- Email: [email protected]
Aseptic discitis in patients with ankylosing spondylitis: a retrospective study of 14 cases
Joint Bone Spine 72 (2005) 248–253 http://france.elsevier.com/direct/BONSOI/
Original article
Aseptic discitis in patients with ankylosing spondylitis: a retrospective study of 14 cases Sandrine Langlois, Jean Pierre Cedoz, Anne Lohse, Eric Toussirot, Daniel Wendling * Service de Rhumatologie, Hôpital Jean Minjoz, CHU, Boulevard Fleming, 25030 Besançon cedex, France Received 8 December 2003; accepted 25 May 2004 Available online 06 October 2004
Abstract Objectives. – To determine whether patients with ankylosing spondylitis (AS) and discitis exhibit specific characteristics as compared to patients who have AS without discitis. Methods. – We retrospectively collected clinical, laboratory, and imaging data from the charts of 79 patients with AS admitted to a rheumatology department. Results. – Of the 79 patients, 14 (18%) had discitis that was not due to infection or trauma. Mean age at the diagnosis of AS was 40 years and mean duration of AS at admission was 10 years, with no difference between the two groups. In three patients, discitis was the inaugural manifestation of AS. Two patients had discitis at two levels. The lumbar and thoracolumbar segments were the most common sites of discitis. Symptoms of discitis were present in 10 of the 14 patients. Stage III sacroiliitis was significantly more common in the patients with discitis. None of the patients experienced neurological compromise. Conclusion. – The frequency of aseptic discitis in patients with AS is probably overestimated as a result of inclusion and exclusion biases. Discitis usually occurs at an advanced stage of AS under the form of an erosive condition related to both mechanical factors and osteoporosis. Inaugural or early discitis can occur, however, as a result of the inflammatory process. © 2004 Elsevier SAS. All rights reserved. Keywords: Ankylosing spondylitis; Discitis
1. Introduction Aseptic discitis is an uncommon manifestation of ankylosing spondylitis (AS) first described by Andersson in 1937. Destructive lesions of the disk and adjacent vertebral bone occur in the absence of infection or trauma. This pattern is characteristic of AS [1]. Studies conducted using modern imaging techniques and pathological examination of the entheses have rekindled interest in aseptic discitis as a manifestation of AS. Nevertheless, few data are available on aseptic discitis in AS. The studies reported by Marie-Bonnin [2] in 1986 and Bouvier et al. [3] in 1987 (43 patients) are among the most recent and largest case-series. Discitis is defined primarily on the basis of imaging features as a variable combi* Corresponding author. Tel.: +33-3-81-66-82-41; fax: +33-3-81-66-86-86. E-mail address: [email protected] (D. Wendling). 1297-319X/$ - see front matter © 2004 Elsevier SAS. All rights reserved. doi:10.1016/j.jbspin.2004.05.015
nation of disk-space narrowing, erosions in the adjacent endplates, and peripheral sclerosis. The objectives of the present study in a population with AS were to look for specific characteristics of patients with discitis as compared to those without discitis, to compare our findings with results in the literature, and to discuss available evidence on the pathophysiology of discitis associated with AS.
2. Patients and methods We reviewed the medical records of all the patients with AS admitted to the rheumatology department of the Besançon Teaching Hospital, Besançon, France, over a 6-year period. We selected the patients who met the modified New York criteria for AS [4]. With this criteria set, the diagnosis is definite in patients who have radiological evidence of grade II to IV sacroiliitis on both sides, or grade III or IV sacroiliitis on one side with at least one of the three following clinical
S. Langlois et al. / Joint Bone Spine 72 (2005) 248–253
249
Table 1 Numerical clinical data on the 14 cases with discitis and the 65 controls without discitis Schöber’s index, cm Finger-to-floor distance, cm Occiput-to-wall distance, cm Chest expansion BASFI BASDAI
Cases 1.25 37 10.3 2.3 87.5 69
Controls 2.60 26.6 4 .9 3.4 57.1 55.8
P value 0.058 0.32 0.10 0.37 0.16 0.29
BASDAI: Bath Ankylosing Spondylitis Disease Activity Index. BASFI: Bath Ankylosing Spondylitis Functional Index.
criteria: low back pain and stiffness for longer than 3 months with the pain abating upon physical activity but persisting at rest, motion range limitation of the lumbar spine in the coronal and sagittal planes, and reduced chest expansion as compared to normal values adjusted for age and sex. A diagnosis of probable AS is given in patients with all three clinical criteria but without the radiological criterion, or vice versa; in this last situation, other possible causes of sacroiliitis should be considered. In this study, the diagnosis of discitis was based on a review of the imaging studies in the medical records. Full-length spinal radiographs to look for multifocal discitis were not available for all the patients. We defined cases as patients with discitis and controls as patients without discitis. We compared clinical, laboratory, and imaging study features in the cases and controls. The following clinical variables were recorded: sex and age; conventional criteria for evaluating AS (peripheral arthritis, ocular involvement, enthesopathy, Schöber index, finger-tofloor distance, chest expansion, occiput-to-wall distance, Bath Ankylosing Spondylitis Disease Activity Index [BASDAI], and Bath Ankylosing Spondylitis Functional Index [BASFI]), comorbidities, and whether the patient was a manual laborer. Laboratory variables abstracted from the records were as follows: tests for inflammation at AS diagnosis and at discitis diagnosis, whether the HLA B27 antigen was present, and whether the serum IgA level was ≥3 g/l. Standard radiographs of the lumbar spine and pelvis were reviewed to determine the presence and stage of sacroiliitis and to look for syndesmophytes. We recorded bone mineral density data when available. In the cases, we recorded the following data: symptoms of discitis and findings from the investigations used to establish the diagnosis of discitis (histology, computed tomography [CT], radionuclide bone scanning, and magnetic resonance imaging [MRI]). Finally, we used the classification scheme developed by Bouvier et al. [3] to classify the patients based on the imaging features of discitis. 3. Results Of the 79 AS patients with valuable medical records who were admitted during the study period, 14 had radiological evidence of discitis. 3.1. Clinical data Mean duration of AS was 10 years in both groups. In three patients, discitis was the inaugural manifestation of AS.
