Annals of the Rheumatic Diseases 1994; 53: 440-443

CASE STUDIES IN DIAGNOSTIC IMAGING Series Editor: V N Cassar-Pullicino*

Sacroiliac oint sepsis

P G White, A M Cooper

University Hospital ofWales, Cardiff, UnitedKingdomDepartment ofRadiologyP G WhiteDepartment ofRheumatologyA M CooperCorrespondence to:Dr A M Cooper,Department ofRheumatology, Heath Park,Cardiff CF4 4XW, UnitedKingdom.

*Department of DiagnosticImaging, The Robert Jonesand Agnes Hunt OrthopaedicHospital, Oswestry,Shropshire SY10 7AG,United Kingdom.

Clinical historyA thirty eight year old woman was admittedwith a two month history of right sciaticaunresponsive to analgesia and bed rest.Examination revealed reduced straight legraising on the right, reduced range ofmovement of the lumbar spine but no specifictenderness or neurological deficit. Plain radio-graphs of the lumbar spine and sacroiliac joints(SIJs) were normal, as was magnetic resonanceimaging (MRI) of the lumbar spine. Hercondition improved after a course of tractionbut two months later she was readmitted withintractable right sacroiliac pain, weight loss,night sweats and rigors. On examination shewas pale and apyrexial with no lympha-

./

Figure 1 Lateral radiograph of the lumbosacral spine showing loss ofdisc height at theL5-S1 level and erosion of the anterior part of the inferior end plate ofL5 (arrowed).

denopathy. There was loss of lumbar lordosis,paraspinal muscle spasm and generalisedtenderness of the lumbosacral spine; there wasno neurological deficit. Haematologicalinvestigations revealed a normochromicnormocytic anaemia with a normal white cellcount, but the erythrocyte sedimentation rate(ESR) was elevated to 95 mm/hour.

Radiological findingsFurther imaging consisted of plain radiographsof the lumbosacral spine and SIJs, a technetium99 m MDP bone scan and computed tomo-graphy (CT) of the SIJs. There was loss of discheight at the L5-S1 level with erosion of theanterior part of the inferior end plate of the L5vertebra (fig 1). There was widening of the rightSIJ with ill defined margins (fig 2). A delayedstatic image from the bone scan with aquantitative profile for the SIJs, revealedincreased uptake at the right SIJ, normalactivity on the left, and mildly increased activityat the L5-S1 level (fig 3). The CT scan (fig 4)showed widening of the right SIJ with erosionof the osseous margins and subchondralsclerosis, more marked on the sacral side. Anormal left SIJ was demonstrated.

Differential diagnosisThe table lists the commonest causes ofsacroiliitis. In Reiter's syndrome, gout, psori-asis and rheumatoid disease, bilateral sym-metrical or asymmetrical change is morefrequent than unilateral sacroiliac disease.Osseous erosion with variable subchondralsclerosis occurs in association with Reiter'ssyndrome and psoriatic arthritis; bony anky-losis occurs less frequently than in ankylosingspondylitis. In osteoarthritis the joint space isnarrowed with adjacent sclerosis and osteo-phytosis, but no erosive change. In hemiplegicpatients unilateral sacroiliac disease occurs onthe paralysed side and can affect thecontralateral SIJ in osteoarthritis of the hip.Sacroiliac involvement in chronic tophaceous

List of the common causes of unilateral and bilateralsacroiliac joint erosions. ++ = common; + = less frequent.(S) = symmetrical, (A) = asymmetricalCommon causes of sacroiliitis Unilateral Bilateral

Ankylosing spondylitis ++ (S)Inflammatory bowel ++ (S)diseaseInfection ++Psoriasis + ++ (A)Reiter's syndrome + ++ (A)Gout + ++ (A)Rheumatoid arthritis + ++ (A)Osteoarthritis + + (A)

440

on April 2, 2021 by guest. P

rotected by copyright.http://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.53.7.440 on 1 July 1994. Dow

nloaded from

http://ard.bmj.com/

Sacroiliac joint sepsis

Figure 2 Anteroposterior radiograph of th,margins of the right sacroiliac joint and wid

a'

L R

Figure 3 Posterior view bone scan of the pelvis withsacroiliac quantitative profile, showing increased uptake atthe right sacroiliac joint, normal activity at the left, andmildly increased uptake at the L5-S1 level.

gout is not common, but large well definederosions with surrounding sclerosis may beseen. Radiographic examination of the hands,feet and spine may reveal characteristicinvolvement of other joints.The radiographic and CT findings, with

unilateral loss of definition of the subchondralbone and widening of the joint space, arestrongly suggestive of infective sacroiliitis inour patient. Aspiration of the right sacroiliacjoint under fluoroscopic or CT guidance wouldbe helpful to confirm infection and allowisolation of the organism.

