Nerve anastomoses in the upper limb: an up-to-date meta-analysis Authors: Dawid Plutecki, Michał Bonczar, Jerzy Walocha, Weronika Chaba-Karnaś, Antonina Nowak, Tomasz Kozioł, Kacper Bąk, Marcel Frankiewicz, Mariusz Trystuła, Mateusz Koziej Published: 2026 ABSTRACT AbstractIntroductionThere are numerous communicating branches (CB) between the nerves of the brachial plexus. Therefore, the objective of this study was to provide comprehensive data on the prevalence, morphometry, and anatomical classifications of upper extremity anastomoses such as: Martin–Gruber anastomosis (MGA), Marinacci anastomosis (MA), Riche-Cannieu anastomosis (CRA), Berrettini anastomosis (BA), median and musculocutaneous nerves anastomoses (MMA), or radial and lateral cutaneous nerve of forearm anastomosis (RLCA).Material and methodsMajor online medical databases such as PubMed, Embase, Scopus, ScienceDirect, Web of Science, SciELO, BIOSIS, Current Content Connect and Korean Journal Database were searched to gather all relevant studies regarding the anatomical characteristics of these structures. Finally, a total of 189 studies were included in this meta-analysis, based on 23,572 upper extremities.ResultsThe prevalence of MGA was 20.74% (cadaveric studies), 18.93% (fetal studies), 16.65% (healthy subjects) and 20.37% (patients with carpal tunnel syndrom). The most prevalent was type II of the Oh et al. classification of MGA, occurring in 58.01%, and pattern I of Rodríguez–Niedenführ et al. classification, which resulted in 96.07%. Nevertheless the pooled prevalence of MA was 2.04%. The BA was found in 67.97% (cadaveric studies), 29.07% (fetal studies) and 11.88% (electrophysiological studies).ConclusionThis study represents the most extensive and up-to-date evaluation of nerve anastomoses in upper limb morphometry and anatomy up to date. Its findings provide an important reference for clinicians, particularly surgeons performing procedures in the upper extremities region.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1007/s00276-026-04001-x.Keywords: Nerve anastomosis, Nerve communication, Nerve connection, Brachial plexus Introduction There are numerous communicating branches (CB) between the nerves of the brachial plexus. The typical brachial plexus is formed by the union of fibers arising from the anterior rami of the lower five cervical (C5–C8) and the first thoracic (T1) spinal nerves. These roots join in a characteristic pattern to form the superior, middle, and inferior trunks, each of which divides into anterior and posterior divisions. The divisions then reorganize to form the lateral, medial, and posterior cords, which continue distally as the terminal branches of the plexus [20, 22, 153]. Communications usually occur between anterior divisions of nerves sharing similar spinal origins, such as the musculocutaneous nerve (MCN) with the median nerve (MN), or the MN with the ulnar nerve (UN) in the forearm and hand. Although less common, connections between posterior division nerves have also been reported, such as the communication between the superficial branch of the radial nerve (SBRN) with the lateral cutaneous nerve of the forearm (LCNF) [197]. Interconnections between the MN and UN commonly occur in the brachial plexus, forearm, or hand [129]. In 1763, the Swedish anatomist Martin first described the morphological characteristics of a transfer of nerve fascicles in the forearm from the proximal median or anterior interosseous nerve to the distal ulnar nerve. This description was later complemented by Gruber in 1870, resulting in the term Martin–Gruber anastomosis (MGA) [39, 54, 161, 164, 218]. In 1964, Marinacci, using electrodiagnostic studies, described a patient who, following trauma to the median nerve in the forearm, exhibited denervation of the forearm flexor muscles but retained function of all median-innervated hand muscles. The Marinacci communication (MA), also known as the „reverse MGA”, carries motor fibers from the proximal ulnar to the distal median nerve in the forearm [127, 135, 161]. Another notable interconnection is the Riche–Cannieu anastomosis (CRA), also known as the „deep palmar communication”. Originally described in two separate anatomical studies by Riche and Cannieu in 1896, this anastomosis represents a connection between the deep branch of the ulnar nerve and the recurrent branch of the median nerve. It results in variable innervation of the thenar muscles by the ulnar nerve [161, 167, 218]. The Berrettini anastomosis (BA), also known as the „superficial palmar communication”, is a sensory connection between the third common palmar digital nerve from the median nerve and the fourth common palmar digital nerve from the superficial branch of the ulnar nerve in the palm. It was named after the Italian painter Pietro Berrettini da Cortona, who first illustrated this communication in 1741[161, 167, 218]. The most common variation of the MCN involves an anastomotic branch with the MN. Anatomical variations of the communication between the MCN and MN are relatively common and are thought to result from developmental differences during embryogenesis [18, 181]. Anatomical variations of peripheral nerve anastomoses are clinically important, as they may lead to diagnostic errors and, consequently, surgical complications due to inaccurate preoperative planning. Additionally, anatomical nerve connections may become compressed or entrapped by muscles or ligaments, leading to focal mononeuropathies [173]. Therefore, the objective of this study was to provide comprehensive data on the prevalence, morphometry, and anatomical classifications of upper extremity anastomoses. To accomplish this, an up-to-date systematic literature search and meta-analysis were performed. Methods and materials Search strategy Major online medical databases such as PubMed, Embase, Scopus, ScienceDirect, Web of Science, SciELO, BIOSIS, Current Content Connect, and Korean Journal Database, were used to gather all relevant studies on upper limb nerve anastomoses (Supplementary Table 1). The study collection ended in September 2025. In agreement with the Boolean technique, the following search terms were employed: (median nerve anastomosis OR median nerve communication OR ulnar nerve anastomosis OR ulnar nerve communication OR Martin Gruber OR Marinacci OR Riche Cannieu OR deep palmar anastomosis OR deep palmar communication OR Berrettini OR Superficial palmar anastomosis OR Superficial palmar communication OR Musculocutaneous nerve anastomosis OR Musculocutaneous nerve communication OR Medial brachial cutaneous nerve anastomosis