Mean age at the diagnosis of AS was 33 years in the cases and 30 years in the controls (P = 0.30). Mean age at the diagnosis of discitis in the cases was 40 years. We found no significant differences between the cases and the controls for the proportion of patients with a family history of AS (33% vs. 17%, P = 0.21), manual labor (64% vs. 54%), ocular involvement, peripheral arthritis, enthesopathy, or numerical indices, although Schöber’s index was lower in the cases (1.25 vs. 2.60 cm) with a P value of borderline significance (0.058) (Tables 1 and 2). Comorbidities in the controls were as follows: psoriasis (1), eosinophilic gastrointestinal disorder akin to chronic inflammatory bowel disease (1), and Reiter syndrome (1). Among the cases, two patients had psoriasis. Symptoms of discitis were present in 12 of the 14 cases, with wide variations in clinical patterns. An injury with direct impact on the spine 4 years earlier was recorded as a possible precipitating factor in one of the cases. Of the 14 cases, 12 had discitis at a single level and two at two levels. The clinical outcome was consistently favorable, without neurological compromise or need for surgery. Follow-up radiographs showed development of vertebral synostosis over a 6-year period in one patient. In patient no. 11, intravenous bisphosphonate therapy was followed by increased bone sclerosis with no change in the erosions. The patient with cervical discitis achieved a full recovery after intravenous bisphosphonate therapy; an MRI scan done after 7 years showed no evidence of discitis. In patients no. 1 and no. 4, the disk space narrowing worsened over the years. 3.2. Laboratory tests We found no significant differences between the cases and controls regarding presence of HLA B27 (83% of cases and 93% of controls), serum IgA elevation, or severity of ASrelated inflammation. Parameters for inflammation were similar at the diagnosis of discitis and at other times during follow-up (Table 3). 3.3. Imaging study findings The cases were more likely to have spinal syndesmophytes, although the difference with the controls was not statistically significant. Stage III sacroiliitis was significantly more common among the cases than among the controls (57% vs. 29%; P = 0.045) (Table 4). Imaging study patterns of discitis were distributed as follows: pseudotuberculous,
250
S. Langlois et al. / Joint Bone Spine 72 (2005) 248–253
Table 2 Clinical features in the 14 cases with ankylosing spondylitis and discitis Spinal segment
Level
Case
Sex Age at diagnosis of discitis, years
Cervical Thoracic
C3-C4 T8-T9 T9-T10 T10-T11 T10-T11
8 12 7 2 6
M M M M F
T11-T12 3 T12-L1 5 T12-L1 9
T12-L1 L1-L2 L3-L4 L3-L4
L3-L4 L4-L5 L4-L5 L5-S1
Thoracolumbar
Lumbar
Lumbosacral
Peripheral Ocular arthritis involvement
Enthesopathy
Manual Fracture Trauma occupation
Clinical manifestations
18 42 22 45 33
Time since the diagnosis of AS (years) 0 3 5 0 11
Yes Yes No No Yes
No No Yes No Yes
Yes No Yes Yes No
Yes Yes Yes Yes No
No No No No No
No No No No No
M M M
51 46 42
27 15 14
No No Yes
Yes No No
No No Yes
Yes Yes No
No No Yes
1
M
30
14
No
No
No
No
No
No No Direct impact 4 years earlier No
Asymptomatic Inflammatory pain Inflammatory pain Mechanical pain Thoracolumbar pain stiffness++ Inflammatory pain Pain (type not specified) Thoracic kyphosis
10 11
M M
44 43
25 15
No No
No No
Yes No
Yes Yes
No No
No No
13 14 4
M M F
42 32 60
5 Yes 9
No No Yes
Yes No No
No No No
Yes No No
No No No
No No No
Table 3 Laboratory test findings at the diagnosis of discitis in the 14 cases Case 1 2 3 4 5 6 7 8 9 10 11 12 13 14
HLA B27 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No Yes Yes
IgA elevation Yes No No No No No Yes Yes No Yes Yes No No No
ESR (mm/h) NA 44 NA 80 NA 40 96 68 NA 24 34 NA 0 40
CRP (mg/l) NA 37 NA NA NA NA 60 32 NA 28 6 NA 0 84
NA: not available; ESR: erythrocyte sedimentation rate; CRP: C-reactive protein.
n = 1; pseudarthrosis-like, n = 2; late erosive, n = 2; early erosive, n = 3; isolated disk space narrowing, n = 3; and unclassifiable, n = 3 (Figs. 1–5). MRI was performed in six patients, of whom four had no typical signs of inflammation (low signal from the disk and adjacent endplates on T1weighted images, high signal on T2-weighted images, and enhancement after gadolinium injection). 3.4. Absorptiometry findings Overall, T-scores indicated osteopenia in the cases as compared to the controls, both at the lumbar spine (mean
Pain (type not specified) Inflammatory pain Inflammatory pain stiffness malalignment Inflammatory pain Inflammatory pain Mechanical low back pain and L5 sciatica
Table 4 Radiological and densitometry data in the 14 cases with discitis Case
Sacroiliitis
Syndesmophytes
T-score
Yes Yes Yes Yes Yes No Yes Yes Yes Yes Yes No Yes NA
NA NA −0.2 NA −4 NA NA −1.3 −4.6 1.7 2.2 NA 0.7 −0.7
II/III 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
NA: not available.
