ORIGINAL ARTICLE

A high incidence of extensor pollicis brevis insertion into the distal phalanx inrheumatoid arthritis patients who required the surgical reconstruction forthumb boutonni!ere deformity

Shunji Okitaa, Keiichiro Nishidaa, Aiji Ohtsukab and Toshifumi Ozakia

aDepartment of Orthopedic Surgery, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, Okayama,Japan; bDepartment of Human Morphology, Okayama University Graduate School of Medicine, Dentistry and Pharmaceutical Sciences,Okayama, Japan

ABSTRACTObjectives: The aim of the current study was to investigate the pattern of extensor pollicis brevis(EPB) insertion macroscopically and histologically using cadaveric thumbs, and to compare the inci-dence of different insertions with that of thumb boutonni!ere deformity in rheumatoid arthritis (RA)patients who required surgical reconstruction.Methods: We examined 103 thumbs of 58 adult cadavers with no evidence of RA, and reviewed thesurgical records of 28 thumbs of 23 RA patients who underwent surgical reconstruction for thumbboutonni!ere deformity. The incidence of different insertion patterns of the cadaveric thumbs and theRA thumbs were compared using the Fisher’s exact test.Results: Macroscopic and histologic examination revealed that the insertion patterns of EPB could bedivided into three groups: insertion into the base of the proximal phalanx (Type P1), integration ofEPB into the dorsal fibrocartilage of the MCP joint (Type P2), and insertion into the distal phalanx(Type D). The incidence of Type D was significantly higher in RA patients with thumb boutonni!eredeformity (64%) than that in the non-RA cadavers (29%; P< .05).Conclusions: EPB is inserted into the distal phalanx more frequently in RA patients who requiresurgery for thumb boutonni!ere deformity than non-RA cadavers, suggesting an additional possiblemechanism of this deformity.

ARTICLE HISTORYReceived 20 July 2018Accepted 23 September 2018

KEYWORDSExtensor pollicis brevis;insertion pattern; thumbboutonni!ere deformity;rheumatoid arthritis

Introduction

Boutonni!ere deformity of the thumb, also known as theNalebuff’s type 1, is the most common deformity seen in50% to 74% of thumb deformities associated with rheuma-toid arthritis (RA) [1–4], and is characterized by metacarpo-phalangeal (MCP) joint flexion and interphalangeal (IP)joint hyperextension. The pathophysiology of boutonni!eredeformity is understood to start with MCP joint synovitis,which causes extensor pollicis brevis (EPB) to stretch andcauses the extensor pollicis longus (EPL) to be displacedulnarward and palmarward. As a result, dorsal support ofthe MCP joint is lost, and volar subluxation of the proximalphalanx develops. The displaced EPL further promotes MCPjoint flexion and IP joint hyperextension [5,6].

Recent studies on the anatomy of the EPB suggest thatthere are several variations of EPB insertion [7,8]. Duringclinical practice in surgery of the rheumatoid thumb, weoften encountered cases where EPB did not insert into thebase of the proximal phalanx, but this variation was notusually described in anatomy textbooks. To date, no studyhas focused on the relationship between the insertion pat-tern of EPB and the development of boutonni!ere deformity

in RA patients. We investigated the pattern of EPB insertionand the incidence of different variations using cadavericthumbs and compared this with that of RA patients withboutonni!ere deformity who required surgical reconstruction.

Patients and methods

The cadaveric study was conducted at our institute fromApril 2015 to May 2016. We observed 118 thumbs (59 rightand 59 left) of 59 adult Japanese cadavers. We excluded twothumbs of one cadaver with history of RA, and 13 thumbsof 13 cadavers because they were dissected for the anatomypractice before our investigation. Finally, we investigated103 thumbs (50 right and 53 left) of 58 cadavers (25 malesand 33 females). The mean age of the cadavers was 83.4years (range, 65–98 years). We used cadavers that had beenfixed in 10% neutral buffered formalin. After skin removaland careful dissection of the superficial fascia on the dorsumof each thumb, insertion levels of EPB were easily identifiedmacroscopically, and recorded. Then, the cadaveric thumbsamples were removed and cut vertically to confirm theinsertion of EPB. After macroscopic observation, the sam-ples for histologic study were fixed in 10% formaldehyde

