.00–15.43 mm) from the midline maxillary suture, 37.32 mm (95% CI = 36.19–38.45 mm) from the incisive foramen, and opposite the third maxillary molar (M3) in 64.9% (58.7–70.7%) of the total population. Conclusion An up-to-date, comprehensive analysis of GPF and GPC clinical anatomy is presented. The results from this evidence-based anatomical study provides a unified set of data to aid clinicians in their practice.


Introduction
The hard palate is formed by the fusion of the palatine processes of the maxilla and the horizontal plates of the palatine bone at the so-called transverse palatine suture [20].
The mucosa of the hard palate is predominantly supplied by the greater palatine artery (GPA), which originates from the descending palatine artery in the pterygopalatine fossa, descends through the greater palatine canal (GPC), and emerges from the greater palatine foramen (GPF) near the posterior border of the hard palate [46,51,52]. The location of the GPF varies, but it can generally be identified by palpation of the palate opposite the third maxillary molar teeth [35,60,74]. Viveka et al. [83] concluded that the utilization of multiple anatomical reference points, such as the incisive foramen, the midline maxillary suture, and the second and third maxillary molars, simplifies identification of the GPF. Adequate identification of the GPF allows for visualization of arterial pulsations, and confirms the location of the GPA.
At the hard palate, the GPA courses anteriorly in close proximity to the alveolar ridge. The greater palatine nerve traverses a groove medial to the artery, from which it is separated by a palpable crest, which can be used by clinicians to localize both structures [13,60]. The main trunk of the GPA-the lateral branch-enters the nasal cavity through the incisive foramen [50,60], where it anastomoses with the posterior septal branch of the sphenopalatine artery to 1 3 supply the anteroinferior portion of the nasal septum. The diameter of the GPA is greatest at the site of its emergence from the GPF, and then decreases gradually as it courses toward the incisive foramen. The GPA gives off most of its branches in the premolar area, and more commonly toward the alveolar side, rather than to the hard palate [28].
An accurate appreciation of the GPA's location and size is essential to avoid its injury and the resulting surgical and post-surgical complications [69]. Bleeding from the GPA can be difficult to control, with the potential to cause significant blood loss and palatal tissue necrosis [16]. The injury itself, or damage caused by attempts to arrest hemorrhage, may lead to postoperative pseudoaneurysms, or injury to the greater palatine nerve, resulting in paresthesia or insufficient anesthesia of the ipsilateral hard palate [17], and in rare cases, transient ophthalmoplegia [21,22].
Injury to the GPA occurs most commonly during subepithelial connective tissue graft harvesting and can result in prolonged intraoperative bleeding and postoperative wound healing complications related to impaired blood flow [14,72]. In fact, the position of the GPA, along with the thickness of the palatal mucosa, are the two main factors that dictate the size of subepithelial connective tissue grafts that can be safely harvested from the hard palate [16].
GPA injury may also be implicated during down-fracture of the maxilla [12], or in other surgical procedures such as osteotomy of the medial and lateral maxillary sinus walls, pterygomaxillary disjunction, endoscopic medial maxillectomy [13], and pterygopalatine fossa infiltration [12]. The last procedure involves injecting either a vasoconstricting agent into the greater palatine canal--to prophylactically induce hemostasis and limit posterior epistaxis during endoscopic sinus surgery and septorhinoplasty--or an anesthetic solution through the greater palatine canal into the pterygopalatine fossa, to achieve anesthesia of the hemi-maxilla during dental procedures by maxillary nerve block [11]. Clinicians can increase the efficiency and safety of these procedures by referring to the anatomical structures in the oral cavity when determining the adequate position, angle, and length of the needle used for pterygopalatine fossa infiltration [16].
Lastly, the morphological parameters discussed are of clinical significance in the mobilization of GPA for closure of oroantral fistula using mucoperiosteal pedicled palatal flaps [16]; radical release of the GPA during cleft palate repair and reconstruction [26]; and endoscopic cauterization of the GPA at the incisive foramen for the purpose of controlling recurrent or uncontrolled anterior epistaxis [15].
We aimed to update and extend the methodology outlined by Tomaszewska to conduct the meta-analysis on the location of the GPF relative to the maxillary molars, by applying it to other anatomical data extracted from the studies. The objective of our review was to update and extend that of Tomaszewska et al. [75] in 2014. The protocol was methodologically planned and followed, although it was not registered. An updated search strategy was utilized to broaden the scope of the research question to include all available anatomical data to synthesize as evidence by introducing more keyword phrases that describe other related anatomical structures than the GPF.
The main objective was to synthesize evidence from all available studies reporting anatomical data, including cadaveric (i.e., dry skulls) and CT-imaging studies of adult patients (i.e., ≥ 21 years old), combining the results into a comprehensive set of readily available data. The primary outcomes to be measured were the pooled mean estimates of the distances between the center of the GPF and five major anatomical reference points, GPF and GPC diameters, and length and angle of the GPC; and the pooled prevalence estimates of the location of the GPF relative to the maxillary molar teeth, morphology of the GPF, and direction of GPF opening into the oral cavity. Secondary outcome measures included subgroup analysis based on the geographical region of the studies included in the analysis, to probe for sources of heterogeneity.

