Entheses: tendon and ligament attachment sites

M. Benjamin1, D. McGonagle2

1School of Biosciences, Cardi! University, Museum Avenue, Cardi! , UK, 2Academic Unit of Musculoskeletal Disease, ChapelAllerton Hospital, Leeds, UKCorresponding author: M. Benjamin, School of Biosciences, Cardi! University, Museum Avenue, Cardi! CF10 3AX, UK.E-mail: benjamin@cardi!.ac.uk

Accepted for publication 14 December 2008

The current brief review focuses on certain issues relating toform–function relationships that are evident at tendon orligament attachment sites (entheses). It evaluates the devel-opment of entheses (both fibrocartilaginous and fibrous) andhighlights again an issue largely ignored for decades – i.e.how entheses attached to the metaphyses of long bonesmanage to keep the same relative position as the bones growin length. Attention is drawn to the manner in which enthesisfibrocartilage prevents direct cell–cell communication be-tween osteocytes and tendon/ligament cells and how (in a

healthy enthesis) it presents a physical barrier separatingthe blood supply of bone from that of tendon/ligament. Thepossibility that the thoracolumbar fascia, with its multitudeof muscular associations and numerous sites of ligamentousattachment could increase stress concentration at entheses israised, the structure and development of enthesophytes(bony spurs) is reviewed as is the concept of a synovio-entheseal complex (SEC). How these functional anatomicalunits (SECs) could trigger pain and inflammation in athletesis briefly discussed.

An enthesis is the site of attachment of a tendon,ligament or joint capsule to the skeleton (Fig. 1). Itsclinical significance is largely threefold – (a) it is acommon region for overuse injuries; (b) it is targetedin rheumatic diseases known collectively as theseronegative spondyloarthropathies – which includeankylosing spondylitis and psoriatic arthritis; (c) itneeds to be reconstituted when a tendon or ligamentis reattached to a bone at surgery – e.g. in an anteriorcruciate ligament reconstruction or a tendon transferprocedure. In addition, entheses have been of longinterest to anthropologists and archeologists, for themarkings left on dried bones by tendons and liga-ments, or the influence of muscle pull on overall boneshape, have been interpreted by some authors asindicators of the type and/or extent of physicalactivity or whether an individual was right or lefthanded (Dutour, 1986; Drapeau, 2008).The current article focuses on aspects of form–

function relationships of entheses manifest at thetissue and organ level and seeks to complementrather than duplicate other recent reviews on en-theses by the authors and their colleagues (Benjamin& McGonagle, 2001; Benjamin et al., 2002, 2008a, b;Milz et al., 2005; Shaw & Benjamin, 2007). Thus, thereview does not attempt to deal in an exhaustive waywith all threads of the topic that might be suggestedby its title. It focuses instead on developmentalmatters, the relationship between tendon and liga-

ment entheses and fascia, the ‘‘barrier’’ characteris-tics of enthesis fibrocartilage at hard–soft tissueinterfaces, the concept of a synovio-entheseal com-plex (SEC) and the structure, development and sig-nificance of enthesophytes (bony spurs at attachmentsites).

Enthesis development

The story of enthesis development and of enthesisfibrocartilage di!erentiation in particular is an ex-cellent illustration of how entheses can to a largeextent, be regarded as self-designing systems wheremorphology is dictated by the influences of mechan-ical load. It is a clear example of how functiondictates structure. A study of enthesis develop-ment helps us to understand how mechanical influ-ences can account for di!erences in the quantity offibrocartilage at di!erent entheses (Evans et al., 1990;Benjamin et al., 1991, 1992), the regional distributionof fibrocartilage within an attachment site and thesite of development of bony spurs (Benjamin et al.,2008a, b; McGonagle et al., 2008). Intriguingly, oneof the great pioneers of enthesis biology, Schneider(1956) commented on how the structure of thebiceps brachii enthesis di!ered in a person who hadspent a lifetime cleaning windows (hence with apronated forearm) compared with a farm hand who

Scand J Med Sci Sports 2009: 19: 520–527& 2009 John Wiley & Sons A/Sdoi: 10.1111/j.1600-0838.2009.00906.x