T-score, −1.2 vs. −0.6) and at the femoral neck (mean T-score, −1.1 vs. −0.6). Two of the cases had severe osteoporosis (T-score<−4) (Table 4). 3.5. Histological findings Biopsies were obtained in three patients. In two of these patients, only disk material was collected, 2 and 6 months after symptom onset, respectively; the findings were normal. In the remaining patient, a biopsy specimen was obtained 2 months after symptom onset and contained both disk and vertebral bone; histology showed tissue repair. No histologi-
S. Langlois et al. / Joint Bone Spine 72 (2005) 248–253
251
Fig. 1. Computed tomography of the lumbar spine with reconstruction: features mimicking tuberculosis. Fig. 3. C3–C4 discitis seen as isolated disc space narrowing without sclerosis. Loss of the normal lumbar lordotic curvature.
Fig. 2. Lateral view of T12-L1: extensive discovertebral destruction predominating anteriorly. Note the neural arch fracture at the level of the discitis (case no 9).
Fig. 4. Anteroposterior view of L3–L4. Late erosive form. Irregular erosions in the adjacent endplates and sclerosis. Lateral syndesmophytes (case no 10).
cal evidence of inflammation was found in any of the three biopsies.
The male predominance among patients with discitis reflects the age distribution of patients with AS and is in agreement with earlier studies, the only exception being a case-series of 19 patients with discitis studied by Bartha and Varju [8], in which females predominated. Age at the diagnosis of discitis was 40 years in our study, in keeping with earlier work [2]. However, patients with pseudotuberculous or pseudarthrosis-like lesions may be younger (35 years in patients with pseudoarthrosis-like discitis in a study by Bouvier et al. [10] and 30 years in our study). The duration of AS at the diagnosis of discitis was 7 years in one study [5] and 28 years in another [9]. We found a tendency toward greater spinal ankylosis manifesting as a lower Schöber index value in the cases, suggesting that discitis possibly occurred in those patients with more advanced disease. Of our 14 patients with discitis, nine were manual laborers. In 1972, Cawley et al. [7] emphasized the major role for manual labor in the occurrence of discitis, pointing out that all their pa-
4. Discussion The prevalence of discitis in our series was 18%, as compared to only 5–10% in earlier studies by Rasker et al. [5], and Agarwal et al. [6] Cawley et al. [7] reported radiological abnormalities in up to 28% of patients, but they seem to have included Romanus lesions. The high prevalence in our study may be ascribable to selection bias: our study patients were admitted to a rheumatology department, suggesting that they possibly represented a subset with severe AS. In some studies, aseptic discitis was more common in patients with severe disease. Furthermore, all our patients had AS, whereas other studies also included patients with other spondyloarthropathies.
252
S. Langlois et al. / Joint Bone Spine 72 (2005) 248–253
Fig. 5. Magnetic resonance imaging, sagittal sections through the lumbar spine. (a) Low signal from the L3–L4 disk and adjacent endplates. (b) High signal on the T2-weighted sequence. (c) Enhancement after gadolinium injection. Note the rim of sclerosis and the normal signal from the disk on sequences (b) and (c).
tients with discitis had occupations involving light or heavy manual labor. In contrast to Louyot et al. [11], who reported predominant involvement of the thoracolumbar levels (75% of cases from T10 to L1), only 38% of our patients had thoracolumbar lesions; 44% had lumbar lesions. Spinal cord compression is uncommon in AS patients with discitis [12] and may result from bone instability or from granulation tissue growing in the epidural space at the level of the lesion [13]. The high prevalence of the HLA B27 antigen in the patients with discitis reflects that in the overall AS population. Severity of inflammation, as assessed by laboratory tests) was not greater in our cases than in our controls. In earlier studies, a single case with syndesmophytes at the level with discitis was reported [3]. Syndesmophytes developed as part of the repair process in two of our 14 patients. The higher proportion of patients with stage III sacroiliitis among the cases than the controls in our study cannot be ascribed to longer AS duration, as this variable was similar in the two groups. Rather, this finding suggests that discitis may occur preferentially in patients with a faster pace of progression of radiological AS lesions. Agarwal et al. reported histological findings in four patients. Reactive bone formation was noted in two patients and lymphocytic infiltration in the other two patients. These different histological patterns may reflect different stages of the disease, with lymphocytic infiltration occurring at the inflammatory stage and reactive bone formation during the healing process [6]. Wise and Irby highlighted the variability of discitis patterns by describing an unusual case characterized by T7-T8 discitis, spinal cord compression, and histological changes consistent with inflammation. The same patient, however, had T12-L1 discitis with imaging features suggestive of a noninflammatory process, although the absence of symptoms was at variance with previous reports [13]. MRI provides the diagnosis of early discitis by showing signal abnormalities characteristic of inflammation: low signal from the disk and adjacent endplates on T1-
weighted sequences, high signal on T2-weighted sequences, and enhancement after gadolinium injection [14]. Subsidence of the inflammatory process over time results in less typical MRI patterns. These are common, as 2 months or more usually elapse before MRI is performed. The MRI changes are consistent with the histological patterns indicating subacute or chronic lesions. MRI is also useful for ruling out other conditions, most notably spinal tuberculosis [15]. The variability of histological and radiological findings suggests that more than one pathophysiological mechanism may be capable of producing discitis in AS patients. The classic culprit is inflammation, which may lead to a variety of patterns, such as lesions mimicking tuberculosis or pseudarthrosis [16]. Discitis resembling pseudarthrosis develops when the nucleus pulposus exerts excessive pressure on the endplates previously weakened by an inflammatory process [17]. Early erosive forms may develop from Romanus lesions [18]. The intervertebral disk has two distinct components, the peripheral annulus fibrosus, which is an enthesis, and the central nucleus pulposus. Romanus lesions are marginal erosions of the anterior vertebral corners related to inflammation of the annulus fibrosus. Each Romanus lesion is seen as a small erosion rapidly enclosed by a rim of sclerosis. Romanus lesions can be viewed as mild forms of spondylodiscitis. Although healing with production of a syndesmophyte is the usual outcome, progression to classic discitis with vertebral bone destruction has been reported, opening up interesting pathogenic possibilities [3,10]. However, as pointed out by Louyot et al. [11], Romanus lesions usually affect multiple levels and discitis a single level, indicating that one or more additional factors are involved [11]. Several lines of evidence support a role for mechanical factors. Discitis arises at mobile levels located between two fused segments [19]. The pseudarthrosis-like pattern is often associated with mobile facet joints or a neural arch fracture [20,21]. However, Gougeon et al. [22] suggested that neural
S. Langlois et al. / Joint Bone Spine 72 (2005) 248–253
arch fractures may be either a consequence or a cause of discitis. Furst et al. [23] introduced the term “pseudopseudarthrosis” to describe abnormal motion between two fused segments resulting in extensive discovertebral destruction without fractures or pseudarthrosis. Cawley et al. [7] underlined the major role for trauma and suggested that endplate osteoporosis might be involved in some forms of discitis. Of the three patients in our study who had osteoporosis, two had the nonunion pattern of discitis. Finally, histological studies have shown features consistent with a callus, further supporting a contribution of mechanical factors. Taken in aggregate, these considerations suggest a multifactorial pathogenic process [24]. This concept could explain the other patterns of discitis, which may reflect intermediate stages in the natural history of discitis associated with AS.
