J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.41.2.97 on 1 February 1978. Downloaded from

Journal ofNeurology, , andPsychiatry, 1978, 41, 97-107

The changing pattern of spinal arachnoiditis

M. D. M. SHAW, J. A. RUSSELL, AND K. W. GROSSART From the Institute of Neurological Sciences, Southern General Hospital, Glasgow

SUMMARY Spinal arachnoiditis is a rare condition. Eighty cases, diagnosed during a period when 7600 spinal contrast investigations were undertaken, have been reviewed. The majority had suffered a previous spinal condition, the most common being lumbar disc disease. There has been a change in the distribution of arachnoiditis with the lumbar region now most frequently involved. This accounts for the persistence of radicular symptoms and the relatively low incidence of paraplegia when compared with earlier series. Surgery does not appear to have any role in the treatment.

It was not until Horsley's (1909) description under Case material the title "chronic spinal " that spinal arachnoiditis became the first type of arachnoiditis INCIDENCE guest. Protected by copyright. to be recognised as a separate disease entity, al- During the period January 1955 to September though both Spiller et al. (1903) and Mendel and 1976 spinal arachnoiditis was diagnosed in 80 Adler in 1908 (quoted by Elkington, 1936) re- patients, 18 of whom presented in the first 10 year ported localised cystic lesions deep to the pia period. The male: female ratio was 2: 1. The mater, and Schwarz (1897) had described it during histogram (Fig. 1) shows the sex distribution his description of syphilitic meningomyelitis according to age in decades, the range being 8 associated with cavitation of the cord. Since then to 73 years. the condition has been described as "adhesive spinal arachnoiditis" (Stookey, 1927), "meningitis 30 serosa circumscripta spinalis" (Elkington, 1936), and "spinal adhesive arachnoiditis" (Mackay, 2 1939). Male Patients 54 Elkington (1951) stated that the natural history Female Patients 26 is for the condition to progress to paraplegia, but he and other authors (Kulowski and Scott, 1934; 18 Elsberg, 1942; Rocovich, 1947) agreed that surgi- 14- cal decompression is the only, if limited, prospect 0. of relief, or of achieving a static state. However, French (1946) and Lombardi et al. (1962) felt Z 8- http://jnnp.bmj.com/ that this was not justifiable as at best the patients gained only a very transient degree of relief. 6- In view of the changing patterns in this con- dition and the continuing dichotomy as to whether 2- direct surgery has any place in the treatment (Ransford and Harries, 1972), we felt that it would 0 10 20 30 40 50 60 70 80 be worthwhile reviewing the experience gained in Aqe (Years) the Institute of Neurological Sciences, Glasgow, Fig. 1 Spinal arachnoiditis: age and sex distribution. on September 25, 2021 by over the past 21 years. Statistical significance was determined by using the chi-square test. Values SITES of 0.05 were accepted as being significant. The lumbosacral region was the sole site affected in 71% of the cases. In a further 20% the Address for reprint requests: Mr J. A. Russell. Division of Ncuro- suirgery, Institute of Neurological Sciences, Southern General Hospitli, thoracic region was al'so involved, and the whole .Glasgow G51 4TF. Scotland. spinal canal in 5%. In only 4% was the thoracic Accepted 21 July 1977 region alone affected, and there were no cases in 97 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.41.2.97 on 1 February 1978. Downloaded from

98 M. D. M. Shaw, J. A. Russell, and K. W. Grossart whom the condition was restricted to the cervical after the onset of symptoms attributable to spinal region. arachnoiditis. The myelographic criteria used to make the diagnosis were: DIAGNOSTIC METHODS 1. Partial or complete block (Fig. 2). All but one of the patients underwent positive 2. Narrowing of the subarachnoid space (Fig. 3). contrast (meglumine iothalamate 3. Obliteration of the nerve root sleeves and 60% w/v (Conray 280: May and Baker) in 15 thickening of the nerve roots (Figs. 3 and 4). patients, and iophendylate (Myodil: Glaxo) in 64) 4. Irregular distribution and loculation of the at the time of diagnosis of the arachnoiditis. The contrast medium (Fig. 4). remaining patient was diagnosed post mortem (the 5. Fixity of previously inserted contrast medium cause of death was found to be an occlusion of (Fig. 5). the left common carotid artery) some 17 years 6. Pseudo-cyst formation. guest. Protected by copyright. http://jnnp.bmj.com/ on September 25, 2021 by

Fig. 3 Water-soluble myelogram showing narrowing of subarachnoid space, adherence of nerve roots to Fig. 2 Myelogram showing block at upper border of arachnoid mater, and obliteration of nerve root first thoracic vertebra. sleeves. J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.41.2.97 on 1 February 1978. Downloaded from

The changing pattern of spinal arachnoiditis 99 guest. Protected by copyright.

