iew for Med Pharmacological Sciences 2005; 9: 53-66 ...deficit: 11 cases had intraspinal granuloma-tous tissue causing neurological dysfunction in the absence of bony destruction,

53

Chronic ostemyelitis may require surgery incase of a development of biomechanical insta-bility and/or a vertebral collapse with progres-sive deformity.

Key words:Vertebral osteomyelitis, Spondylodiscitis, Pyogenic os-

teomyelitis, Skeletal tuberculosis.

Introduction

Haematogenous vertebral osteomyelitis(HVO) is a relatively rare disorder which ac-counts for 2-4% of all cases of infectious bonedisease1. In recent years, the incidence of spinalinfections has seemed to increase according tothe growing number of intravenous drug usersin young people and in the elderly with the useof intravenous access devices, genitourinarysurgery and manipulation. Males are more fre-quently affected than females with an averageage of onset in the fifth and sixth decade. Theonset of symptoms is typically insidious withneck or back pain often underestimated by thepatient. The early diagnosis is also difficult dueto the non-specific nature of laboratory and ra-diographic findings. The frequent observationof back pain also makes the diagnosis a chal-lenge in most cases. Several studies in the liter-ature report an average delay in the diagnosisof haematogenous vertebral osteomyelitis from2 to 6 months after the beginning of the symp-toms2,3,4. In this article we review the clinicalfeatures and the diagnostic approach tohaematogenous vertebral osteomyelitis in or-der to optimise treatment strategies and fol-low-up assessment.

Abstract. – This article review the clinicalfeatures and the diagnostic approach tohaematogenous vertebral osteomyelitis in orderto optimise treatment strategies and follow-upassessment. Haematogenous spread is consid-ered to be the most important route: the lumbarspine is the most common site of involvementfor pyogenic infection and the thoracic spine fortuberculosis infection. The risk factors for devel-oping haematogenous vertebral osteomyelitisare different among old people, adults and chil-dren: the literature reports that the incidenceseems to be increasing in older patients. Thesource of infection in the elderly has been relat-ed to the use of intravenous access devices andthe asymptomatic urinary infections. In youngpatients the increase has been correlated withthe growing number of intravenous drugabusers, with endocarditis and with immigrantsfrom areas where tuberculosis is still endemic.The onset of symptoms is typically insidiouswith neck or back pain often underestimated bythe patient. Fever is present in 10-45% of pa-tients. Spinal infections may cause severe neu-rological compromise in few cases, but mildneurological deficit, limited to one or two nerveroots, was detected in 28-35% of patients. Thediagnosis of haematogenous vertebral os-teomyelitis may be very difficult, as the symp-toms can be sometimes not specific, vague oralmost absent. The usual delay in diagnosis hasbeen reported to be two to four months, despitethe use of imaging techniques: in the early diag-nosis of vertebral ostemyelitis is important therole of bone scintigraphy. The general principlesfor the management of spine infections are nonoperative, consisting of external immobilizationand intravenous antibiotics, followed by oral an-tibiotics. Indications for surgery should be givenin case of absence of clinical improvement after2-3 weeks of intravenous antibiotics, persistentback pain and systemic effects of chronic infec-tion and with presence or progression of neuro-logical deficit in elderly or in cervical infection.

European Review for Medical and Pharmacological Sciences 2005; 9: 53-66

Clinical features, diagnostic and therapeutic approaches to haematogenous vertebral osteomyelitisAL. GASBARRINI, E. BERTOLDI, M. MAZZETTI*, L. FINI***, S. TERZI, F. GONELLA, L. MIRABILE, G. BARBANTI BRÒDANO, A. FURNO**, A. GASBARRINI***, S. BORIANI*

Department of Orthopaedics and Traumatology, Maggiore Hospital “C.A. Pizzardi” - Bologna (Italy)*Department of infections disease, Maggiore Hospital “C.A. Pizzardi” - Bologna (Italy)**Nuclear Medicine, Maggiore Hospital “C.A. Pizzardi” - Bologna (Italy)***Internal Medicine, Catholic University - Rome (Italy)

54

Etiopathogenesis

Spinal infections may affect the vertebralbody, the intervertebral disc, the neural archor the posterior elements but most commonlythey involve the anterior and middlecolumns1. The infection can involve and crosscortical bone and longitudinal ligaments,leading to soft tissue abscesses. Epidural ab-scesses may arise adjacent to the area of os-teomyelitis or, less commonly, occur de novo.

Haematogenous spread by means of thearterial system is considered to be the mostimportant route, because the vertebral bodyis richly supplied by an arterial network, es-pecially in the anterior subchondral regionnear the anterior longitudinal ligament5. Discspace infections that occurs in adults are usu-ally associated with prior surgical disruptionof the disc, while the most common site of denovo infection in children is within the disc.Indeed, histological analyses have confirmedthat an endarteriolar supply to the disc is pre-sent until childhood but is then slowly oblit-erated in the first three decades of life. Thus,adult intervertebral disc is usually not primar-ily involved, since it is avascular1.

As a result of haematogenous spread, mul-tiple foci of infection can occur. A complexvalveless venous drainage, known as Batson’sparavertebral venous plexus, may also act asa potential route of infection, particularly forspread from the pelvic organs especially in in-stances of sepsis originating in the urinarybladder, bowel and female pelvic organs5.

The lumbar spine is the most common siteof involvement for pyogenic infection fol-lowed by the thoracic, cervical and sacral re-gions1,5-7. The thoracic spine is the most com-monly affected site in tuberculosis infection8,which may be explained by the frequent in-volvement of mediastinal lymphnodes andpleura in pulmonary tuberculosis, from wheremicroorganism can reach the vertebral bonethrough the limphatic route8.

