Neurological complications of brucella spondylitis Mousa AM, Bahar RH, Araj GF, Koshy TS, Muhtaseb SA, Al-Mudallal D S, Marafie AA. Neurological complications of brucella spondylitis. Acta Neurol Scand 1990: 81: 16-23. Twenty-two patients with brucella spondylitis and neurobrucellosis were studied during a 2-year period. The diagnosis was based on history of exposure, compatible signs and symptoms, high antibody titre and/or positive culture of a clinical specimen(s). Spondylitis was confirmed by plain radiographs, bone scan, CT and in some cases by histology. Neurobrucellosis was confirmed by CSF examination and culture, myelography, NCV, EMG and CT head. The spondylitis was early in 4 cases, chronic active in 12, smouldering ‘‘partially healed” in 3 and healed in 3 cases. Of these, 15 patients (68 %) had neurological complications of various types. Plain radiographs were not a good index of activity of spondylitis. Tc99 bone scan was not specific and it remained positive long after the completion of therapy. CT was superior in revealing details of bone destruction, soft tissue swelling and entrapment of nerve roots and cord. The 3 modalities were complementary. Spondylitis is commonly associated with neurobrucellosis and symptoms of one may over shadow those of the other and in some cases neurobrucellosis may be subclinical. In all cases of spondylitis, a thorough search for neurobrucellosis should be made and vice versa. Prolonged treatment with a combination of 3 antibrucella drugs is recommended and prolonged follow-up is necessary.
Brucellosis is endemic in the Arabian Peninsula and of late there has been a sharp increase in the number of reported human cases from the region (1, 2). Of several species, Br. melitensis causes the more severe disease (3). This variety appears to be endemic in Kuwait (1, 4, 5). Osteoarticular complications of brucellosis are common (10-70%) and spondylitis rates highest in incidence (2-53 % ) (3, 6-8). Neurological complications of brucellosis have long been recognized and their details frequently reported (3, 9, lo). Subacute and chronic meningitis exist in association with encephalitis, cranial neuritis, myelitis and myelopathy and are more frequently seen than “pure” acute brucellar meningitis (11). Osteoarticular and neurological complications are the main cause of morbidity in brucellosis (3, 11-13). It is generally accepted that the organisms gain access to the central nervous system (CNS) via the bloodstream (11) and predisposition to CNS brucellosis by spondylitis is well known (8, 10, 14), but there have been very few reports on the subject for almost 2 decades. Several studies on neurobrucellosis and on spondylitis have recently emerged from the Arabian Gulf region (4,15, 16) and elsewhere, but none of these 16
A. M. Mousa, R. H. Bahar, G. F. Araj, T. S. Koshy, S. A. Muhtaseb, D. S. Al-Mudallal, A. A. Marafie Al-Adan Hospital, Ministry of Public Health, Riqa and Faculty of Medicine, Kuwait University, Jabriya, Kuwait
Key words: brucellosis; neurobrucellosis; neurological manifestation of brucellosis; osteoarticular complication of brucellosis: spondylitis caused by brucellosis; undulant fever Dr Abdul Rahman Al-Mahdi Mousa, MRCP-UK, Consultant Physician, Al-Adan Hospital, Ministry of Public Health, P.O. Box 46940, 64020 Fahaheel, Kuwait Accepted for publication May 18, 1989
has looked closely into the association between spondylitis and CNS brucellosis. Localization of brucella spondylitis by bone scan, details of various radiological appearances and delineation of extent of bone and soft tissue damage by computer-assisted tomography (CT) have been reported (17-19). Although helpful in diagnosis, radiology is not always informative of activity nor pathognomonic of aetiology nor is it capable of detecting CNS brucellosis, the diagnosis of which necessitates a high degree of suspicion and special investigations. This prospective study was therefore carried out to highlight the association of the various stages of brucella spondylitis with manifest and subclinical CNS brucellosis. Material and methods
From October 1985 to December 1987, all adult patients presenting with symptoms suggestive of brucella spondylitis and/or neurobrucellosis were studied as part of an ongoing study on the disease. In each patient, the diagnosis of brucellosis was based on a positive history of exposure, compatible signs and symptoms, high antibody titre on a single
Brucella spondylitis
specimen of blood or a 4-fold rise in titre by 2 or more of the following tests: tube agglutination test (TAT), indirect immunofluorescence test (IF) and enzymelinked immunosorbent assay (ELISA), as previously described (1,4). In every patient suspected to have spondylitis, a plain AP and lateral radiographs and isotope bone scan using 99m Tc-MDP as described before (19) were performed. When either or both showed bone disease, CT of the affected bone was done. Where CT was not diagnostic and the radiologist was unable to confirm spondylitis, needle aspiration cytology was performed. According to the symptoms, their duration, signs and radiological appearances, patients were classified into 3 groups : Those suspected to have early spondylitis present with recent (less than 3 months) back pain of any degree of severity, with or without muscle spasm, restriction of movements, localized tenderness or neurological deficit. Their plain radiographs will either be normal or show straightening of spine, blurring of vertebral surfaces and/or narrowing of intervertebral disc space (IVD) but no active bone destruction or displacement of vertebral axis. Their isotope scan will show low to moderate uptake by the whole vertebral body and/or appendages. Their CT will either be normal or shows rarefaction of bone. Those suspected to have chronic active spondylitis present with longstanding (3-6 months) back pain and/or restriction of movements of any degree of severity,with or without local deformity, tenderness or neurological deficit. Their plain radiographs will show various degrees of bone involvement with narrowing of IVD, epiphysitis, geodes or gross destruction, patchy sclerosis of vertebral surfaces, remodeling of vertebral borders, syndesmophytes with or without paravertebral non-calcified soft tissue swelling or displacement of vertebral axis. Their isotope scan will show marked isotope uptake by the whole or part of the vertebral body but not by the appendages only. Their CT will show active bone destruction with associated healing. 3) Those suspected to have late spondylitis present with mild to moderate longstanding (over 6 months) back pain or stiffness with very little or no local tenderness and with or without spine deformity or neurological deficit. Their plain radiographs will show changes as in the chronic active cases but with dense sclerosis of whole or part of the vertebra, narrowing or disappearance of IVD and displacement of axis or calcification in the paravertebral soft tissue may be present. Bone scan subdivides these patients into: i) smouldering or partially healed spondylitis, in which isotope uptake will be scanty and patchy by the whole apparently diseased vertebra but not
by the appendages only; and ii) healed spondylitis, in which no apparent isotope uptake will be seen by the apparently deformed vertebra. In both smouldering and healed spondylitis, CT will show dense sclerosis with no apparent ongoing destruction. If the presence of spondylitis was confirmed, lumbar puncture was performed and the CSF examined as described before (4), whether the patient had neurological manifestations or not. If the CSF was abnormal, or the patient had neurological manifestations, CT head was done. In all patients presenting primarily with CNS manifestations, CSF was examined and CT head, spine radiographs, bone scan and where indicated CT of the affected vertebrae were done whether the patient had symptoms of spondylitis or not. Myelography, NCV and EMG were done where indicated. All the patients had complete haemograms, ESR, blood chemistry, urine analysis, and chest radiograms performed. Three to six blood cultures and at least one CSF culture for brucellosis were done for every patient. Serology and culture for other relevant diseases were done where necessary. Histological examination of needle aspirates was done in 5 cases and culture in 3 cases. Treatment and follow-up
All the patients were treated with 1 g/day streptomycin for 4-8 weeks, plus 2 of the following drugs for a minimum of 16-24 weeks: trimethoprim sulphamethoxazole (TMP-SMZ) 160 mg TMP and 800 mg SMZ every 12 h, doxycycline 200 mg/day or tetracycline 1 g every 12 h and rifampicin 600 mg per day. If compliance was good, treatment was given continuously; otherwise, a period of rest ranging from 2 to 4 weeks was allowed after every 8 weeks course of therapy. Patients were followed up after 2, 4 and 6 months and then every 6 months. Serology and at least one blood culture were repeated in every follow-up visit but radiology was repeated after 2 and 6 months and every 6 months thereafter. In addition, patients were instructed to report at the earliest sign of a relapse. Results
A total of 22 patients (14 males and 8 females) were seen during the period of the study. Their mean age was 51 years, and the mean duration of illness was 70 weeks (range 1 to 260 weeks). All females and 10 males were Arabs while the remaining 4 males were non-Arab Asians, Two of the male patients were veterinarians and the remaining males and females were shepherds or breeders or had frequently con17
Mousa et a1 Table 1. Clinical data and presentations of 22 patients with brucellar spondylitis and neurobrucellosis Radiological diagnosis
Case number Sex Agelyears)
Group 8 active
Group A early
Group C smouldering
healed
1
2
3
4
1
2
3
4
5
6
7
8
9
1 0 1 1 12
1
2
3
1
2
3
M
M
M
F
M
F
F
M
M
F
F
M
M
M
M
M
F
F
M
M
F
M
65 60 65 45 75 70 50 40 65 28 100 35
27 41 32 18
45 46 70
56 43 50
Clinical meningitis Subclinical meningitis Compressive myelopathy Non-compressive myelopathy Radiculopathy Psychiatric disorder Backache Other CNS disease Fever, chills, sweats Splenomegaly Hepatomegaly High ESR
F: female:
M: male
sumed raw milk. Table 1 shows the clinical data and the main presentations of the patients. Three groups were identified according to the presence and activity of spondylitis. Group A . Those with suspected early spondylitis. This group of young adults (mean age 30 years) had a mean duration of illness of 54 weeks. Group B. Those with active chronic spondylitis (Tables 1, 2). This group consisted of older individuals (mean age 58 years) who had a longer duration of illness (mean 72 weeks). Group C.Patients with healed or partially healed (smouldering) spondylitis. Their mean age was 51.7 years and their mean duration of illness was longest, 89 weeks.