5. Conclusion
[6]
[7]
[8] [9]
[10]
[11] [12]
[13] [14]
The occurrence of aseptic discitis in patients with AS is probably underestimated, in part because screening is not performed and in part because the symptoms may be ascribed to the underlying joint disease. In AS patients, discitis is associated with clinical signs reflecting advanced disease and with a significantly greater likelihood of stage III radiological sacroiliitis. One or more levels may be involved. Discitis may be the inaugural manifestation of AS. The outcome is usually favorable. The variability in radiological and histological patterns may reflect different phases in the natural course of the lesion.
[15]
[16] [17]
[18] [19]
References
[20]
[1]
[21] [22]
[2] [3] [4] [5]
Hunter T. The spinal complications of ankylosing spondylitis. Semin Arthritis Rheum 1989;19:172–82. Marie-Bonnin A. Les spondylodiscites de la pelvispondylite rhumatismale. Paris: Th.: Méd.; Lariboisière; 1986 p. 2104. Bouvier M, Tebib J, Colson F. Évolution des spondylodiscopathies de la spondylarthrite ankylosante. Rev Rhum 1987;54:229–33. Wendling D. Spondylarthrite ankylosante. EMC, APP LOC, 14-230A-10, 1998;14. Rasker JJ, Prevo RL, Lanting DJ. Spondylodiscitis in ankylosing spondylitis, inflammation or trauma? A description of six cases. Scand J Rheumatol 1996;25:52–7.
[23]
[24]
253
Agarwal AK, Reidbord HE, Kraus DR, Eisenbeis Jr. CH. Variable histopathology of discovertebral lesion (spondylodiscitis) of ankylosing spondylitis. Clin Exp Rheumatol 1990;8:67–9. Cawley MID, Chalmers TM, Kellgren HJ, Ball J. Destructive lesions of vertebral bodies in ankylosing spondylitis. Ann Rheum Dis 1972; 31:345–58. Bartha A, Varju T. Spondylodiscitis in ankylosing spondylitis. Ann Rheum Dis 2002;61(Suppl 1):221. KabasakalY, Garett S, Colin A. The epidemiology of spondylodiscites in ankylosing spondylitis. A controlled study. Br J Rheumatol 1996; 35:660–3. Bouvier M, Le Jeune E, Roulilot M, Chappaz JC. Les formes pseudopottiques et pseudodystrophiques des spondylodiscites de la SPA. Rev Rhum 1980;47:21–9. Louyot P, Gaucher A, Mathieu J, Miquel G. La spondylodiscite de la spondylarthrite ankylosante. Rev Rhum 1963;30:263–74. Jobanputra P, Kirkham B, Duke O, Crockard A, Panayi GS. Discovertebral destruction in ankylosing spondylitis complicated by spinal cord compression. Ann Rheum Dis 1988;47:344–77. Wise CM, Irby WR. Spondylodiscitis in ankylosing spondylitis: variable presentations. J Rheumatol 1983;10:1004–6. Eschelman DJ, Beers GJ, Naimark A, Yablon I. Pseudarthrosis in ankylosing spondylitis mimicking infectious diskitis: MR appearance. AJNR Am J Neuroradiol 1991;12:1113–4. Petterson T, Laasonen L, Leirisalo-Repo M, Tervahartiala P. Spinal pseudarthrosis complicating ankylosing spondylitis: a report of two patients. Br J Rheumatol 1996;84:1319–23. Doury P, Pattin S, Eulry F. Les spondylodiscopathies de la spondylarthite ankylosante. Presse Med 1984;13:446. Martorell J, Dubost JJ, Baron O, Michel JL, Sauvezi B, Rampon S. Spondylodiscopathies pseudodystrophiques de la spondylarthrite ankylosante. Presse Med 1984;13:1703. Guerin C, Cassou M. Les spondylodiscites de la pelvispondylite rhumatismale. Actualités Rhumatologiques 1981:35–42. Barozzi L, Olivieri I, De Matteis M, Podula A, Pavlica P. Seronegative spondylarthropathies: imaging of spondylitis, enthesis and dactylitis. Eur J Radiol 1998;27(Suppl 1):512–7. Peh CGW, Luk DKK. Pseudoarthrosis in ankylosing spondylitis. Ann Rheum Dis 1994;53:206–10. Martel W. Spinal pseudarthrosis. Arthritis Rheum 1978;21:485–90. Gougeon J, Rampon S, Deshayes P, Bussière JL, Seignon B, Le Loët X, et al. Discopathies postfracturaires et lésions vertébrales destructrices au cours de la pelvispondylite rhumatismale. Rev Rhum 1977;44:17–25. Furst Sheldon R, Kindynis P, Gundry C, Haghighi P, Sartris DJ, Resnick D. Pseudopseudarthrosis in a patient with ankylosing spondylitis. J Rheumatol 1990;17:258–62. Fonteille J, Bouvier M, Bonvoisin B, Lejeune E. Spondylarthrite ankylosante : pathogénie des lésions destructrices discovertébrales. Quinze observations. Presse Med 1984;13:15–28.