520

had spent his time milking cows (with a supinatedforearm)!Our understanding of the di!erentiation of fibro-

cartilage during enthesis development is aided by theimmunohistochemical study of Gao et al. (1996) onthe femoral attachment of the medial collateralligament (MCL) of the rat knee joint. In the fetusor newborn rat, the tendon/ligament initially at-taches to the hyaline cartilage that forms the anlagenfor the future bone. As ossification occurs and bonereplaces hyaline cartilage, the cartilage at the attach-ment site initially escapes erosion to leave a cartilagedisc at the attachment site. However, this plug ofhyaline cartilage is eventually completely resorbedand replaced by fibrocartilage (Fig. 2). The resorp-tion and replacement occur hand in hand and a keypoint is that the fibrocartilage develops by metaplasiain the tendon or ligament. Thus, cells that wereinitially di!erentiating into fibroblasts now changephenotype and become fibrocartilage cells. If thetendon/ligament fibroblasts were arranged in long-itudinal rows between parallel collagen fibers at theenthesis, then the fibrocartilage cells become ar-ranged likewise (Fig. 1). It seems highly probablethat the signal for metaplasia is mechanical loadingand in particular, elevated levels of compression and/or shear – in-line with general principles relatingto the development of fibrocartilage elsewhere

(Benjamin & Ralphs, 1998). Thus, in the rat kneejoint, the enthesis fibrocartilage of the MCL replacesthe original hyaline cartilage rudiment between 30and 45 days post-partum (Gao et al., 1996) – i.e. aftera period of physical activity and at a time corre-sponding to an increase in the mechanical strength ofthe ligament (Booth & Tipton, 1970; Tipton et al.,1978). Further evidence supporting the importanceof mechanical loading for enthesis fibrocartilagedevelopment comes from the experimental work of(Thomopoulos et al., 2007) on the supraspinatusenthesis of the mouse. When the left shoulders ofnewborn mice were paralyzed by intramuscular in-jections of botulinum A toxin, fibrocartilage devel-opment was strikingly delayed compared with thecontralateral shoulder, which received a saline injec-tion alone. This work contrasts strikingly with thefindings of Zumwalt (2006) who investigated therelationship between enthesis morphology of sixdi!erent well-defined tendons, and correspondingmuscle action/size in adult sheep. Her interest largelyrelated to attempts by archeologists to gain an insightinto likely activity levels from skeletal remains. Thus,it must be recognized that her evaluations of enthesismorphology were entirely restricted to the surfacefeatures of adult macerated bones. Nevertheless, it isworth noting that ‘‘endurance-trained sheep’’ (exer-cised for 90 days) did not have a macroscopic bonemorphology at attachment sites that di!ered from

Fig. 2. A diagrammatic representation of the developmentof fibrocartilage at the femoral attachment of the medialcollateral ligament of the rat knee joint, based on the work ofGao et al. (1996). (a) Initially, the tendon or ligament[characterized by di!erentiating fibroblasts (F)] attaches tothe hyaline cartilage (HC) that preceded the bone rudiment.(b) As ossification ensues, the hyaline cartilage is eroded, butfibrocartilage (FC) appears in the tendon or ligament be-cause of metaplasia of the fibroblasts. (c) The metaplasia isprobably driven by mechanical stimuli associated withmovement of the tendon or ligament relative to the bone(B) at the insertion site. (d) Eventually all the originalcartilage rudiment is resorbed, but further fibrocartilageforms in the tendon or ligament by continuing metaplasia.

Fig. 1. The olecranon enthesis of triceps brachii in man. Asthe tendon approaches the bone (B), it becomes fibrocarti-laginous. A broader zone of uncalcified fibrocartilage (UF) isseparated from a thinner region of calcified fibrocartilage(CF) by a tidemark (T). This represents the outer limit ofcalcification and is thus the hard–soft (i.e. mechanical)boundary of the attachment site. Its straight charactercontrasts with the irregular interface (*) of the true tissueboundary between bone and tendon. Note that the cells inthe zone of uncalcified fibrocartilage (arrows) are arrangedin longitudinal rows between parallel collagen fibers (C) andthat in this non-pathological specimen, there are no bloodvessels visible in the fibrocartilage zones. Masson’s tri-chrome. Scale bar5 100 mm.