Serology and blood culture results
All the patients had significantly raised antibody titre by 2 or more of the tests used (Table 3). TAT titres ranged from less than 20 to 5260. IF titres ranged from IgG 320 to 5120, IgM from less than 40 to 640 and IgA from less than 40 to 1280. ELISA titres ranged from IgG 1600 to over 12,800, IgM from less than 100 to 6400 and IgA from less than 100 to 12,800. A rising titre was shown by 3 cases each in groups A and B. The lowest titres of all specific antibodies were found in patients from group C. Br. melitensis was isolated from the blood in 2 cases from group A and 3 cases from group C. In group B, Br. melitensis was isolated from both blood and CSF in one patient.
Table 2. Presentations, bone scan and CT findings in patients with chronic active brucella spondylitis and neurobrucellosis (n = 12) Bone scan Case
Presentation
1 meningitis
2 3 4 5 6
meningitis hemiplegia tetraparesis tetraplegia paraplegia 7 prolonged fever 8 backache/sciatica
9 backache/sciatica 10 backache 1 1 fever t renal failure 12 backache/sciatica total patients/vertebrae % patients/vertebrae
Cervical
Dorsal
-
D4,9,10
D12,,1,-*
C5.6
-
-
DT12 Dl,
c3.4
-
-
(C,,)*
-
-
3/4 25/12
-
-
D,(7,8)** Dl,(9-1 IAP)
-
6/12 50/35
Lumbar L45 L1,2,4,5 L5
L4,Ap(L,,,)* * ILJ* * L3.4
(L3,4)* *
-
L4.5
8/16 67/47
Sacral (S,)
S,
-
S,
-
Other osteoarticular proximal femor, lower 1/3rd tibia. none elbows, tars, knees lower end femor, SIJ, STCJ, patella, ankles STCJ humerus elbows none ankles, knees STCJ, SIJ, tibias, wrists, hands hips, knee, shoulder, STCJ none
CT C/RC PTS
-
-
- + + + + +
+- +-
-
-
+ - + -
Head atrophy BGC infarct Normal Normal ND Normal Normal IH ND Normal Normal
2/2
2516
* healed by CT; * * active by CT, not detected by bone scan; C/RC: cord/root compression; PTS: paravertebral soft tissue swelling; Ap: vertebral appendages; SIJ: sacroiliac joint; STCJ: sternoclavicular joint; 8GC: basal ganglia calcification; IH: internal hydrocephalus; ND: not done.
18
Brucella spondylitis Table 3. Blood and CSF serology, SCF findings and treatment in 22 patients with brucella spondylitis and neurobrucellosis(TAT, ELlSA/lF reciprocal of titre)
CSF*
Blood Group case number A
8
1
80
2
1280
3
640
4 1
1280 320
2
640
3
1280
4
40
5
320
6
40
7
2560
8
640
9
640
10
c
TAT
2560
11
640
12
160
1
320
2
640
3
320
1
10
2
40
3
10
IgG
IgM
IgA
1600 320 12800 1280 12800 1280 12800 12800 1280 12800 1280 12800 1280 12800 1280 12800 1280 6400 160 12800 5120 12800 2560 12800 2560 6400 160 12800 1280 6400 640 800 1280 6400 320 6400 320 800 1280 1600 640 6400 160
200 80 6400 640 1600 160 1600 400 40 100 40 800 160 800 40 100 40 200 40 800 80 6400 160 800 160 800 160 1600 80 3200 640 200 40 400 40 100 40 800 40 200 40 200 80
1600 160 400 3200 12800 6400 1280 12800 12800 1600 6400 -
12800 12800 3200 12800 100 40 400 3200 400 6400 3200 200 100 3200 -
IgG
100 40 100 40 6400 320 6400 1600 160 12800 1280 3200 160 100 40 6400 160 800 40 3200 40 100 40 12800 640 200
* CSF sugar and protein mmol/ 1 and mg/ 1 respectively; doxycycline; rif = rifampicin.
-
6400 320 100 100 20 100 20 100 20 100 20 6400 160 100 20
**
IgM
IgA
TMP/ strep tetra SMZ
W8C %lymph sugar
prot
-
3.5
484
5
16
-
4.2
522
14
83
1.9
1788
4
-
89
3.9 2.5
90
rif
remarks
2
16
16
-
7
initially progressed, later improved* * improved.
17
6
17
229 2687
4 16 4 1 0
4 -
2.6
5200
3
6
13
initially 2 weeks gentamysin & progressed. * 19 improved. - reinfected 2 years* * later, same presentation. 18 improved.
-
3.5
459
6
16
-
16 improved.
-
2.8
1224
4
16
-
16
improved.
-
3
2317
8
16
6
17
improved.
80
3.6
331
-
-
-
-
died.
80
2.15
720
7
12
6
10 improved.
-
3.6
5200
4
17
-
10 improved.
21
2
1960
5
16
-
16
-
3.2
223
4
18
-
18 improved.
16
100
2
600
_
_
_
-
died.
2
-
3.8
700
7
16
-
16
improved.
2
-
3.6
355
4
18
-
16
5
-
4.6
575
4
16
-
16
relapsed after 13 weeks therapy.* * improved.
-
-
3.2
500
7
18
-
24
improved.
-
-
4.8
270
-
-
-
-
refused treatment.
52
90
2.8
1460
8
20
24
16 improved.* *
-
-
6
269
5
9
-
100 100 4 40 100 100 20 20 400 3200 128 80 100 100 100 100 122 40 320 100 800 440 40 100 800 1 40 40 100 100 5 40 40 100 400 6 40 100 800 136 40 100 1600 36 40 100 100 1 40 400 6400 300 40 100 100 4 -
Treatment (weeks1
relapsed, later improved.
-
100 3200 40 100 100 100 100 20 100 100 20 100 100 20 100 100 20 100 200 20 100 100 20
positive blood culture;
Cerebrospinal fluid (Table 3)
Group A . Only one patient had clinical meningitis with lymphocytic pleocytosis (128 WBC/mm3 with 84% lymphocytes), hypoglycorrhachia (1.9 mmolil) and high CSF protein (1788 mg/l). CSF specific antibody titre IgG, IgM and IgA were also elevated. In the remaining 3 patients who had other CNS manifestations, CSF was normal except for slight elevation of protein in 2 patients. Group B. Of the 12 patients in this group, 6 had meningitic CSF, of whom only 2 presented with
4
erratic course, compliance doubtful.**
* * * positive CSF culture. strep = streptomycin; tetra = tetracycline or
manifest meningitis. Four patients presented with paraplegia, prolonged fever with anicteric hepatitis, headache and radiculopathy and fever with renal failure respectively. CSF cell count ranged from 16 to 440 WBC/mm3, protein from 331 to 5200 mg/l and sugar from 2 to 3.6 mmol/l. CSF specific antibodies, mainly IgG and IgA, were significantly elevated in 6 cases. Of the remaining 6 cases, 5 had significantly raised CSF protein (500-5200 mg/l) and 3 of these cases had raised specific antibody titres as well. Br. melitensis was isolated from the CSF of only one patient. 19
Mousa et a1 Group C. CSF was normal in 5 of the 6 cases. One patient with healed spondylitis, however, had meningitic CSF with high brucella specific antibody titre (IgG and IgA) in association with CT-detected basal ganglia calcification. Haemogram, ESR, liver and renal functions
The haemoglobin was slightly low in group B. The WBC counts showed no significant differences in the 3 groups. The ESR was slightly raised in group A and significantly raised in group B. The mean alkaline phosphatase was significantly raised in groups A and B (121.6 and 215.6 IU/1 respectively). This was without liver function derangement in most of the cases. Renal failure was present in 5 cases in group B (reversible in 4 and irreversible in 1). The liver was enlarged in 4 cases in group B and the spleen was enlarged in 2 cases each in groups A and B. Histological examination of vertebral needle aspirates in 5 cases in group B showed nonspecific inflammatory reaction and the culture was negative in the 3 cases where it was done. Radiological findings
Group A . Plain radiographs of the spine were normal in all cases. Initial bone scan was abnormal in 3 cases and it became abnormal during the course of therapy in the fourth case. Increased isotope uptake was seen in several thoracic appendages in all 4 cases and in one of them it was also seen in both sterno-
Fig. 1. Osteoporosis with almost complete collapse of D,, with faint surfaces and preserved disc space; high isotope uptake by D, and D,, was seen on bone scan; CT Fig. 3 (case 6, group B, Tables).