Matsui et al.9 noted that the degree of de-struction may depend on the patients’s bio-logical reaction relating to their physical con-dition or age. Buchelt et al.10, instead, sug-gested that it was influenced by the species ofpathogen, reporting that more than two seg-ments were involved in 23% of patients withtuberculous and 9% of patients with pyogenicinfection. Others studies have reported multi-

ple level involvement in pyogenic infection in5 to 18% of patients7,11-13.

The commonest cause of osteomyelitis isStaphilococcus Aureus. In adults entericGram negative organism are second cause,while H. Influenzae is more frequent inyoung infants14 and Group B Streptococci inneonates15. In a retrospective analysis16 in apopulation of 72 patients, the organism morefrequently isolated were Staphilococcus Au-reus and Epidermidis, Brucella Melitensis, E.Coli, Mycobacterium Tuberculosis. Strepto-coccus pneumoniae, S. agalactiae, S. viridans,S. faecalis, Proteus mirabilis, PseudomonasAeruginosa, Candida Glabrata were uncom-mon. Hidatidosis, actinomicosis, aspergillosisand fungal infections rarely cause os-teomyelitis, while only one case ofHaemophilus paraphrophilus15 and Pastourel-la dogmatis17 vertebral osteomyelitis havebeen reported. Serratia marcenscens andPseudomonas Aeruginosa can be responsibleof osteomyelitis in intravenous drugabusers18.

The risk factors for developing HVO aredifferent among old people, adults and chil-dren18. The literature reports that the inci-dence of HVO seems to be increasing in old-er patients. This may be related to the greaterlongevity of the general population as well asto the more intensive treatment of serious ill-nesses and the use of chemotherapy to treatcancer and immunological disorders.Carrage7 reviewed 111 patients: sixty-one(55%) were sixty years old or more and forty-four patients (40%) had an impaired immunesystem. In a multicenter study, which includ-ed 219 adult patients, Colmenero et al.19 re-ported similar figures concerning age and as-sociated disease.

The source of infection in the elderly hasbeen related to the use of intravenous accessdevices, with resultant nosocomial bacter-aemia20. Others source were respiratory ororal infection, skin ulceration, genitourinarysurgery, placement of indwelling bladdercatheters or ureteral stents, or both. Typical-ly, HVO caused by Gram-negativepathogens (mainly Escherichia coli and Pro-teus Mirabilis) derive from the urinary tractand it should be considered when elderly pa-tients complain of back pain. In fact, asymp-tomatic urinary infections are very commonin old people21 due to urinary incontinence,

Al. Gasbarrini, E. Bertoldi, M. Mazzetti, L. Fini, S. Terzi, F. Gonella, L. Mirabile, et al

use of catheters, comorbid illness such as di-abetes or neoplasms, vaginal bacterial colo-nization in postmenopausal period16. Renalinsufficiency, chronic hepatic disease, alco-holism, recent surgery, haemodialisis areother recognised risk factor in the elderly.Recent studies demonstrated a relationshipbetween surgical procedure, ageing and im-mune dysfunction. In fact, monocyte pheno-type and function can be altered as well asthe bactericidal activity of neutrophils fol-lowing on surgical stress. A reduction of T-cell responsiveness and NK cell function isassociated with ageing16.

In young patients the increased incidenceof HVO has been correlated with the grow-ing number of intravenous drug abusers andwith endocarditis11,22,23. Sometimes endocardi-tis may be complicated by septic arthritis andvertebral osteomyelitis. An early diagnosis isgenerally difficult because those patients areusually treated for fever, bone pain and stiff-ness, even underestimating or masking theendocarditis24.

The recent increase in the incidence of tu-berculosis may be related to a growing num-ber of immunocompromised patients, whichmay be caused by the growing number of im-migrants from areas where tuberculosis is stillendemic25.

Physical Findings

Back pain and paravertebral muscle spasmare the most common clinical find-ings7,12,19,23,26 in HVO. Some authors reportthe presence of fever in 10-45% of patients,even in pyogenic osteomyelitis7,11,19,25. Thisfact frequently allow clinicians to suspect thepossibility of infection. The absence of feverwas significantly more frequent in spinal tu-berculosis, with a greater presence of spinaldeformity. The latter is in close relation withthe considerable destructive character ofcaseating granuloma and is an important di-agnostic clue8,19. Others reported a slight butpersistent fever in spinal infections in 65% to90% of the cases13,27.

Mild neurological deficit, limited to oneor two nerve roots, was detected in 28-35%of patients7,11,13,19,27. Rarely spinal infectionsmay cause severe deficits such as complete

or incomplete paraplegia5,7. Neurologicaldeficit (in particular, paralysis) are frequent-ly associated with epidural abscesses. Hadji-pavlou et al6 reported 33 cases of epiduralabscesses as a complication of spondy-lodiscitis in study of 101 cases of pyogenicspinal infection. Out of this group of pa-tients 15 had paraparesis or paraplegia. In aretrospective analysis of 29 cases of spinaltuberculosis, Nussbaum et al.25 noted that 22patients (76%) presented with neurologicaldeficit: 11 cases had intraspinal granuloma-tous tissue causing neurological dysfunctionin the absence of bony destruction, 2 caseshad intramedullary tuberculomas, 9 hadmarked bone collapse with neurological in-jury. The possibility of serious neurologicalcomplication was higher in the thoracic andcervical spine as opposed to the lumbarspine. Therefore, when cephalic levels areinvolved, more caution should be exercisedin assessing possible epidural abscess forma-tion and preventing its neurological seque-lae6,28. The greater diagnostic delay and thefrequent existence of spinal deformity par-tially explain why neurological deficits weresignificantly more frequent in tuberculosisosteomyelitis19,25,26.

Kyphosis is a rare complication and occursmore commonly in tuberculous spondylitis. Apositive straight leg raising test, sinus tractformation and subcutaneous abscesses arepresent only in a few percent of thepatients29. Colmenero et al19 noted paraverte-bral masses in 49.7% of cases, epidural ab-scesses in 36.5% and psoas abscesses in10.9%. Paravertebral and epidural masseswere present respectively in 78% and 68% oftuberculosis infection with a statistically sig-nificant difference with respect to pyogenicand brucellar osteomyelitis.

