Distinct topography of erosion and new bone formation in achilles tendon enthesitis: Implications for understanding the link between inflammation and bone formation in spondylarthritis




This study combined ultrasonography of the Achilles tendon enthesis at different stages of spondylarthritis (SpA) with microanatomic studies of normal cadaveric entheses, with the aim of exploring the relationship between bone erosion and new bone formation in enthesitis.


Thirty-seven patients with SpA and Achilles tendon enthesitis (20 with early SpA and 17 with chronic SpA) and 10 normal control subjects underwent ultrasound scanning. The presence of bone erosion and spur formation was recorded at 3 sites: the proximal and distal halves of the enthesis and the adjacent calcaneal superior tuberosity. Parallel histologic analysis was performed on cadaveric Achilles tendon entheses to determine whether regional variations in bone density and trabecular architecture in relation to fibrocartilage distribution are related to disease patterns.


Bone erosion in patients with early SpA occurred at either the proximal insertion or the superior tuberosity (11 of 20 patients; P < 0.001 versus distal enthesis). Very small spurs, which were present almost exclusively at the distal enthesis, were evident in patients with early SpA and in normal control subjects. However, large spurs were evident distally only in patients with chronic SpA (9 of 17 patients, compared with none of 20 patients with early SpA; P < 0.0001). Histologic studies showed that aged normal individuals had small spurs at the corresponding location. The bone-to-marrow ratio was also significantly lower in the regions prone to erosions (P < 0.05).


Bone erosion in association with Achilles tendon enthesitis in SpA is anatomically uncoupled from bone formation—the 2 processes are topographically and temporally distinct. We thus conclude that disease patterns in SpA are related to normal enthesis structure and biomechanics.

The spondylarthritides (SpA) are a heterogeneous group of diseases characterized by inflammation at the enthesis, the adjacent bone, and the synovium. Insertional inflammation or enthesitis is a characteristic feature of SpA, but the relationship between enthesitis, new bone formation, and bone erosion remains poorly understood. In rheumatoid arthritis, inflammation is associated with both focal bone erosion and generalized osteoporosis, with little new bone formation. In SpA, the general trend of inflammation leading to bone loss is less obvious, because inflammation is also associated with joint fusion or ankylosis (1).

With the advent of anti–tumor necrosis factor (anti-TNF) therapy, there is a critical need to understand the relationship between inflammation, erosion, and new bone formation. Despite the use of anti-TNF therapy in SpA, patients continue to experience radiographic progression (2, 3); this observation was mirrored in experimental studies showing continuing ankylosis even after the inhibition of TNF (4). However, studies have shown that anti-TNF treatment prevents erosive disease in psoriatic arthritis patients with hand involvement (5).

It has been suggested that a better appreciation of enthesitis in SpA will come from understanding Achilles tendon disease (6). We previously described how this insertion is part of an enthesis organ, and thus how fibrocartilage is present not just at the insertion site itself, but also next to the bone–tendon junction (7). Within the enthesis itself, fibrocartilage is more prominent proximally, reflecting higher compressive loading at this site. In the present study, we hypothesized that the anatomy and biomechanics of the normal enthesis could directly influence bone erosion and bony spur formation. This hypothesis was explored with ultrasound scanning and microanatomic studies.



Twenty patients with early SpA (mean duration 11 months [range 2–24 months]), as defined by the European Spondyloarthropathy Study Group (8), and posterior ankle pain relating to the Achilles tendon region were recruited. The mean age of the patients was 43 years (range 26–64 years); 12 patients were male, and 8 patients were female. One patient had ankylosing spondylitis (AS), 13 had psoriatic arthritis, and 6 had reactive arthritis. Approval to scan patients was obtained from the local ethics committees (in both the UK and France), and all patients gave informed consent.

In order to investigate erosion and spur formation in chronic SpA, 17 patients with SpA-related Achilles tendon enthesitis were also scanned. The mean age of patients in this group was 52.8 years (range 27–70 years); 9 patients were male, and 8 patients were female. Thirteen patients in this group had AS, 3 had psoriatic arthritis, and 1 had reactive arthritis; the mean disease duration was 272.6 months (range 38–612 months). The Achilles tendons of 10 asymptomatic control subjects were also scanned. The mean age of the control subjects was 44 years (range 25–66 years); 6 of these individuals were male, and 4 were female.

Ultrasound scanning protocol.