20
Fig. 2 . Reduced height of C, with irregular bone destruction of its inferior surface and narrowingof relevant disc space. Superior surfaces of C, and C, are apparently intact. Note marked prevertebral soft tissue swelling (Case 5, group B, Tables).
clavicular joints. CT spine was normal in 3 cases, and in the fourth, early formation of Schmorl's nodes without bone destruction was found. Group B . Plain radiographs showed varying degrees of spondylitis. Activity was suggested by the lesser degree and patchier nature of associated sclerosis and/or presence of uncalcified paravertebral soft tissue swelling. In 8 cases, suspected active spondylitis in one region was associated with suspected healed spondylitis in another. Healed spondylitis only was suspected in 3 cases and the diagnosis was undetermined in 5 cases (Fig. 1). Paravertebral soft tissue swelling was seen in 4 cases. This was large in the cervical region (Fig. 2) and small in the lumbar region in 2 cases each. Cervical paravertebral soft tissue swelling was the only radiological evidence of activity in one case in which disc space narrowing was associated with sclerosed adjacent vertebral margins, suggesting a healed lesion. Bone scan (Table 2) showed active spondylitis in 10 cases and smouldering spondylitis in 3. CT was positive in all the cases and it delineated details of bone destruction, size and anatomic site of soft tissue swelling, disc herniation and spinal cord or root compression. CT was also a good indicator of ongoing healing, as reflected by sclerosis in and adjacent to affected bone. Where activity was missed by plain radiographs, CT showed either centrally placed destruction of bone with surrounding dense sclerosis (in 3 cases) or destruction with too little sclerosis to be detected by plain radiographs (in one case) (Fig. 3). Spondylitis confirmed by CT was most common in the lumbar region (67%) and lesser in the thoracic (50%), cervical (25%) and sacral (25%). There was no direct correlation between the
Brucella spondylitis
Discussion
Fig. 3. CT of D,, (Fig. 1); note marked destruction of bone with secondary cystic changes and obliteration with soft tissue of the subarachnoid space.
anatomic site of active spondylitis or its extent and the finding of a meningitic CSF. Group C. Plain radiographs showed lumbar or lumbosacral spondylitis in all the cases. Bone scan showed active spondylitis in one case and smouldering spondylitis in 2 others. CT, however, showed dense sclerosis with no evidence of activity in the 6 cases. CT head. This showed calcification of the basal ganglia, cerebral infarction and internal hydrocephalus in one case each in group B. Basal ganglia calcification only was seen in one case from group C. Treatment and outcome
One case refused treatment and 2 from group B died before specific therapy was started. Of these, one case died of terminal renal failure and the other of a possible cardiac arrhythmia. The fate of the patient who refused treatment is not known. Initially, the disease progressed in spite of specific treatment in 2 cases in group A (Cases 1 and 3) (Tables 1, 3). In case 3 who developed hyperoesinophilia of 45 % , Br. melitensis was resistant to streptomycin. Two other patients relapsed 13 weeks after completing a 12-week course of treatment with a combination of streptomycin, tetracycline and rifampicin (case 1, smouldering spondylitis, group C and case 9 group B). These and all the other patients responded well to prolonged drug therapy and bed rest followed by physiotherapy. In those followed up for 16-24 months, plain radiographs and CT of spine showed considerable healing with resolution of soft tissue swelling, resolution of cord compression and sclerosis of bone. Bone scan, however, continued to show varying degrees of isotope uptake.
This study has confirmed the close association between spondylitis and CNS complications of brucellosis. Of 19 cases with early, chronic active or late spondylitis, 14 (74%) had some form of neurological complications. In chronic active spondylitis alone, 83 % had neurological complications. Such complications were not as related to the extent of bone involvement or proximity to the cranium as they were to the age of the patient at onset of illness and chronicity of the spinal lesion. The high rate of neurological complications of brucellosis has been documented before (3) and it has been emphasized (8) that meningomyelitis in neurobrucellosis should not be diagnosed unless brucella myelopathy caused by compression has been excluded. Neither the haemogram nor the ESR were useful indicators of the presence or stage of spondylitis or neurobrucellosis. Patients with chronic disease, however, had lower haemoglobin and those with active spondylitis had higher ESR. Similar observations and the usefulness of a high ESR in differentiating between simple disc prolapse and brucellar disc disease have been reported (18, 20). Serology was most useful in diagnosis, as it was positive in all the cases. Blood culture was positive in 23% of cases, mostly in those with acute illness, extensive bone involvement or severe relapse. Similar rates of isolation from blood have been reported (2 l), but higher rates were obtained from culture of liver or bone marrow aspirates (22, 23). Positive CSF culture was found in 1 of 9 cases with clinical or subclinical meningitis (1 1%). CSF biochemistry and cell counts varied widely, a finding well recognized in brucellosis (9, 11, 24, 25). In one of our cases not reported here, clinically manifest meningitis was associated with normal CSF cytology and chemistry and a positive culture. Similar rare cases have been reported previously (1 1, 25-28). Predisposition to neurobrucellosis by spondylitis and the tendency of infection to localize in the meninges are well known (8, 11, 14,29, 30). Myelitis and myelopathy may be caused by compression or inflammation and the size of the granuloma or disc herniation determines whether it can be diagnosed radiologically (8, 11, 17, 31). This could explain the finding of isolated high CSF protein in some cases who had no meningitis or radiologically evident spinal block. In some of our cases who had otherwise normal CSF, the ratio of CSF titre to serum titre was higher than expected from passive transfer (32). Inflammatory vasculitis may lead to local spinal damage or breakage of the blood-brain barrier, with resultant leakage into the CSF of organisms, their toxins or other substances causing septic or aseptic meningitis or immunologically induced CNS disease. 21
Mousa et al The finding of high CSF brucella specific antibody titre in 4 of our cases may have occurred in this manner or it may have resulted from chronic meningitis in which the other CSF parameters have returned to normal. It is generally accepted that CNS brucellosis is caused by haematogenous spread (1 1) and the same probably applies to spondylitis, as was evidenced by the high incidence (67%) of associated asymptomatic involvement of other bones and joints in most of our cases, which was revealed only by bone scan. It appears that both spondylitis and meningitis may occur simultaneously during the acute stage and symptoms of one may mask those of the other and the diagnosis may be difficult. Also spondylitis and/or radiculitis may coexist with simple disc disease, giving rise to further diagnostic difficulties and pitfalls that have frequently been pointed out (2,3,6, 8, 11, 18, 20, 23, 31). The earliest brucellar spondylitic lesions begin in bone (6, 7) rather than in the intervertebral discs (IVD) (3, 13, 18, 20, 24), whose lesions, because of poor regeneration, develop faster, giving a false impression that the IVD were involved first. Plain radiographs and CT may be completely normal during the acute phase, but bone scan may show various patterns of increased isotope uptake that may be reversed by treatment (19). Disc lesions are very common in brucellosis but they may be so slight that they appear radiologically normal (6, 8, 17). Pure disc lesions are probably very rare, as apparently normal adjacent vertebrae may be shown to be the site of disease by bone scan and/or CT, as occurred in some of our cases. It is probable that brucellar spondylitis occurs among young people as frequently as it does among old people, but the infection is cleared from bone successfully in the vast majority of cases. Only in a minority of cases, small pockets of infection may persist as a source of relapses or of local bone or soft tissue disease. Among old people, bone regeneration becomes poor and the bone becomes locus minoris resistentiae, for which brucellae have a special predeliction (33). Because of the inadequate healing, clearing of infection from bone is difficult and the disease progresses into a chronic active or smouldering symptomatic or asymptomatic spondylitis. In the latter condition, because of heavy bone sclerosis, activity is detected only by bone scan and formation of large granulomata is unlikely, but the lesions may remain the source of relapses or microscopic abscess formation. Meningitis may not be commonly seen with this form of spondylitis, but radiculopathy caused by root irritation or compression by tiny abscess or disc herniation may occur, as was seen in 2 of our cases. Culture positive microabscesses and root hernia22