Entheses

521

that of controls. Whether this reflects the type ofmuscle fiber activity promoted by endurance trainingor the use of adult rather than juvenile animals isunclear. It is also uncertain whether endurancetraining influences enthesis structure at a cell, tissueor molecular level. Although the e!ects of endurancetraining on enthesis structure are unknown, a greatercross-sectional area of the Achilles tendon in distancerunners than non-runners has been noted (Magnus-son & Kjaer, 2003). The same authors stated that theincrease in cross-sectional area a!ected the distalpart of the tendon, but it was unclear whether thisembraces the precise site of tendon attachment to thecalcaneus (Magnusson & Kjaer, 2003). It would thusbe interesting to know if the ‘‘footprint’’ of thetendon is correspondingly expanded in runners. Ifso, this suggests appositional growth of the attach-ment site – implying fibrillogenesis around the cir-cumference of the ‘‘trained enthesis’’ in order todissipate stress concentration in-line with the in-creased loading on the tendon. However, if thefootprint is not increased, stress concentration perunit area must be greater and this could have abearing on the incidence of insertional tendinopa-thies. Such considerations are entirely conjectural,but worthy of further investigation in a suitableanimal model.There is an intriguing issue concerning the devel-

opment of entheses that attach to the shafts of longbones that also highlights the importance of mechan-ical loading. Such tendons or ligaments need tomigrate along a bone as it grows, in order that theycan maintain the same relative position with respectto the growth plate. A number of authors haveaddressed this issue over the years, but perhaps themost important contributions are those of Dorfl(1980a, b). As Dorfl points out, the migration reflectsthe fact that interstitial growth does not contribute tothe elongation of diaphyses – which can only grow inlength by the activity of the epiphyseal growth plate.Thus, if a tendon/ligament does not migrate at a ratematching the overall growth in length of the longbone to which it is attached, it will eventually beattached nearer to the middle of the shaft by adult-hood (Dorfl, 1980a). The necessary migration can bepromoted by an attachment of the growing tendon/ligament to the periosteum rather than to the boneitself. As the periosteum is stretched by the growth inlength of the long bone and/or by muscular traction,the entheses are ‘‘dragged’’ with it (Dorfl, 1980a).The rate of migration depends on the position of theenthesis relative to the growth plate and to the‘‘neutral center.’’ LaCroix (1949) defines this as theplace at which the periosteum is pulled equally in twoopposite directions by the activity of the proximal-and distal-growth plates. The location of the neutralcenter varies from bone to bone, according to the

proportional contribution to growth in length attri-butable to the epiphyseal growth plates at either endof the long bone. The farther away an enthesis isfrom the neutral center, the greater the rate ofmigration. The periosteum drags the enthesis withit and is itself pulled by the epiphyses as these arepushed apart by growth plate activity (Dorfl, 1980a).Intriguingly Dorfl argues that the details of themechanism of migration vary and depend on whetherthe tendon/ligament is attaching to a resorptive sur-face, an osteogenic surface or a surface that is bothresorptive and osteogenic. Thus the tendon/ligamentdoes not necessarily have to slide over the bonesurface – its migration can be facilitated by inter-stitial growth of the enthesis itself if the developingenthesis is characterized by an osteogenic zone(Dorfl, 1980a). Dorfl hypothesizes that the collagenfibers lengthen at the insertion site where an osteo-genic zone is present, so that the attachment itself isalways maintained. Where the attachment is to anabsorptive surface, fibrillogenesis occurs and there ispossibly also a shortening of existing fibers (presum-ably reflecting matrix metalloproteinase activity) thathelps to maintain a continuous bony attachmentduring growth. The molecular basis of such eventsis completely unexplored.The tension within the collagen fibers that facil-

itates migration could be regarded as producingdefined traction lines directing enthesis migration ina macroscopic equivalent of traction lines directingcell migration in a culture dish in vitro. It is also agood example of a tensegrity system operating at thetissue level. Interestingly, the experimental studies ofDorfl (1980b) show that if the growth plates areprevented from functioning properly and thus ten-sion does not develop in the periosteum, the enthesesdo not migrate. He confirmed this by insertingstaples into rabbit long bone to arrest growth plateactivity. Without normal bone lengthening, the peri-osteum is not stretched, and thus the periosteumcannot drag any entheses with it (Dorfl, 1980b).Grant et al. (1980) also has recognized the impor-tance of tension developing in the periosteum as aresult of growth plate activity. He referred to theperiosteum at such locations as an elastic sleeve andenvisaged soft tissues (muscles, tendons, ligaments)attaching to it as being like hitchhikers carried alongan expanding periosteum (Grant et al., 1980).

