The role of biomechanical factors and HLA-B27 in magnetic resonance imaging-determined bone changes in plantar fascia enthesopathy


  • Professor Emery is an Arthritis Research Campaign Professor in Rheumatology.

  • Drs. McGonagle and Marzo-Ortega contributed equally to this work.



To study the role of biomechanical factors and HLA-B27 in plantar fasciitis.


T1-weighted and T2 spectral presaturation with inversion recovery (fat suppressed) magnetic resonance imaging (MRI) sequences of the plantar fascia insertion and adjacent bone were performed on 28 patients with plantar fasciitis; 17 had spondylarthropathy (SpA)-associated disease, and 11 had mechanically induced disease. The relationship between the degree of bone edema, scored on a semiquantitative scale (from absent to severe), and the patient's HLA-B27 status was determined.


On MRI, edema within the soft tissue at the enthesis was evident in both groups. Bone edema in the adjacent calcaneum was evident in 64.7% (11 of 17) of patients with SpA and in 45% (5 of 11) of those with mechanically induced disease (P = 0.441). HLA-B27 was identified in 9 (53%) of the patients with SpA but in none (0%) of those with mechanically induced disease. All 6 of the SpA patients with extensive bone edema but none of the 5 SpA patients with mild bone edema were HLA-B27 positive (P = 0.002).


The association of HLA-B27 with bone pathology in early enthesitis may have implications for a better understanding of the pathogenesis of SpA.

The enthesis is the point of insertion of ligament, tendon, or joint capsule to bone. Inflammatory involvement of the enthesis, or enthesitis, is a characteristic feature of the spondylarthropathies (SpA) and is regarded as the primary lesion, especially in ankylosing spondylitis (AS) (1–3). Magnetic resonance imaging (MRI) in patients with inflammatory enthesitis of SpA has demonstrated diffuse bone edema adjacent to the enthesis at the earliest stages of disease (4). The pathogenesis of this bone edema (histologically, an osteitis) (5) is unknown, although immunologic (6) and biomechanical (7) factors may be important. The best-characterized immunologic factor in SpA is the striking HLA-B27 association, but the relationship between HLA-B27 and enthesitis has not previously been reported.

Although enthesitis is regarded as the primary skeletal lesion in AS and SpA, some of the bone changes (e.g., sacroiliitis and spinal osseous pathology) are difficult to explain solely in terms of enthesopathy. However, many of the bone changes seen on MRI appear to be closely juxtaposed with insertion points or other sites rich in fibrocartilage. Our primary hypothesis for the present study was that biomechanical factors are important in the bone pathology that occurs adjacent to entheseal insertions and that the microtrauma induced at these sites leads to inflammation with subsequent repair. Based on this, we further hypothesized that in patients with SpA, where bone microtrauma adjacent to insertions is evident, the presence of the HLA-B27 gene determines the extent of these bone changes. We therefore used MRI to investigate the role of biomechanical factors in the bone pathology associated with enthesopathy at the plantar fascia, a common site of disease induced by mechanical and inflammatory factors, and we explored the relationship between the HLA-B27 gene and these bone changes.


Patients. Twenty-eight patients were recruited into the study; 17 had early SpA and 11 had early mechanically induced enthesopathy (mechanical group). All patients had clinical symptoms suggestive of plantar fasciitis, with pain and tenderness over that site. The patients in the inflammatory group fulfilled the European Spondylarthropathy Study Group (ESSG) criteria for SpA (8). The mechanical group was defined by a suggestive history involving walking or standing on the feet for long periods of time, the absence of pain at other sites, and the absence of a personal or family history suggestive of inflammatory arthropathy, including psoriasis, colitis or uveitis, low back pain, and lack of involvement of other joints in the 6 months following the study assessment. None of these patients fulfilled the ESSG criteria for SpA.

Five of the patients in the inflammatory group, but none in the mechanical group, had concomitant Achilles enthesitis. The demographic characteristics of both groups are summarized in Table 1.