tions in association with apparently healed spondylitis have been reported (8,13, 15,27). Radiculitis may also occur as a result of arachnoiditis, localized neuritis or demyelination (13, 3 1). Syndromes of cervicobrachialgia and lumbosciatica without disc prolapse have been reported (1 1). Isolated pronounced atrophy of the shoulder muscles, as seen in one of our cases (case 3, smouldering spondylitis) (Table l), is reportedly a very rare complication of brucellosis (34). Patients with symptomatic smouldering spondylitis should therefore be investigated and treated as energetically as patients with chronic active spondylitis. In chronic active spondylitis, bone destruction continues, with varying degrees of ineffective healing with characteristic radiological appearances that may appear months after the onset of infection (3, 6-8, 33). Of these, paravertebral soft tissue swelling is probably the only consistent indicator of activity. Differentiation from osteoarthrosis, osteoporosis, osteomalacia, actinomycosis, Scheuermann’s disease and cancer deposits may at times be difficult (6, 8, 17, 18). Spinal tuberculosis (TB, Pott’s disease) presents a real challenge, as the 2 conditions are so similar that brucellar spondylitis has been aptly called pseudo-Pott’s disease (33). History, social environment and a positive tuberculin test may be helpful but their value is limited in areas endemic for both diseases. In brucellosis, localized, referred and girdle pains are more pronounced; patients may therefore present earlier, sometimes before radiological changes become apparent. In TB, pain is less marked and in the vast majority of cases advanced radiological changes are apparent at presentation, although in TB such changes take longer to manifest themselves after the onset of infection. The course in TB is more protracted, destruction and general rarefaction of bone are more extensive, healing takes longer to set in and sclerosis and parrot-beak osteophytes are not a feature. In TB, usually a single vertebra and disc may be involved, very rarely multiple and contiguous vertebral lesions occur but skip lesions and the involvement of more than one spinal region at a time are not known. Gibbus and collapse of vertebral bodies and later ankylosing are more common in TB. Paravertebral soft tissue abscesses are commoner in TB, occurring in 50% of cases, and they are usually larger than in brucellosis. Bone scan in brucellosis may reveal multiple involvement of other bones and joints, which is not seen in TB. The age incidence is different, as TB is more commonly seen in younger patients (6-8, 17, 18, 35). Plain radiographs may therefore not be pathognomonic and may miss active disease. Bone scan is not specific but is a good indicator of activity and extent of skeletal involvement. It is not useful, however, for short-term follow-up, as it remains positive
Brucella spondylitis
for a long time (6). CT may also not be diagnostic, but it is superior in delineating details of soft tissue, bone, cord and root lesions. Minor degrees of sclerosis missed by plain radiographs and centrally placed bone lesions are also shown better by CT. The 3 radiological modalities are complementary. In a few cases, diagnosis may not be possible without histological examination of bone or recovering the organism. In this series, both early and chronic active spondylitis were associated with a large variety of CNS lesions. Acute meningitis was mainly associated with early spondylitis affecting young adults. Subacute, subclinical and chronic meningitis, parkinsonism, deafness, papilloedema, basal ganglia calcification, cerebral vasculitis with TIA, paraplegia, tetraplegia, radiculitis and psychiatric manifestations were seen in association with chronic active and smouldering spondylitis affecting mainly middle-age and elderly patients. All these are well recognized complications of brucellosis (8-1 l), and the finding of CNS manifestations in brucellosis should therefore prompt a search for possible underlying spondylitis and vice versa. Treatment of patients should be prolonged since the eradication of the organism from bone and CSF may be difficult (13, 15,33). Our present policy is to use a combination of 3 drugs for a minimum of 16-24 weeks (1, 4, 5). Non-compliance or reinfection usually causes failure of therapy. References 1. MOUSAAM, ELHAGKM, KHOGALIM, MARAFIEAA. The nature of human brucellosis in Kuwait: a study of 379 cases. Rev Infect Dis 1988: 10: 211-217. 2. NORTONWL. Brucellosis and rheumatic syndromes in Saudi Arabia. Ann Rheum Dis 1984: 43: 810-815. 3. SPINKWW. The nature of brucellosis. Minneapolis: University of Minnesota Press, 1956: 145-170. 4. MOUSA AM, KOSHY TS, ARAI GF, MARAFIEAA, MUHTASEBSA, AL-MUDALLAL DS, BUSHARATULLA MS. Brucella meningitis: presentation, diagnosis and treatment: aprospective study often cases. Q J Med (New Series) 1986: 233: 873-885. 5. MOUSAAM, MUHTASEB SA, AL-MUDALLAL DS, KHODEIR SM, MARAFIEAA. Osteoarticular complications of brucellosis: a study of 169 cases. Rev Infect Dis 1987: 9: 531-543. J et al. Brucella spondylitis; 6. ARIZAJ, GUDIOLF, VALVERDE a detailed analysis based on current findings. Rev Infect Dis 1985: 7: 656-664. J. Osteoarticular sites of brucellosis. Ann 7. ROTES-QUEROL Rheum Dis 1957: 16: 63-68. 8. GANADOW, CRAIGAJ. Brucellosis myelopathy. J Bone Joint Surg (Am) 1958: 40: 1380-1388. RW, SAHSAL, JOYNT RJ.Protean manifestations 9. FINCHAM of central nervous system brucellosis: case histories of a wide variety of clinical forms. JAMA 1963: 184: 269-275. A. Neurological complications of undulant 10. NELSON-JONES fever. Clinical picture. Lancet 1951: 1: 495-497. 11. SAHS AL. Brucellosis (Malta fever, undulant fever). In:
VINKENPJ, BRUYNGW, eds. Handbook of clinical neurology. Amsterdam: North Holland, 1978: 33: 305-326. 12. LAM K, SILVERSTEINLM, CARLISLERJ, BAYER AS. Disseminated brucellosis initially seen as sternoclavicular arthropathy. Arch Intern Med 1982: 142: 1193-1 194. 13. YOUNGEJ. Human brucellosis. Rev Infect Dis 1983: 5: 821-843. NS, SACHDEVA JR, CHOPRA JS. Brucella menin14. PATHANIA gitis and spondylitis. J Indian Med Assoc 1966: 47: 290-292. 15. BASHIRR, AL-KAWIZ, HARDEREJ, JINKINSJ. Nervous system brucellosis: diagnosis and treatment. Neurology 1985: 35: 1567-1581. RA, AL-DINAS, ARAI GF, LULUAR, MOUSAAM, 16. SHAKIR SAADA MA. Clinical categories ofneurobrucellosis. A report of nineteen cases. Brain 1987: 110: 213-223. F. Undulant fever spondylitis. Br J Radio1 1958: 3 1: 17. ZAMMIT 683-690. MMS. Brucellosis of the spine. Br J Surg 1976: 18. GLASGOW 63: 283-288. 19 BAHARRH, AL-SUHAILIAR, MOUSAAM, NAWAZMK, KHADDAHN, ABDUL-DAYEMHM. Brucellosis: appearances on skeletal imaging. Clin Nucl Med 1988: 13: 102-106. 20. KELLYPJ, MARTINkJ7 SCHIRGERA, WEED LA. Brucellosis of bones and joints. Experiences with 36 patients. JAMA 1960: 174: 347-353. 21. WILSONGS, MILES AA. Topley & Wilson principles of bacteriology, virology and immunity, 5th ed. London: Edward Arnold, 1964. 22. BUCHANAN TM, FABER LC, FELDMAN RA. Brucellosis in the United States 1960-1972:an abattoir-associated disease. I. Clinical features and therapy. Medicine (Baltimore) 1974: 53: 403-413. 23. LIFESORM, HARDERE, MCCORKELL SJ. Spinal brucellosis. J. Bone Joint Surg (Br) 1985: 67: 345-351. 24. JOGLEKARMD, NAGALOTIMATH SJ. Neurobrucellosis. J Assoc Physicians India 1976: 24: 183-186. 25. LEYSDG. Abortus fever. Br Med J 1943: 1: 187-189. 26. NICJOLSE. Meningoencephalitis due to brucellosis: report of a case in which Brucella abortus was recovered from the CSF and a review of the literature. Ann Intern Med 1951: 35: 673-693. 27. MCCULLAGHEP, CLODFELTER HM. Encephalitis due to undulant fever; report of four cases. Ann Intern Med 1937: 10: 1508-1513. 28. BLACKRA. Brucella meningitis in an infant. Arch Paediat 1937: 54: 702-713. JR. Melitensis meningo29. HANSMANNGH & SCHENKEN encephalitis. Mycotic aneurysm due to Brucella melitensis var porcine. Am J Pathol 1932: 8: 435-444. 30. SHARPWB. Pathology of undulant fever. Arch Pathol 1934: 18: 72-107. J. Epidemiology and neurological 31. DEVILLAFANE-LASTRA manifestations of brucellosis. In: VAN BOGAERTL, KAFER JP, POCHGF, eds. Tropical neurology. Buenos Aires: Lopez Libreros, 1963: 302-325. WW, MA BU. Multiple sclerosis: the blood 32. TOURTELLOITE brain barrier and the measurement of de novo central nervous system IgG synthesis. Neurology (NY) 1978: 28 (SUPPI):76-83. 33. AGUILARJA, ELVIDGEAR. Intervertebral disc disease caused by the brucella organism. J Neurosurg 1961: 18: 27-33. V. Vorlautige mitteilung uber einen fall von febris 34. JOHNSSON undulans bang mit neurologischen komplikationen. Med Klin 1929: 25: 389. A, GAINES VD. The radiology of osteo35. BONAKDER-POUR myelitis. Orthopaed Clin North Am 1983: 14: 21-37.