Ultrasound scanning was performed independently by 2 experienced ultrasonographers (RJW and MAD) at 2 hospitals, using different machines (the Philips HDI 5000 system [Philips, Best, The Netherlands] and an Esaote Technos scanner [Esoate, Genoa, Italy) employing linear multifrequency transducers (15-8 MHz and 14-10 MHz, respectively). Subjects were asked to lie down in the prone position on a couch, with the foot at a 90-degree angle to the leg. Images were obtained in the sagittal and transverse planes, to confirm the presence of a cortical erosion, as previously described (9, 10), or to confirm the presence of a bony spur. At the time of ultrasound scanning, the ultrasonographers were not aware of the hypothesis being tested but simply documented and stored digital images of any abnormalities that were noted, as previously described (11).

Methods of scoring erosions and spurs on ultrasound scans.

Erosions were defined according to the Outcome Measures in Rheumatology Clinical Trials definition of a bone erosion (12). This is a cortical break visible in at least 2 perpendicular planes. Bone spurs were defined as cortical protrusions seen in at least 2 perpendicular planes. The presence or absence of erosions and spurs was recorded in 3 separate regions within the enthesis organ complex (Figure 1), as follows: 1) the fibrous distal half of the insertion, which is predominantly subject to tensile forces, 2) the fibrocartilaginous proximal half of the insertion, a site that is subject to both tension and compression, and 3) the fibrocartilage-covered superior tuberosity. The sizes of the spurs were documented using a 0–3-point scale, where 1 = minimal, 2 = moderate, and 3 = large.

Figure 1.

a, Longitudinal ultrasound image through the posterior heel of a normal individual. Inset, Position of the ultrasound probe. The heel was divided into 3 zones (the superior tuberosity [ST], the proximal part of the enthesis [PE], and the distal part of the enthesis [DE]) and scored for spur formation or bone erosion. In this subject, a spur was noted in the distal part of the enthesis only (solid arrows), but there was no evidence of erosion. The tendon (AT) appeared normal along its length. The numbers on the vertical axis represent distance in centimeters. b, Longitudinal ultrasound image through the posterior heel of a patient with psoriatic arthritis. There is mild thickening and hypoechogenicity of the tendon (within the region indicated by thin arrows). Erosions are evident in the proximal tendon and in the region of the superior tuberosity (open arrows), with a small bone spur (thick arrow) present in the distal enthesis.

The presence of hypoechoic thickening or loss of normal fibrillar architecture at the entheses was also scored. Power Doppler ultrasound was not included as part of the study. All ultrasound images were read in a blinded manner by an experienced ultrasonographer (RJW) and an experienced rheumatologist (DM). The scoring was performed by consensus between the 2 readers.


The Achilles tendon enthesis was removed from 10 male dissecting room cadavers (ages 62–91 years) donated to Cardiff University, in accordance with the 1984 Anatomy Act and the 1961 Human Tissues Act. The cadavers were devoid of gross musculoskeletal defects, but no medical history was available. All material was decalcified in 5% nitric acid, dehydrated in graded alcohols, cleared in xylene, and embedded in paraffin wax. The blocks were sagittally sectioned at 8 μm, and sections were collected at 1-mm intervals. At each 1-mm sample point, 12 slides were stained with Masson's trichrome, toluidine blue, or Hall-Blunt quadruple stain. In the central, medial, and lateral thirds of the enthesis, digital photomicrographs of the bone were obtained at each 1-mm sample point in 3 regions: the superior tuberosity and the proximal and distal parts of the enthesis itself (Figure 2a). In each region, the bone-to-marrow volume ratio (i.e., the proportion of bone tissue to marrow tissue within the calcaneal regions sampled) was calculated using algorithms developed in our laboratory with MatLab software, version 7.1.246, release 14 (The MathWorks Ltd. UK, Cambridge, UK). A higher ratio indicated a greater quantity of bone at a site.

Figure 2.

Histology of the Achilles tendon enthesis organ in a dissecting room cadaver. a, Low-power view of a sagittal section, showing the 3 regions in which the bone-to-marrow ratio of the enthesis organ was analyzed (the superior tuberosity of the calcaneus [ST], the proximal part of the enthesis [PE], and the distal part of the enthesis [DE]). Note that the superior tuberosity is covered by thick periosteal fibrocartilage (PF). Also note that the distal enthesis has a greater trabecular density, which was reflected in the greater bone-to-marrow ratio. Bar = 3 mm. b, Small bony spur (arrow) that developed in the distal (more fibrous) part of the enthesis. Note the thick wedge of uncalcified fibrocartilage (FC) in the proximal part of the enthesis. Bar = 4 mm. c, Evidence of erosion (arrow) of the periosteal fibrocartilage at the base of the superior tuberosity. Bar = 1 mm. All sections were stained with Masson's trichrome. RB = retrocalcaneal bursa; SF = sesamoid fibrocartilage.