OR Medial brachial cutaneous nerve communication OR medial antebrachial cutaneous nerve anastomosis OR medial antebrachial cutaneous nerve communication OR Radial nerve anastomosis OR radial nerve communication). The search terms were individually adapted to each database to minimize potential bias. Neither date, language, article type, nor text availability conditions were applied. An additional search was conducted through the references of the identified studies at the end of the search stage to ensure the accuracy of the process. During the study, the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed. Furthermore, the Critical Appraisal Tool for Anatomical Meta-analysis (CATAM) and Anatomical Quality Assessment Tool (AQUA) were used to provide the highest quality findings [49, 50, 79, 200, 219]. Eligibility assessment The database search and the manual search identified a total of 13 442 studies that were initially evaluated by two independent reviewers. After removing duplicates and irrelevant records, a total of 368 articles were qualified for a full-text evaluation. To minimize potential bias and maintain accurate statistical methodology, articles such as case reports, case series, conference reports, reviews, letters to editors, and studies that provided incomplete or irrelevant data were excluded. The inclusion criteria consisted of original studies with extractable numerical data regarding the topic of this study. Finally, a total of 189 studies were included in this meta-analysis—Table 1 [1–6, 8–10, 12–17, 19, 21, 23–30, 32–48, 51–78, 81–90, 92–117, 119–126, 128, 130–132, 134–142, 145–152, 154–159, 162–166, 168–172, 174, 175, 177–179, 183–186, 188–196, 198, 199, 201, 202, 205–214, 216, 217, 220–223]. Table 1Characteristics of included studies. MGA—Martin Gruber anastomosis; BA—Berrettini anastomosis; MMA—anastomosis between median and musculocutaneous nerves; CRA—Cannieu-Riché anastomosis; MA—Marinacci anastomosis; RLCA—Radial Nerve and Lateral Cutaneous Nerve of Forearm AnastomosisFirst authorYearRegionCountryType of studyNumber of specimensAnastomosis typeDaza L [51]2025AfricaSouth AfricaCadaveric study162CRA, BAGrados J de JO [70]2025North AmericaMexicoCadaveric study80MGAMendoza M [136]2025North AmericaUSACadaveric study104BATran P [201]2025North AmericaUSACadaveric study101MGACaetano EB [32]2024South AmericaBrazilCadaveric study20CRAKaiser P [93]2024EuropeAustriaCadaveric study104BADémoulin F [52]2023EuropeFranceCadaveric study24RA & LCNFErtekin A [61]2023AsiaTurkeyElectrophysiological study506MGAKhadanovich A [104]2023EuropeCzech republicCadaveric study102RA & LCNFKhadanovich A [103]2024EuropeCzech republicCadaveric study93RA & LCNFMoasses Z [137]2024AsiaIranCadaveric study26MCN & MNPalimar V [147]2023AsiaIndiaCadaveric study54BARiveros A [162]2023South AmericaChileCadaveric study30MGAAlpaslan AT [6]2022AsiaTurkeyFetuses study52MGAMohamed HE-S [138]2022AfricaEgyptElectrophysiological study100MGAGhosh B [68]2022AsiaIndiaCadaveric study60MCN & MNMarton A [130]2022EuropeUKCadaveric study27BAAl-Sobhi MG [8]2023AsiaSaudi ArabiaCadaveric study40MCN & MNDi Stefano V [188]2021EuropeItalyElectrophysiological study1882MGADuarte JM [56]2021South AmericaArgentinaElectrophysiological study338BADukan R [57]2021EuropeFranceCadaveric study8BAQuintero ID [158]2021South AmericaColombiaCadaveric study106BASagar TN [169]2021AsiaIndiaCadaveric study58MMASur A [194]2021AsiaIndiaElectrophysiological study280MGABallesteros-Acuña LE [17]2020South AmericaColombiaCadaveric study106CRAHefny M [78]2020AfricaEgyptElectrophysiological study280MGASeidel ME [174]2018North AmericaUSAElectrophysiological study411BAShruthi B [177]2020AsiaIndiaCadaveric study60MMACaetano EB [36]2019South AmericaBrazilCadaveric study80CRACetkin HE [40]2019AsiaTurkeyCadaveric study40RA & LCNFChandrika PV [41]2019AsiaIndiaCadaveric study80MMADiz-Díaz J [54]2019EuropeSpainCadaveric study50MGAHodzic R [82]2020EuropeBosnia & HercegovinaCadaveric study60MGAPriya A [157]2019AsiaIndiaCadaveric study60MMAWon S-Y [217]2010AsiaKoreaCadaveric study28MGACaetano EB [35]2018South AmericaBrazilCadaveric study80CRAKara AB [94]2018AsiaTurkeyFetuses study100MMA, MGA, BAKhake S [105]2018AsiaIndiaCadaveric study60MMASthapak E [189]2018AsiaIndiaCadaveric study100MMALazo Velásquez JC [116]2018South AmericaPeruElectrophysiological study254MGA, MA, CRA, BAEhler E [59]2017EuropeCzech republicElectrophysiological study457MGAHayashi M [77]2017AsiaJapanCadaveric study130MMAPatel ZK [150]2017AsiaIndiaCadaveric study60MMAPrates L [156]2003South AmericaBrazilCadaveric study64MGA, MAAziz Saba EK [13]2017AfricaEgyptElectrophysiological study200MGAMat Taib CN [132]2017AsiaMalaysiaCadaveric study44MMA, MGAYamamoto M [220]2018AsiaJapanCadaveric study22MMAAhadi T [2]2016AsiaIranElectrophysiological study46CRACaetano EB [33]2016South AmericaBrazilCadaveric study100MGACaetano EB [34]2016South AmericaBrazilFetuses study40MMACavalheiro CS [39]2016South AmericaBrazilCadaveric study100MGADepukat P [53]2017EuropePolandCadaveric study100BAGopal K [69]2016AsiaIndiaCadaveric study40MMAKaur N [96]2016AsiaIndiaCadaveric study60BAKaur N [97]2016AsiaIndiaCadaveric study60MGASchmitt C [172]2016South AmericaBrazilCadaveric study100MGAVanisree S [211]2016AsiaIndiaCadaveric study60MMAYang H [221]2016AsiaKoreaCadaveric study90MGA, CRA, BABalachandra N [14]2015AsiaIndiaCadaveric study20MMABallesteros LE [15]2015South AmericaColombiaCadaveric study106MMAKhosrawi S [106]2015AsiaIranElectrophysiological study457MGAPoublon AR [154]2015EuropeNetherlandsCadaveric study20RA & LCNFSulaiman S [192]2015EuropeUKCadaveric study150RA & LCNFSulaiman S [193]2016EuropeUKCadaveric study98BABallesteros LE [16]2014South AmericaColombiaCadaveric study108MGAJanani Y [90]2014AsiaIndiaCadaveric study64MMA, BAKate N [95]2015AsiaIndiaElectrophysiological study300MGAZolin SD [223]2014North AmericaUSACadaveric study30BAChaudhary P [43]2013AsiaIndiaCadaveric study40MMAKumar N [113]2014AsiaIndiaCadaveric study50MMAKwiatkowska M [114]2014EuropePolandCadaveric study20BAMariya [128]2013AsiaIndiaCadaveric study76MMARupanjali B [168]2013AsiaIndiaElectrophysiological study200MGADahiphale V [48]2012AsiaIndiaCadaveric study40MMAFelippe M [63]2012South AmericaBrazilElectrophysiological study30MGA, MANasr AY [142]2012AsiaSaudi ArabiaCadaveric study60MMASawant S [171]2012AsiaIndiaCadaveric study100MMABudhiraja V [30]2011AsiaIndiaCadaveric study116MMAHodzic R [83]2011EuropeBosnia & HercegovinaElectrophysiological study120MGAKervancioglu P [102]2011AsiaTurkeyFetuses study20MMAPawar S [151]2011AsiaIndiaElectrophysiological