Table 1. Demographics of the study patients*
Mechanical(n = 11)Inflammatory (SpA)(n = 17)
  • *

    The mechanical group represents patients with mechanically induced enthesopathy. The normal value for C-reactive protein is <5 mg/liter. SpA = spondylarthropathy; NA = not applicable.

Male, %2770
Age, mean (range) years43 (30–69)34 (19–53)
Duration of plantar fasciitis, mean (range) weeks32 (12–104)34.6 (12–104)
Duration of underlying diagnosis, mean (range) yearsNA2.32 (0.25–8.3)
Clinical sacroiliitis, %041
C-reactive protein, mean mg/liter7.420
HLAhyphen;B27 positive, %053

MRI. MRI was performed with Philips 0.5T and 1.5T Gyroscanners ACS NT (Philips Medical Systems, Best, The Netherlands), with image acquisition using a Philips volume head coil placed on the sole of the foot. Sagittal T1-weighted spin-echo pulse sequences and T2-weighted turbo spin-echo or fat-suppressed sequences were obtained on all patients.

For the sagittal T1-weighted spin-echo pulse sequences (1.5T scanner), the acquisition parameters were as follows: repetition time (TR) 382 msec, time to echo (TE) 14 msec, matrix 205 × 256 pixels, field of view (FOV) 160 mm, slice thickness 4.0 mm, slice gap 0.4 mm, number of signals averaged (NSA) 2, and acquisition time 2 minutes 38 seconds. For the T2-weighted turbo spin-echo or T2 fat-suppressed (FS) sequences (1.5T scanner), the acquisition parameters were as follows: TR 3,296 msec, TE 90 msec, echo train length 1/1 msec, matrix 203 × 256 pixels, FOV 180 mm, slice thickness 4.0 mm, slice gap 0.4 mm, NSA 4, and acquisition time 4 minutes.

Images were scored independently by 2 observers, a musculoskeletal radiologist (PO) and a rheumatologist (DM), who were blinded to the patients' details and clinical diagnoses. Bone edema, soft tissue edema, and calcaneal spurs were analyzed as follows. Bone edema, defined as regions of high or intermediate marrow signal on T2-weighted FS images, was assessed in the calcaneum, adjacent to the insertion of the plantar fascia. Bone edema was scored on a semiquantitative scale as absent, mild (only a small region of bone adjacent to insertion was abnormal), moderate (up to half of the adjacent calcaneum showed bone edema), or severe (more than half of the calcaneum affected). Soft tissue edema was defined as high signal intensity on T2-weighted FS images within the plantar fascia or the superficial and deep soft tissues surrounding the insertion of the plantar fascia. Its presence or absence was recorded in each case. Calcaneal spurs were identified on T1-weighted images, and their presence or absence was recorded.

Statistical analysis. All categorical data were analyzed using the chi-square test, with Fisher's exact test when appropriate. For assessment of reliability in the MRI scoring, kappa statistics were applied to calculate intraobserver and interobserver reproducibility. Kappa values express agreement beyond chance (9), taking into account the presence or absence of absolute agreement, and not the level of disagreement among scores. Kappa values can be interpreted as follows: 0.0–0.20 = poor; 0.20–0.40 = fair; 0.40–0.60 = moderate; 0.60–0.80 = good; and 0.80–1.00 = excellent.


The MRI scans of the heel of all 28 study patients were suitable for analysis. The clinical diagnoses in the SpA group were reactive arthritis (ReA) in 3 patients, psoriatic arthritis (PsA) in 4, AS in 3, and undifferentiated SpA (uSpA) in 7. The reliability of MRI scoring was good, as shown by the kappa values for intraobserver (plantar fascia edema scored by PO, κ = 1.000, P = 0.000; scored by DM, κ = 0.48, P = 0.038) and interobserver (plantar fascia bone edema κ = 0.908, P = 0.000) reproducibility.