The diagnosis of HVO may be very diffi-cult, as the symptoms can be sometimes notspecific, vague or almost absent. The onset ofsymptoms is insidious and often underesti-mated by the patient. Therefore, misdiagno-sis or delayed diagnosis are very frequent.

The usual delay in diagnosis has been re-ported to be two to four months, despite theuse of CT and scintigraphic bonescanning30,31. Carragee7,32 presented a studywith MRI scan: diagnosis of infection was ob-tained in a median time of less than 3 weeksafter the onset of spinal symptoms. It is likely

55

Haematogenous vertebral osteomyelitis

56

that the use of MRI has made possible to di-agnosis this disease in the early stages27,33-35.Buchelt et al10 reported that the mean inter-val between onset of symptom and diagnosiswas significantly longer for patients with tu-berculosis. This fact can be explained by theslower progression of tuberculosis and shouldbe evaluated as an anamnestic factor in dif-ferential diagnosis. Active or previously diag-nosed extraspinal tuberculosis was showed in33% to 52% of the cases10,19,25. The diagnosticdelay in pyogenic infections was significantlyshorter, which may reflect the higher clinicalexpression of this group of patients. Delay indiagnosis may result in vertebral destructionor perforation of the spinal canal36.

Misdiagnosis are more common in 60-70years old patients and the most frequent local-ization of osteomyelitis seems the lumbarspine. Spinal infections are often confused withmetastatic carcinoma, spinal stenosis, herniat-ed nucleus pulposus and back strain. In severalanalysis, misdiagnosis are significantly associat-ed with the age of the patients, absence offever and positive straight raising test37.

Diagnostic approaches

Laboratory testsThe value of laboratory tests in HVO is

still unclear. The leukocyte count is typicallynot elevated in spinal infection. According toreports in the literature, leukocyte countsrange from 13% to 60%. Levels of ESR aremore commonly elevated, ranging from 73%to 100%6,11,19,23,26,27,31. However, when leucocy-tosis, neutrophilia and high values of ESRand CRP are present, they strongly suggest apyogenic infection19. In the presence of anepidural abscess these tests are more sensi-tive. As reported by previous studies11,12,19,23,26,blood culture was the most useful routinetest, providing microbiological diagnosis in30% to 50% of cases12,19. During a feverspike, a higher percentage of cultures will bepositive than during chronic phases of infec-tion. If the cause of septicemia is known orthe blood cultures are positive, patients areless likely to be referred for an additional in-vasive procedure. However, there is a chancethat a secondary organism may be missed ifbiopsy is not performed12. Aetiological diag-

nosis of vertebral osteomyelitis is frequentlydifficult when blood cultures are negative. Inthese cases, spinal specimens were obtainedin 30% to 70% of the patients by closed per-cutaneous or open surgical biopsy19,26.

Differential diagnosis of the etiology be-tween tuberculous and pyogenic infectionmay be difficult. A negative Mantoux testindicates non-specific etiology. A positiveMantoux test is not patognomonic for thediagnosis and bacteriologic or histologicaltests should be performed7. Belzunegui etal16 suggest that even the isolation of My-cobacterium tuberculosis in other tissue, orfluid sample or histological evidence ofcaseating granuloma may be enough for thediagnosis of HVO. Nussbaum et al25 report-ed 21% of cases with tuberculous spondylitiswho had no previous or concurrent diagno-sis of extraspinal tuberculosis, no family his-tory of tuberculosis and a negative tuber-culin skin test. The yield of routine bacterio-logical and serological tests (positive seroag-glutination at titres 1/60) was very high inbrucellosis infection19.

Imaging TechniquesInfection of the spine must be differentiat-

ed from degenerative disease, non-infectiousinflammatory lesions and spinal neoplasm.The infection can affect the vertebrae, the in-tervertebral disc, the paraspinal soft tissue,the epidural space, the meninges, and/or thespinal cord. Imaging plays an important rolein the overall evaluation of these lesions andthe ideal technique is expected to provide in-formation that will help characterize and de-lineate the disease process, guide biopsyand/or drainage procedures, suggest themethod of treatment (medical vs surgical)and assess the response to therapy37.

Plain Radiography (Figure 1). Radi-ographs of the spine show no signs of spondy-lodiscitis in the early stages of the disordersand only subtle changes, including endplatedemineralisation and/or irregularity, may benoted38,39. The earliest radiographic sign inpyogenic infection appears in the third weekof the disease: it is a slight narrowing of theintervertebral space and a loss of definition inthe superior endplate. This is followed byprogressive disc space narrowing, gradual de-velopment of osteolysis with irregularity of


the vertebral body margins and further de-struction of the subchondral plate. In thefourth week, signs of vertebral destructionmay be seen in some cases and after sixweeks, they are always evident.

In the differential diagnosis of infectionfrom cancer, the appearance of a destructivevertebral bone lesion associated with a well-preserved disc space with sharp endplatesfavours a diagnosis of neoplastic infiltration,whereas the destructive bone lesion associat-ed with a poorly defined vertebral bony end-plate with or without loss of disc height sug-gests infection with a better prognosis. Discspace narrowing can be due to coincidentaldegenerative disc disease independent of anyinfectious process5.

Bone scintigraphy (Figure 2). The role ofscintigraphy in the early diagnosis of vertebralostemyelitis is important. First of all, bonescintigraphy is widely available, it is easilyperformed, safe to use and rapidly completed.