Loss of disc height at L5-S 1 and increasedscintigraphic activity at this level may be dueto degenerative disc disease, infective discitis or

lningbofthejoint (arrowed). rheumatoid arthritis. End plate erosions arestrongly suggestive of infection and furtherevaluation of the L5-S 1 disc with MRI isindicated.

Follow up and final diagnosisSoon after readmission the patient developedevidence of compression of the right SI nerveroot, with sciatica and a reduced ankle jerk.MRI with axial Ti weighted images showed amass in the anterior epidural space at the L5and S 1 levels, consistent with an epiduralabscess (fig 5). Ti weighted images also

Figure 4 Axial CT image of the sacroiliac joints showing erosions of the osseous marginswith subchondral sclerosis on the right (arrowed) and a normal left sacroiliac joint.

Figure 5 Axial Ti weighted MRI of the sacrum showinganterior epidural abscess (arrowed), right sided sacroiliitisand inhomogenous signalfrom the vertebral marrow.

441

on April 2, 2021 by guest. P

rotected by copyright.http://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.53.7.440 on 1 July 1994. Dow

nloaded from

http://ard.bmj.com/

White, Cooper

'E

Figure 6 Coronal proton density MRI of the sacrum showing right sided sacroiliitis withmarrow oedema in the sacrum (arrow).

showed reduced signal intensity of thevertebral marrow, indicating sclerosis oroedema, and swelling of the prevertebral softtissues. Coronal proton density images clearlydemonstrated the extent of sacroiliac infection(fig 6). Sagittal Ti weighted images showedthe appearances of discitis at the L5-S1 leve]with destruction of the margins between thedisc and the adjacent end plates; T2 weightedimages showed increased signal intensity fromthe same areas. Following neurosurgicalreferral laminectomy and L5-S1 discectomywas performed. Infected granulation tissue wasremoved from the epidural space andstaphylococcus aureus was subsequentlyisolated from this material. After a six week

Figure 7 Lateral radiograph of the lumbosacral spineshowing complete loss of disc height at the L5-S1 level andend plate destruction (arrowed).

course of intravenous antibiotics the patientmade a full recovery. Radiographs of thelumbosacral spine and pelvis taken six monthslater showed destruction ofthe L5-S 1 vertebralend plates with complete loss of disc height (fig7) and sclerosis of the right SIJ (fig 8). The finaldiagnosis was infective right sacroiliitis andchronic epidural abscess with discitis.

DiscussionSpinal and sacroiliac infections mostcommonly result from haematogenous spread,usually by the arterial route. Batson's vertebralvenous plexus is less important in the spreadof infection than was previously thought.2 Inthis patient infection may have been startedeither at the sacroiliac joint or the L5-S1 disc.The arrangement of the microcirculation inboth the subchondral bone of the ilium' andthe vertebral end plates3 predisposes to slowblood flow, making these areas prone tohaematogeneous infection.The clinical presentation of both spinal

infection and infective sacroiliitis is non-r specific. In vertebral osteomyelitis neurologicalt examination is usually normal at the time ofy presentation, and the delay in diagnosisi average three months for pyogenic infectionsI and two years for tuberculosis.4 EpiduralI abscess may present acutely or, as in this case,e as a chronic infection secondary to discitis. InI both types of presentation laminectomy is

indicated,3 but in chronic infection granulationI tissue is usually found rather than frank pus.2y The ESR is elevated in 73% of cases3 but onlys 35% of patients have a raised white cell count.

k Radiological evaluationThis has an important role in the diagnosis ofboth spinal and sacroiliac infections. A rangeof imaging modalities is available.

A) PLAIN RADIOGRAPHSIn discitis plain film changes may be seen afterone to three weeks,' with loss of definition ofthe end plates and reduction in disc height,followed by areas of destruction in the adjacentvertebral body. In infective sacroiliitis the firstplain film changes develop after two to threeweeks. Erosions usually start on the iliac sideof the joint, probably due to the greaterthickness of the cartilage over the sacral jointsurface.5 Soft tissue swelling with or withoutabscess formation is more easily discernible inthe spine than in the SIJ. In this patient theerosive changes were more marked on thesacral side, suggesting that infection may haveinvolved the joint by direct extension from afocus in the sacrum.