Fascia and entheses

The importance of fascial expansions of entheses toadjacent structures (including other entheses) as amechanism for stress dissipation has been empha-sized in recent reviews (Benjamin et al., 2008a, b).Here we focus attention on the converse point of

Benjamin & McGonagle

522

whether fascial connections may actually increase thestress levels to which entheses are exposed. Considerthe thoracolumbar fascia (TLF). This very extensiveand complex collection of sheets of dense connectivetissue is arranged at various depths in the thoracicand lumbar regions of the spine (Bogduk & Macin-tosh, 1984; DeRosa & Porterfield, 2007). It providesan important surface for muscle attachment and itsextensive nature means that seemingly disparatemuscles of the trunk and extremities are mechani-cally linked to each other by virtue of their commonfascial connections (Vleeming et al., 1995; DeRosa &Porterfield, 2007). Thus, the TLF e!ectively couplestogether gluteus maximus, latissimus dorsi, erectorspinae, multifidus, quadratus lumborum, the rhom-boids, biceps femoris and the flat muscles of theabdominal wall! This is an impressive diversity ofmuscles spanning the trunk, pelvis and extremities.As the fascia also attaches firmly to many ligaments(each with its own enthesis), it means that stressconcentration at an individual bone–ligament junc-tion can be influenced by events some considerabledistance away. For example, stress concentration at afacet joint capsule could theoretically be influencedby, e.g. multifidus tone, for the TLF fuses withthe supraspinous ligaments which in turn are con-tinuous with the interspinous ligaments anchoring tothe facet joint capsules (Willard, 2007). Such aninterspinous–supraspinous–thoracolumbar ligamentcomplex creates a central support system for thelumbar spine (Willard, 2007).By a similar reasoning, many muscles can generate

force that is transferred to the sacroiliac joint (SIJ)and its entheses via the TLF – potentially increasingstress concentration at these sites. It should be notedthat the TLF attaches to the ligamentous stockingthat surrounds the lumbar spine and expands infer-iorly to form a strong and elaborate, posteriorcapsule for the SIJ (Willard, 2007). Evidently, musclecontraction stabilizes the lumbospinal region in away that ligamentous restraints alone cannot (De-Rosa & Porterfield, 2007). Thus, van Wingerden etal. (2004) have shown that isometric contractions ofbiceps femoris, gluteus maximus or erector spinaecan increase the sti!ness of the SIJ. They suggestedthat the muscle-induced sti!ness reduced the levels ofshear in the SIJ in preparation for load transfer fromthe spine to the legs. If the suggestion is correct, it notonly means that shear is reduced in the synovial partof the joint, but also in the fibrous part – i.e. at theentheses of the interosseous ligament. Nevertheless,the tensile loading of attachment sites is likely to beincreased and thus it seems possible to argue thatmuscle tone could both increase and decrease aspectsof stress concentration at SIJ joint entheses.Francois et al. (2000) have cautioned that enthesi-

tis is not the sole contributory factor to SIJ joint

disease in ankylosing spondylitis but that synovitisand subchondral bone marrow changes may o!er abetter explanation for the joint destruction thatoccurs in this disease. Nevertheless, they recognizethat enthesitis is a feature and thus the influence ofmuscle tone and myofascial sti!ness on the SIJ andits ligaments merits further consideration. In thiscontext, it should be noted that Masi et al. (2005)have drawn attention to the forgotten ‘‘bowstringsign’’ of Jacques Forestier in early ankylosing spon-dylitis. Such patients have firm, contracted dorso-lumbar muscles on the concave side when the patientflexes his/her spine laterally. This is not a feature ofnormal individuals, but the observation is not widelyknown by rheumatologists today.