Bone edema at the insertion of the plantar fascia. Bone edema in the calcaneum at the insertion of the plantar fascia was more common in the inflammatory group: 11 of 17 SpA patients (64.7%) had bone edema, compared with 5 of 11 patients (45%) in the mechanical group. The difference did not reach statistical significance (P = 0.441). Six of the 11 SpA patients with bone edema had moderate or severe edema (Figure 1A); the other 5 patients had mild edema. In addition, bone edema at the Achilles tendon insertion was seen in all 5 patients with clinical enthesitis at this site. In all but 1 patient, edema at the Achilles tendon insertion was clearly distinguished from bone edema at the insertion of the plantar fascia. In the latter case, the abnormalities were so extensive that they merged. In the mechanical group, all 5 patients with bone edema had mild bone edema (Figure 1B).

Figure 1.

Magnetic resonance images of the heel using T2 fat-suppressed sagittal sequences in 2 patients with acute plantar fasciitis. A, HLA-B27-positive patient with spondylarthropathy (SpA), showing severe bone edema in the calcaneum (∗). There is also soft tissue edema in the superficial (white arrow) and deep (black arrow) soft tissues surrounding the insertion of the plantar fascia. B, HLA-B27-negative patient with mechanically induced plantar fasciitis, showing soft tissue edema in the superficial (white arrow) and deep (black arrow) soft tissues surrounding the insertion of the plantar fascia, but less extensive than that in the SpA patient. The extent of bone edema in these patients was comparable to that seen in HLA-B27-negative patients with SpA.

Effect of HLA-B27 on the severity of plantar fascia bone edema. Nine of the 17 patients in the SpA group (53%), but none of the 11 patients in the mechanical group were HLA-B27 positive. There was no correlation between B27 positivity and the presence or absence of bone edema (P = 0.245). However, in the 11 patients with SpA who had bone edema, there was a strong correlation between its severity and B27 positivity (Table 2). All 6 SpA patients with moderate or severe edema were B27 positive, and none of the 5 SpA patients with mild edema were B27 positive (P = 0.002). Four of the 5 patients with Achilles enthesitis were B27 positive (3 with AS and 1 with uSpA) and had moderate or extensive bone edema at the plantar fascia insertion. One B27-negative SpA patient (with ReA) had mild edema at that site.

Table 2. Clinical and MRI findings in the 17 patients with spondylarthropathies*
PatientDiagnosisHLA-B27ClinicalsacroiliitisRadiologicsacroiliitisAchillestendinitisMRI edema
Achilles tendonPlantar fascia
  • *

    None of the patients in the group with mechanically induced enthesopathy had sacroiliitis, and all of them were HLA–B27 negative. MRI = magnetic resonance imaging; uSpA = undifferentiated spondylarthropathy; PsA = psoriatic arthritis; ReA = reactive arthritis; NA = not available; AS = ankylosing spondylitis.


Soft tissue edema at the insertion of the plantar fascia. Soft tissue edema was evident at the plantar fascia insertion in 16 of the 17 patients in the SpA group and in all 11 patients in the mechanical group. As previously noted, the edema was more pronounced in the tissues immediately adjacent to the enthesis rather than within the enthesis itself (Figures 1A and B). Because changes within the soft tissues were less extensive than those in bone, it was not possible to ascertain differences in intensity between the groups. Soft tissue changes were also evident at the Achilles tendon insertion in all patients with Achilles enthesitis.

Calcaneal spurs. Calcaneal spurs were seen in 29% (n = 5) of the patients in the SpA group and in 27% (n = 3) of the patients in the mechanical group.


The present study used MRI to investigate the role of biomechanical factors and HLA-B27 in the osseous changes associated with enthesopathy. It showed that mechanically induced and inflammatory enthesitis have similar MRI appearances, with soft tissue and bone edema at the plantar fascia insertion. However, the degree of bone edema in SpA significantly correlated with the presence of HLA-B27, suggesting that the effect of this gene may be mediated within the bone adjacent to entheseal insertions. These findings have implications for a better understanding of enthesitis in human SpA.

Further work is needed to address some issues raised in this study. First, AS is largely a disease of the spine and not the heel. The relationship between B27 and bone changes in the spine needs to be addressed. However, this is not possible in the British population because most patients with SpA and spinal disease are B27 positive. One of the reasons we assessed the heel is because it is a commonly affected site in both strongly (AS, uSpA, ReA) and weakly (PsA) B27-associated SpA.