Fig. 5 Plain radiograph showing fixity of previously Fig. 4 Water-soluble myelogram showing irregular inserted contrast medium.

distribution of contrast medium and thickened nerve http://jnnp.bmj.com/ roots. films were lost) were submitted to surgery which allowed confirmation of the diagnosis. At the time of the study the myelograms of 77 patients were reviewed. Two patients' films had PREVIOUS SPINAL CONDITIONS been mislaid but each had been diagnosed as spinal Ninety-four per cent of the patients had under- arachnoiditis at the time of the original reporting gone a previous illness or treatment involving the of the myelogram. Of the 77 patients, two of the spinal column or its contents. The most common above myelographic features were found in all condition was intervertebral disc disease which on September 25, 2021 by the patients except one in whom the diagnosis was had been diagnosed in 64% of all the patients, radiologically uncertain. The frequency with and of these, all but one suffered from lumbar which the radiographic criteria was found is disc disease. Of the latter, 10% also developed a shown in Table 1. second spinal disease (Table 2). Primary in- Thirty-one of the patients (including the one fections, which had all required multiple lumbar in whom, on review of the radiology, the diagnosis punctures for the administration of intrathecal was felt to be uncertain, and one of those whose antibiotics had been diagnosed in 15% of the J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.41.2.97 on 1 February 1978. Downloaded from

100 M. D. M. Shaw, J. A. Russell, and K. W. Grossart Table 1 Spinal arachnoiditis-radiographic findings Of the 36 patients with arachnoiditis who underwent myelography during assessment of Arachnoiditis treatment groups their lumbar disc disease, 24 were demonstrated to have a lesion at one level, eight at multiple levels, Radiographic finidings No. Operation No operation and four were negative. Only one of these patients Block 41 21 20 was investigated using meglumine iothalamate Globulation 59 25 34 60% w/v. Nerve thickening 25 9 16 Fixed contrast 56 20 36 Fifty-one patients had previously undergone Narrow subarachnoid space 65 23 42 spinal surgery: two for tumours (both dermoid No radiographs 1 - 1 Lost radiographs 2 1 1 tumours in the lumbar canal), one for spinal Uncertain diagnosis 1 1 0 subarachnoid haemorrhage in whom the only Total number of patients 80 31 49 finding was of a swollen cervical cord, and one for an anterolateral cordotomy undertaken for pain after a traumatic amputation of the right Table 2 Spinal arachnoiditis-diseases occurring arm. Forty-seven patients underwent surgery for together with lumbar disc disease lumbar disc disease: 77% were submitted to a fenestration procedure, 11% to a laminectomy Second disease Nutmber of patients (60% of which were hemilaminectomies), and in Injury 2 the remainder the approach was unknown. Postoperative infection I Spinal angioma I Multiple level exploration was undertaken in 62% Multiple sclerosis I of the 47 patients (79% of which were at two