Enthesis fibrocartilage as a barrier to direct cellularcommunication between tenocytes and osteocytes

It is well known that osteocytes have numerouselongated processes that extend into canaliculi, per-meating the bone matrix and communicating witheach other via gap junctions. Equally, gap junctionalcommunication is a feature of tendons, for confocalmicroscopy demonstrates that these cells also havelong, delicate processes that contact those of neigh-boring cells (McNeilly et al., 1996). Gap junctions arepresent both between tendon/ligament cells withinthe same longitudinal row and between cells inadjacent rows (McNeilly et al., 1996) and theirlocation is shown diagrammatically in Fig. 3. Criti-cally, however, the cells of enthesis fibrocartilage donot express connexins (Ralphs et al., 1998; Fig. 3)and thus as in hyaline cartilage, intercellular com-munication must be indirect – via cell–matrix inter-actions. This creates a barrier to direct cell–cellcommunication between tendon/ligament and bone,complementing the vascular barrier stemming fromthe lack of blood vessels in the enthesis fibrocartilageof a healthy attachment (Fig. 1). The significance ofthis is unclear but it may be related to preventingbone tissue growing into tendons or ligaments.This is supported by the observation that bonyspurs typically develop in the most fibrous regi-ons of fibrocartilaginous entheses (Benjamin et al.,2008a, b).

Synovio-entheseal complexes

Benjamin & McGonagle (2001) coined the term‘‘enthesis organ’’ to capture the idea that there ismore to an enthesis than simply the attachment siteitself. Many tendons or ligaments attach to bone atthe bottom of a shallow pit or adjacent to a slightlyelevated region of bone (e.g. a tuberosity) so thatthere is the potential for a degree of mechanical

Entheses

523

leverage between the tendon/ligament and the boneimmediately adjacent to the attachment. This helpsto dissipate stress concentration away from the en-thesis itself. A common component of many enthesisorgans is synovium. This may line a subtendinousbursa, form a tendon sheath that extends close to theattachment site itself or be part of an adjacentsynovial joint. Thus a region vulnerable to mechan-ical damage (the enthesis), that is essentially anti-inflammatory and poorly vascularized is juxtaposedto a rather di!erent tissue (synovium) that in contrastis pro-inflammatory and highly vascular. In order toemphasize the importance of recognizing this anato-mical proximity, we have coined the term synovio-entheseal complex (McGonagle et al., 2007) anddocumented examples of SECs in a follow-up paper(Benjamin & McGonagle, 2007). The SECs asso-

ciated with the insertions of the subscapularis tendonon the lesser tuberosity of the humerus and theAchilles tendon on the calcaneus are illustrated inFigs 4 and 5.We have argued that synovial membranes near

entheses ‘‘walk an immunological tightrope’’ in somuch as tissue breakdown products resulting fromwear and tear at an attachment site could trigger aninflammatory reaction in the adjacent synovium(McGonagle et al., 2007). Such a concept is particu-larly relevant to the seronegative spondyloarthropa-thies (SpA) where enthesitis and synovitis are oftenviewed as independent pathologies. It suggests thepossibility that bursitis could lead to enthesitis andvice versa – a suggestion made earlier by Rufai et al.(1995). Thus the term ‘‘synovio-entheseal complex’’is an attempt to increase awareness that (a) enthesitisand synovitis may not necessarily be independentpathologies (b) degenerative changes at an enthesiscould trigger inflammation in a synovial membranethat is not part of a synovial joint.

Fig. 3. A diagrammatic representation of fibroblasts (F) andfibrocartilage cells (FC) at an enthesis. The presence of gapjunctions (GJ) facilitating intercellular communication is afeature of the former only. The gap junctions are presentboth between adjacent cells in the same longitudinal row (leftside of the figure) and between cells in adjacent rows (rightside of the figure). Note the irregular shape of the fibroblaststhat is evident in confocal images of tendons/ligaments cut incross-section (see McNeilly et al., 1996 for confocal images).It is the presence of long finger-like and sheet-like cellprocesses that enables the cells to contact each other. Thiscontrasts with the more rounded shape of the fibrocartilagecells that isolates them within the extracellular matrix. Thecollagen fibers are included in the longitudinal section butomitted from the transverse section to highlight the positionof the gap junctions.

Fig. 4. A low-power view of the insertional enthesis (E) ofthe tendon of subscapularis in man showing the existence ofa synovio-entheseal complex. The enthesis is closely asso-ciated with the synovium both of the adjacent bursa (B) andof the tendon sheath (TS) surrounding the biceps tendon(BT). Masson’s trichrome. Scale bar5 5mm.