Tc-99m-MDP (methylene diphosphonate)with SPECT (single photon emission computedtomography) shows a sensitivity of 90% in theearly detection of osteomyelitis5,12,22,35,40,41. Nor-mal Tc-99m-MDP bone images of the vertebraexclude osteomyelitis with a high degree of cer-

tainty42. Further sites of infection not clinicallysuspected can also detected by an increased up-take in other part of the skeleton. An initialwhole-body scan is crucial to the diagnosiswhen symptoms are difficult to localise and/orthe patient is old, confused and febrile5,43. Un-fortunately, the study is not specific35,40,42,44.Specificity depends on the underlying conditionof the bone. In adults with normal radiographsand no reason for increased bone turnover, thespecificity of the scan is higher. When bone re-modelling is increased by fractures, tumors, ac-tivated ostheoarthritis, non-infectious inflam-matory lesions or pseudoarthrosis, the specifici-ty of the bone scan is reduced5. False-negativeresults of bone scans have been observed in el-derly patients, presumably because of the re-gional ischemia secondary to arterioscleroticdisease. This suggests that a negative result of abone scan may not reliably exclude infection,especially in older persons40.

Recently the use of immunoscintigraphywith 99 Tc-labelled antigranulocyte antibod-ies Fab’ fragments in the management ofneonates and infants with fever of unknownorigin has been reported and this procedurehas shown high sensitivity and specificity inthe diagnosis of osteomyelitis and soft tissueinfections45.

57


Figure 1. F.D., male, 51 years of age, spondylodiscitisat L3-L4 due to Staphilococcus aureus, latero-lateral x-ray.

Figure 2. B.A., male, 28 years of age, tubercularspondylodiscitis at L3- L4. Tecnetium-99 total-bodybone scintigraphy.

58

Ga-67 Citrate Imaging. Several studiesconfirmed the utility of Ga-67 imaging toidentify vertebral osteomyelitis41,42,46. Ga-67imaging is often used as a complement tobone scintigraphy to enhance the specificityof the study and detect extraosseous sites ofinfection40,43,46. Ga-67 citrate SPECT is able toidentify unsuspected causes of endocarditis,paravertebral abscess, subaxillary soft tissueabscess and other additional sites of infec-tion. Unlike 99 Tc-SPECT imaging or planarGa-67 citrate imaging, Ga-67 citrate SPECTcan estimate the severity of infection42.

Despite the excellent results achieved, thedual-tracer technique has its disadvantages.The procedure requires two different tracersand multiple prolonged imaging sessions ondifferent days. There is also an increased costand inconvenience to patients, many ofwhom are elderly or debilitated, or both.

Ga-67 imaging may be a better tool in thefollow-up of the response to treatement sinceis less sensitive to bone remodelling and givesa more accurate degree of the infectiousprocess activity. 99 Tc-scintigraphy remainsactive until complete healing takes place, andmay remain active after infection has becomequiescent due to its sensitivity to bone re-modelling and repair5. For this reason, it re-mains positive for a long time compared withGa-67 scan, which will become positive onlywhen the infection is in a active phase47.

Computed Tomography. CT scans yieldpositive findings in the early stages, becausethe involved disc shows small hypodense ar-eas. CT also shows the disc flattening and thevertebral endplate destruction, which are notvisible on conventional radiographs in theearly stage13. The extent of the inflammatoryprocess is defined. Paravertebral abscesseswith psoas involvement are easily identifiedafter contrast administration. Intraspinal ex-tension of the process with an epidural ab-scesses is better defined by MRI (Figure 3)13.

The CT scan may also be used more di-rectly in diagnosis by assisting with needleaspiration of a suspected lesion (Figure 4)1.In particular even when the diagnosis of in-fectious discitis is established by means ofdiagnostic imaging, a specific microbiologicdiagnosis is highly desirable for definitivemedical treatment with the antibiotics towhich the pathogens are sensitive. In any

event, without the identification of a specificpathogen, the choice of antibiotic therapywould remain empiric. This technique maybe used in the thoracic and lumbar spine,but is generally too dangerous to attempt inthe cervical spine because of the surround-ing structures1. Chew et al48 reported thatCT-guided needle aspiration is an accuratemethod for identifying bacterial or granulo-matous infection of the disc space, but allfalse negative results originate from cases offungal infection. Previously reported seriesof imaging-guided needle biopsy in sponta-neous infection have been variable in theiryield of microbiologically identifiedpathogens. Chew et al48 reported a 91%yield (39 out of 43 patients with infection),Perronne et al26 reported a 74% yield (29out of 39), Carragee7 reported a 61% yield(27 out of 44). Three explanations have beengiven when negative cultures were obtained:concurrence of antibiotic administration be-fore biopsy, small-bore biopsy needle pre-cluding a satisfactory retrieval of tissue sam-ples, and the natural healing of interverte-bral disc infection as postulated by Fraser49.According to Fraser’s theory49,50, vasculargranulation tissue from the vertebral sub-chondral plate invades and resorbs the in-fected disc space, enabling the infected re-


Figure 3. B.A., male, 28 years of age, tubercularspondylodiscitis at L3-L4. Magnetic resonance.

gion to heal spontaneously after approxi-mately 6 weeks, thus leading to negative cul-tures. This also explains why spondylodisci-tis can have a self-limited course in uncom-promised hosts. The inability to identify apathogen does not necessarily indicate thatno infection is or was present or that suchpatients would not benefit from empiricallychosen antibiotic therapy.

MR Imaging (Figures 3 to 5). With an ac-curacy rate of 90%, Magnetic resonanceImaging (MRI) is the preferred diagnosticimaging method for spinal osteomyelitis.MRI permits early diagnosis of infection andprovides direct visualization of the spinalcord, subarachnoid space, extradural soft tis-sue and spinal column, without intrathecalcontrast35,40. This technique may not be suit-able for patients with movement disorders,orthopedic hardware, pacemakers and cer-tain prosthetic heart valves. MRI cannot al-ways distinguish osteomyelitis from severedegenerative arthritis40,43. The earliest sign ofan infectious process on MRI is altered mar-row signal within the vertebra, caused by ac-cumulation of oedema in the vertebral body.The inflammatory tissue and the ischemic ar-

eas, which replace the bone marrow, lead to along relaxation time. MRIs are obtained withT1-weighted image and T2-weighted imagespin-echo pulse sequences. Plain or contrast-

59


Figure 4. F.A., female, 48 years of age, CT-guided nee-dle biopsy: A, insertion of needle through the vertebralpedicle; B, sagittal reconstruction; C, 3D reconstruction.