Search strategy
The authors strictly followed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) [57] guidelines throughout the literature search [Online Supplementary File 1].
The major electronic databases (PubMed, Embase, Scien-ceDirect, and Web of Science) were searched extensively to identify articles eligible for inclusion in our meta-analysis up to July 2022. No lower date limit was applied. The following search terms: "greater palatine artery", "greater palatine canal", "greater palatine foramen", "pterygopalatine fossa anatomy", "pterygopalatine canal", and "descending palatine canal" were used in different combinations, as shown in Table 1. The references in all included articles were searched manually to identify any further relevant publications. We included only published studies, relying on the journal review process as one step of quality control.

Eligibility
Study eligibility for inclusion in our meta-analysis was assessed independently by two reviewers (J.R. and W.R.). Studies were considered eligible for inclusion if they (1) were cadaveric or imaging studies, and (2) reported relevant and extractable data on the clinical anatomy of the greater palatine artery, foramen, or canal. The reviewers did not 1 3 consider (1) case reports, systematic reviews, animal studies, letters to editors, or meta-analyses, (2) studies that provided missing, unclear, or incomplete results, and (3) studies that did not clearly define (by text or figures) the descriptive anatomy used in the study [33]. Review of full-text articles was limited to the ones published in English language. All differences of opinion among the reviewers concerning the eligibility of the studies were resolved by consensus through consultation with a third reviewer (D.K.).

Data extraction
The studies were analyzed looking for all numerical parameters that could be directly compared between studies. This meant that the same parameter was used in at least two different studies and measured with a comparable degree of precision. The following parameters were included: 1. Distance between the GPF and the posterior nasal spine (GPF-PNS) 2. Distance between the GPF and the posterior border of hard palate (GPF-PBHP) 3. Distance between the GPF and the anterior nasal spine (GPF-ANS) 4. Distance between the GPF and the midline maxillary suture (GPF-MMS) 5. Distance between the GPF and the incisive foramen (GPF-IF) 6. Location of the GPF in relation to the second (M2) and third (M3) maxillary molars 7. Diameter of the GPF in anteroposterior (AP) and lateromedial (LM) dimensions 8. Shape of the GPF 9. Direction of GPF opening into the oral cavity 10. Angle of the GPC relative to the vertical plane and to the transverse plane 11. Length of the GPC 12. Diameter of the GPC upper opening in the anteroposterior (AP) dimension