Benjamin & McGonagle

524

Enthesophytes

Bony spurs (enthesophytes) are common features ofattachment sites. They include the syndesmophytesthat develop in the annulus fibrosus (and which bytheir growth can lead eventually to spinal ankylosisin SpA patients) and bony spurs that grow at or nearnumerous other entheses [particularly the Achillestendon (Fig. 5) and plantar fascia]. They are ob-viously closely related to osteophytes that develop at

the periphery of synovial joints in association withdegenerate articular cartilage in patients withosteoarthritis. Indeed it has been suggested thatthe development of both types of spurs is due tothe fact that some people are intrinsic ‘‘bone for-mers’’ (Rogers et al., 1997). They even suggest thatdi!erences in the propensity for spur formationbetween individuals could contribute to the mixedclinical outcomes that certain treatment protocolsmay produce. Intriguingly, Menz et al. (2008) alsoprovide evidence supporting a link between osteoar-thritis and the formation of plantar calcaneal enthe-sophytes. In our studies we have noted that youngnormal subjects often have small Achilles tendonspurs (McGonagle et al., 2008).We have recently analyzed the structure of enthe-

sophytes of di!erent sizes, at a wide variety ofattachment sites in order to try to piece together anaccount of how they could develop (Benjamin et al.,2008a, b). It seems that in man at least, their devel-opment can embrace a number of methods of ossi-fication – endochondral, intramembranous andchondroidal. Evidence of endochondral ossificationis suggested by the presence of calcified cartilageremnants in the cores of some larger spurs. Cur-iously, however, signs of chondrocytic hypertrophyare far less striking than they are in the classicendochondral ossification that typifies the growthplate of a long bone. Although it is possible thatmarked hypertrophy does indeed occur in entheso-phytes that form with age in man and has simplybeen missed because of its short duration, it is alsoworth considering that the absence of significanthypertrophy reflects a much slower rate of bonedevelopment in a spur compared with a growth plate.Whatever the explanation, it is certainly the case thatenthesophytes usually develop at the edge of anenthesis – generally the most fibrous (and thus leastcartilaginous) part of a fibrocartilaginous attachmentsite (Benjamin et al., 2008a, b). This observation is in-line with the widespread view that enthesophytes aregenerally traction spurs. Nevertheless, we ourselveshave argued that plantar fascial spurs cannot betraction spurs, for they must lie on the surface ofthe plantar fascia rather than within it (Kumai &Benjamin, 2002). We have suggested that they de-velop in response to degenerative changes at theenthesis. Our histological observations on the loca-tion of plantar fascial spurs are supported andextended by the radiological studies of Abreu et al.(2003). According to these authors, heel spurs devel-oping on the plantar surface of the calcaneus canarise at several di!erent locations – notably theattachments of local intrinsic muscles, but rarely inthe plantar fascia itself. More recently, Li andMuehleman (2007) have also expressed the view,based on their radiological, gross anatomical and

Fig. 5. The synovio-entheseal complex at the insertion ofthe human Achilles tendon (AT) onto the calcaneus (C).Note that the enthesis (E) lies adjacent to a bursa (B) intowhich the synovial-covered tip of a fat pad (FP) protrudes.There is a prominent spur in the distal part of the enthesis.Toluidine blue. Scale bar5 4mm.

Entheses

525

histological studies, that plantar spurs do not de-velop in response to traction. An important pointmade by these authors is that the trabecular organi-zation of the spurs is more suggestive of compressiveforces acting on the spur than tensile loading alongits long axis. The importance of compression in thedevelopment of plantar heel spurs is further sup-ported by the recent work of Menz et al. (2008) –who found that patients with such spurs are morelikely to be obese than those without.

Perspectives

A sound understanding of form–function relation-ships of entheses underpins any appreciation of what

may be happening in pathological conditions –whether the enthesopathies are associated with thedevelopment of overuse injuries or with the spondy-loarthritides. In relationship to overuse injuries, wehave noted that even normal attachments are proneto microdamage and microscopic inflammation. Itappears that overuse exaggerates this response andleads to clinically manifesting disease. Following onfrom this, it would seem that overuse injuries includ-ing bursitis are likely to be (at least in part) enthesisorgan damage manifesting as adjacent synovitis.