Total 5 levels, 14% at more than two levels and in the guest. Protected by copyright. others, though multiple, the actual number is unknown) and 26% were bilateral explorations. patients. Seventy-five per cent of these infections In order to compare this incidence with that were tuberculous and the remainder pyogenic. occurring in a general population undergoing Injury accounted for a further 9% of the patients, lumbar disc surgery, the cases operated upon dur- while tumours (3%), subarachnoid haemorrhage ing six years selected at random from the period (1%), spinal anaesthetic (1%), and previous under study, were analysed. Exploration at more cordotomy (1%) were all uncommon. than one level was carried out in 50% of these Sixty-four patients had undergone a previous patients, and 3% were explored bilaterally. either for diagnostic or thera- Of the seven patients who ultimately became peutic reasons. Twelve of these patients had no paraplegic, two had sustained injuries, two had record of whether blood was present in the cere- had , two dermoid tumours, brospinal fluid either macroscopically or micro- and one a prolapsed lumbar intervertebral disc scopically. In the remaining 52 patients, blood removed. was present in 8% (75% of which was only detected on microscopy). TIME INTERVAL During the course of investigation of their The time intervals, between investigation and/or initial complaint 43 patients had been submitted treatment of the initial spinal condition and the to myelography, but in only one case was a water- onset of symptoms which were attributable to the http://jnnp.bmj.com/ soluble contrast medium used. The myelograms arachnoiditis varied from immediately to 27 were examined for evidence of technical diffi- years (Table 3). Symptoms developed within one culties, and 39% showed such evidence. Only two of these 17 myelograms had evidence suggesting Table 3 Spinal arachnoiditis-time interval from leakage of cerebrospinal fluid alone; the remainder primary disease to symptomatic onset showed subdural or extradural extravasation of Time interval Number ofpatients % total contrast medium. In order to compare this inci- dence with that occurring during a series of ILess than I month 16 20 on September 25, 2021 by 1-6 months 13 16 myelograms, 50 consecutive myelograms under- 6-12 months 10 13 taken at the Institute were reviewed. In this 12-24 months 8 10 the subdural 2-5 years 13 16 series, leakage of the dye outside 5-10 years 3 4 space was found in 6%. In a further 2% leakage 10-20 years 9 11 occurred and resulted in concentric constriction More than 20 years 3 4 No primary disease 5 6 in the lumbar sac at the level of the spinal puncture. Total 80 100 J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.41.2.97 on 1 February 1978. Downloaded from

The changing pattern of spinal arachnoiditis 101 year in 49%, but in 15% the interval was greater Table 5 Spinal arachnoiditis-follow-up than 10 years. D:aration (yr) Nimnber ofpatients CLINICAL FEATURES I 12 In 21 patients root signs were either absent or 1-2 14 2-5 26 unilateral, but most of these (90%) complained 5-10 12 of bilateral sciatica. Only five patients showed 10-20 9 20 1 either no root signs or had signs limited to one None 6 root (Table 4). The remaining 59 patients had Total 80 both bilateral symptoms and signs. Three patients were paraplegic and a further 11 paraparetic. The bladder of one of the latter was paralysed and in a second there was difficulty emptying completely. cause unknown; after three years from a cardiac Among the other 66 patients only one had im- infarct; and after 14 months, sudden death under paired bladder function. Straight leg raising was criminal circumstances. Only in the second patient recorded in 68 patients at the time of diagnosis was the arachnoiditis progressive: in the other of arachnoiditis; 33 were abnormal and 35 normal. three it remained static. Progression of the disease occurred in 25% of all the patients (Table 6), four further patients becoming paraplegic (Table Table 4 Spinal arachnoiditis-presence of clinical 4). In addition to the seven paraplegic patients, features incontinence of urine was present in three other

patients, two who had difficulties at the time of guest. Protected by copyright. At time of Clinicalfeatiures diagnosis At outcome or death diagnosis and the third with a spastic quadriparesis which progressed over four years. The progressive Paraplegia 3 7 group also included one patient whose symptoms Quadriparesis 1 1 Paraparesis 11 10 had been static for 13 years until, after a back Multiple roots 60 45 injury, his pain became more severe. Fifty per cent Single root 2 3 Pain alone of the patients remained in a static condition, Back alone l I and this includes the three patients who were Sciatica Unilateral - - already paraplegic at the time of diagnosis. All Bilateral 2 3 the patients who were paraparetic were able to Pain free - 4 walk, though two required walking sticks and one No follow-up - 6 experienced claudication which was not due to Total 80 80 peripheral vascular disease, on walking 200 yards. Of the four patients who were symptomless, two had undergone decompressive laminectomies and, TREATMENT in addition, in one case a prolapsed intervertebral Thirty-one patients were treated by decompressive disc which was thought to be contributing to the laminectomy. In this group there were significantly symptomatology was removed. The two con- more patients with a spinal block than in the servatively managed patients had been free from non-surgical groups (Table 1) x2=5.5097 P<0.02). symptoms for two and three years respectively, http://jnnp.bmj.com/ The was left open in eight instances, and those after surgery for one and two years. and a prolapsed intervertebral disc was removed from one case. Nerve root decompression was carried out in a further two patients as it was Table 6 Spinal arachnoiditis-results of surgery and felt that, in addition to the arachnoiditis, com- conservative treatment pression by the superior facet was at least in part responsible for the symptoms. An arachnoidal cyst Sywnptoms Total Operation No operationi was decompressed in another case. No symptoms 4 2 2 on September 25, 2021 by Only six patients were treated with courses of Static Paraplegia 3 3 - steroids and no patient with radiotherapy. Other symptoms 37 10 27 Progressive 20 9 11 OUTCOME Improved Good 5 3 2 The duration of follow-up achieved is shown in Moderate 3 3 - Table 5, and the final condition of the patients in No follow-up 6 1 5 Follow-up too short 2 - 2 Table 4. There were four deaths: after 17 years Total 80 31 49 from carotid artery occlusion; after four years, J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.41.2.97 on 1 February 1978. Downloaded from