A

C

B

Figure 5. M.A.F., female, 29 years of age; the magneticresonance shows a tubercular abscess at T10-T11 withspinal cord compression.

60

enhanced sagittal (for bony structures) andtransverse (for paravertebral abscesses) stud-ies are performed. Infection is considered if:(A) the T1-weighted sequence shows de-creased signal intensity of the vertebral mar-row and intervertebral disc space, togetherwith an inability to discern a margin betweenthe disc and the adjacent vertebral marrow;(B) the T2-weighted sequence shows in-creased signal intensity of the vertebral mar-row adjacent to the involved disc and an in-crease in signal intensity from the disc itself;(C) the intravenous Gd-DTPA shows a ho-mogeneous contrast enhancement of the discand the vertebral bodies are visible. MRI al-most invariably demonstrates paravertebralsoft tissue swelling. This can extend posteri-orly into the epidural space, and postero-lat-erally into the intervertebral foramina.Epidural abscesses are defined clearly onsagittal planes as hyperintense extradural le-sions in T2-weighted images. The epiduralabscess and the dural sac are usually separat-ed by a hypointense stria, which probablycorresponds to the leptomeninges13. Abnor-mal soft tissue within the epidural space isdue to an epidural inflammatory mass ofgranulation tissue or an epidural abscess. Pre-cise definition of the epidural mass is possibleonly with the administration of intravenousGd-DTPA. A homogeneous enhancementcorresponds to inflammatory tissue withoutpurulent collection; a peripheral enhance-ment with a central hypointense area corre-sponds with a true abscess with fluid purulentcontenent13. Intravenous Gd-DTPA can high-light abnormalities within the cord. Infectiousmyelitis is uncommon. It is usually associatedwith neurological deficit without extraduralcompression and is enhanced following intra-venous Gd-DPTA.

Tuberculous spondylitis is characterized bysome peculiar findings, including a normalsignal of the intervertebral disc space, thepresence of a paraspinal soft-tissue mass, theinvolvement of many vertebral bodies and lo-calization in the posterior vertebral bodiesand arches. The size of the paraspinal mass isusually larger in tuberculosis than in otherpyogenic infections13,35. MRI is useful for thedifferential diagnosis between tuberculous in-fection and other forms of spondylodiscitis inthe chronic stage: the first case shows a slight-ly high signal on T1-weighted images in the

late stage, whereas the second shows low-sig-nal intensity13. Other differential diagnoseswere obtained with different spinal patholo-gies, including postoperative changes, spinaldegenerative disease, vertebral metastases. Inmetastatic spine localizations, the interverte-bral discs are not involved and the affectedvertebral bodies do not show contrast en-hancement13.

Therapeutic approaches

Antibiotic therapyThe therapy in osteomyelitis is significantly

addressed to eradicate established bone in-fection and to prevent the progression. Thosespecific aims are obtained by administrationat of appropriate antimicrobial drugs and, ifnecessary, by surgical approach.

Since the medical treatment must be initi-ated before identifying the responsible organ-ism and its in vitro sensitivity, the therapymust be chosen on the basis of the most com-mon causes of the infection. Therefore Beta-lactam antibiotics represent the first choicebecause they are safe, able to penetrate tothe site of infection and active against themost common pathogens causing os-teomyelitis. The broad spectrum treatmentmay be changed to a specific antimicrobialtherapy based on bone cultures51.

The optimal duration of therapy for pyo-genic spinal infections has been debated inthe literature, with several studies recom-mending 6 to 8 weeks of intravenous therapyand others recommending only 4 weeks7,11.This is the main reason for the long hospitalstay. However in account of the cost, the in-convenience of maintaining an intravenousline and danger of nosocomial infection, otherroutes can be evaluated. Anyway, parenteraldrugs, for at least a week, are often neces-saries before using only oral therapy. Howev-er there are some studies showing that aparental treatment of less than 4 weeks dura-tion is associated with a 25% relapse rate28,52.For this reason, Hadjipavlou et al6 proposed 6weeks of intravenous antibiotics followed by 6weeks of oral medication to provide an extramargin of safety. Antituberculous chemother-apy, consisting of at least two medications,should be administered for a period of at least12 months. The addition of a third agent forthe initial 2 to 6 months is also advised25.


ESR levels should be taken into account inthe first critical month, when the decision toundergo to a surgical treatment is usuallymade. Carragee et al53 suggest that as a gener-al trend, a decreasing ESR during the firstmonth of non-surgical treatment is a goodprognostic sign. However a rapid response ofless than 50% of the ESR is rarely associatedwith treatment failure and successful of con-servative treatment is seen in 40% of caseswith persistently elevated or rising ESR. Eval-uation of the clinical indicators, including areduction in back pain and constitutional re-covery, risk factors, age, immunosuppressionand stability of neurological status against theESR results may be helpful in order to assessthe efficacy of the medical treatment.

When osteomyelitis become recurrent, sur-gical approaches can be more important thanantibiotic management and it may be neces-sary to use local aggressive therapy or to re-move infected bone.

Some reports have suggest also that elimi-nation of risk factors, supplementation withcalcium, bisphosphonates, and/or vitamin D,and the treatment with testosterone and/orestrogen may promote bone repair51.

Surgical treatment The general principles for the management

of spine infections are non operative, consist-ing of external immobilization and intravenousantibiotics, followed by oral antibiotics6.