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758

5960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116CONTACT Keiichiro Nishida knishida@md.okayama-u.ac.jp Department of Orthopedic Surgery, Okayama University Graduate School of Medicine, Dentistry

and Pharmaceutical Sciences, 2-5-1 Shikata-cho, Kita-ku, Okayama City, Okayama 700-8558, Japan.! 2018 Japan College of Rheumatology

MODERN RHEUMATOLOGYhttps://doi.org/10.1080/14397595.2018.1532484

overnight and decalcified in formic acid solution for 2–5days. They were then embedded in paraffin, cut into 10-mmsections, stained with Masson’s trichrome and observed bylight microscopy.

Next, we reviewed the surgical records of 28 thumbs (19right and nine left) of 23 RA patients (22 females and onemale) with thumb boutonni!ere deformity who underwentsurgical reconstruction of the thumb between April 2014and May 2018. All patients were Japanese, and the meanage of the patients was 61.6 (range, 27–80) years. The surgi-cal procedures included silicone implant arthroplasty in 21thumbs, and soft tissue reconstruction with synovectomy atthe MCP joint in seven thumbs. All the records had descrip-tions of EPB insertion patterns and indicated whether ornot EPB extended to the distal phalanx. We did not obtainpathological samples of the bone–tendon unit of EPB duringboutonni!ere deformity reconstruction.

The incidence of the different insertion patterns of EPBthat were identified macroscopically in the cadaveric thumbsand the RA thumbs were compared using Fisher’s exact test.A P-value<.05 was considered statistically significant. Thestudy was approved by the institutional review board (1607-019), and all patients provided written informed consent.

Results

Macroscopically, the EPB insertion pattern was classifiedinto two types: Type P: EPB ended at the level of the MCPjoint; and Type D: EPB ended at the level of the IP joint(Figure 1). In Type D, EPB runs distally along EPL, like aconjoint tendon. Direct pulling of EPB resulted in extensionof the MCP joint to a different extent in Type P, and inextension of both the MCP joint and the IP joint inType D.

EPB was present in all the cadaveric thumbs and dou-bling was found in 4.9% (n¼ 5). Seventy-one percent(n¼ 73) of the thumbs were Type P thumbs and 29%(n¼ 30) were Type D. EPB inserted not only into the distalphalanx but into the base of the proximal phalanx in sevenof the thumbs classified as Type D. Macroscopic variance ofEPL were not found in the current study, and all the EPLinserted into the distal phalanx.

Histological examination revealed that the Type Pthumbs consisted of 52% (n¼ 54) of Type P1 with EPBfirmly attached to the periosteum at the dorsal base of theproximal phalanx of the thumbs, and 18% (n¼ 19) of TypeP2 with EPB attached to the dorsal capsule–fibrocartilagecomplex in the thumbs (Table 1). In thumbs of Type D,EPB was firmly attached to the periosteum at the dorsalbase of the distal phalanx in each case (Figure 2).

All the surgical records described whether EPB extendeddistally or not because inflammation of RA made it difficultto distinguish Type P1 from Type P2. In the RA patientswith thumb boutonni!ere deformity, 36% (n¼ 10) of thethumbs were Type P, while 64% (n¼ 18) of the thumbswere Type D (Figure 3). The incidence of Type D was thussignificantly higher (P< .05) in the thumbs of patients with

RA and boutonni!ere deformity than in the cadaveric thumbs(Table 1).

Histological analysis revealed that the dorsal fibrous com-plex of the MCP joint of the thumb was mainly composedof three layers: the superficial layer with transverse fibersthat are interchangeably referred to as the extensor hood,the intermediate layer consisting of tendon, and the deeplayer with the fibrous capsule integrated into the dorsalfibrocartilage in the vertical plane. Although a hood hadvariable thicknesses and sub-hood space, it lay upon theextensor in all thumbs and there was no insertion of EPBinto a hood (Figure 4).