study200MGALoukas M [121]2011North AmericaUSACadaveric study50CRAPark BK [148]2010AsiaKoreaCadaveric study19RLCARopars M [165]2010EuropeFranceCadaveric study30RLCABhattarai C [23]1970AsiaNepalCadaveric study32MMAFolberg CR [65]2009South AmericaBrazilCadaveric study23RLCAGuerri-Guttenberg RA [73]2009South AmericaArgentinaCadaveric study & Fetuses56MMAMaeda S [124]2009AsiaJapanCadaveric study9MGAMaeda S [125]2009AsiaJapanCadaveric study453MMAUnver Dogan N [207]2010AsiaTurkeyFetuses study200BAUnver Dogan N [206]2010AsiaTurkeyFetuses study200MMA, MAJoshi S [92]2008AsiaIndiaCadaveric study170MMAKwolczak-McGrath A [115]2008EuropePolandFetuses study40MMAUysal II [208]2009AsiaTurkeyFetuses study140MMABiafora SJ [24]2007North AmericaUSACadaveric study50BAChitra R [45]2007AsiaIndiaCadaveric study50MMAHuanmanop T [87]2007AsiaThailandCadaveric study79RLCAKrishnamurthy A [112]2007AsiaIndiaCadaveric study44MMALoukas M [122]2007North AmericaUSACadaveric study200BAShen Y [175]2007AsiaChinaCadaveric study105MGATagil SM [196]2007AsiaTurkeyCadaveric study30BAHong JT [84]2006AsiaKoreaCadaveric study38BAKazakos KJ [100]2005EuropeGreeceCadaveric study163MGALee K-S [117]2005AsiaKoreaCadaveric study102MGALoukas M [120]2005North AmericaUSACadaveric study258MMAMarur T [131]2005AsiaTurkeyCadaveric study37MGAVicente DP [213]2005EuropeSpainCadaveric study46MMABeheiry EE [21]2004AfricaEgyptCadaveric study60MMADon Griot JPW [55]2004EuropeNetherlandSubject study43BAIkiz ZAA [88]2004AsiaTurkeyCadaveric study48RLCAKawashima T [98]2004AsiaJapanCadaveric study169BAMurata K [140]2004North AmericaUSACadaveric study35BATryfonidis M [202]2004EuropeUKCadaveric study20RLCAAmoiridis G [10]2003EuropeGreeceElectrophysiological study50MGA, MAMeenakshi-Sundaram S [135]2003AsiaIndiaElectrophysiological study200MAUysal İİ [209]2003AsiaTurkeyFetuses study200MMASarikcioglu L [170]2003AsiaTurkeyCadaveric & Electrophysiological study90MGA, MAChoi D [46]2002EuropeUKCadaveric study276MMAErdem HR [60]2002AsiaTurkeyElectrophysiological study100MGAGriot JPWD [71]2002EuropeNetherlandsCadaveric study26BARodriguez-Niedenführ M [164]2002EuropeSpainCadaveric study140MGARodriguez-Niedenführ M [163]2002EuropeSpainCadaveric study236MGAAktan ZA [4]2001AsiaTurkeyCadaveric study48MMAHasegawa O [76]2001AsiaJapanElectrophysiological study212MGAOlave E [146]2001South AmericaChileCadaveric study56BAPastor Gómez J [149]2001EuropeSpainElectrophysiological study362MGASimonetti S [179]2001EuropeItalyElectrophysiological study24MGABoutros S [26]2000North AmericaUSACadaveric study8RLCAPeter J [152]2000EuropeNetherlandsCadaveric study53BAPeter J [152]2000EuropeNetherlandsCadaveric study53BAPrasada Rao PV [155]2000AfricaZimbabweCadaveric study24MMAAlmeida JA de [5]1999South AmericaPeruCadaveric study40MGABas H [19]1999North AmericaUSACadaveric study30BABudak F [29]1999AsiaTurkeyElectrophysiological study216MGAChantelot C [42]1999EuropeFranceCadaveric study50MGAShu HS [178]1999AsiaChinaCadaveric study72MGAStančić MF [186]1999EuropeCroatiaCadaveric study100BACanovas F [38]1998EuropeFranceCadaveric study10MGAChiarapattanakom P [44]1998AsiaThailandCadaveric study112MMAVenieratos D [212]1998EuropeGreeceCadaveric study158MMAEglseder WA [58]1997North AmericaUSACadaveric study108MMAStallings SP [185]1997North AmericaUSAElectrophysiological study600MGATaams KO [195]1997AfricaSouth AfricaCadaveric study112MGA, MAGert Van Dijk J [67]1997EuropeNetherlandsElectrophysiological study150MGAAjmani ML [3]1996AsiaIndiaCadaveric study68CRAYang Z-X [222]1995AsiaSingaporeCadaveric study24MMAAbrams RA [1]1994North AmericaUSACadaveric study20RLCAAuerbach DM [12]1994North AmericaUSACadaveric study20RLCABudak F [28]1994AsiaTurkeyCadaveric study216CRAOkuno Y [145]1994AsiaJapanCadaveric study34MGA, CRANakashima T [141]1993AsiaJapanCadaveric study108MGAAmoiridis G [9]1992EuropeGermanyElectrophysiological study100MGA, MAHoogbergen MM [85]1992EuropeNetherlandsCadaveric study25BAKosugi K [111]1992AsiaJapanCadaveric study75MMAUchida Y [205]1992AsiaJapanElectrophysiological study268MGAFerrari GP [64]1991EuropeFranceCadaveric study50BAKayamori R [99]1987AsiaJapanElectrophysiological study1200MGAKimura I [108]1987AsiaJapanElectrophysiological study220MGA, MA, CRAKosugi K [110]1986AsiaJapanCadaveric study110MMABonnel F [25]1985EuropeFranceCadaveric study50BAFalconer D [62]1985North AmericaUSACadaveric study10CRAHoriguchi M [86]1985AsiaJapanCadaveric study67MMAKimura I [107]1983North AmericaUSAElectrophysiological study150CRA, MAMeals RA [134]1983North AmericaUSACadaveric study50BAStreib E [191]1983EuropeGermanyElectrophysiological study150MGASrinivasan R [184]1981North AmericaUSAFetuses study260MGACrutchfield CA [47]1980North AmericaUSAElectrophysiological study100MGAGutmann L [74]1977North AmericaUSAElectrophysiological study63MGAIyer V [89]1976North AmericaUSAElectrophysiological study65MGAKimura J [109]1976North AmericaUSAElectrophysiological study656MGAWilboum A [216]1976North AmericaUSAElectrophysiological study108MGASouza O [183]1975South AmericaBrazilCadaveric study60CRARosen A [166]1973North AmericaUSAElectrophysiological study96MAHarness D [75]1971AsiaIsraelCadaveric study35CRAMannerfelt L [126]1966EuropeSwedenCadaveric study56BA, MGAMachida N [123]1961AsiaJapanCadaveric study50BABuch-Hansen K [27]1955EuropeGermanyCadaveric study75MMAMonden M [139]1942AsiaJapanCadaveric study120MMAHirasawa K [81]1931AsiaJapanCadaveric study186MMA, BAVallois H [210]1922EuropeFranceCadaveric study200MMALinell EA [119]1921EuropeUKCadaveric study20BAKerr AT [101]1918North AmericaUSACadaveric study175MMAStopford JS [190]1918EuropeUKCadaveric study51BARiché P [159]1897EuropeFranceCadaveric study23CRAThomson A [199]1893EuropeUKCadaveric study406MGAVillar M [214]1888EuropeFranceCadaveric study37MMATestut L [198]1883EuropeFranceCadaveric study105MMAGruber W [72]1870EuropeGermanyCadaveric study250MGAGegenbaur C [66]1867EuropeGermanyCadaveric study41MMA Characteristics of included studies. MGA—Martin Gruber anastomosis; BA—Berrettini anastomosis; MMA—anastomosis between median and musculocutaneous nerves; CRA—Cannieu-Riché anastomosis; MA—Marinacci anastomosis; RLCA—Radial Nerve and Lateral Cutaneous Nerve of Forearm Anastomosis Data extraction Two independent reviewers extracted data from qualified studies (DP, TK). Qualitative data, such as year of publication, country and continent of origin, data collection methodology, and information on nerve anastomose in the studied groups, were collected. Quantitative data, such as sample size, numerical data on anatomical characteristics, morphometry, and classifications of the nerve anastomoses, were