Original article
Aseptic discitis in patients with ankylosing spondylitis: a retrospective study of 14 cases Sandrine Langlois, Jean Pierre Cedoz, Anne Lohse, Eric Toussirot, Daniel Wendling * Service de Rhumatologie, Hôpital Jean Minjoz, CHU, Boulevard Fleming, 25030 Besançon cedex, France Received 8 December 2003; accepted 25 May 2004 Available online 06 October 2004
Abstract Objectives. – To determine whether patients with ankylosing spondylitis (AS) and discitis exhibit specific characteristics as compared to patients who have AS without discitis. Methods. – We retrospectively collected clinical, laboratory, and imaging data from the charts of 79 patients with AS admitted to a rheumatology department. Results. – Of the 79 patients, 14 (18%) had discitis that was not due to infection or trauma. Mean age at the diagnosis of AS was 40 years and mean duration of AS at admission was 10 years, with no difference between the two groups. In three patients, discitis was the inaugural manifestation of AS. Two patients had discitis at two levels. The lumbar and thoracolumbar segments were the most common sites of discitis. Symptoms of discitis were present in 10 of the 14 patients. Stage III sacroiliitis was significantly more common in the patients with discitis. None of the patients experienced neurological compromise. Conclusion. – The frequency of aseptic discitis in patients with AS is probably overestimated as a result of inclusion and exclusion biases. Discitis usually occurs at an advanced stage of AS under the form of an erosive condition related to both mechanical factors and osteoporosis. Inaugural or early discitis can occur, however, as a result of the inflammatory process. © 2004 Elsevier SAS. All rights reserved. Keywords: Ankylosing spondylitis; Discitis
1. Introduction Aseptic discitis is an uncommon manifestation of ankylosing spondylitis (AS) first described by Andersson in 1937. Destructive lesions of the disk and adjacent vertebral bone occur in the absence of infection or trauma. This pattern is characteristic of AS [1]. Studies conducted using modern imaging techniques and pathological examination of the entheses have rekindled interest in aseptic discitis as a manifestation of AS. Nevertheless, few data are available on aseptic discitis in AS. The studies reported by Marie-Bonnin [2] in 1986 and Bouvier et al. [3] in 1987 (43 patients) are among the most recent and largest case-series. Discitis is defined primarily on the basis of imaging features as a variable combi* Corresponding author. Tel.: +33-3-81-66-82-41; fax: +33-3-81-66-86-86. E-mail address: [email protected] (D. Wendling). 1297-319X/$ - see front matter © 2004 Elsevier SAS. All rights reserved. doi:10.1016/j.jbspin.2004.05.015
nation of disk-space narrowing, erosions in the adjacent endplates, and peripheral sclerosis. The objectives of the present study in a population with AS were to look for specific characteristics of patients with discitis as compared to those without discitis, to compare our findings with results in the literature, and to discuss available evidence on the pathophysiology of discitis associated with AS.
2. Patients and methods We reviewed the medical records of all the patients with AS admitted to the rheumatology department of the Besançon Teaching Hospital, Besançon, France, over a 6-year period. We selected the patients who met the modified New York criteria for AS [4]. With this criteria set, the diagnosis is definite in patients who have radiological evidence of grade II to IV sacroiliitis on both sides, or grade III or IV sacroiliitis on one side with at least one of the three following clinical
S. Langlois et al. / Joint Bone Spine 72 (2005) 248–253
249
Table 1 Numerical clinical data on the 14 cases with discitis and the 65 controls without discitis Schöber’s index, cm Finger-to-floor distance, cm Occiput-to-wall distance, cm Chest expansion BASFI BASDAI
Cases 1.25 37 10.3 2.3 87.5 69
Controls 2.60 26.6 4 .9 3.4 57.1 55.8
P value 0.058 0.32 0.10 0.37 0.16 0.29
BASDAI: Bath Ankylosing Spondylitis Disease Activity Index. BASFI: Bath Ankylosing Spondylitis Functional Index.
criteria: low back pain and stiffness for longer than 3 months with the pain abating upon physical activity but persisting at rest, motion range limitation of the lumbar spine in the coronal and sagittal planes, and reduced chest expansion as compared to normal values adjusted for age and sex. A diagnosis of probable AS is given in patients with all three clinical criteria but without the radiological criterion, or vice versa; in this last situation, other possible causes of sacroiliitis should be considered. In this study, the diagnosis of discitis was based on a review of the imaging studies in the medical records. Full-length spinal radiographs to look for multifocal discitis were not available for all the patients. We defined cases as patients with discitis and controls as patients without discitis. We compared clinical, laboratory, and imaging study features in the cases and controls. The following clinical variables were recorded: sex and age; conventional criteria for evaluating AS (peripheral arthritis, ocular involvement, enthesopathy, Schöber index, finger-tofloor distance, chest expansion, occiput-to-wall distance, Bath Ankylosing Spondylitis Disease Activity Index [BASDAI], and Bath Ankylosing Spondylitis Functional Index [BASFI]), comorbidities, and whether the patient was a manual laborer. Laboratory variables abstracted from the records were as follows: tests for inflammation at AS diagnosis and at discitis diagnosis, whether the HLA B27 antigen was present, and whether the serum IgA level was ≥3 g/l. Standard radiographs of the lumbar spine and pelvis were reviewed to determine the presence and stage of sacroiliitis and to look for syndesmophytes. We recorded bone mineral density data when available. In the cases, we recorded the following data: symptoms of discitis and findings from the investigations used to establish the diagnosis of discitis (histology, computed tomography [CT], radionuclide bone scanning, and magnetic resonance imaging [MRI]). Finally, we used the classification scheme developed by Bouvier et al. [3] to classify the patients based on the imaging features of discitis. 3. Results Of the 79 AS patients with valuable medical records who were admitted during the study period, 14 had radiological evidence of discitis. 3.1. Clinical data Mean duration of AS was 10 years in both groups. In three patients, discitis was the inaugural manifestation of AS.