Quality assessment
The authors used the AQUA tool to evaluate both the quality and accuracy of the anatomical studies incorporated into this meta-analysis, as well as to properly classify their quality and risk of biases [32]. The assessment covers five domains: (1) objective(s) and study characteristics, (2) study design, (3) methodology characterization, (4) descriptive anatomy, and (5) reporting of results. The potential risk for bias in each domain is appraised by judging it as "low," "high," or "unclear" using the signaling questions with answers "yes," "no," or "unclear," respectively. In other words, all queries answered with "yes" place the corresponding domain in the "low" risk of bias category, whereas all queries answered with "no" place the corresponding domain in the "high" risk of bias category. Inadequate data that did not allow for clear scrutiny were placed in the "unclear" risk of bias category.

Statistical analysis
The extracted data were pooled into a meta-analysis using R software, with the 'meta' package (R Foundation for Statistical Computing). The inverse-variance, random-effects model was used to calculate the pooled effect size estimate across the studies, and the DerSimonian-Laird method was used to estimate the between-study variance, τ 2 . Statistical heterogeneity was assessed using the I 2 statistic and interpreted according to the guidelines in Chapter 9.5.2 of the Cochrane Handbook (Higgins 2011). This statistic expresses (greater AND palatine AND artery) OR (greater AND palatine AND canal) OR (greater AND palatine AND foramen) OR (pterygopalatine AND fossa AND anatomy) OR (pterygopalatine AND canal) OR (descending AND palatine AND canal) ScienceDirect ("greater palatine artery") OR ("greater palatine canal") OR ("greater palatine foramen") OR ("pterygopalatine fossa anatomy") OR ("pterygopalatine canal") OR ("descending palatine canal") Web of Science Core Collection/SciELO/BIOSIS/Current Content Connect/Korean Journal Database/Russian Citation Index (((((ALL = (greater palatine artery)) OR ALL = (greater palatine canal)) OR ALL = (greater palatine foramen)) OR ALL = (pterygopalatine fossa anatomy)) OR ALL = (pterygopalatine canal)) OR ALL = (descending palatine canal) 1 3 the percentage of variation across studies. Heterogeneity of I 2 < 25% was considered low, between 25 and 75% was considered moderate, and > 75% was considered high. Subgroup analyses based on the geographic regions in which the studies were performed were conducted to detect sources of heterogeneity. To assess statistically significant differences between two or more subgroups, confidence intervals were compared. If the confidence intervals overlapped, then the differences were considered statistically insignificant [33].

Study identification
The study identification process is presented in Fig. 1

Characteristics of included studies
The characteristics of included studies are presented in Table 2. A total of 75 studies (n = 22,202 subjects) were considered eligible and were included in the meta-analysis. In total, there were 29 imaging studies and 46 cadaveric studies. The studies spanned the years 1982 to 2022 and originated from Africa, Asia, Europe, North America, and South America.