also gathered. Any discrepancies between the studies identified by the two reviewers were resolved by contacting the authors of the original studies wherever possible or by consensus with a third reviewer (JW). Statistical analysis To perform the meta-analyses, MetaXL version 5.3 software (EpiGear International Pty Ltd, Wilston, Queensland, Australia), and Comprehensive Meta-analysis version 3.0 software (Biostat Inc., Englewood, NJ, USA) were used. A random-effects model was used in all analyses. The heterogeneity among the studies was evaluated using both the Chi-squared test and the I-squared statistic [80]. The I-squared statistic was interpreted as follows: 0–40% as “might not be important”; 30–60% as “may represent moderate heterogeneity; 50–90% as “may represent substantial heterogeneity”; 75–100% as “may represent considerable heterogeneity”. P-value of <0.05 and the confidence intervals (95% CI) were used to find statistically significant differences between the studied groups. In cases of overlapping confidence intervals, the differences were considered statistically insignificant. Results Search results Initially, a total of 13 442 (13 364 through records identified from databases and 78 through additional search) articles were identified in the databases. A total of 12 770 records were removed because they were duplicates or articles irrelevant to the studied topic. Furthermore, 672 records were screened. Following another evaluation, an additional 304 records were excluded due to a lack of data or irrelevance to the present study. Subsequently, a total of 368 studies were qualified for a full-text assessment, during which 63 studies were removed because they were case reports or case series, 96 articles were excluded due to lack of data, 19 papers were excluded because they were review articles, and 1 studies were excluded because it was an animal study. Finally, a total of 189 studies were included in this meta-analysis, based on 23 572 upper extremities. The overall data collection process is presented in Fig. 1. Additionally, the characteristics of all the submitted studies are summarized in Table 1. Flow diagram presenting process of collecting data included in this meta-analysis Martin Gruber anastomosis The pooled prevalence of MGA was 20.74% (95% CI 16.43–25.42%) for cadaveric was 18.93% (95% CI 9.76%–30.14%) for fetal studies, based on 2891 and 612 upper limbs. Divided by region, the MGA was found most often in North America–37.64% (95% CI 30.71%–44.82%, with 181 specimens). Although, for electrophysiological studies, the pooled prevalence of MGA was 16.65% (95% CI 13.47–20.09%) for healthy subjects and 20.37% (95% CI 8.36–35.61%) for adults with carpal tunnel syndrome, according to 4407 and 2852 upper extremities. In both subgropus, the most prevalent MGA was in Europe—25.10% (95% CI 17.13–33.99%, based on 727 specimens) and 38.93% (95% CI 0.00–100%, depend on 1906 specimens). The analysis of the Oh et al. classification of MGA was established, considering a total of 589 upper limbs (electrophysiological studies). The most prevalent was type II, communication innervating the first dorsal interosseus muscle, occurring in 58.01% (95% CI 47.92–67.78%). However, according to Rodríguez-Niedenführ et al. classification, based on 319 specimens (cadaveric & fetal studies), proved that, the most common was pattern I (single communication), which result in 96.07% (95% CI 92.13%—98.76%). To be more precise, I can be divided into 3 subgroups, the most typical was type Ic (communication arose from the anterior interosseous nerve) at 75.16% (95% CI 65.17%—83.97%). The mean MGA diameter was set at 7.94 mm (SE=1.28), and the length was 44.62 mm (SE=7.59). All results mentioned before and more detailed ones are gathered in Table 2 and Figs. 2, 3. Table 2Statistical results of this meta-analysis regarding the Martin Gruber anastomosis (MGA). LCI—lower confidence interval. HCI—higher confidence interval. Q—Cochran’s Q. Oh et al. Classification: Type I—communication innervating the hypothenar muscles, Type II—communication innervating the first dorsal interosseus muscle, Type III—communication innervating the thenar muscles. Rodríguez-Niedenführ et al. Classification: Pattern I—single communicating branch, Type Ia—communication emerges from the branch of the median nerve to the superficial forearm flexor muscles, Type Ib—communication was between the origin of the branch to the superficial forearm flexors and the origin of the anterior interosseous nerve, Type Ic—communication arose from the anterior interosseous nerve, Pattern II—double communicating branchesCategoryMeanSEVarianceLower limitUpper limitZ-valuep-valueMGA diameter [mm]7.941.281.635.4410.446.230,00MGA length [mm]44.627.5957.5629.7559.495.880,00CategoryNPooled prevalenceLCIHCIQI2Cadaveric studiesMGA prevalence [Overall]289120.74%16.43%25.42%246.7788.25MGA prevalence [Asia]71023.15%13.18%34.85%115.1791.32MGA prevalence [Europe]134614.36%10.26%19.00%34.2176.61MGA prevalence [North America]18137.64%30.71%44.82%0.000.00MGA prevalence [South America]54221.21%15.48%27.56%17.3865.49Fetal studiesMGA prevalence [Overall]61218.93%9.76%30.14%27.7689.19Electrophysiological studies—healthy adultsMGA prevalence [Overall]440716.65%13.47%20.09%161.3488.22MGA prevalence [Africa]58020.27%17.09%23.64%0.290.00MGA prevalence [Asia]204413.21%9.16%17.87%75.5088.08MGA prevalence [Europe]72725.10%17.13%33.99%15.0480.05MGA prevalence [North America]105614.35%5.93%25.38%31.2093.59Electrophysiological studies—adults with carpal tunnel syndromeMGA prevalence [overall]285220.37%8.36%35.61%341.6797.95MGA prevalence [Asia]51817.72%2.65%40.00%56.0496.43MGA prevalence [Europe]190638.93%0.00%100.00%100.4199.00MGA prevalence [North America]42814.50%7.31%23.50%7.8774.59Types prevalence according to oh et al. classificationType I58918.00%14.21%22.13%13.1631.59Type II58.01%47.92%67.78%49.8781.95Type III18.54%13.29%24.44%24.6963.54Type I+II2.34%0.11%6.48%47.8181.18Type II+III1.85%0.00%5.04%49.6681.88Types prevalence according to Rodríguez-Niedenführ et al. classificationPattern I31996.07%92.13%98.76%26.5150.97Pattern II3.93%1.24%7.87%26.5150.97Type Ia30613.06%6.09%22.01%50.5174.26Type Ib13.57%7.04%21.70%41.5468.71Type Ic75.16%65.17%83.97%43.6770.23 Statistical results of this meta-analysis regarding the Martin Gruber anastomosis (MGA). LCI—lower confidence interval. HCI—higher confidence interval. Q—Cochran’s Q. Oh et al. Classification: Type I—communication innervating the hypothenar muscles, Type II—communication innervating the first dorsal interosseus muscle, Type