Discussion

Most anatomy textbooks describe EPB as generally originat-ing from the dorsal aspect of the radius and from the inter-osseous membrane, inserting distally into the base of theproximal phalanx of the thumb. Its action is described as anextension of the MCP joint [9]. However, several cadavericstudies have reported anatomical variations of EPB[8,10–16], and of these, some reports have described inser-tion of EPB into the extensor hood at the level of the MCPjoint [8,15,16]. Shigematsu et al. [7] macroscopically classi-fied EPB insertion into eight types after an anatomical studyof cadavers. They reported that EPB was completely insertedinto the extensor hood in 29% (41 out of 144) of hands;completely inserted into the base of the proximal phalanx in22% (32 out of 144) of hands; partially inserted into the

117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174

175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232

Figure 1. The appearance of the dorsum of left cadaveric thumbs after skinremoval. (a) The Type P EPB ends at the level of the MCP joint. (b) The Type DEPB inserts into the distal phalanx.

Table 1. Comparison of the insertion pattern of EPB between cadavers andRA patients with thumb boutonni!ere deformity.

Cadavers RA patients with boutonni!ere deformityType n¼ 103 n¼ 28

Type P 73 (71 %) 10 (36 %)Type P1 54 (52 %) –Type P2 19 (18 %) –

Type D 30 (29 %) 18 (64 %)

Number of fingers (%).

2 S. OKITA ET AL.

base of the proximal phalanx as well as into the extensorhood and extended to the base of the distal phalanx in 9%(n¼ 13) of hands; and inserted into the extensor hood andthe base of the distal phalanx in 9% (n¼ 13) of hands.When we classified these results according to our classifica-tion criteria, 70% (n¼ 101) of the thumbs were Type Pthumbs and 30% (n¼ 43) were Type D, those were verysimilar to our results. However, our histological findingsindicated that the hood is a superficial transverse fiber,which runs over the dorsal surface of the MCP joint andthat there were no instances of EPB insertion into the hood.

Several studies have described the anatomy of the dorsalfibrous complex of the MCP joint of the thumb [17–19].Bade et al. [17], in a microscopic study, reported that thedorsal connective tissue of the thumb forms different layersof collagen lamella as a peritendinous system around EPLand EPB. Hunter-Smith et al. [18] described the dorsal tri-angular fibrocartilage of the MCP joint filling the dorsal

space between the proximal phalanx and the metacarpalhead to stabilize and match the congruity of the joint. Joshiet al. [19] described loose connective tissue connectionsbetween the dorsal triangular fibrocartilage and the extensortendon. Our histological findings support these studies, andwe found the dorsal fibrous complex of the MCP joint ofthe thumb was mainly composed of three layers: the exten-sor hood, the tendon, and the fibrous capsule integratedwith the dorsal fibrocartilage in the vertical plane. Thus, inthis study, we classified the insertion pattern of EPB intothree Types; P1, P2, and D.

In RA patients with thumb boutonni!ere deformity, 36%(n¼ 10) of the thumbs were Type P. In Type P2, EPB had a

233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290

291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348

Figure 2. Histology of vertical sections of EPB of cadavers (Masson’s trichromestain). (a) Type P1 insertion of EPB at the MCP joint. The dorsal fibrous complexof the MCP joint is composed of three layers: the superficial layer with trans-verse fibers that are interchangeably referred to as the extensor hood (whitearrowheads), the intermediate layer with EPB (black arrowheads), and the deeplayer with the fibrous capsule integrated into the dorsal fibrocartilage (straightblack arrows). The EPB inserts into the base of the proximal phalanx. (b) TypeP2 insertion of EPB at the MCP joint. Arrows and arrowheads denote the samestructures as in Figure 2(a). EPB attaches to the base of the proximal phalanxwith loose connective tissue and the end of the EPB is integrated into the dor-sal fibrocartilage. The extensor hood is seen on the dorsal surface of MCP joint.(c) Vertical section of Type D insertion of EPB. Arrows and arrowheads denotethe same structures as in Figure 2(a). The extensor hood is difficult to identifyin this low-power field picture. EPB has no insertion into the proximal phalanxand inserts to the distal phalanx.