Second, the diffuse bone pathology that is seen could be associated with a particular type of SpA, for example, AS. However, extensive edema was also seen in other subtypes, such as uSpA. At 1-year followup, only 1 of these patients fulfilled criteria for AS. Although the number of study patients was small, we believe that our findings suggest that B27 per se, and not the underlying SpA, may determine the severity of bone changes.

Last, the study group comprised patients with early disease, and the B27 and non-B27 SpA groups had comparable durations of plantar fasciitis (a mean of 37 weeks and 30 weeks, respectively). These findings are therefore unlikely to reflect the effect of disease chronicity. Further studies are needed to determine whether bone changes persist and to determine the relationship of HLA-B27 to persistent bone edema.

Five patients with plantar fasciitis in the SpA group, but none in the mechanical group, also had Achilles enthesitis. This was apparent at the time of scoring the MRI scans and could have influenced the results. However, regions of bone edema at the plantar fascia were usually discrete from those at the Achilles enthesis. Furthermore, severe edema at the plantar fascia also occurred without concomitant Achilles enthesitis. Extensive bone edema at this insertion was evident only in HLA-B27-positive subjects. This suggests that the effect of HLA-B27 on bone edema is not confined to the plantar fascia, but is generalized.

The degree of abnormality in the soft tissue at insertion points was modest in comparison to the entheseal bone changes and was comparable in both mechanically induced and inflammatory disease. This lack of specificity may reflect the anatomy of the tendinous part of the entheses, where fibroblasts are tightly crosslinked with little scope for accumulation of water. Hence, the MRI changes within the fibrous part of the entheses are less marked in comparison to changes in the bone marrow. In the absence of severe bone changes, MRI could not distinguish between inflammatory and mechanically induced disease in the present study.

HLA-B27 is the best-characterized genetic factor in SpA and probably plays a major role in disease pathogenesis. The relationship between B27 and osteitis (e.g., sacroiliitis, which is characteristic of SpA) has not previously been investigated. Although in the present study, there was no association between the presence or absence of bone edema per se and B27 (which could be due to the small number of study subjects), there was a strong correlation between the extent of bone pathology and B27.

We propose the following model to explain these observations. First, inflammatory enthesopathy is associated with a degree of trauma at the soft tissues at the point of insertion. This is likely, since the normal enthesis is a site of increased biomechanical stress and even clinically asymptomatic large entheseal insertions are prone to microtrauma with evidence of tissue damage at these sites (10). In some but not all cases, this may be associated with a degree of microtrauma in the underlying bone, in both a purely mechanically induced and an inflammatory enthesopathy. When this occurs in mechanically induced plantar fasciitis or in non-B27-related SpA, then the degree of bone pathology may be limited. However, in B27-positive subjects with SpA, these bone changes trigger autoimmunity to a native protein. Whether this is directed toward the fibrocartilage adjacent to the bone, a bone protein, or both remains to be determined.

The presence of HLA-B27 at sites of bone inflammation in SpA could have 3 consequences: more severe, more extensive, and more chronic inflammation. Such a scenario could help explain both the role of B27 in disease, as well as the osseous changes evident in SpA. Finally, it is possible that inflammation at entheseal insertions and adjacent bone in SpA is influenced by interactions between microtrauma and bacteria, or their constituent adjuvant, at these sites (11). Therefore, where there is minimal bone microtrauma in the presence of microbes or their constituent adjuvants, a combination of these factors could contribute to immune activation.

In conclusion, this study investigated the enthesitis lesion and showed that it is associated with osseous pathology, whether related to inflammation or to biomechanical factors, and that the extent of these changes strongly correlates with the presence of HLA-B27 in SpA. These findings may be important for determining the factors that contribute to enthesitis and for elucidating the role of HLA-B27 in disease.


We thank Dorothy Faheem and Jane Cullingworth for their help with scanning and Dr. Richard J. Wakefield and Paul Astin for statistical analysis.