Key words: insertion sites, enthesopathy, develop-ment, fibrocartilage, fascia.

References

Abreu MR, Chung CB, Mendes L,Mohana-Borges A, Trudell D, ResnickD. Plantar calcaneal enthesophytes:new observations regarding sites oforigin based on radiographic, MRimaging, anatomic, andpaleopathologic analysis. SkeletalRadiol 2003: 32: 13–21.

Benjamin M, Evans EJ, Rao RD, FindlayJA, Pemberton DJ. Quantitativedi!erences in the histology of theattachment zones of the meniscal hornsin the knee joint of man. J Anat 1991:177: 127–134.

Benjamin M, Kaiser E, Milz S. Structure–function relationships in tendons: areview. J Anat 2008a: 212: 211–228.

Benjamin M, Kumai T, Milz S, BoszczykBM, Boszczyk AA, Ralphs JR. Theskeletal attachment of tendons –tendon ‘‘entheses’’. Comp BiochemPhysiol 2002: 133: 931–945.

Benjamin M, McGonagle D. Theanatomical basis for diseaselocalisation in seronegativespondyloarthropathy at entheses andrelated sites. J Anat 2001: 199: 503–526.

Benjamin M, McGonagle D.Histopathologic changes at ‘‘synovio-entheseal complexes’’ suggesting anovel mechanism for synovitis inosteoarthritis and spondylarthritis.Arthritis Rheum 2007: 56: 3601–3609.

Benjamin M, Newell RL, Evans EJ,Ralphs JR, Pemberton DJ. Thestructure of the insertions of thetendons of biceps brachii, triceps andbrachialis in elderly dissecting roomcadavers. J Anat 1992: 180: 327–332.

Benjamin M, Ralphs JR. Fibrocartilagein tendons and ligaments – anadaptation to compressive load. J Anat1998: 193 : 481–494.

Benjamin M, Toumi H, Suzuki D,Hayashi K, McGonagle D. Evidencefor a distinctive pattern of boneformation in enthesophytes. AnnRheum Dis 2008b.

Bogduk N, Macintosh JE. The appliedanatomy of the thoracolumbar fascia.Spine 1984: 9: 164–170.

Booth FW, Tipton CM. Ligamentousstrength measurements in pre-pubescent and pubescent rats. Growth1970: 34: 177–185.

DeRosa C, Porterfield J. Anatomicallinkages and muscle slings of thelumbopelvic region. In: Vleeming A,Mooney V, Stoeckart R, eds.Movement, stability & lumbopelvicpain Integration of research andtherapy. Edinburgh: ChurchillLivingstone Elsevier, 2007:47–62.

Dorfl J. Migration of tendinousinsertions. I. Cause and mechanism.J Anat 1980a: 131: 179–195.

Dorfl J. Migration of tendinousinsertions. II. Experimentalmodifications. J Anat 1980b: 131:229–237.

Drapeau MS. Enthesis bilateralasymmetry in humans and Africanapes. Homo 2008: 59: 93–109.

Dutour O. Enthesopathies (lesions ofmuscular insertions) as indicators ofthe activities of neolithic Saharanpopulations. Am J Phys Anthropol1986: 71: 221–224.

Evans EJ, Benjamin M, Pemberton DJ.Fibrocartilage in the attachment zonesof the quadriceps tendon and patellarligament of man. J Anat 1990: 171:155–162.

Francois RJ, Gardner DL, Degrave EJ,Bywaters EG. Histopathologicevidence that sacroiliitis in ankylosing

spondylitis is not merely enthesitis.Arthritis Rheum 2000: 43: 2011–2024.

Gao J, Messner K, Ralphs JR, BenjaminM. An immunohistochemical study ofenthesis development in the medialcollateral ligament of the rat knee joint.Anat Embryol 1996: 194: 399–406.

Grant PG, Buschang PH, Drolet DW,Pickerell C. Invariance of the relativepositions of structures attached to longbones during growth: cross-sectionaland longitudinal studies. Acta Anat1980: 107: 26–34.

Kumai T, Benjamin M. Heel spurformation and the subcalcanealenthesis of the plantar fascia. JRheumatol 2002: 29: 1957–1964.