23
Brucellosis is endemic in the Arabian Peninsula and of late there has been a sharp increase in the number of reported human cases from the region (1, 2). Of several species, Br. melitensis causes the more severe disease (3). This variety appears to be endemic in Kuwait (1, 4, 5). Osteoarticular complications of brucellosis are common (10-70%) and spondylitis rates highest in incidence (2-53 % ) (3, 6-8). Neurological complications of brucellosis have long been recognized and their details frequently reported (3, 9, lo). Subacute and chronic meningitis exist in association with encephalitis, cranial neuritis, myelitis and myelopathy and are more frequently seen than “pure” acute brucellar meningitis (11). Osteoarticular and neurological complications are the main cause of morbidity in brucellosis (3, 11-13). It is generally accepted that the organisms gain access to the central nervous system (CNS) via the bloodstream (11) and predisposition to CNS brucellosis by spondylitis is well known (8, 10, 14), but there have been very few reports on the subject for almost 2 decades. Several studies on neurobrucellosis and on spondylitis have recently emerged from the Arabian Gulf region (4,15, 16) and elsewhere, but none of these 16
A. M. Mousa, R. H. Bahar, G. F. Araj, T. S. Koshy, S. A. Muhtaseb, D. S. Al-Mudallal, A. A. Marafie Al-Adan Hospital, Ministry of Public Health, Riqa and Faculty of Medicine, Kuwait University, Jabriya, Kuwait
Key words: brucellosis; neurobrucellosis; neurological manifestation of brucellosis; osteoarticular complication of brucellosis: spondylitis caused by brucellosis; undulant fever Dr Abdul Rahman Al-Mahdi Mousa, MRCP-UK, Consultant Physician, Al-Adan Hospital, Ministry of Public Health, P.O. Box 46940, 64020 Fahaheel, Kuwait Accepted for publication May 18, 1989
has looked closely into the association between spondylitis and CNS brucellosis. Localization of brucella spondylitis by bone scan, details of various radiological appearances and delineation of extent of bone and soft tissue damage by computer-assisted tomography (CT) have been reported (17-19). Although helpful in diagnosis, radiology is not always informative of activity nor pathognomonic of aetiology nor is it capable of detecting CNS brucellosis, the diagnosis of which necessitates a high degree of suspicion and special investigations. This prospective study was therefore carried out to highlight the association of the various stages of brucella spondylitis with manifest and subclinical CNS brucellosis. Material and methods
From October 1985 to December 1987, all adult patients presenting with symptoms suggestive of brucella spondylitis and/or neurobrucellosis were studied as part of an ongoing study on the disease. In each patient, the diagnosis of brucellosis was based on a positive history of exposure, compatible signs and symptoms, high antibody titre on a single
Brucella spondylitis
specimen of blood or a 4-fold rise in titre by 2 or more of the following tests: tube agglutination test (TAT), indirect immunofluorescence test (IF) and enzymelinked immunosorbent assay (ELISA), as previously described (1,4). In every patient suspected to have spondylitis, a plain AP and lateral radiographs and isotope bone scan using 99m Tc-MDP as described before (19) were performed. When either or both showed bone disease, CT of the affected bone was done. Where CT was not diagnostic and the radiologist was unable to confirm spondylitis, needle aspiration cytology was performed. According to the symptoms, their duration, signs and radiological appearances, patients were classified into 3 groups : Those suspected to have early spondylitis present with recent (less than 3 months) back pain of any degree of severity, with or without muscle spasm, restriction of movements, localized tenderness or neurological deficit. Their plain radiographs will either be normal or show straightening of spine, blurring of vertebral surfaces and/or narrowing of intervertebral disc space (IVD) but no active bone destruction or displacement of vertebral axis. Their isotope scan will show low to moderate uptake by the whole vertebral body and/or appendages. Their CT will either be normal or shows rarefaction of bone. Those suspected to have chronic active spondylitis present with longstanding (3-6 months) back pain and/or restriction of movements of any degree of severity,with or without local deformity, tenderness or neurological deficit. Their plain radiographs will show various degrees of bone involvement with narrowing of IVD, epiphysitis, geodes or gross destruction, patchy sclerosis of vertebral surfaces, remodeling of vertebral borders, syndesmophytes with or without paravertebral non-calcified soft tissue swelling or displacement of vertebral axis. Their isotope scan will show marked isotope uptake by the whole or part of the vertebral body but not by the appendages only. Their CT will show active bone destruction with associated healing. 3) Those suspected to have late spondylitis present with mild to moderate longstanding (over 6 months) back pain or stiffness with very little or no local tenderness and with or without spine deformity or neurological deficit. Their plain radiographs will show changes as in the chronic active cases but with dense sclerosis of whole or part of the vertebra, narrowing or disappearance of IVD and displacement of axis or calcification in the paravertebral soft tissue may be present. Bone scan subdivides these patients into: i) smouldering or partially healed spondylitis, in which isotope uptake will be scanty and patchy by the whole apparently diseased vertebra but not
by the appendages only; and ii) healed spondylitis, in which no apparent isotope uptake will be seen by the apparently deformed vertebra. In both smouldering and healed spondylitis, CT will show dense sclerosis with no apparent ongoing destruction. If the presence of spondylitis was confirmed, lumbar puncture was performed and the CSF examined as described before (4), whether the patient had neurological manifestations or not. If the CSF was abnormal, or the patient had neurological manifestations, CT head was done. In all patients presenting primarily with CNS manifestations, CSF was examined and CT head, spine radiographs, bone scan and where indicated CT of the affected vertebrae were done whether the patient had symptoms of spondylitis or not. Myelography, NCV and EMG were done where indicated. All the patients had complete haemograms, ESR, blood chemistry, urine analysis, and chest radiograms performed. Three to six blood cultures and at least one CSF culture for brucellosis were done for every patient. Serology and culture for other relevant diseases were done where necessary. Histological examination of needle aspirates was done in 5 cases and culture in 3 cases. Treatment and follow-up
All the patients were treated with 1 g/day streptomycin for 4-8 weeks, plus 2 of the following drugs for a minimum of 16-24 weeks: trimethoprim sulphamethoxazole (TMP-SMZ) 160 mg TMP and 800 mg SMZ every 12 h, doxycycline 200 mg/day or tetracycline 1 g every 12 h and rifampicin 600 mg per day. If compliance was good, treatment was given continuously; otherwise, a period of rest ranging from 2 to 4 weeks was allowed after every 8 weeks course of therapy. Patients were followed up after 2, 4 and 6 months and then every 6 months. Serology and at least one blood culture were repeated in every follow-up visit but radiology was repeated after 2 and 6 months and every 6 months thereafter. In addition, patients were instructed to report at the earliest sign of a relapse. Results
A total of 22 patients (14 males and 8 females) were seen during the period of the study. Their mean age was 51 years, and the mean duration of illness was 70 weeks (range 1 to 260 weeks). All females and 10 males were Arabs while the remaining 4 males were non-Arab Asians, Two of the male patients were veterinarians and the remaining males and females were shepherds or breeders or had frequently con17
Mousa et a1 Table 1. Clinical data and presentations of 22 patients with brucellar spondylitis and neurobrucellosis Radiological diagnosis
Case number Sex Agelyears)
Group 8 active
Group A early
Group C smouldering
healed
1
2
3
4
1
2
3
4
5
6
7
8
9
1 0 1 1 12
1
2
3
1
2
3
M
M
M
F
M
F
F
M
M
F
F
M
M
M
M
M
F
F
M
M
F
M
65 60 65 45 75 70 50 40 65 28 100 35
27 41 32 18
45 46 70
56 43 50
Clinical meningitis Subclinical meningitis Compressive myelopathy Non-compressive myelopathy Radiculopathy Psychiatric disorder Backache Other CNS disease Fever, chills, sweats Splenomegaly Hepatomegaly High ESR
F: female:
M: male
sumed raw milk. Table 1 shows the clinical data and the main presentations of the patients. Three groups were identified according to the presence and activity of spondylitis. Group A . Those with suspected early spondylitis. This group of young adults (mean age 30 years) had a mean duration of illness of 54 weeks. Group B. Those with active chronic spondylitis (Tables 1, 2). This group consisted of older individuals (mean age 58 years) who had a longer duration of illness (mean 72 weeks). Group C.Patients with healed or partially healed (smouldering) spondylitis. Their mean age was 51.7 years and their mean duration of illness was longest, 89 weeks.