Figure 3. The dorsal aspect of left thumb boutonni!ere deformity secondary toRA during surgery. (a) Type P insertion of EPB that ends at the level of the MCPjoint. (b) Type D insertion of EPB that inserts into the distal phalanx.

Figure 4. High-power field of the vertical sections of the metacarpophalangealregion of EPB (Masson’s trichrome stain). (a) The hood (white arrowheads) isfound on the dorsal surface of the MCP joint. Loose connective tissue spacewas found between the hood and EPB. (b) The thin extensor hood (whitearrowheads) is found on the dorsal surface of EPB. No connective tissue spaceis found between EPB and the hood.

MODERN RHEUMATOLOGY 3

weaker attachment to the proximal phalanx than Type P1,with firm attachment of EPB to the proximal phalanx, sosynovitis not only attenuates EPB but also can easily causedetachment of EPB from the proximal phalanx with dorsalfibrocartilage, causing thumb boutonni!ere deformity.Therefore, we speculated that in RA patients with thumbboutonni!ere deformity with Type P insertion, the incidenceof Type P2 EPB might be higher than Type P1 (Figure 5).

We assumed that it is congenitally determined whetherthe EPB is Type P or Type D, and the incidence of differentinsertion is not affected by inflammation of RA. However, itis difficult to distinguish Type P1 from Type P2 because theinflammation of RA may cause the disruption of the TypeP1 insertion (Table 1). In the present study, the rate of

Type D was higher in thumb boutonni!ere deformity com-pared with the normal population. If a patient with RA hasType D EPB, thumb boutonni!ere deformity might be easilydeveloped as a result of synovitis and swelling at the MCPjoint, because the Type D EPB insertion has no bony attach-ment around the MCP joint. This may explain the signifi-cantly higher incidence of Type D in hands with thumbboutonni!ere deformity among our RA patients who requiredsurgery.

There are some limitations to this study. First, we didnot take any samples from RA patients, because in mostcases EPB was ruptured or had lost the insertion to theproximal phalanx. During surgical reconstruction of EPBinsertion into the proximal phalanx, it is difficult to obtainsamples of the bone–tendon unit. Second, the cadavers andRA patients used in our study were all Japanese. As Jabiret al. [9] observed, the anatomy and variations of EPB seemto show ethnic differences. Third, the small RA patient sam-ple size reduced the power of the study. Finally, this studyhas a statistical weakness in comparing the result of RApatients with boutonni!ere deformity who required surgicalreconstruction with that of cadavers. We have no dataregarding the insertion pattern of RA patients with thedeformity who did not require surgery or of RA patientswithout boutonni!ere deformity. Further studies on EBP vari-ation in RA cases undergoing surgery or investigations ofEPB in RA patients with boutonniere deformity by otherimaging techniques will be needed to enhance our under-standing of the mechanism of the deformity.

Conflict of interests

None

References

1. Ratliff AH. Deformities of the thumb in rheumatoid arthritis.Hand. 1971;3(2):138–43.

2. Brumfield RH, Conaty JP. Reconstructive surgery of the thumbin rheumatoid arthritis. Orthopedics. 1980;3:529–33.

3. Nalebuff EA. Diagnosis, classification and management ofrheumatoid thumb deformities. Bull Hosp Joint Dis. 1968;29(2):119–37.