LaCroix P. L’organisation des Os. Paris:Masson, 1949.

Li J, Muehleman C. Anatomicrelationship of heel spur tosurrounding soft tissues: greatervariability than previously reported.Clin Anat 2007: 20: 950–955.

Magnusson SP, Kjaer M. Region-specificdi!erences in Achilles tendon cross-sectional area in runners and non-runners. Eur J Appl Physiol 2003: 90:549–553.

Masi AT, Sierakowski S, Kim JM.Jacques Forestier’s vanished bowstringsign in ankylosing spondylitis: a call totest its validity and possible relation tospinal myofascial hypertonicity. ClinExp Rheum 2005: 23: 760–766.

McGonagle D, Lories RJ, Tan AL,Benjamin M. The concept of a‘‘synovio-entheseal complex’’ and itsimplications for understanding jointinflammation and damage in psoriaticarthritis and beyond. Arthritis Rheum2007: 56: 2482–2491.

McGonagle D, Wakefield RJ, Tan AL,D’Agostino MA, Toumi H, Hayashi

Benjamin & McGonagle

526

K, Emery P, Benjamin M. Distincttopography of erosion and newbone formation in achilles tendonenthesitis: implications for under-standing the link between inflam-mation and bone formation inspondylarthritis. Arthritis Rheum2008: 58: 2694–2699.

McNeilly CM, Banes AJ, Benjamin M,Ralphs JR. Tendon cells in vivo form athree dimensional network of cellprocesses linked by gap junctions. JAnat 1996: 189: 593–600.

Menz HB, Zammit GV, Landorf KB,Munteanu SE. Plantar calcaneal spursin older people: longitudinal traction orvertical compression? J Foot Ankle Res2008: 1: 7.

Milz S, Benjamin M, Putz R. Molecularparameters indicating adaptation tomechanical stress in fibrous connectivetissue. Advances Anat, Embryo CellBiol 2005: 178: 1–71.

Ralphs JR, Benjamin M, Waggett AD,Russell DC, Messner K, Gao J.Regional di!erences in cell shape andgap junction expression in rat Achillestendon: relation to fibrocartilage

di!erentiation. J Anat 1998: 193:215–222.

Rogers J, Shepstone L, Dieppe P. Boneformers: osteophyte and enthesophyteformation are positively associated.Ann Rheum Dis 1997: 56: 85–90.

Rufai A, Ralphs JR, Benjamin M.Structure and histopathology of theinsertional region of the humanAchilles tendon. J Orthop Res 1995: 13:585–593.

Schneider H. Zur Struktur derSehnenansatzzonen. Z AnatEntwicklung 1956: 119: 431–456.

Shaw HM, Benjamin M. Structure–function relationships of entheses inrelation to mechanical load andexercise. Scand J Med Sci Sports 2007:17: 303–315.

Thomopoulos S, Kim HM, RothermichSY, Biederstadt C, Das R, Galatz LM.Decreased muscle loading delaysmaturation of the tendon enthesisduring postnatal development. JOrthop Res 2007: 25: 1154–1163.

Tipton CM, Matthes RD, Martin RK.Influence of age and sex on the strengthof bone–ligament junctions in knee

joints of rats. J Bone Joint Surg 1978:60: 230–234.

van Wingerden JP, Vleeming A, BuyrukHM, Raissadat K. Stabilization of thesacroiliac joint in vivo: verification ofmuscular contribution to force closureof the pelvis. Eur Spine J 2004: 13: 199–205.

Vleeming A, Pool-Goudzwaard AL,Stoeckart R, van Wingerden JP,Snijders CJ. The posterior layer of thethoracolumbar fascia. Its function inload transfer from spine to legs. Spine1995: 20: 753–758.

Willard FH. The muscular, ligamentous,and neural structure of thelumbosacrum and its relationship tolow back pain. In: Vleeming A,Mooney V, Stoeckart R, eds.Movement, stability & lumbopelvicpain integration of research andtherapy. Edinburgh: ChurchillLivingstone Elsevier, 2007:5–45.

Zumwalt A. The e!ect of enduranceexercise on the morphology of muscleattachment sites. J Exp Biol 2006: 209:444–454.

Entheses

527