Mean age at the diagnosis of AS was 33 years in the cases and 30 years in the controls (P = 0.30). Mean age at the diagnosis of discitis in the cases was 40 years. We found no significant differences between the cases and the controls for the proportion of patients with a family history of AS (33% vs. 17%, P = 0.21), manual labor (64% vs. 54%), ocular involvement, peripheral arthritis, enthesopathy, or numerical indices, although Schöber’s index was lower in the cases (1.25 vs. 2.60 cm) with a P value of borderline significance (0.058) (Tables 1 and 2). Comorbidities in the controls were as follows: psoriasis (1), eosinophilic gastrointestinal disorder akin to chronic inflammatory bowel disease (1), and Reiter syndrome (1). Among the cases, two patients had psoriasis. Symptoms of discitis were present in 12 of the 14 cases, with wide variations in clinical patterns. An injury with direct impact on the spine 4 years earlier was recorded as a possible precipitating factor in one of the cases. Of the 14 cases, 12 had discitis at a single level and two at two levels. The clinical outcome was consistently favorable, without neurological compromise or need for surgery. Follow-up radiographs showed development of vertebral synostosis over a 6-year period in one patient. In patient no. 11, intravenous bisphosphonate therapy was followed by increased bone sclerosis with no change in the erosions. The patient with cervical discitis achieved a full recovery after intravenous bisphosphonate therapy; an MRI scan done after 7 years showed no evidence of discitis. In patients no. 1 and no. 4, the disk space narrowing worsened over the years. 3.2. Laboratory tests We found no significant differences between the cases and controls regarding presence of HLA B27 (83% of cases and 93% of controls), serum IgA elevation, or severity of ASrelated inflammation. Parameters for inflammation were similar at the diagnosis of discitis and at other times during follow-up (Table 3). 3.3. Imaging study findings The cases were more likely to have spinal syndesmophytes, although the difference with the controls was not statistically significant. Stage III sacroiliitis was significantly more common among the cases than among the controls (57% vs. 29%; P = 0.045) (Table 4). Imaging study patterns of discitis were distributed as follows: pseudotuberculous,
250
S. Langlois et al. / Joint Bone Spine 72 (2005) 248–253
Table 2 Clinical features in the 14 cases with ankylosing spondylitis and discitis Spinal segment
Level
Case
Sex Age at diagnosis of discitis, years
Cervical Thoracic
C3-C4 T8-T9 T9-T10 T10-T11 T10-T11
8 12 7 2 6
M M M M F
T11-T12 3 T12-L1 5 T12-L1 9
T12-L1 L1-L2 L3-L4 L3-L4
L3-L4 L4-L5 L4-L5 L5-S1
Thoracolumbar
Lumbar
Lumbosacral
Peripheral Ocular arthritis involvement
Enthesopathy
Manual Fracture Trauma occupation
Clinical manifestations
18 42 22 45 33
Time since the diagnosis of AS (years) 0 3 5 0 11
Yes Yes No No Yes
No No Yes No Yes
Yes No Yes Yes No
Yes Yes Yes Yes No
No No No No No
No No No No No
M M M
51 46 42
27 15 14
No No Yes
Yes No No
No No Yes
Yes Yes No
No No Yes
1
M
30
14
No
No
No
No
No
No No Direct impact 4 years earlier No
Asymptomatic Inflammatory pain Inflammatory pain Mechanical pain Thoracolumbar pain stiffness++ Inflammatory pain Pain (type not specified) Thoracic kyphosis
10 11
M M
44 43
25 15
No No
No No
Yes No
Yes Yes
No No
No No
13 14 4
M M F
42 32 60
5 Yes 9
No No Yes
Yes No No
No No No
Yes No No
No No No
No No No
Table 3 Laboratory test findings at the diagnosis of discitis in the 14 cases Case 1 2 3 4 5 6 7 8 9 10 11 12 13 14
HLA B27 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No Yes Yes
IgA elevation Yes No No No No No Yes Yes No Yes Yes No No No
ESR (mm/h) NA 44 NA 80 NA 40 96 68 NA 24 34 NA 0 40
CRP (mg/l) NA 37 NA NA NA NA 60 32 NA 28 6 NA 0 84
NA: not available; ESR: erythrocyte sedimentation rate; CRP: C-reactive protein.
n = 1; pseudarthrosis-like, n = 2; late erosive, n = 2; early erosive, n = 3; isolated disk space narrowing, n = 3; and unclassifiable, n = 3 (Figs. 1–5). MRI was performed in six patients, of whom four had no typical signs of inflammation (low signal from the disk and adjacent endplates on T1weighted images, high signal on T2-weighted images, and enhancement after gadolinium injection). 3.4. Absorptiometry findings Overall, T-scores indicated osteopenia in the cases as compared to the controls, both at the lumbar spine (mean
Pain (type not specified) Inflammatory pain Inflammatory pain stiffness malalignment Inflammatory pain Inflammatory pain Mechanical low back pain and L5 sciatica
Table 4 Radiological and densitometry data in the 14 cases with discitis Case
Sacroiliitis
Syndesmophytes
T-score
Yes Yes Yes Yes Yes No Yes Yes Yes Yes Yes No Yes NA
NA NA −0.2 NA −4 NA NA −1.3 −4.6 1.7 2.2 NA 0.7 −0.7
II/III 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
NA: not available.
T-score, −1.2 vs. −0.6) and at the femoral neck (mean T-score, −1.1 vs. −0.6). Two of the cases had severe osteoporosis (T-score<−4) (Table 4). 3.5. Histological findings Biopsies were obtained in three patients. In two of these patients, only disk material was collected, 2 and 6 months after symptom onset, respectively; the findings were normal. In the remaining patient, a biopsy specimen was obtained 2 months after symptom onset and contained both disk and vertebral bone; histology showed tissue repair. No histologi-
S. Langlois et al. / Joint Bone Spine 72 (2005) 248–253
251
Fig. 1. Computed tomography of the lumbar spine with reconstruction: features mimicking tuberculosis. Fig. 3. C3–C4 discitis seen as isolated disc space narrowing without sclerosis. Loss of the normal lumbar lordotic curvature.