III—communication innervating the thenar muscles. Rodríguez-Niedenführ et al. Classification: Pattern I—single communicating branch, Type Ia—communication emerges from the branch of the median nerve to the superficial forearm flexor muscles, Type Ib—communication was between the origin of the branch to the superficial forearm flexors and the origin of the anterior interosseous nerve, Type Ic—communication arose from the anterior interosseous nerve, Pattern II—double communicating branches Schematic illustrations of types I, II, and III of the Martin–Gruber anastomosis Schematic illustrations of types Ia, Ib, and Ic of the Martin–Gruber anastomosis Berrettini anastomosis The pooled prevalence of BA was 67.97% (95% CI 56.62%—78.35%) for cadaveric, and 29.07% (95% CI 24.06%—34.34%) for fetal studies, based on 2243 and 300 upper limbs. Divided by region, the BA was found most often in South America—87.72% (95% CI 65.36%—100%, with 162 specimens). Although, for electrophysiological studies, the pooled prevalence of BA was 11.88% (95% CI 3.70%—23.37%), according to 1003 upper extremities. BA was usually unilateral—71.52% (95% CI 53.36%—86.84%), depending on 256 specimens. The analysis of the Sulamain et al. classification of BA was established, considering a total of 1144 upper limbs. The most common was type I—74.62% (95% CI 63.98–83.99%)—CB attach proximally into the fourth common digital nerve located between the little and ring fingers, which originated from the ulnar nerve and joined the third common digital nerve located between the ring and long fingers, which originated from the median nerve distally. The mean BA length was set at 19.70 mm (SE=2.69). All results mentioned before and more detailed ones are gathered in Table 3 and Fig. 4. Table 3Statistical results of this meta-analysis regarding the Berrettini anastomosis (BA). LCI—lower confidence interval. HCI—higher confidence interval. Q—Cochran’s Q. Sulamain et al. classification: Type I—communicating branch attach proximally into the fourth common digital nerve located between the little and ring fingers, which originated from the ulnar nerve and joined the third common digital nerve located between the ring and long fingers, which originated from the median nerve distally, Type II—communicating branch attached proximally to the common digital nerve from the median nerve between the ring and long fingers, and joined the common digital nerve from the ulnar nerve between the little and ring fingers distally, Type III—communicating branch was found to travel perpendicular to both the common digital nerve between the little and ring fingers, and the common digital nerve between the ring and long fingers, Type IV—communicating branch, having multiple attachment points to the common digital nerve between the little and ring fingers, and the common digital nerve between the ring and long fingers in a diffuse manner, Type V—the common digital nerve between the little and ring fingers, and the common digital nerve between the ring and long fingers, gave branches that merged and continued distally in the mid-palm to supply the radial side of the ring fingerCategoryMeanSEVarianceLower limitUpper limitZ-valuep-valueBA length [mm]19.702.697.2414.4224.977.320,00CategoryNPooled prevalenceLCIHCIQI2Cadaveric studiesBA prevalence [Overall]224367.97%56.62%78.35%1007.9396.73BA prevalence [Asia]63549.84%23.08%76.64%351.7597.73BA prevalence [Europe]78575.46%54.91%91.62%477.8597.07BA prevalence [North America]49972.25%62.31%81.21%29.1779.43BA prevalence [South America]16287.72%65.36%100.00%12.3391.89Fetal studiesBA prevalence [Overall]30029.07%24.06%34.34%0.060.00Electrophysiological studiesBA prevalence [Overall]100311.88%3.70%23.37%44.3095.49Bilateral vs unilateralUnilateral BA prevalence [Overall]25671.52%53.36%86.84%53.6886.96Bilateral BA prevalence [Overall]28.48%13.16%46.64%53.6886.96Types prevalence according to Sulamain et al. classificationType I113074.62%63.98%83.99%292.2092.81Type II10.31%6.36%15.05%106.8380.34Type III7.54%2.63%14.45%263.1492.02Type IV3.01%1.39%5.20%62.2266.25Type V0.93%0.41%1.64%23.058.91 Statistical results of this meta-analysis regarding the Berrettini anastomosis (BA). LCI—lower confidence interval. HCI—higher confidence interval. Q—Cochran’s Q. Sulamain et al. classification: Type I—communicating branch attach proximally into the fourth common digital nerve located between the little and ring fingers, which originated from the ulnar nerve and joined the third common digital nerve located between the ring and long fingers, which originated from the median nerve distally, Type II—communicating branch attached proximally to the common digital nerve from the median nerve between the ring and long fingers, and joined the common digital nerve from the ulnar nerve between the little and ring fingers distally, Type III—communicating branch was found to travel perpendicular to both the common digital nerve between the little and ring fingers, and the common digital nerve between the ring and long fingers, Type IV—communicating branch, having multiple attachment points to the common digital nerve between the little and ring fingers, and the common digital nerve between the ring and long fingers in a diffuse manner, Type V—the common digital nerve between the little and ring fingers, and the common digital nerve between the ring and long fingers, gave branches that merged and continued distally in the mid-palm to supply the radial side of the ring finger Schematic illustrations of the different types of the Berrettini anastomosis Median and Musculocutaneous nerves anastomoses The pooled prevalence of MMA was 23.90% (95% CI 20.11–27.92%) for cadaveric, and 10.30% (95% CI 4.17–18.55%) for fetal studies, based on 4546 and 740 upper limbs.. Divided by region, the MMA was found most often in North America—34.99% (95% CI 21.53–49.73%, with 541 specimens). MMA was usually unilateral—74.12% (95% CI 62.23–84.48%), depending on 294 specimens. On the other hand, MMA was fusion in 13.68% (95% CI 6.27–23.20%), based on 540 upper limbs. However, according to Loukas et al. classification, based on 523 upper extremities, proved that, the most common was type I—CB proximal to the point of entry of the MCN into the coracobrachialis, in 80.76% (95% CI 68.61–90.61%). The mean MMA length was set at 52.13 mm (SE=4.57). All results mentioned before and more detailed ones are gathered in Table 4 and Fig. 5. Table 4Statistical results of this meta-analysis regarding the anastomosis between Median and Musculocutaneous nerves (MMA). LCI—lower confidence interval. HCI—higher confidence