However, despite the effectiveness of med-ications, some patients require surgery.

Indications to operative debridementshould be given in case of absence of clinicalimprovement after 2-3 weeks of intravenousantibiotics, with presence of persistent backpain and systemic effects of chronic infectionsuch as malnutrition and cathexis.

Presence or progression of neurologicalcompromise may result from abscess or ver-tebral collapse, with highest risk if the age in-creased and in cervical infection28. An abscessformation is an indication for surgery also be-cause antibiotics are generally ineffective anda drainage is necessary (Figure 6)55. As theinfection becomes chronic, despite adequatemedical treatment, other two complicationsmay require surgery: a development of bio-mechanical instability and related chronicpain and/or a vertebral collapse with progres-sive deformity56.

In surgical management the infected tissueshould be thoroughly debrided and the in-fected area receive adequate blood flow to al-low for tissue healing. Spinal stabilizationshould be maintained with bone fusion or re-stored if compromised by either preoperativeinfectious process or by iatrogenically in-duced instability following decompression.

Because the infection involves disc spaceand subjacent vertebral body, an anterior ap-proach is warranted56; moreover it allowsplacement of a structural bone graft to recon-stitute lost height as well as anterior loadsharing57 (Figure 7). A posterior decompres-sion fail a direct access to the affected area(anterior spinal elements). Further posteriorbone and ligamentous structures, essential inmaintaining biomechanical, are usually unin-volved and removes this structures con-tributes for progressive deformity. Adding aposterior stabilization is indicate to betterachieves spinal stability and to correct thekyphotic deformity57. Stabilization of thespine using posterior pedicle screw has the ad-vantage of restoration of spinal alignmentwithout any communication with the infectionsite, which usually is anterior56,58. Several stud-ies54,57 demonstrate the efficacy of autologousbone grafts to achieve a solid fusion. Tradi-tional autologous bone grafts include iliaccrest, rib or fibula. Further structural allografthas used in the reconstruction of the anteriorspinal column with high fusion rates and awith a advantage of avoided the morbidity as-sociated with a second operative site59.

Early surgical decompression results inrapid improvement of neurological deficit,decrease in kyphotic deformities and stabi-lization with bony fusion. Same-day simulta-neous anterior and posterior approaches witharthrodesis and internal fixation can be suc-cessfully used in the treatment of spinal infec-tion (Figure 7). The presence of active infec-tion does not preclude the use of internal fix-ation56,57.

Follow-up strategies

After 70 days of therapy, MR imaging, TCscans and bone scintigraphy must be repeted.

If Bone Scintigraphy is still positive for in-fection, another two months of immobiliza-

61


62

tion and antibiotics are required. Laboratorytest and bone scan will be repeated until theyare negative.

If bone scintigraphy is negative for infec-tion, SPECT Ga-67 scan is performed to con-firm the recovery and if SPECT Ga-67 is nega-tive, the patient has recovered, oral antibioticsare discontinued and mobilization with orthe-sis is permitted. Clinical findings and plain ra-diography are checked after 3 and 12 months(Figure 8). If Ga-67 scan is positive, another

two months of immobilization and antibioticsare required. Laboratory test and Ga-67 scanwill be repeated until they are negative.

With clinical and imaging worsening or af-ter 4 months of conservative treatment with-out recovery, surgical approach should beconsidered.


Figure 7. B.A., female,68 years of age, spondy-lodiscitis at L3-L4; CTscan for decompressionand posterior stabiliza-tion, arthrodesis with al-lograft and plate. A,transverse; B, frontal re-construction.

A B

Figure 6. M.A.F. female, 29 years of age; tubercularspondylodiscitis at T10-T11; the tubercular abscess hasbeen sucked dry by applying a circumferential drainage,the thoracic spine being acceded via posterior ap-proach.

Figure 8. C.L., male, 53 years of age, spondylodiscitis aL1-L2. The latero-lateral x-ray shows a full vertebral fu-sion following conservative treatment, a sign that theinfection has healed.

63


Figure 9. Algorithm for diagnosis and treatment of vertebral osteomyelitis.

64

The validity of MR imaging in followingthe therapeutic response remains to be de-fined because the clinical findings may notcorrelate with the MRI findings32,34. Despitethe clinical improvement of patients, MRImay improve, worsen, or stay the same withinthe first weeks after institution of therapy.Caution should be used in interpreting thefollow-up MR-images in assessing therapeu-tic response. For this reason, there is no justi-fication for routine use of MRI to follow apatient’s response to antibiotic therapy. Thedevelopment of new radiculopathy or caudaequina syndrome should remain an indicationfor repeat MRI, to exclude the developmentof an abscess or other space-occupying lesion.

In conclusion, most of the studies in the lit-erature identified haematogenous vertebralosteomyelitis as a challenge for the physician:symptoms are not specific and sub-acute orchronic presentation is most common. In gen-eral, a delay in diagnosis is a rule rather thanthe exception. This is an easily missed infec-tious process, particularly in old people, inwhom degenerative radiographic changes andcondition resulting in back pain, such as os-teoporotic fractures or spinal metastases, arecommon and signs of sepsis may not manifest.However, persisting localized back pain andtenderness with elevated ESR should promptthe physician to also consider haematogenousvertebral osteomyelitis, although fever andleukocytosis may often not be present.

Once a haematogenous vertebral os-teomyelitis is suspected, a long series of imag-ing and laboratory tests, and if necessary sur-gical procedure, must be started. The purposeof the study is to formulate a systematic,comprehensive and simple approach to themanagement of this disease following the di-agnostic algorithm suggested (Figure 9).

References

1) THE ADULT SPINE: Principles and Practice. 2nd Edi-tion, J.W. Frymoyer. Editor-in-Chief. Lippincott-Raven Publishers, Philadelphia, 1997.