Statistical analysis.

McNemar's test was used to compare the proportion of abnormalities in the 3 different regions of the enthesis examined by ultrasound scanning. A Bonferroni correction was performed, and P values less than 0.017 were considered significant. Fisher's exact test was used to compare the proportion of abnormalities between the groups, and P values less than 0.05 were considered significant. All ultrasound data analyses were performed with SPSS version 12.0 software (SPSS, Chicago, IL). For the histology data, means and SDs were calculated from the measurements of the bone-to-marrow ratio in the 10 specimens. The differences between variables were analyzed using a generalized linear model repeated-measures analysis of variance. When significant differences were observed, Fisher's post hoc tests were used. In each case, P values less than 0.05 were considered significant. All statistical analyses for the histology data were performed with MatLab software (version 7.1.246, release 14).


Ultrasound scanning results.

At the time of ultrasound scanning, 16 of the 20 patients with early SpA had diffuse hypoechoic thickening of the Achilles tendon enthesis that was suggestive of acute enthesitis. Erosions were seen only in the 2 more fibrocartilaginous areas—either the proximal part of the insertion site proper (11 of 20 cases; P < 0.001 versus the distal enthesis insertion) or the superior tuberosity (the same 11 cases; P < 0.001 versus the distal enthesis), and none were observed in the distal part of the enthesis. Erosions in the 2 fibrocartilaginous areas frequently coalesced and could not be distinguished satisfactorily, and were thus scored as single erosions, the site of which was scored according to where the erosion was largest. No erosions were observed at the border of the distal and middle regions. No erosions were evident in any of the normal subjects.

Bony spurs were seen exclusively in the distal half of the Achilles tendon and were present in 8 of 20 patients with early SpA (P < 0.0001 versus the proximal regions). Six of the 8 spurs were small, and 2 were medium-sized (Figure 1). Similarly, 8 of the 10 normal subjects had spurs at the distal part of the insertions; 4 of the spurs were small, and 4 were medium-sized. The same pattern was noted in patients with chronic SpA, in whom spurs were noted mainly distally. Among these 17 patients, 15 [88.2%] had spurs in the distal enthesis, and 2 [11.8%] had spurs in the proximal enthesis. However, large spurs were evident in 9 of the 17 patients with chronic SpA, compared with none of the 20 patients with early SpA (P < 0.0001). Also, 7 of the 17 patients with chronic SpA had small spurs, and 1 of these patients had a medium spur. Erosion formation was noted in 7 patients (at the proximal insertion in 3 patients and at the superior tuberosity in 4 patients).

Histologic findings.

The paucity of enthesis fibrocartilage was confirmed in the distal part of the enthesis, but the tissue was more conspicuous proximally (Figure 2). The periosteal fibrocartilage covering the superior tuberosity is also evident in Figure 2a. A greater density of bone trabeculae was evident in the distal region of the Achilles tendon, and the trabeculae were orientated along the line of pull of the tendon. Conversely, fewer trabeculae were in the proximal part of the enthesis, and their orientation was less regular. The variations in bone density were reflected by a marked difference in the bone-to-marrow ratio in the 3 enthesis organ regions, with significantly more bone tissue compared with marrow tissue in the distal part of the enthesis (mean ± SD 0.78 ± 0.04) compared with either the proximal part (mean ± SD 0.51 ± 0.04; P < 0.05) or the superior tuberosity (mean ± SD 0.38 ± 0.07; P < 0.05). Furthermore, the mean value for the proximal part of the enthesis was significantly different from that for the superior tuberosity (P < 0.05).

Microscopic focal erosions of periosteal fibrocartilage and/or the underlying bone were seen on the superior tuberosity in 8 of the 10 specimens (Figure 2c). Bony spurs (some of which were small and may not have been clinically recognizable) were present in 7 of the 10 specimens, predominantly in the fibrous (distal) part of the enthesis (Figure 2b), and are likely to be a normal variant that can be seen on high-resolution ultrasound imaging. In 1 specimen, a small erosion was present in the distal part of the attachment.