interval. Q—Cochran’s Q. Loukas et al. classification: Type I—communications proximal to the point of entry of the MCN into the coracobrachialis, Type II—communications were distal to the point of entry of the MCN into the coracobrachialis, Type III—the MCN did not pierce the coracobrachialis, Type IV—communications were proximal to the point of entry of the MCN into the coracobrachialis and additional communication took place distally (Type I and II coexisted)CategoryMeanSEVarianceLower limitUpper limitZ-valuep-valueMMA length [mm]52.134.5720.8643.1861.0911.420,00CategoryNPooled prevalenceLCIHCIQI2Cadaveric studiesMMA prevalence [Overall]454623.90%20.11%27.92%467.1789.30MMA Prevalence [Africa]8416.92%0.00%51.32%10.1390.13MMA prevalence [Asia]287721.61%17.38%26.17%287.7887.49MMA prevalence [Europe]93832.57%21.50%44.69%93.3592.50MMA prevalence [North America]54134.99%21.53%49.73%22.7391.20Fetal studiesMMA prevalence [Overall]74010.30%4.17%18.55%53.7988.85Bilateral vs unilateralUnilateral MMA prevalence [Overall]29474.12%62.23%84.48%50.2974.15Bilateral MMA prevalence [Overall]25.88%15.52%37.77%50.2974.15Median and musculocutaneous nerves fused in patients with MMAFusion prevalence [Overall]54013.68%6.27%23.20%82.4886.66Types prevalence according to loukas et al. classificationType I52380.76%68.61%90.61%145.4189.00Type II7.03%0.99%16.84%168.0490.48Type III2.31%0.63%4.88%33.0751.62Type IV8.66%6.27%11.38%17.176.82 Statistical results of this meta-analysis regarding the anastomosis between Median and Musculocutaneous nerves (MMA). LCI—lower confidence interval. HCI—higher confidence interval. Q—Cochran’s Q. Loukas et al. classification: Type I—communications proximal to the point of entry of the MCN into the coracobrachialis, Type II—communications were distal to the point of entry of the MCN into the coracobrachialis, Type III—the MCN did not pierce the coracobrachialis, Type IV—communications were proximal to the point of entry of the MCN into the coracobrachialis and additional communication took place distally (Type I and II coexisted) Schematic illustrations of the different types of anastomosis between median and musculocutaneous nerves Other anastomoses Relying on cadaveric studies, the pooled prevalence of CRA was 50.98% (95% CI 26.11–75.62%), based on 1050 specimens. Divided by region, the CRA was found most often in South America—93.74% (95% CI 68.41–100%, with 240 specimens). Moreover RLCA was founded in 40.98% (95% CI 27.80–54.82%), based on 696 specimens. Predicated on electrophysiological studies, the pooled prevalence of CRA was 62.68% (95% CI 14.26–100%), based on 670 upper extremities, nevertheless the pooled prevalence of MA was 2.04% (95% CI 0.33–4.88%), based on 1316 upper limbs. All results mentioned before and more detailed ones are gathered in Table 5 and Figs. 6, 7. Table 5Statistical results of this meta-analysis regarding other anastomoses. LCI—lower confidence interval. HCI—higher confidence interval. Q—Cochran’s QCategoryNPooled prevalenceLCIHCIQI2Cannieu-Riché anastomosis (CRA)Cadaveric studiesCRA prevalence [Overall]105050.98%26.11%75.62%860.5098.37CRA prevalence [Asia]44350.07%19.57%80.54%139.9897.14CRA prevalence [Europe]3916.08%6.02%29.40%0.240.00CRA prevalence [North America]6018.81%8.87%31.19%1.109.23CRA prevalence [South America]24093.74%68.41%100.00%95.2596.85Electrophysiological studiesCRA prevalence [Overall]67062.68%14.26%100.00%396.4399.24Marinacci Anastomosis (MA)Electrophysiological studiesMA prevalence [Overall]13162.04%0.33%4.88%58.3086.28Radial nerve and lateral cutaneous nerve of forearm anastomosis (RLCA)Cadaveric studiesRLCA prevalence [Overall]69640.98%27.80%54.82%172.9191.90RLCA prevalence [Asia]18640.46%23.02%59.15%17.6583.00RLCA prevalence [Europe]43943.28%23.03%64.67%103.6794.21RLCA prevalence [North America]4849.62%0.00%100.00%26.5292.46 Statistical results of this meta-analysis regarding other anastomoses. LCI—lower confidence interval. HCI—higher confidence interval. Q—Cochran’s Q Schematic illustration of the Cannieu-Riché anastomosis Schematic illustration of the Marinacci anastomosis Discussion Embryological basis of upper limb innervation and anatomical variants The nerve fibres extend into the myotomes at stage 13 of embryological development [176]. Following the vascularization by a single central artery, the nerves of spinal levels C5 to T1 connect and form a brachial plexus at stages 14–16 [7, 118, 176]. At stage 16, the brachial plexus differentiates, allowing for its short branches to be distinguished. Simultaneously, the median, ulnar and radial nerves enter into hand plate. Subsequently, as the forearm skeleton and musculature develop, the UN and MN separate into divisions innervating the forearm muscles. By stages 20–21, the pattern of upper limb innervation closely resembles that of the adult. Incomplete separation and aberrant distribution of nerve fibers during embryological development are considered the basis of inter-nerve communications. The fibers of one peripheral nerve may transiently course with another nerve trunk and subsequently separate distally, thereby forming neural anastomoses [187]. Such connections most commonly occur between MCN and MN, MN and UN, or SBRN and LCNF and may be found in the arm, forearm and hand. Anatomical variants of upper limb innervation may coexist with arterial variations. Antonopoulos et al. described a case of type I BA with concomitant incomplete superficial palmar arch [11]. In that report, the superficial palmar arch was formed solely by superficial branch of the ulnar artery, whereas the neural anastomosis linked the third and fourth common digital nerves. Furthermore, BA may exhibit a plexiform pattern, with the superficial palmar arch perforating the anastomotic network Neural anastomosis may also coexist. Nimma et al., in an electrophysiological study, reported a case of bilateral MGA accompanying reverse MGA (known also as MA) in the same patient Additionally, Natsis et al. depicted an atypical ulnar-median nerve communication connecting palmar cutaneous branch of the UN with the third common palmar digital nerve, coexisting with four distinct recurrent motor branches of the MN. Finally, Tsakotos et al. described a case of MMA coexisting with bilateral symmetric accessory coracobrachialis muscle. They implied that the impaired muscle development may contribute to abnormal innervation [143, 144, 180, 203]. Not a novel concept, but the most comprehensive data The present study investigates in detail the CB between the nerves derived from the brachial plexus. It is not the only one to