Fig. 2. Lateral view of T12-L1: extensive discovertebral destruction predominating anteriorly. Note the neural arch fracture at the level of the discitis (case no 9).
Fig. 4. Anteroposterior view of L3–L4. Late erosive form. Irregular erosions in the adjacent endplates and sclerosis. Lateral syndesmophytes (case no 10).
cal evidence of inflammation was found in any of the three biopsies.
The male predominance among patients with discitis reflects the age distribution of patients with AS and is in agreement with earlier studies, the only exception being a case-series of 19 patients with discitis studied by Bartha and Varju [8], in which females predominated. Age at the diagnosis of discitis was 40 years in our study, in keeping with earlier work [2]. However, patients with pseudotuberculous or pseudarthrosis-like lesions may be younger (35 years in patients with pseudoarthrosis-like discitis in a study by Bouvier et al. [10] and 30 years in our study). The duration of AS at the diagnosis of discitis was 7 years in one study [5] and 28 years in another [9]. We found a tendency toward greater spinal ankylosis manifesting as a lower Schöber index value in the cases, suggesting that discitis possibly occurred in those patients with more advanced disease. Of our 14 patients with discitis, nine were manual laborers. In 1972, Cawley et al. [7] emphasized the major role for manual labor in the occurrence of discitis, pointing out that all their pa-
4. Discussion The prevalence of discitis in our series was 18%, as compared to only 5–10% in earlier studies by Rasker et al. [5], and Agarwal et al. [6] Cawley et al. [7] reported radiological abnormalities in up to 28% of patients, but they seem to have included Romanus lesions. The high prevalence in our study may be ascribable to selection bias: our study patients were admitted to a rheumatology department, suggesting that they possibly represented a subset with severe AS. In some studies, aseptic discitis was more common in patients with severe disease. Furthermore, all our patients had AS, whereas other studies also included patients with other spondyloarthropathies.
252
S. Langlois et al. / Joint Bone Spine 72 (2005) 248–253
Fig. 5. Magnetic resonance imaging, sagittal sections through the lumbar spine. (a) Low signal from the L3–L4 disk and adjacent endplates. (b) High signal on the T2-weighted sequence. (c) Enhancement after gadolinium injection. Note the rim of sclerosis and the normal signal from the disk on sequences (b) and (c).
tients with discitis had occupations involving light or heavy manual labor. In contrast to Louyot et al. [11], who reported predominant involvement of the thoracolumbar levels (75% of cases from T10 to L1), only 38% of our patients had thoracolumbar lesions; 44% had lumbar lesions. Spinal cord compression is uncommon in AS patients with discitis [12] and may result from bone instability or from granulation tissue growing in the epidural space at the level of the lesion [13]. The high prevalence of the HLA B27 antigen in the patients with discitis reflects that in the overall AS population. Severity of inflammation, as assessed by laboratory tests) was not greater in our cases than in our controls. In earlier studies, a single case with syndesmophytes at the level with discitis was reported [3]. Syndesmophytes developed as part of the repair process in two of our 14 patients. The higher proportion of patients with stage III sacroiliitis among the cases than the controls in our study cannot be ascribed to longer AS duration, as this variable was similar in the two groups. Rather, this finding suggests that discitis may occur preferentially in patients with a faster pace of progression of radiological AS lesions. Agarwal et al. reported histological findings in four patients. Reactive bone formation was noted in two patients and lymphocytic infiltration in the other two patients. These different histological patterns may reflect different stages of the disease, with lymphocytic infiltration occurring at the inflammatory stage and reactive bone formation during the healing process [6]. Wise and Irby highlighted the variability of discitis patterns by describing an unusual case characterized by T7-T8 discitis, spinal cord compression, and histological changes consistent with inflammation. The same patient, however, had T12-L1 discitis with imaging features suggestive of a noninflammatory process, although the absence of symptoms was at variance with previous reports [13]. MRI provides the diagnosis of early discitis by showing signal abnormalities characteristic of inflammation: low signal from the disk and adjacent endplates on T1-
weighted sequences, high signal on T2-weighted sequences, and enhancement after gadolinium injection [14]. Subsidence of the inflammatory process over time results in less typical MRI patterns. These are common, as 2 months or more usually elapse before MRI is performed. The MRI changes are consistent with the histological patterns indicating subacute or chronic lesions. MRI is also useful for ruling out other conditions, most notably spinal tuberculosis [15]. The variability of histological and radiological findings suggests that more than one pathophysiological mechanism may be capable of producing discitis in AS patients. The classic culprit is inflammation, which may lead to a variety of patterns, such as lesions mimicking tuberculosis or pseudarthrosis [16]. Discitis resembling pseudarthrosis develops when the nucleus pulposus exerts excessive pressure on the endplates previously weakened by an inflammatory process [17]. Early erosive forms may develop from Romanus lesions [18]. The intervertebral disk has two distinct components, the peripheral annulus fibrosus, which is an enthesis, and the central nucleus pulposus. Romanus lesions are marginal erosions of the anterior vertebral corners related to inflammation of the annulus fibrosus. Each Romanus lesion is seen as a small erosion rapidly enclosed by a rim of sclerosis. Romanus lesions can be viewed as mild forms of spondylodiscitis. Although healing with production of a syndesmophyte is the usual outcome, progression to classic discitis with vertebral bone destruction has been reported, opening up interesting pathogenic possibilities [3,10]. However, as pointed out by Louyot et al. [11], Romanus lesions usually affect multiple levels and discitis a single level, indicating that one or more additional factors are involved [11]. Several lines of evidence support a role for mechanical factors. Discitis arises at mobile levels located between two fused segments [19]. The pseudarthrosis-like pattern is often associated with mobile facet joints or a neural arch fracture [20,21]. However, Gougeon et al. [22] suggested that neural
S. Langlois et al. / Joint Bone Spine 72 (2005) 248–253
arch fractures may be either a consequence or a cause of discitis. Furst et al. [23] introduced the term “pseudopseudarthrosis” to describe abnormal motion between two fused segments resulting in extensive discovertebral destruction without fractures or pseudarthrosis. Cawley et al. [7] underlined the major role for trauma and suggested that endplate osteoporosis might be involved in some forms of discitis. Of the three patients in our study who had osteoporosis, two had the nonunion pattern of discitis. Finally, histological studies have shown features consistent with a callus, further supporting a contribution of mechanical factors. Taken in aggregate, these considerations suggest a multifactorial pathogenic process [24]. This concept could explain the other patterns of discitis, which may reflect intermediate stages in the natural history of discitis associated with AS.