Distance between the greater palatine foramen and selected anatomical landmarks
The results of the meta-analysis regarding the distance between the greater palatine foramen and surrounding anatomical landmarks are presented in Table 4. A total of 8 studies [8,11,29,35,56,63,77,83] (n = 2358 subjects) reported data on the distance from the greater palatine foramen to the posterior nasal spine (GPF-PNS  Table displaying characteristics of the 75 studies that were included in the meta-analysis, sorted in alphabetical order (A-Z) by last names of the first authors of the studies. The study characteristics included country of study origin, whether the study subjects were either dry skulls (i.e., cadaveric studies) or CBCT (cone-beam computed tomography) scans (i.e., imaging studies), number of subjects in the studies, and the number of sides included in each study  High Low Low Low Low Aoun et al. [5] Low Low Low Low Low Aoun and Nasseh [3] Low Low Low Low Low Aoun et al. [4] Low Low Low Low Low Apinhasmit et al. [6] Low Low Low Low High Ashwini and Jaishree [7] High Low Low Low Low Awad et al. [8] Low Low Low Low Low Ayoub et al. [9] Low Low Low Low Low Badshah et al. [10] High Low High Low Low Bahşi et al. [11] Low Low Low Low Low Beetge et al. [12] High Low Low Low Low Cagimni et al. [16] High Low Low Low High Campbell et al. [17] High Low High Low Low Cheung et al. [18] High High High High Low Chopra et al. [19] High Low Low Low High Chrcanovic and Custódio [20] High Low Low Low Low Das et al. [22] High Low High Low Low Dave et al. [23] High Low High High Low D'Souza et al. [21] Low Low Low Low Low Douglas and Wormald [24] High Low Low Low Low Duruel et al. [25] Low Low Low Low Low Fonseka et al. [27] Low Low Low Low Low Fu et al. [28] High Low Low Low High Gibelli et al. [29] Low Low Low Low Low Hassanali and Mwaniki [31] Low Low Low Low Low Howard-Swirzinski et al. [34] Low Low Low Low Low Hwang et al. [35] Low Low Low Low Low Ikuta et al. [36] Low Low Low Low Low Ilayperuma et al. [37] High Low Low Low Low Jaffar and Hamadah [38] High Low High Low Low Kaffe et al. [39] High High Low Low Low Kang et al. [40] High A total of 4 studies [5,27,40,59] (n = 365 subjects) reported data on the distance from the greater palatine foramen to the anterior nasal spine (GPF-ANS). The pooled mean, across the four studies, was calculated to be 46.24 mm (95% CI = 44.30-48.18 mm). The Q test showed high heterogeneity (Q = 90.68; p < 0.0001), which was confirmed by the I 2 test (96.7%). Subgroup analysis based on geographical region was not performed due to the low number of studies; there were only two possible subgroups and one of these contained only one study, precluding the possibility of pooling the mean using meta-analysis.

Location of the greater palatine foramen in relation to maxillary molars
The results of the meta-analysis regarding the location of the GPF in relation to the maxillary molar teeth are presented in Table 5. Only two studies [19,27] (n = 284 subjects) reported data on the prevalence of the greater palatine foramen being located "anterior to the 2nd maxillary molar teeth". The pooled prevalence, across the two studies, was calculated to be 3.27% (95% CI = 0.45-20.29%). The statistical significance of the Q test (Q = 6.67, df = 1, p = 0.0098) allowed the null hypothesis of homogeneity to be rejected. The I 2 test showed moderate to high heterogeneity (I 2 = 85.0%; 95% CI = 58.5-94.6%).

Morphology of the greater palatine foramen
The results of the meta-analysis regarding the morphology of the GPF are presented in Table 6. A total of 6 studies [6-8, 19, 41, 48, 60, 64] (n = 986 subjects) reported data on the prevalence of the greater palatine foramen being "oval/ ovoid" in shape. The pooled prevalence, across the six studies, was calculated to be 77.8% (95%CI = 57.6-90.0%). The statistical significance of the Q test (Q = 150.22, p < 0.0001) allowed the null hypothesis of homogeneity to be rejected. The I 2 test showed high heterogeneity (I 2 = 96.7%; 95% CI = 94.7-97.9%).