summarize neural anastomoses in the upper limb; however, it is based on the greatest sample size and offers the most rigorous analysis to date. Notably, the current research comprises 189 studies, evaluating not only MN and UN connections but also anastomoses involving the radial, musculocutaneous, and medial cutaneous nerves of the arm and forearm. Moreover, this is the first systematic review to analyze occuring of MGA, BA, and MMA in fetal subgroup. The prevalence of anastomoses was established separately for the two study types. The first group consisted of cadaveric, and fetal studies, further categorized into morphological types, whereas the second group comprised electrophysiological studies. This stratified approach was adopted due to heterogeneity in the presence of intranervous connections when both groups were analyzed together. The median and ulnar nerve anastomoses have been previously assessed by Roy et al., based on 58 studies included in their analysis [167]. Both studies systematically searched several databases; the present study screened nine, in comparison to five assessed in the second paper. Remarkably, the present study applied the AQUA guidelines to ensure a thorough and standardized assessment of the methodological quality of the included studies [200]. In contrast, the compared review did not employ any objective tool for quality appraisal. Furthermore, our search strategy incorporated broader search terms to enhance sensitivity. However, Roy et al. imposed no restrictions regarding language, date range, or article type and additionally screened reference lists to maximize the completeness of their results. Moreover, the present study involved a search for additional anastomoses. The first anastomosis analyzed in both studies was MGA. In the present paper, its prevalence was higher in the morphological study types, reaching 20.74% (cadaveric studies) and 18.93% (fetal studies), and lower in the electrophysiological group at 16.65% (healthy subjects) and 20.37% (patients with carpal tunnel syndrom). Moreover, MGA presence was most frequent in North America regarding the former group. Roy et al. did not divide geographical distributions by study type and noted that North America was the most common region for MGA prevalence. The mean length of MGA was 44.62 mm 土 7.59 mm. The least frequent anastomosis was the MA, known also as „reverse MGA”. Its prevalence was reported exclusively based on electrophysiological studies, reaching 2.04%, compared with 0.8% in the second review. Another analysed variant was the connection between the ulnar and the median nerves on the palmar surface of the hand, known as BA. Roy et al. reported an overall prevalence of 60.9%, whereas in the present study, the BA was found in 67.97% in cadaveric studies, 29.07% in fetal studies and 11.88% in electrophysiological studies ones. For the classification of this anastomosis, the system proposed by Sulamain et al. was applied. Type I was the most common variant, accounting for 74.22% of cases in our study, which is comparable to the referenced study reporting such type prevalence of 86.2%. The mean length of the BA did not differ substantially between the studies, measuring approximately 19–20 mm in both. Finally, the prevalence of CRA observed in the present study was comparable to that reported by Roy et al. (50.98% vs. 55.5%). Similar discrepancies in connection frequency between morphological and electrophysiological studies were also noted. In recent days, a new systematic review with meta-analysis [91] has been published addressing communications between the MN and UN in the forearm and hand. While this study provides valuable insights into the prevalence and clinical implications of MGA, its methodological approach appears to rely on relatively narrowly defined search terms. As a consequence, the number of included studies (20 out of 879 studies in contrast to 189 studies of 13 442 records in our meta-analysis) is limited, reflecting a highly restrictive selection process. Additionally, it should be noted that the time frame of database searching in the published study was similar to that applied in our analysis (September 2025). Subgroup analyses demonstrated that the prevalence of MGA was greater among healthy individuals (approximately 25% vs 16.65% in our meta-analysis) than in patients with carpal tunnel syndrome (around 8–10% vs 20.37% in our meta-analysis). Rajadurai et al. stated that compression of the median nerve interferes with the electrophysiological identification of MGA, and cadaver-based studies showed lower detection rates (approximately 6% vs 20.74%), reflecting methodological differences. Our findings do not reveal any such discrepancies. Importantly, this is the first systematic review to analyze MMA and RLCA. The first anastomosis constitutes a connection between median and musculocutaneous nerves. Its pooled prevalence was relatively high and reached 23.90% (cadaveric studies) and 10.30% (fetal studies). The most common communication pattern was Type I, located proximal to the point of entry into the coracobrachial muscle. The second variant connects radial nerve with lateral cutaneous nerve of the forearm. The pooled prevalence of RLCA, reported solely from cadaveric studies, amounted to 40.98%. Neural anastomoses are less frequent in fetuses (MGA, BA, and MMA) than in cadaveric group, primarily due to the immaturity of the nervous system, as many connections develop and stabilize later during postnatal growth and functional maturation. Additionally, limited sensory and motor activity in utero, along with the absence of environmental and adaptive stimulations, reduces the formation and persistence of such connections compared to adults. Clinical connotations The upper limb nervous anastomoses are not indicators for intervention; their presence may pose a risk for atypical manifestations of musculoskeletal pathologies, such as carpal tunnel syndrome, neuropathies and cubital tunnel syndrome [182]. In regard to MGA, excluding this variant prior to establishing a diagnosis of ulnar neuropathy at the elbow is essential [61]. Maya-González et al. outlined the case of a 70-year-old patient with bilateral median and ulnar nerve involvement [133]. The patient presented with progressive pain and paresthesia in both hands over a 4-month period. Nerve conduction examination revealed the coexistence of a bilateral MGA and MA. Additionally, mildly prolonged distal sensory responses of both median nerves and absent sensory response of the right ulnar nerve to the little finger indicated bilateral