5. Conclusion
[6]
[7]
[8] [9]
[10]
[11] [12]
[13] [14]
The occurrence of aseptic discitis in patients with AS is probably underestimated, in part because screening is not performed and in part because the symptoms may be ascribed to the underlying joint disease. In AS patients, discitis is associated with clinical signs reflecting advanced disease and with a significantly greater likelihood of stage III radiological sacroiliitis. One or more levels may be involved. Discitis may be the inaugural manifestation of AS. The outcome is usually favorable. The variability in radiological and histological patterns may reflect different phases in the natural course of the lesion.
[15]
[16] [17]
[18] [19]
References
[20]
[1]
[21] [22]
[2] [3] [4] [5]
Hunter T. The spinal complications of ankylosing spondylitis. Semin Arthritis Rheum 1989;19:172–82. Marie-Bonnin A. Les spondylodiscites de la pelvispondylite rhumatismale. Paris: Th.: Méd.; Lariboisière; 1986 p. 2104. Bouvier M, Tebib J, Colson F. Évolution des spondylodiscopathies de la spondylarthrite ankylosante. Rev Rhum 1987;54:229–33. Wendling D. Spondylarthrite ankylosante. EMC, APP LOC, 14-230A-10, 1998;14. Rasker JJ, Prevo RL, Lanting DJ. Spondylodiscitis in ankylosing spondylitis, inflammation or trauma? A description of six cases. Scand J Rheumatol 1996;25:52–7.
[23]
[24]
253
Agarwal AK, Reidbord HE, Kraus DR, Eisenbeis Jr. CH. Variable histopathology of discovertebral lesion (spondylodiscitis) of ankylosing spondylitis. Clin Exp Rheumatol 1990;8:67–9. Cawley MID, Chalmers TM, Kellgren HJ, Ball J. Destructive lesions of vertebral bodies in ankylosing spondylitis. Ann Rheum Dis 1972; 31:345–58. Bartha A, Varju T. Spondylodiscitis in ankylosing spondylitis. Ann Rheum Dis 2002;61(Suppl 1):221. KabasakalY, Garett S, Colin A. The epidemiology of spondylodiscites in ankylosing spondylitis. A controlled study. Br J Rheumatol 1996; 35:660–3. Bouvier M, Le Jeune E, Roulilot M, Chappaz JC. Les formes pseudopottiques et pseudodystrophiques des spondylodiscites de la SPA. Rev Rhum 1980;47:21–9. Louyot P, Gaucher A, Mathieu J, Miquel G. La spondylodiscite de la spondylarthrite ankylosante. Rev Rhum 1963;30:263–74. Jobanputra P, Kirkham B, Duke O, Crockard A, Panayi GS. Discovertebral destruction in ankylosing spondylitis complicated by spinal cord compression. Ann Rheum Dis 1988;47:344–77. Wise CM, Irby WR. Spondylodiscitis in ankylosing spondylitis: variable presentations. J Rheumatol 1983;10:1004–6. Eschelman DJ, Beers GJ, Naimark A, Yablon I. Pseudarthrosis in ankylosing spondylitis mimicking infectious diskitis: MR appearance. AJNR Am J Neuroradiol 1991;12:1113–4. Petterson T, Laasonen L, Leirisalo-Repo M, Tervahartiala P. Spinal pseudarthrosis complicating ankylosing spondylitis: a report of two patients. Br J Rheumatol 1996;84:1319–23. Doury P, Pattin S, Eulry F. Les spondylodiscopathies de la spondylarthite ankylosante. Presse Med 1984;13:446. Martorell J, Dubost JJ, Baron O, Michel JL, Sauvezi B, Rampon S. Spondylodiscopathies pseudodystrophiques de la spondylarthrite ankylosante. Presse Med 1984;13:1703. Guerin C, Cassou M. Les spondylodiscites de la pelvispondylite rhumatismale. Actualités Rhumatologiques 1981:35–42. Barozzi L, Olivieri I, De Matteis M, Podula A, Pavlica P. Seronegative spondylarthropathies: imaging of spondylitis, enthesis and dactylitis. Eur J Radiol 1998;27(Suppl 1):512–7. Peh CGW, Luk DKK. Pseudoarthrosis in ankylosing spondylitis. Ann Rheum Dis 1994;53:206–10. Martel W. Spinal pseudarthrosis. Arthritis Rheum 1978;21:485–90. Gougeon J, Rampon S, Deshayes P, Bussière JL, Seignon B, Le Loët X, et al. Discopathies postfracturaires et lésions vertébrales destructrices au cours de la pelvispondylite rhumatismale. Rev Rhum 1977;44:17–25. Furst Sheldon R, Kindynis P, Gundry C, Haghighi P, Sartris DJ, Resnick D. Pseudopseudarthrosis in a patient with ankylosing spondylitis. J Rheumatol 1990;17:258–62. Fonteille J, Bouvier M, Bonvoisin B, Lejeune E. Spondylarthrite ankylosante : pathogénie des lésions destructrices discovertébrales. Quinze observations. Presse Med 1984;13:15–28.