Serology and blood culture results
All the patients had significantly raised antibody titre by 2 or more of the tests used (Table 3). TAT titres ranged from less than 20 to 5260. IF titres ranged from IgG 320 to 5120, IgM from less than 40 to 640 and IgA from less than 40 to 1280. ELISA titres ranged from IgG 1600 to over 12,800, IgM from less than 100 to 6400 and IgA from less than 100 to 12,800. A rising titre was shown by 3 cases each in groups A and B. The lowest titres of all specific antibodies were found in patients from group C. Br. melitensis was isolated from the blood in 2 cases from group A and 3 cases from group C. In group B, Br. melitensis was isolated from both blood and CSF in one patient.
Table 2. Presentations, bone scan and CT findings in patients with chronic active brucella spondylitis and neurobrucellosis (n = 12) Bone scan Case
Presentation
1 meningitis
2 3 4 5 6
meningitis hemiplegia tetraparesis tetraplegia paraplegia 7 prolonged fever 8 backache/sciatica
9 backache/sciatica 10 backache 1 1 fever t renal failure 12 backache/sciatica total patients/vertebrae % patients/vertebrae
Cervical
Dorsal
-
D4,9,10
D12,,1,-*
C5.6
-
-
DT12 Dl,
c3.4
-
-
(C,,)*
-
-
3/4 25/12
-
-
D,(7,8)** Dl,(9-1 IAP)
-
6/12 50/35
Lumbar L45 L1,2,4,5 L5
L4,Ap(L,,,)* * ILJ* * L3.4
(L3,4)* *
-
L4.5
8/16 67/47
Sacral (S,)
S,
-
S,
-
Other osteoarticular proximal femor, lower 1/3rd tibia. none elbows, tars, knees lower end femor, SIJ, STCJ, patella, ankles STCJ humerus elbows none ankles, knees STCJ, SIJ, tibias, wrists, hands hips, knee, shoulder, STCJ none
CT C/RC PTS
-
-
- + + + + +
+- +-
-
-
+ - + -
Head atrophy BGC infarct Normal Normal ND Normal Normal IH ND Normal Normal
2/2
2516
* healed by CT; * * active by CT, not detected by bone scan; C/RC: cord/root compression; PTS: paravertebral soft tissue swelling; Ap: vertebral appendages; SIJ: sacroiliac joint; STCJ: sternoclavicular joint; 8GC: basal ganglia calcification; IH: internal hydrocephalus; ND: not done.
18
Brucella spondylitis Table 3. Blood and CSF serology, SCF findings and treatment in 22 patients with brucella spondylitis and neurobrucellosis(TAT, ELlSA/lF reciprocal of titre)
CSF*
Blood Group case number A
8
1
80
2
1280
3
640
4 1
1280 320
2
640
3
1280
4
40
5
320
6
40
7
2560
8
640
9
640
10
c
TAT
2560
11
640
12
160
1
320
2
640
3
320
1
10
2
40
3
10
IgG
IgM
IgA
1600 320 12800 1280 12800 1280 12800 12800 1280 12800 1280 12800 1280 12800 1280 12800 1280 6400 160 12800 5120 12800 2560 12800 2560 6400 160 12800 1280 6400 640 800 1280 6400 320 6400 320 800 1280 1600 640 6400 160
200 80 6400 640 1600 160 1600 400 40 100 40 800 160 800 40 100 40 200 40 800 80 6400 160 800 160 800 160 1600 80 3200 640 200 40 400 40 100 40 800 40 200 40 200 80
1600 160 400 3200 12800 6400 1280 12800 12800 1600 6400 -
12800 12800 3200 12800 100 40 400 3200 400 6400 3200 200 100 3200 -
IgG
100 40 100 40 6400 320 6400 1600 160 12800 1280 3200 160 100 40 6400 160 800 40 3200 40 100 40 12800 640 200
* CSF sugar and protein mmol/ 1 and mg/ 1 respectively; doxycycline; rif = rifampicin.
-
6400 320 100 100 20 100 20 100 20 100 20 6400 160 100 20
**
IgM
IgA
TMP/ strep tetra SMZ
W8C %lymph sugar
prot
-
3.5
484
5
16
-
4.2
522
14
83
1.9
1788
4
-
89
3.9 2.5
90
rif
remarks
2
16
16
-
7
initially progressed, later improved* * improved.
17
6
17
229 2687
4 16 4 1 0
4 -
2.6
5200
3
6
13
initially 2 weeks gentamysin & progressed. * 19 improved. - reinfected 2 years* * later, same presentation. 18 improved.
-
3.5
459
6
16
-
16 improved.
-
2.8
1224
4
16
-
16
improved.
-
3
2317
8
16
6
17
improved.
80
3.6
331
-
-
-
-
died.
80
2.15
720
7
12
6
10 improved.
-
3.6
5200
4
17
-
10 improved.
21
2
1960
5
16
-
16
-
3.2
223
4
18
-
18 improved.
16
100
2
600
_
_
_
-
died.
2
-
3.8
700
7
16
-
16
improved.
2
-
3.6
355
4
18
-
16
5
-
4.6
575
4
16
-
16
relapsed after 13 weeks therapy.* * improved.
-
-
3.2
500
7
18
-
24
improved.
-
-
4.8
270
-
-
-
-
refused treatment.
52
90
2.8
1460
8
20
24
16 improved.* *
-
-
6
269
5
9
-
100 100 4 40 100 100 20 20 400 3200 128 80 100 100 100 100 122 40 320 100 800 440 40 100 800 1 40 40 100 100 5 40 40 100 400 6 40 100 800 136 40 100 1600 36 40 100 100 1 40 400 6400 300 40 100 100 4 -
Treatment (weeks1
relapsed, later improved.
-
100 3200 40 100 100 100 100 20 100 100 20 100 100 20 100 100 20 100 200 20 100 100 20
positive blood culture;
Cerebrospinal fluid (Table 3)
Group A . Only one patient had clinical meningitis with lymphocytic pleocytosis (128 WBC/mm3 with 84% lymphocytes), hypoglycorrhachia (1.9 mmolil) and high CSF protein (1788 mg/l). CSF specific antibody titre IgG, IgM and IgA were also elevated. In the remaining 3 patients who had other CNS manifestations, CSF was normal except for slight elevation of protein in 2 patients. Group B. Of the 12 patients in this group, 6 had meningitic CSF, of whom only 2 presented with
4
erratic course, compliance doubtful.**
* * * positive CSF culture. strep = streptomycin; tetra = tetracycline or
manifest meningitis. Four patients presented with paraplegia, prolonged fever with anicteric hepatitis, headache and radiculopathy and fever with renal failure respectively. CSF cell count ranged from 16 to 440 WBC/mm3, protein from 331 to 5200 mg/l and sugar from 2 to 3.6 mmol/l. CSF specific antibodies, mainly IgG and IgA, were significantly elevated in 6 cases. Of the remaining 6 cases, 5 had significantly raised CSF protein (500-5200 mg/l) and 3 of these cases had raised specific antibody titres as well. Br. melitensis was isolated from the CSF of only one patient. 19
Mousa et a1 Group C. CSF was normal in 5 of the 6 cases. One patient with healed spondylitis, however, had meningitic CSF with high brucella specific antibody titre (IgG and IgA) in association with CT-detected basal ganglia calcification. Haemogram, ESR, liver and renal functions
The haemoglobin was slightly low in group B. The WBC counts showed no significant differences in the 3 groups. The ESR was slightly raised in group A and significantly raised in group B. The mean alkaline phosphatase was significantly raised in groups A and B (121.6 and 215.6 IU/1 respectively). This was without liver function derangement in most of the cases. Renal failure was present in 5 cases in group B (reversible in 4 and irreversible in 1). The liver was enlarged in 4 cases in group B and the spleen was enlarged in 2 cases each in groups A and B. Histological examination of vertebral needle aspirates in 5 cases in group B showed nonspecific inflammatory reaction and the culture was negative in the 3 cases where it was done. Radiological findings
Group A . Plain radiographs of the spine were normal in all cases. Initial bone scan was abnormal in 3 cases and it became abnormal during the course of therapy in the fourth case. Increased isotope uptake was seen in several thoracic appendages in all 4 cases and in one of them it was also seen in both sterno-
Fig. 1. Osteoporosis with almost complete collapse of D,, with faint surfaces and preserved disc space; high isotope uptake by D, and D,, was seen on bone scan; CT Fig. 3 (case 6, group B, Tables).