Direction of opening of the greater palatine foramen
The results of the meta-analysis regarding the direction of opening of the GPF into the oral cavity are presented in Table 7. A total of 10 studies [1,2,7,37,43,65,66,68,77,82] (n = 4,534 subjects) reported data on the prevalence of the greater palatine foramen opening into the oral cavity in the inferior-anterior-lateral direction. The pooled prevalence, across the ten studies, was calculated to be 14.41% (95% CI = 4.91-35.43%). The statistical significance of the Q test (Q = 873.78, p < 0.0001) allowed the null hypothesis of homogeneity to be rejected. The I 2 test showed high heterogeneity (I 2 = 99.0%; 95% CI = 98.7-99.2%). Table 6 Size and shape of the greater palatine foramen CI confidence interval, AP anteroposterior, LM lateromedial a Two studies [12,25] were excluded from the subgroup analysis because they were the only studies in their own respective subgroups b One study [60] was excluded from the subgroup analysis c Meta-analysis for the greater palatine foramen shapes "Round", "Slit/lancet", and "Other" were not followed up by subgroup analysis due to the low number of studies A total of 15 studies [1,2,7,20,31,37,38,41,43,55,65,66,68,77,82] (n = 5,312 subjects) reported data on the prevalence of the greater palatine foramen opening in the inferior-anterior-medial direction into the oral cavity. The pooled prevalence, across the 15 studies, was calculated to be 54.54% (95% CI = 40.53-67.87%). The statistical significance of the Q test (Q = 886.74, p < 0.0001) allowed the null hypothesis of homogeneity to be rejected. The I 2 test showed high heterogeneity (I 2 = 98.4%; 95% CI = 98.0-98.7%).
A total of 2 studies [35,52] (n = 310 subjects) reported data on the measured angle between the transverse plane and the axis of the greater palatine canal. The pooled mean, across the two studies, was calculated to be 62.63° (95% CI = 53.32-71.94°). The statistical significance of the Q test (Q = 141.84, df = 1, p < 0.0001) allowed the null hypothesis of homogeneity to be rejected. The I 2 test showed high heterogeneity (I 2 = 99.29%; 95% CI = 98.75-99.60%).