2) MALAWSKI SK, LUKAWSKI S. Pyogenic infection ofspine. Clin Orthop 1991; 272: 58-66.

3) CAHILL DW, LOVE LC, RECHTINE GR. Pyogenic os-teomyelitis of the spine in the elderly. J Neurosurg1991; 74: 878-886.

4) EMERY S, CHAN D, WOODWARD H. Treatment ofhematogenous pyogenic vertebral osteomyelitiswith anterior debridement and primary bone graft-ing. Spine 1989; 14: 284-291.

5) TYRRELL PNM, CASSAR-PULLICINO VN, MC CALL IW.Spinal infection. Eur Radiol 1999; 9: 1066-1077.

6) HADJIPAVLOU AG, MADER JT, NECESSARI JT, MUFFOLETTO

AJ. Hematogenous pyogenic spinal infections andtheir surgical management. Spine 2000; 13:1668-1679.

7) CARRAGEE E. Pyogenic vertebral osteomyelitis. JBone Joint Surg 1997; 79-A: 874-880.

8) HAAS DW, DES PREZ RW. Mycobacterium tuberculosis.In: Mandell GL, Bennett JE, Dolin R, eds. Principlesand Practice of Infectious Disease. 4th ed. NewYork: Churchill Livingstone, 1995, 2213-2250.

9) MATSUI H, HIRANO N, SAKAGUCHI Y. Vertebral os-teomyelitis: an analysis of 38 surgically treatedcases. Eur Spine J 1998; 7: 50-54.

10) BUCHELT M, LACK W, KUTSCHERA HP, ET AL. Comparisonof tuberculous and pyogenic spondylitis. Clin Or-thop 1993; 296: 192-199.

11) NOLLA JM, ARIZA J, GOMEZ-VAQUERO C, ET AL. Sponta-neus pyogenic vertebral osteomyelitis in nondrugusers. Semin Arthritis Rheu 2002; 31: 271-278.

12) PATZAKIS MJ, RAO S, WILKINS J, MOORE TM, HARVEY

PJ. Analysis of 61 cases of vertebral os-teomyelitis. Clin Orthop 1991; 264: 178-183.

13) MAIURI F, IACONETTA G, GALLICCHIO B, MANTO A, BRIG-ANTI F. Clinical and magnetic resonance diagno-sis. Spine 1997; 22: 1741-1746.

14) RASOOL MN. Primary subacute haematogenousosteomyelitis in children. J Bone Joint Surg Br2001; 8: 93-98.

15) WILSON CM, REISS-LEVY EA, Au TC. Haemophilusparaphrophilus vertebral osteomyelitis. Med JAust 1994; 160: 512-514.

16) BELZUNEGUI J, INTXAUSTI JJ, DE DIOS JR, ET AL.Haematogenous vertebral osteomyelitis in the el-derly. Clin Rheumatol 2000; 19: 344-347.

17) SORBELLO AF, O’DONNELL J, KAISER SMITH J, FIZHAIR J,DONESON S. Infective endocarditis due to Pas-teurella dagmatis: a case report and review. ClinInfect Dis 1994; 18: 336-338.

18) WILBER RB. Beta-lactam therapy of osteomyelitisand septic arthritis. Scand J Infect Dis 1984; 42(Suppl): 155-168.

19) COLMENERO JD, JIMENEZ-MEJIAS ME, SANCHEZ-LORA FJ,ET AL. Pyogenic, tuberculous and brucellar vertebralosteomyelitis: a descriptive and comparative studyof 219 cases. Ann Rheum Dis 1997; 56: 709-715.

20) TORDA AJ, GOTTLIEB T, BRADBURY R. Pyogenic verte-bral osteomyelitis: analysis of 20 cases and re-view. Clin Infect Dis 1995; 20: 320-328.

21) NICOLLE LE. Asymptomatic bacteriuria in the elder-ly. Infect Dis Clin North Am 1997; 11: 647-652.


22) CAHILL DW, LOVE LC, RECHTINE GR. Pyogenic os-teomyelitis of the spine in the elderly. J Neurosurg1991; 74: 878-886.

23) SAPICO FL, MONTGOMERIE JZ. Vertebral osteomyelitisin intravenous drug abusers: report of three cas-es and review of the literature. Rev Infect Dis1980; 2: 196-206.

24) SEXTON DJ, SPELMAN D. Current best practise andguidelines. Assessment and management ofcomplication in infective endocarditis. Infect DisNorth Am 2002; 16: 507-521.

25) NUSSBAUM ES, ROCKSWOLD GL, BERGMAN TA, ERICKSON

DL, SELJESKOG EL. Spinal tubercolosis: a diagnosticand management challange. J Neurosurg 1995;83: 243-247.

26) Perronne C, Saba J, Behloul Z, et al. Pyogenicand tuberculous spondylodiscistis in 80 adult pa-tients. Clin Infect Dis 1994; 19: 746-750.

27) CHELSOM J, SOLBERG CO. Vertebral osteomyelitis ata Norvegian University Hospital 1997: clinical fea-tures, laboratory findings and outcome. Scand JInfect Dis 1998; 30: 147-151.

28) EISMONT FJ, BOHLMAN HH, SONI PL, GOLDBERG VM,FREEHAFER AA. Pyogenic and fungal vertebral os-teomyelitis with paralysis. J Bone Joint Surg1983; 65: 19-29.

29) BURANAPATIKIT B, KIRIRATNIK T. Clinical manifestationof tuberculousus and pyogenic spine infection. JMed Assoc Thai 2001; 84: 1522-1526.

30) BONFIGLIO M, LANGE TA, KIM YM. Pyogenic verte-bral osteomyelitis. Disc space infection. Clin Or-thop 1973; 96: 234-247.

31) DIGBY JM, KERSLEY JB. Pyogenic non-tuberculousspinal infection. J Bone Joint Surg 1979; 61-B:47-55.