4. Terrono A, Millender L, Nalebuff E. Boutonniere rheumatoidthumb deformity. J Hand Surg Am. 1990;15(6):999–1003.

5. Nalebuff EA. The rheumatoid thumb. Clin Rheum Dis. 1984;10(3):589–607.

6. Inglis AE. Rheumatoid arthritis in the hand. Am J Surg. 1965;109:368–74.

7. Shigematsu S, Shimizu H, Beppu M, Hirata K. Anatomy of theextensor pollicis brevis associated with an extension mechanismof the thumb metacarpophalangeal joint. Hand Surg. 2014;19(02):171–9.

8. Kulshreshtha R, Patel S, Arya AP, Hall S, Compson JP.Variations of the extensor pollicis brevis tendon and its inser-tion: a study of 44 cadaveric hands. J Hand Surg Eur Vol. 2007;32(5):550–3.

9. Jabir S, Lyall H, Iwuagwu FC. The extensor pollicis brevis: areview of its anatomy and variations. Eplasty. 2013;13:e35

10. Stein AH. Jr., Variations of the tendons of insertion of theabductor pollicis longus and the extensor pollicis brevis. AnatRec. 1951;110(1):49–55.

349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406

407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464

Figure 5. (a) Schematic representation of the classical mechanism of the devel-opment of boutonni!ere deformity in Type P1 thumbs. Synovitis of the MCPjoint causes detachment of EPB from the proximal phalanx (straight blackarrow), and the displaced EPL promotes MCP joint flexion and IP joint hyper-extension. However, the EPB inserts firmly into the proximal phalanx, so thatsynovitis cannot easily cause detachment of EPB from the proximal phalanx. (b)Schematic representation of the mechanism of development of boutonni!eredeformity in Type P2 thumbs. EPB inserts into the dorsal capsule–fibrocartilagecomplex. Synovitis not only attenuates the EPB but also easily detaches thedorsal capsule–fibrocartilage complex from the proximal phalanx (straight blackarrow). The dislocated EPL leads to thumb boutonni!ere deformity. (c)Schematic representation of the mechanism of development of boutonni!eredeformity in Type D thumbs. EPB mainly inserts into the distal phalanx, whichcannot extend the MCP joint independently but extends the MCP and IP jointscollectively. The dislocated EPL leads to thumb boutonni!ere deformity.

4 S. OKITA ET AL.

11. Yoshida Y. Anatomical study on the extensor digitorum profun-dus muscle in the Japanese. Okajimas Folia Anat Jpn. 1990;66(6):339–53.

12. Aydinlio"glu A, Sakul BU, Diyarbakirli S. A rare insertion site forabductor pollicis longus and extensor pollicis brevis muscles.Acta Anat (Basel). 1998;163(4):229–32.

13. Abdel-Hamid GA, El-Beshbishy RA, Abdel Aal IH. Anatomicalvariations of the hand extensors. Folia Morphol (Warsz). 2013;72(3):249–57.

14. Aydinlioglu A, Tosun N, Keles P, Akpinar F, Diyarbakirli S.Variations of abductor pollicis longus and extensor pollicis brevismuscles: surgical significance. Kaibogaku Zasshi 1998;73:19–23.

15. Brunelli GA, Brunelli GR. Anatomy of the extensor pollicis bre-vis muscle. J Hand Surg Br. 1992;17(3):267–9.

16. Dawson S, Barton N. Anatomical variations of the extensor pol-licis brevis. J Hand Surg Br. 1986;11(3):378–81.

17. Bade H, Krolak C, Koebke J. Fibrous architecture of the dorsalaponeurosis of the thumb. Anat Rec. 1995;243(4):524–30.

18. Hunter-Smith DJ, Slattery PG, Rizzitelli A, Hunter-Smith SR,Fairbank S, Rozen WM, Findlay MW. The dorsal triangularfibrocartilage of the metacarpophalangeal joint: a cadavericstudy. J Hand Surg Am. 2015;40(7):1410–5.

19. Joshi SS, Joshi SD, Athavale SA, Kishve PS, Jadhav SD. Dorsaldigital expansion of thumb. J Anat Soc India. 2008;57:135–9.

465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522

523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580

MODERN RHEUMATOLOGY 5