carpal tunnel syndrome and ulnar neuropathy at Guyon’s canal. Recognition of these anatomical variants was essential for establishing the correct diagnosis. Furthermore, with MGA presence, there is a risk of iatrogenic injury that may produce unexpected ulnar sensory deficits. Riechers et al. described a case in which, after repair of a ruptured biceps tendon, injury to ulnar sensory fibers resulted in paresthesia of the fourth and fifth digits despite preserved hand motor function [160]. Considering BA, it can alter the expected sensory distribution in the hand by allowing the UN to innervate digits typically attributed to the MN. As a consequence, median nerve injury may present with preserved sensation on the hand, and the results of a conduction trial can be misinterpreted, provided the anastomosis is not recognized. Wakode et al. reported a case of a patient who sustained a laceration of the median nerve [215]. However, the patient did not exhibit significant sensory deficits, as subsequent nerve conduction studies demonstrated BA, with ulnar sensory fibers supplying all digits except the thumb. Lastly, Bölükbaşi et al. report a case of a patient with bilateral RCA [204]. Upon median nerve stimulation compound muscle potential from the abductor pollicis brevis and flexor pollicis muscles was noted from ulnar, rather than median nerve, with normal sensory responses, thereby diagnosis of RCA was stated. Additionally, the patient experienced dull pain in the neck region, with no radiation to upper extremities for five years. Indisputably, the cause of neck pain was cervical spondylosis, however the authors outline that the presence of an ulnar-median nerve anastomosis can lead to the misinterpretation of a complete median nerve lesion as partial. Strengths and limitations The main strength of our study lies in the abundance of analyzed upper limbs’anasomoses, with regard to type of study, geographical region and morphological features, such as the mean length and diameter of neural connection. Furthermore, variant classifications were systematically reported. Overall, the present study provides a comprehensive analysis of upper limb innervation variants, adding innovative contributions to the current anatomical knowledge and providing valuable practical information for clinical practice. Nevertheless, several limitations warrant addressing. Firstly, substantial heterogeneity was observed among the included studies, reflected by the wide range of confidence intervals. Although electrophysiological studies were not combined with cadaveric, fetal and intraoperative ones, the heterogeneity persisted. This might have appeared due to use of various classification systems in the included studies or discrepancies in the electrophysiological procedures themselves. Secondly, different places of measurement of nerves’ length and diameter may have introduced bias. Finally, regarding classification systems, only a single system was chosen for each anastomosis. Incorporating more of them could provide additional detail; however, it would likely introduce more complexity, potentially reducing the clarity of the analysis. Future directions There is a need to unify various classification systems of the aforementioned anastomoses. For instance, the MGA has at least five categorization schemes, depending on the site of connection [31, 54], the number of branches forming the anastomosis [96], the muscle innervated [78] or the nerve of origin [167]. Consequently, no consensus has been reached, which may lead to misunderstandings and reluctance to apply these classifications in clinical practice. Furthermore, the coexistence of different anastomoses should be addressed in future studies. These connections are not uncommon and, as demonstrated in the case reports, they may occur in pairs. Original studies exploring this pattern would be beneficial for clinicians, as they could clarify the authentic prevalence of such configurations, improve anatomical awareness during diagnostic and surgical procedures, ultimately leading to reduced risk of misinterpretation or iatrogenic injury. Conclusions The nerve anastomoses exhibit a high degree of variability regarding its anatomy, especially their classification types. The prevalence of MGA was 20.74% (cadaveric studies), 18.93% (fetal studies), 16.65% (healthy subjects) and 20.37% (patients with carpal tunnel syndrom). The most prevalent was type II of the Oh et al. classification of MGA, communication innervating the first dorsal interosseus muscle, occurring in 58.01%, and pattern I (single communication) of Rodríguez-Niedenführ et al. classification, which resulted in 96.07%. The BA was found in 67.97% (cadaveric studies), 29.07% (fetal studies) and 11.88% (electrophysiological studies). Based on BA classification—the most common was type I of the Sulamain et al. classification—74.62%. The pooled prevalence of MMA was 23.90% (cadaveric studies) and 10.30% (fetal studies), with type I—communications proximal to the point of entry of the MCN into the coracobrachialis, at 80.76%, based on Loukas et al. classification. CRA was established at 50.98% (cadaveric studies), and 62.68% (electrophysiological studies), nevertheless the pooled prevalence of MA was 2.04%. Supplementary Information Below is the link to the electronic supplementary material. Supplementary file1 (DOCX 24 KB) Abbreviations Median nerve Ulnar nerve Musculocutaneous nerve Superficial branch of radial nerve Lateral cutaneous nerve of the forearm Martin Gruber anastomosis Berrettini anastomosis Anastomosis between median and musculocutaneous nerves Cannieu-Riché anastomosis Marinacci anastomosis Radial nerve and lateral cutaneous nerve of forearm anastomosis Author Contribution DP—Search, Extraction, Literature, Statistical Analysis, Writing, Figures MB—Statistical Analysis, Tabels, Writing JW—Literature, Writing WCh-K—Literature, Writing AN—Figures, Writing TK—Search, Extraction, Writing KB—Literature, Writing MF—Literature, Writing MT—Literature, Writing MK—Statistical Analysis, Writing. Funding The authors received no financial support for the research, authorship, and/or publication of this article. Data availability No datasets were generated or analysed during the current study. Conflict interests The authors declare no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Informed consent Not applicable. Ethical approval Not applicable. Footnotes Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary file1 (DOCX 24 KB) Data Availability Statement No datasets were generated or analysed during the current study.