In this study, we investigated bone erosion in early SpA, using the Achilles tendon enthesis organ as a model. The large size of this enthesis organ (which can span 3 cm or more from proximal to distal) and its superficial location beneath the skin make it a most suitable insertion for ultrasound study. An important consideration underpinning the value of the present study is that anti-TNF therapy in SpA can prevent erosion formation (5), but the effect on new bone formation is less well understood (3, 4). In the present study, we observed that erosions typically occurred in the proximal portion of the enthesis in patients with SpA. We also noted that fibrocartilage was abundant at erosion-prone sites in dissecting room cadavers, and that the quantity of bone tissue (as suggested by the bone-to-marrow ratio) was lower at these sites. Although the histologic findings were from elderly subjects with unknown medical histories, they nevertheless support the idea that enthesis microanatomy plays a major role in the phenotypic expression of what is considered to be primarily an immunologically mediated disease.

We also looked at spur formation in SpA and noted no evidence of pathologic spur formation in patients with early disease. Small spurs were present distally in normal individuals and in dissecting room cadavers and were also at the same location in patients with SpA. Given that this is a site of a higher entheseal bone-to-marrow ratio (i.e., a denser network of cancellous bone spicules), we propose that small spurs are a normal anatomic variation that are evident on high-resolution imaging. Because large spurs were seen only in patients with chronic SpA, the present findings raise the possibility that spur formation is not only anatomically uncoupled from erosion but also is temporally uncoupled, because both processes may be occurring at different times in the disease process. This suggests that spurs might develop quite slowly and mature once inflammation abates. This latter theory needs to be ascertained in a longitudinal manner but could be very important for understanding the long-term consequences of suppressing entheseal inflammation with the anti-TNF class of drugs.

Because the majority of patients (13 of 17) in the group with chronic SpA had AS, compared with only 1 patient in the group with early SpA, it is also plausible that the larger spurs observed in patients with chronic SpA could be attributed to the disease type. However, it was interesting to note that although all of the patients in the group with chronic SpA (including 4 non-AS patients) had spurs, the 1 patient with AS in the group with early SpA did not have any spurs. It is also likely that anatomic and biomechanical factors underlie spur formation or erosion that may be seen in a similar distribution in degenerative enthesopathies such as supraspinatus tendinitis.

Mechanistically, fibrocartilage on a bone surface is a functional adaptation to withstanding compression and/or shear: if bone is directly subjected to compression, it is resorbed, and cartilage formation is enhanced (13, 14). Similarly, bony spurs were typical of fibrous regions of the enthesis; where tensile loading is likely to be greater, bone was more abundant, and the trabeculae were aligned along the direction of tendon pull in accordance with Wolff's law. It seems reasonable to suggest that the intermittent compression of the tendon against the calcaneal tuberosity during foot movements might inhibit spur formation in much the same way as movement discourages fracture healing (15) and aortic pulsatility in Forestier disease prevents bone bridging on the left side of the spine (16). Thus, spur formation may be typical of the distal rather than the proximal part of the enthesis organ because of a lesser degree of inhibition at such sites. Such anatomic factors within the enthesis organ that relate to the normal insertional structure and biomechanics may ultimately dictate the expression of disease, rather than immune factors per se.

A limitation of the current study is that the histologic studies could be performed only on dissecting room cadaveric material obtained from elderly individuals, whereas the ultrasound studies were conducted on much younger persons. However, although the absolute values for the bone-to-marrow ratio in the 3 regions of the enthesis examined may vary with age, there is no reason to expect any differential effect of age on the proximal or distal enthesis or the superior tuberosity.

We conclude that the phenotypic expression of SpA at different parts of the Achilles tendon enthesis is driven by inflammation, regional microanatomy, and biomechanics, which determine the phenotypic expression of disease. Erosions develop in regions of compression, but spur formation occurs in regions where tensile forces are likely to be higher.


Dr. McGonagle had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. McGonagle, D'Agostino, Emery, Benjamin.

Acquisition of data. McGonagle, Wakefield, Tan, D'Agostino, Hayashi, Emery, Benjamin.

Analysis and interpretation of data. McGonagle, Wakefield, Tan, D'Agostino, Toumi, Hayashi, Benjamin.

Manuscript preparation. McGonagle, Wakefield, Tan, D'Agostino, Hayashi, Emery, Benjamin.

Statistical analysis. Tan, Toumi.


We thank Dr. Elizabeth Hensor for advice regarding statistics.