32) CARRAGEE EJ. The clinical use of magnetic reso-nance imaging in pyogenic vertebral os-teomyelitis. Spine 1997; 22: 780-785.

33) SHARIF HS. Role of MR Imaging in the manage-ment of spinal infection. Am J Roentgenol 1992;158: 1333-1345.

34) ENZMANN DR. Infection and inflammation. In: Mag-netic Resonance of the Spine. Enzmann DR, DeLaPaz RL, Rubin JB, Louis S, Eds. Mosby CV,Louis S, 1990: pp 260-300.

35) MODIC MT, FEIGLIN DH, PIRAINO DW, ET AL. Vertebralosteomyelitis: assessment using MR. Radiology1985; 157: 157-166.

36) BERONIUS M, BERGMAN B, ANDERSSON R. Vertebralosteomyelitis in Goteborg, Sweden: a retrospec-tive study of patients during 1990-95. Scand J In-fect Dis 2001; 33: 527-532.

37) BURANAPANITKIT B, LIM A, GEATER A. Misdiagnosisin vertebral osteomyelitis problems and fac-tors. J Med Assoc Thai 2001; 1984: 1743-1750.

38) PRICE AC, ALLEN JH, EGGERS FM, SHAFF MI, EVERETTE

JA. Intervertebral disk-space infection: CTchanges. Radiology 1983; 149: 725-729.

39) SHARIF HS, OSARUGUE AA, CLARKE DC, ET AL. Brucel-lar and Tuberculous Spondylitis: Comparativeimaging features. Radiology 1989; 171: 419-425.

40) LOVE C, PATEL M, LONNER BS, TOMAS MB, PALESTRO

CJ. Diagnosing spinal osteomyelitis. Clin NuclMed 2000; 25: 963-977.

41) GRATZ S, DORNER J, OESTMANN JW, ET AL. Ga-67 Cit-rate and Tc-99m MDP for estimating the severityof vertebral osteomyelitis. Nucl Med Commun2000; 21: 111-120.

42) ADATEPE MH, POWELL OM, ISAACS GH, NICHOLS K, CE-FOLA R. Hematogenous pyogenic vertebral os-teomyelitis: diagnostic value of radionuclide boneimaging. J Nucl Med 1986; 27: 1680-1685.

43) MCEWAN L, WONG JC. Nuclear medicine imaging inearly vertebral osteomyelitis: still of clinical utility.Australas Radiol 2000; 44: 454-457.

44) BAYER AS, BOLGER AF, TAUBERT KA, ET AL. Diagnosisand managenment of infectious endocarditis andits complications. Circulation 1998; 98: 2936-2948.

45) GRATZ G, BEHN T, HERMANN A, ET AL. Immunoscintig-raphy in neonates and infants with fever of un-known origin. Nucl Med Commun 1998; 19: 1037-1045.

46) LISBONA R, DERBEKYAN V, NOVALES-DIAZ J, VEKSLER A.Gallium-67 scintigraphy in tuberculous and non-tuberculous infectious spondylitis. J Nucl Med1993; 34: 853-859.

47) RESNICK D. Bone and Joint Imaging. 2nd Edition.W.B. Saunders Company, Philadelphia 2000.

48) CHEW FS, KLINE MJ. Diagnostic yield of CT-guidedpercutaneus aspiration procedures in suspectedspontaneus infectious discitis. Radiology 2001;218: 211-214.

49) FRASER RD, OSTI OL, VERNON-ROBERTS B. Discitis af-ter discography. J Bone Joint Surg (Br) 1987; 68:26-35.

50) FRASER RD, OSTI OL, VERNON-ROBERT B. Iatrogenicsdiscitis: the role of intravenous antibiotics in pre-vention and treatment. An experimental study.Spine 1989; 14: 1025-1032.

51) MADER JT, SHIRTLYFF ME, BERQQUIST S, CALOHOM JH.Bone and joint infections in the elderly: practicaltreatment guidelines. Drug Aging 2000; 16: 67-80.

52) SAPICO FL, MONTGOMERIE JZ. Pyogenic vertebral os-teomyelitis: report of nine cases and review of theliterature. Rev Infect Dis 1979; 1: 754-776.

53) CARRAGEE EJ, KIM D, VAN DER VLUGT T, VITTUM D.The clinical use of erythrocyte sedimentation ratein pyogenic vertebral osteomyelitis. Spine 1997;22: 2089-2093.

65


66

54) MCGUIRE RA, EISMONT FJ. The fate of autogenousbone graft in surgically treated pyogenic vertebralosteomyelitis. J Spinal Disord 1994; 7: 206-215.

55) PRZYBYLSKI GJ, ASHWINI DS. Single-stage autoge-nous bone grafting and internal fixation in the sur-gical management of pyogenic discitis and verte-bral osteomyelitis. J Neurosurg 2001; 94: 1-7.

56) SAFRAN O, NAHSHON R, KAPLAN L, SAGIV S, FLOMAN

Y. Sequential or simultaneous, same-day anteri-or decompression and posterior stabilization inthe management of vertebral osteomyelitis ofthe lumbar spine. Spine 1998; 23: 1885-1890.

57) FARAJ AA, WEBB JK. Spinal instrumentation for pri-mary pyogenic infection. Report of 31 patients.Acta Orthop Belg 2000; 66-3: 242-247.

58) SCHUSTER JM, AVELLINO AM, MANN FA, ET AL. Useof structural allografts in spinal osteomyelitis:a review of 47 cases. J Neurosurg 2000; 93: 8-14.

59) REZAI AR, WOO HH, ERRICO TJ, COOPER PR. Contem-porary management of spinal osteomyelitis. Neu-rosurgery 2000; 44: 1018-1025.


Documents

iew for Med Pharmacological Sciences 2005; 9: 53-66 ...deficit: 11 cases had intraspinal granuloma-tous tissue causing neurological dysfunction in the absence of bony destruction,