B) RADIOISOTOPE BONE SCANNINGThis is a more sensitive indicator ofboth spinaland sacroiliac infections than plain radiographywith increased uptake early in the course of thedisease;' specificity of this technique isimproved by three-phase scanning.6 Sensi-

442

on April 2, 2021 by guest. P

rotected by copyright.http://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.53.7.440 on 1 July 1994. Dow

nloaded from

http://ard.bmj.com/

Sacroiliac joint sepsis

Figure 8 Anteroposterior radiograph of the pelvis showing sclerosis of right sacroiliac jointand lumbosacraljunction (both arrowed).

tivities of 92% with specificity of 100% havebeen obtained in the detection of vertebralosteomyelitis using combined gallium andbone scan studies.7

c) COMPUTED TOMOGRAPHY

CT is well suited for imaging the SIJs,8 but anunderstanding of the anatomy is important (fig9) as the appearance of the postero-superiorligamentous part of the joint can mimic thechanges of inflammatory sacroiliitis. CT is alsouseful in spinal infection, giving a goodassessment of the extent of bony involvementand showing extension into the spinal canal3and para-vertebral tissues especially along thepsoas outlines.

D) MAGNETIC RESONANCE IMAGING

MRI is now the technique of choice in spinalinfection. Modic et al7 found a sensitivity of

lliolumbar ligamentInterosseous sacro-iliac

Lumbosacral ligament ligament

Anteriorsacroiliac Synovial membraneligament

Figure 9 The normal anatomy of the sacroiliac joint.

96% and specificity of 92% for the detectionof vertebral osteomyelitis. MRI allows directvisualisation of the epidural space, theca, nerveroots and spinal cord. Epidural infection ifmore readily detected on T1 than T2 weightedimages as on T2 the signal intensity of anepidural abscess is similar to that of CSF. Oncontrast enhanced Ti weighted images homo-genous enhancement suggests the presence ofinfected granulomatous tissue whereasperipheral enhancement indicates a liquefiednecrotic abscess.9 Overall, MRI is also superiorto CT in the evaluation of sacroiliitis,'°particularly for the assessment of sacroiliaccartilage and erosive change. Osteophyteformation, bony ankylosis and subchondralsclerosis are best imaged by CT.

E) INTERVENTIONAL TECHNIQUESIn both discitis and infective sacroiliitisaspiration or needle biopsy under fluoroscopicor CT guidance may facilitate indentificationof the infecting organism. One study revealedthat a positive culture result was obtained in50% of infective vertebral lesions followingneedle biopsy,"1 previous antibiotic therapybeing a frequent cause of negative results. Themost frequent pyogenic organism isolated inspinal and sacroiliac infections is Staphylococcusareus, but gram-negative organisms arebecoming more frequent in drug abusers3 12and tuberculosis remains an important cause ofinfection at both sites.

1 Resnick D, Niwayana G. Diagnosis ofbone andjoint disorders.Philadelphia: Saunders, 1981: 2132-51.

2 Mark A S. Non-degenerative, non-neoplastic diseases of thespine and spinal cord. In: Atlas S W, ed. Magneticresonance imaging of the brain and spine. New York: Raven,1991: 1001-7.

3 Wisneski R J. Infectious disease of the spine: diagnostic andtreatment considerations. Orthop Clin North Am 1991; 22:491-501.

4 Finkenberg J G. Pyogenic and non-pyogenic infections ofthe spine: indications and treatment. Cur Opin Orthop1993; 4: 177-85.

5 Rosenberg D, Baskies A M, Deckers P J, Lester B E, OrdiaJ I, Yablon I G. Pyogenic sacroiliitis. Clin Orthop 1984;184: 128-32.

6 Mawer A H, Chen D C P, Camargo E E, Wong D F, WagnerH N, Alderson P 0. Utility of three-phase skeletalscintigraphy in suspected osteomyelitis: concisecommunication. JNucl Med 1981; 22: 941-9.

7 Modic M T, Feiglin D H, Pirairo D W, et al. Vertebralosteomyelitis: assessment using MR. Radiology 1985; 157:157-66.

8 Forrester D M. Imaging of the sacroiliac joints. Radiol ClinNorth Am 1990; 28: 1055-72.

9 Sandhu F S, DillonW P. Spinal epidural abscess: evaluationwith contrast-enhanced MR imaging. AJR 1992; 158:405-11.

10 Murphey M D, Wetzel L H, Bramble J M, Levine E,Simpson K M, Lindsley H B. Sacroiliitis: MR imagingfindings. Radiology 1991; 180: 239-44.

11 Stoker D J, Kissim C M. Percutaneous vertebral biopsy: areview of 135 cases. Clin Radiol 1985; 33: 569-77.

12 Guyot D R, Manoli A, Kling G A. Pyogenic sacroiliitis inIV drug abusers. AJR 1987; 149: 1209-11.

443

on April 2, 2021 by guest. P

rotected by copyright.http://ard.bm

j.com/

Ann R

heum D

is: first published as 10.1136/ard.53.7.440 on 1 July 1994. Dow

nloaded from

http://ard.bmj.com/