20
Fig. 2 . Reduced height of C, with irregular bone destruction of its inferior surface and narrowingof relevant disc space. Superior surfaces of C, and C, are apparently intact. Note marked prevertebral soft tissue swelling (Case 5, group B, Tables).
clavicular joints. CT spine was normal in 3 cases, and in the fourth, early formation of Schmorl's nodes without bone destruction was found. Group B . Plain radiographs showed varying degrees of spondylitis. Activity was suggested by the lesser degree and patchier nature of associated sclerosis and/or presence of uncalcified paravertebral soft tissue swelling. In 8 cases, suspected active spondylitis in one region was associated with suspected healed spondylitis in another. Healed spondylitis only was suspected in 3 cases and the diagnosis was undetermined in 5 cases (Fig. 1). Paravertebral soft tissue swelling was seen in 4 cases. This was large in the cervical region (Fig. 2) and small in the lumbar region in 2 cases each. Cervical paravertebral soft tissue swelling was the only radiological evidence of activity in one case in which disc space narrowing was associated with sclerosed adjacent vertebral margins, suggesting a healed lesion. Bone scan (Table 2) showed active spondylitis in 10 cases and smouldering spondylitis in 3. CT was positive in all the cases and it delineated details of bone destruction, size and anatomic site of soft tissue swelling, disc herniation and spinal cord or root compression. CT was also a good indicator of ongoing healing, as reflected by sclerosis in and adjacent to affected bone. Where activity was missed by plain radiographs, CT showed either centrally placed destruction of bone with surrounding dense sclerosis (in 3 cases) or destruction with too little sclerosis to be detected by plain radiographs (in one case) (Fig. 3). Spondylitis confirmed by CT was most common in the lumbar region (67%) and lesser in the thoracic (50%), cervical (25%) and sacral (25%). There was no direct correlation between the
Brucella spondylitis
Discussion
Fig. 3. CT of D,, (Fig. 1); note marked destruction of bone with secondary cystic changes and obliteration with soft tissue of the subarachnoid space.
anatomic site of active spondylitis or its extent and the finding of a meningitic CSF. Group C. Plain radiographs showed lumbar or lumbosacral spondylitis in all the cases. Bone scan showed active spondylitis in one case and smouldering spondylitis in 2 others. CT, however, showed dense sclerosis with no evidence of activity in the 6 cases. CT head. This showed calcification of the basal ganglia, cerebral infarction and internal hydrocephalus in one case each in group B. Basal ganglia calcification only was seen in one case from group C. Treatment and outcome
One case refused treatment and 2 from group B died before specific therapy was started. Of these, one case died of terminal renal failure and the other of a possible cardiac arrhythmia. The fate of the patient who refused treatment is not known. Initially, the disease progressed in spite of specific treatment in 2 cases in group A (Cases 1 and 3) (Tables 1, 3). In case 3 who developed hyperoesinophilia of 45 % , Br. melitensis was resistant to streptomycin. Two other patients relapsed 13 weeks after completing a 12-week course of treatment with a combination of streptomycin, tetracycline and rifampicin (case 1, smouldering spondylitis, group C and case 9 group B). These and all the other patients responded well to prolonged drug therapy and bed rest followed by physiotherapy. In those followed up for 16-24 months, plain radiographs and CT of spine showed considerable healing with resolution of soft tissue swelling, resolution of cord compression and sclerosis of bone. Bone scan, however, continued to show varying degrees of isotope uptake.
This study has confirmed the close association between spondylitis and CNS complications of brucellosis. Of 19 cases with early, chronic active or late spondylitis, 14 (74%) had some form of neurological complications. In chronic active spondylitis alone, 83 % had neurological complications. Such complications were not as related to the extent of bone involvement or proximity to the cranium as they were to the age of the patient at onset of illness and chronicity of the spinal lesion. The high rate of neurological complications of brucellosis has been documented before (3) and it has been emphasized (8) that meningomyelitis in neurobrucellosis should not be diagnosed unless brucella myelopathy caused by compression has been excluded. Neither the haemogram nor the ESR were useful indicators of the presence or stage of spondylitis or neurobrucellosis. Patients with chronic disease, however, had lower haemoglobin and those with active spondylitis had higher ESR. Similar observations and the usefulness of a high ESR in differentiating between simple disc prolapse and brucellar disc disease have been reported (18, 20). Serology was most useful in diagnosis, as it was positive in all the cases. Blood culture was positive in 23% of cases, mostly in those with acute illness, extensive bone involvement or severe relapse. Similar rates of isolation from blood have been reported (2 l), but higher rates were obtained from culture of liver or bone marrow aspirates (22, 23). Positive CSF culture was found in 1 of 9 cases with clinical or subclinical meningitis (1 1%). CSF biochemistry and cell counts varied widely, a finding well recognized in brucellosis (9, 11, 24, 25). In one of our cases not reported here, clinically manifest meningitis was associated with normal CSF cytology and chemistry and a positive culture. Similar rare cases have been reported previously (1 1, 25-28). Predisposition to neurobrucellosis by spondylitis and the tendency of infection to localize in the meninges are well known (8, 11, 14,29, 30). Myelitis and myelopathy may be caused by compression or inflammation and the size of the granuloma or disc herniation determines whether it can be diagnosed radiologically (8, 11, 17, 31). This could explain the finding of isolated high CSF protein in some cases who had no meningitis or radiologically evident spinal block. In some of our cases who had otherwise normal CSF, the ratio of CSF titre to serum titre was higher than expected from passive transfer (32). Inflammatory vasculitis may lead to local spinal damage or breakage of the blood-brain barrier, with resultant leakage into the CSF of organisms, their toxins or other substances causing septic or aseptic meningitis or immunologically induced CNS disease. 21
Mousa et al The finding of high CSF brucella specific antibody titre in 4 of our cases may have occurred in this manner or it may have resulted from chronic meningitis in which the other CSF parameters have returned to normal. It is generally accepted that CNS brucellosis is caused by haematogenous spread (1 1) and the same probably applies to spondylitis, as was evidenced by the high incidence (67%) of associated asymptomatic involvement of other bones and joints in most of our cases, which was revealed only by bone scan. It appears that both spondylitis and meningitis may occur simultaneously during the acute stage and symptoms of one may mask those of the other and the diagnosis may be difficult. Also spondylitis and/or radiculitis may coexist with simple disc disease, giving rise to further diagnostic difficulties and pitfalls that have frequently been pointed out (2,3,6, 8, 11, 18, 20, 23, 31). The earliest brucellar spondylitic lesions begin in bone (6, 7) rather than in the intervertebral discs (IVD) (3, 13, 18, 20, 24), whose lesions, because of poor regeneration, develop faster, giving a false impression that the IVD were involved first. Plain radiographs and CT may be completely normal during the acute phase, but bone scan may show various patterns of increased isotope uptake that may be reversed by treatment (19). Disc lesions are very common in brucellosis but they may be so slight that they appear radiologically normal (6, 8, 17). Pure disc lesions are probably very rare, as apparently normal adjacent vertebrae may be shown to be the site of disease by bone scan and/or CT, as occurred in some of our cases. It is probable that brucellar spondylitis occurs among young people as frequently as it does among old people, but the infection is cleared from bone successfully in the vast majority of cases. Only in a minority of cases, small pockets of infection may persist as a source of relapses or of local bone or soft tissue disease. Among old people, bone regeneration becomes poor and the bone becomes locus minoris resistentiae, for which brucellae have a special predeliction (33). Because of the inadequate healing, clearing of infection from bone is difficult and the disease progresses into a chronic active or smouldering symptomatic or asymptomatic spondylitis. In the latter condition, because of heavy bone sclerosis, activity is detected only by bone scan and formation of large granulomata is unlikely, but the lesions may remain the source of relapses or microscopic abscess formation. Meningitis may not be commonly seen with this form of spondylitis, but radiculopathy caused by root irritation or compression by tiny abscess or disc herniation may occur, as was seen in 2 of our cases. Culture positive microabscesses and root hernia22