Discussion
To date, the leading anesthesiology and surgery textbooks have offered only general descriptions regarding clinical localization of the greater palatine foramen (GPF) and greater palatine canal (GPC), often leading to inconsistencies in physician training [82]. Though a large number of studies have been conducted concerning the location and morphometric characteristics of the GPF and GPC, many of these publications report an ongoing difficulty in localizing these structures, and therefore identifying the GPA in clinical settings [30]. Locating the GPF in relation to maxillary molar teeth remains a fast and effective way for clinicians to estimate the location of the GPF. Our findings were consistent with those of a similar review by Tomaszewska et al. [75], which also revealed that the GPF is most commonly located opposite the third maxillary molar (M3). Our results add substantial value to the findings of Tomaszewska et al. [75]. The analysis of the prior review contained only 23 studies (n = 6927 subjects) and the pooled prevalence was estimated to be 63.9% with a 95% confidence interval ranging from 56.5 to 70.9%. Our review, which contained a total of 38 studies (n = 9,754 subjects) and a pooled prevalence of 64.9%, with a 95% confidence interval from 58.7 to 70.7% strengthens the validity of the findings of Tomaszewska et al. [75][76][77] with the addition of 15 studies, adding significantly to the overall sample size, and narrowing the 95% confidence interval.
An additional aspect to consider when referencing the GPF to the maxillary molars is the size and shape of the GPF. Our meta-analysis revealed that the GPF has an anteroposterior (AP) diameter of 5.34 mm and lateromedial (LM) diameter of 2.82 mm, representing the major and minor axes, respectively. This is consistent with our other findings that the GPF was described as "oval or ovoid" in shape in 77.78% of the population. A possible explanation for such AP elongation of the GPF, is that the AP dimension of the palate increases with the eruption of the posterior teeth.
In edentulous patients, the location of the GPF can be accurately triangulated using measured distances to easily identifiable landmarks, the most reliable of which are the median maxillary suture (MMS), the posterior border of the hard palate (PBHP), and the incisive foramen (IF), rather than the posterior nasal septum (PNS) and the anterior nasal septum (ANS). The topography of the hard palate with reference to the anatomical landmarks is of clinical importance also when obtaining free gingival and connective tissue grafts [42], where the distance from the GPF to the incisive foramen (GPF-IF) is used to estimate the possible length of the graft [20,42,64].
Furthermore, using GPF-IF and GPF-MMS, it is possible to derive the angle between the MMS and the line from the IF to the GPF, which Tomaszewska et al. [75] called the MMS-IF-GPF angle. Utilizing our findings for GPF-IF and GPF-MMS, we found the MMS-IF-GPF angle to be 24.07 degrees, which is consistent with the angle calculated  [20].
Knowing the MMS-IF-GPF angle may also be useful in determining the angle to be made by the needle for anesthetic infiltration into the GPF [20,64].
In the setting of maxillary nerve block and hemostasis using the GPC approach, the length of the GPC is particularly relevant. For anesthesia, the needle must advance 30 mm, while for hemostasis, specifically during sinus surgery, it is recommended to infiltrate the needle as deep as 25 mm [75].
Our meta-analysis results suggest that anatomical variation of the direction of opening of the GPF may occur more frequently than previously thought. An inferior-anterior-medial (I-A-M) opening relative to the sagittal plane was found 54.54% of the time, considerably less than that was previously estimated at 82.1% [75]. The second most common direction of opening was in the anterior direction, occurring 30.11% of the time in our study, in stark contrast to 7.6% in the 2015 study by Tomaszewska et al. [77] The most common method of administering anesthesia via the GPF was to bend the needle to an angle of 30-45 degrees. In light of our findings, it may be advisable to administer anesthesia to the maxillary nerve by bending the needle to an angle closer to 30 degrees, as the smaller angle would mitigate the risk of puncturing the hard palate soft tissue in the case that the GPF opens in the anterior direction.
One notable variation in GPF anatomy, as shown in ultrasonographic imaging studies [22], is a bony ledge that partially covers the opening of the foramen; in the presence of this variation, the data collected and pooled on the direction of opening becomes a clinically difficult statistic, and represents another challenge the clinician must be aware of when inserting a needle into the GPF.
We met with several limitations during our systematic review which was the lack of studies which directly described the anatomy of the GPA; this prevents us from making conclusions about the course of the artery itself, at least distal to the GPF. Another issue was the heterogeneity of the included studies, both in terms of the parameters measured and the modalities used to measure them (e.g., imaging versus cadaveric studies). For instance, as mentioned above, different studies used different categories to report the location of the GPF in relation to the maxillary molars, as well as its shape (see also Tables 3,4). On the other hand, some parameters-such as the distance between the GPF and the nasal spines-were only reported in a small number of studies ( Table 2). The main limitation of the meta-analysis was the substantial heterogeneity among the included studies, which persisted even after subgroup analysis based on geographical region. The included studies featured little information on individual patient characteristics, such as gender, precluding a more detailed subgroup analysis. The majority of the studies were performed on dry adult skulls, and consequently, the majority of these studies also did not report gender or age, which posed a limitation when probing for possible sources of heterogeneity.
We propose that the maxillary molar teeth, midline maxillary suture, posterior border of the hard palate, and the incisive foramen are the most reliable anatomical landmarks to accurately locate the GPF. Clinicians may expect to locate the foramen 15.00-15.44 mm from the midline maxillary suture, 1.90-3.22 mm from the posterior border of the hard palate, and 36.19-38.45 mm from the incisive foramen. The main findings are summarized in Fig. 3.

Acknowledgements
The illustration (Fig. 3) is an original work done by the corresponding author of this study, Dong Woon Kim.
Author contributions DWK: data collection or management, data analysis, manuscript writing/editing. JT: manuscript writing/editing. JS: manuscript writing/editing. JR: data collection or management. WR: data collection or management. IŚ: data collection or management. JRP: data collection or management. IMT: protocol/project development.
Funding The authors declare that they did not receive any funding.
Data availability All of the data that were extracted from the included studies during the data collection process and used to perform our analyses are stored in a repository and publicly available on Open Science Framework (https:// osf. io/ 64thm/? view_ only= d4724 58ec3 084d9 da9ef efdfd 396b6 05).

Conflict of interest
The authors declare that they have no conflict of interest.

Ethical approval Not applicable.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.