tions in association with apparently healed spondylitis have been reported (8,13, 15,27). Radiculitis may also occur as a result of arachnoiditis, localized neuritis or demyelination (13, 3 1). Syndromes of cervicobrachialgia and lumbosciatica without disc prolapse have been reported (1 1). Isolated pronounced atrophy of the shoulder muscles, as seen in one of our cases (case 3, smouldering spondylitis) (Table l), is reportedly a very rare complication of brucellosis (34). Patients with symptomatic smouldering spondylitis should therefore be investigated and treated as energetically as patients with chronic active spondylitis. In chronic active spondylitis, bone destruction continues, with varying degrees of ineffective healing with characteristic radiological appearances that may appear months after the onset of infection (3, 6-8, 33). Of these, paravertebral soft tissue swelling is probably the only consistent indicator of activity. Differentiation from osteoarthrosis, osteoporosis, osteomalacia, actinomycosis, Scheuermann’s disease and cancer deposits may at times be difficult (6, 8, 17, 18). Spinal tuberculosis (TB, Pott’s disease) presents a real challenge, as the 2 conditions are so similar that brucellar spondylitis has been aptly called pseudo-Pott’s disease (33). History, social environment and a positive tuberculin test may be helpful but their value is limited in areas endemic for both diseases. In brucellosis, localized, referred and girdle pains are more pronounced; patients may therefore present earlier, sometimes before radiological changes become apparent. In TB, pain is less marked and in the vast majority of cases advanced radiological changes are apparent at presentation, although in TB such changes take longer to manifest themselves after the onset of infection. The course in TB is more protracted, destruction and general rarefaction of bone are more extensive, healing takes longer to set in and sclerosis and parrot-beak osteophytes are not a feature. In TB, usually a single vertebra and disc may be involved, very rarely multiple and contiguous vertebral lesions occur but skip lesions and the involvement of more than one spinal region at a time are not known. Gibbus and collapse of vertebral bodies and later ankylosing are more common in TB. Paravertebral soft tissue abscesses are commoner in TB, occurring in 50% of cases, and they are usually larger than in brucellosis. Bone scan in brucellosis may reveal multiple involvement of other bones and joints, which is not seen in TB. The age incidence is different, as TB is more commonly seen in younger patients (6-8, 17, 18, 35). Plain radiographs may therefore not be pathognomonic and may miss active disease. Bone scan is not specific but is a good indicator of activity and extent of skeletal involvement. It is not useful, however, for short-term follow-up, as it remains positive
Brucella spondylitis
for a long time (6). CT may also not be diagnostic, but it is superior in delineating details of soft tissue, bone, cord and root lesions. Minor degrees of sclerosis missed by plain radiographs and centrally placed bone lesions are also shown better by CT. The 3 radiological modalities are complementary. In a few cases, diagnosis may not be possible without histological examination of bone or recovering the organism. In this series, both early and chronic active spondylitis were associated with a large variety of CNS lesions. Acute meningitis was mainly associated with early spondylitis affecting young adults. Subacute, subclinical and chronic meningitis, parkinsonism, deafness, papilloedema, basal ganglia calcification, cerebral vasculitis with TIA, paraplegia, tetraplegia, radiculitis and psychiatric manifestations were seen in association with chronic active and smouldering spondylitis affecting mainly middle-age and elderly patients. All these are well recognized complications of brucellosis (8-1 l), and the finding of CNS manifestations in brucellosis should therefore prompt a search for possible underlying spondylitis and vice versa. Treatment of patients should be prolonged since the eradication of the organism from bone and CSF may be difficult (13, 15,33). Our present policy is to use a combination of 3 drugs for a minimum of 16-24 weeks (1, 4, 5). Non-compliance or reinfection usually causes failure of therapy. References 1. MOUSAAM, ELHAGKM, KHOGALIM, MARAFIEAA. The nature of human brucellosis in Kuwait: a study of 379 cases. Rev Infect Dis 1988: 10: 211-217. 2. NORTONWL. Brucellosis and rheumatic syndromes in Saudi Arabia. Ann Rheum Dis 1984: 43: 810-815. 3. SPINKWW. The nature of brucellosis. Minneapolis: University of Minnesota Press, 1956: 145-170. 4. MOUSA AM, KOSHY TS, ARAI GF, MARAFIEAA, MUHTASEBSA, AL-MUDALLAL DS, BUSHARATULLA MS. Brucella meningitis: presentation, diagnosis and treatment: aprospective study often cases. Q J Med (New Series) 1986: 233: 873-885. 5. MOUSAAM, MUHTASEB SA, AL-MUDALLAL DS, KHODEIR SM, MARAFIEAA. Osteoarticular complications of brucellosis: a study of 169 cases. Rev Infect Dis 1987: 9: 531-543. J et al. Brucella spondylitis; 6. ARIZAJ, GUDIOLF, VALVERDE a detailed analysis based on current findings. Rev Infect Dis 1985: 7: 656-664. J. Osteoarticular sites of brucellosis. Ann 7. ROTES-QUEROL Rheum Dis 1957: 16: 63-68. 8. GANADOW, CRAIGAJ. Brucellosis myelopathy. J Bone Joint Surg (Am) 1958: 40: 1380-1388. RW, SAHSAL, JOYNT RJ.Protean manifestations 9. FINCHAM of central nervous system brucellosis: case histories of a wide variety of clinical forms. JAMA 1963: 184: 269-275. A. Neurological complications of undulant 10. NELSON-JONES fever. Clinical picture. Lancet 1951: 1: 495-497. 11. SAHS AL. Brucellosis (Malta fever, undulant fever). In:
VINKENPJ, BRUYNGW, eds. Handbook of clinical neurology. Amsterdam: North Holland, 1978: 33: 305-326. 12. LAM K, SILVERSTEINLM, CARLISLERJ, BAYER AS. Disseminated brucellosis initially seen as sternoclavicular arthropathy. Arch Intern Med 1982: 142: 1193-1 194. 13. YOUNGEJ. Human brucellosis. Rev Infect Dis 1983: 5: 821-843. NS, SACHDEVA JR, CHOPRA JS. Brucella menin14. PATHANIA gitis and spondylitis. J Indian Med Assoc 1966: 47: 290-292. 15. BASHIRR, AL-KAWIZ, HARDEREJ, JINKINSJ. Nervous system brucellosis: diagnosis and treatment. Neurology 1985: 35: 1567-1581. RA, AL-DINAS, ARAI GF, LULUAR, MOUSAAM, 16. SHAKIR SAADA MA. Clinical categories ofneurobrucellosis. A report of nineteen cases. Brain 1987: 110: 213-223. F. Undulant fever spondylitis. Br J Radio1 1958: 3 1: 17. ZAMMIT 683-690. MMS. Brucellosis of the spine. Br J Surg 1976: 18. GLASGOW 63: 283-288. 19 BAHARRH, AL-SUHAILIAR, MOUSAAM, NAWAZMK, KHADDAHN, ABDUL-DAYEMHM. Brucellosis: appearances on skeletal imaging. Clin Nucl Med 1988: 13: 102-106. 20. KELLYPJ, MARTINkJ7 SCHIRGERA, WEED LA. Brucellosis of bones and joints. Experiences with 36 patients. JAMA 1960: 174: 347-353. 21. WILSONGS, MILES AA. Topley & Wilson principles of bacteriology, virology and immunity, 5th ed. London: Edward Arnold, 1964. 22. BUCHANAN TM, FABER LC, FELDMAN RA. Brucellosis in the United States 1960-1972:an abattoir-associated disease. I. Clinical features and therapy. Medicine (Baltimore) 1974: 53: 403-413. 23. LIFESORM, HARDERE, MCCORKELL SJ. Spinal brucellosis. J. Bone Joint Surg (Br) 1985: 67: 345-351. 24. JOGLEKARMD, NAGALOTIMATH SJ. Neurobrucellosis. J Assoc Physicians India 1976: 24: 183-186. 25. LEYSDG. Abortus fever. Br Med J 1943: 1: 187-189. 26. NICJOLSE. Meningoencephalitis due to brucellosis: report of a case in which Brucella abortus was recovered from the CSF and a review of the literature. Ann Intern Med 1951: 35: 673-693. 27. MCCULLAGHEP, CLODFELTER HM. Encephalitis due to undulant fever; report of four cases. Ann Intern Med 1937: 10: 1508-1513. 28. BLACKRA. Brucella meningitis in an infant. Arch Paediat 1937: 54: 702-713. JR. Melitensis meningo29. HANSMANNGH & SCHENKEN encephalitis. Mycotic aneurysm due to Brucella melitensis var porcine. Am J Pathol 1932: 8: 435-444. 30. SHARPWB. Pathology of undulant fever. Arch Pathol 1934: 18: 72-107. J. Epidemiology and neurological 31. DEVILLAFANE-LASTRA manifestations of brucellosis. In: VAN BOGAERTL, KAFER JP, POCHGF, eds. Tropical neurology. Buenos Aires: Lopez Libreros, 1963: 302-325. WW, MA BU. Multiple sclerosis: the blood 32. TOURTELLOITE brain barrier and the measurement of de novo central nervous system IgG synthesis. Neurology (NY) 1978: 28 (SUPPI):76-83. 33. AGUILARJA, ELVIDGEAR. Intervertebral disc disease caused by the brucella organism. J Neurosurg 1961: 18: 27-33. V. Vorlautige mitteilung uber einen fall von febris 34. JOHNSSON undulans bang mit neurologischen komplikationen. Med Klin 1929: 25: 389. A, GAINES VD. The radiology of osteo35. BONAKDER-POUR myelitis. Orthopaed Clin North Am 1983: 14: 21-37.
23