102 M. D. M. Shaw, J. A. Russell, and K. W. Grossart In comparing surgical and non-surgical results has continued, 71% of our cases involving the (Table 6), marginally more patients were improved lumbosacral region alone, and in only 4% was in the operation group than in the conservative there no involvement of the lumbosacral region. group (x2=4.9123 P<0.05). The improvement in The arachnoiditis was limited to one vertebral the former group has been maintained for periods level in only 4% of all our patients. from six months to over four years, and in the Malmros (1976) analysed the incidence of latter for six months and two years. There was no arachnoiditis in all the patients reported by Autio other significant difference between the groups. et al. (1972), Ahlgren (1973), Radberg and Wenn- The presence or absence of a spinal block did not berg (1973), and Irstam and Rosencrantz (1973, influence the outcome, nor was there any relation- 1974), who had undergone water-soluble myelo- ship between the cerebrospinal fluid protein level graphy and subsequently returned for reinvestiga- (which varied from normal to 12 g/l) and the tion. The high incidence found (49%) was probably outcome: neither was there any relationship with due to the acceptance of blunting or amputation of the cerebrospinal fluid white cell counts which the root sleeve alone as radiological evidence of were normal except in two cases. arachnoiditis. We do not feel that this is sufficient evidence to make the diagnosis. Indeed of Autio Discussion et al.'s (1972) original six patients fitting this criterion, one has died of malignancy and on histo- Symptomatic spinal arachnoiditis is a relatively logical examination no evidence of arachnoiditis rare condition in which there is tissue within was found (Irstam and Rosencrantz, 1974; and between the leptomeninges. During the period Malmros, 1976). We have, therefore, like other under study 7600 spinal contrast investigations and workers (Seaman et al., 1953; Lombardi and guest. Protected by copyright. 2900 explorations for lumbar disc disease have Passerini, 1964; Austin, 1972) only accepted the been carried out at the Institute. Arachnoiditis more rigid criteria described above. Based on the has been diagnosed more frequently in the second facts that 7600 myelograms have been under- 10 year period than in the first. The number of taken, that in a study of lumbar disc disease positive contrast spinal investigations has risen Jennett and Shaw (1976) were able, in the area from 250 per annum in 1961 to 678 in 1976. The served by the Institute, to achieve a follow-up latter consists of 330 iophendylate myelograms, rate of 67%, and that 43 of the patients in the 256 water soluble myelograms (virtually all using present series had been submitted to previous meglumine iothalamate 60% w/v), and 92 myelography, the incidence of symptomatic spinal reinvestigations. arachnoiditis after this procedure would seem to The incidence (Fig. 1) of spinal arachnoiditis be approximately 1%. We have no experience of remains highest in the 30-50 year age group the technique of air myelography used by Jacob- (Elkington, 1936; French, 1946; Lombardi et al., sen and Lester (1969) for confirmation of the 1962; Lombardi and Passerini, 1964). No previous diagnosis, but it is important, before undertaking report was found in the English literature describ- a second myelogram, to obtain plain radiographs ing the condition in patients under 10 years of age: of the spine (Fig. 5) which may show the one of the present patients was 8 years old and characteristic fixation of residual contrast medium a second, 9. Elkington (1936) reported a male: which is deposited in droplets and streaks female ratio of 3: 1, but Lombardi and Passerini (Christensen, 1942). http://jnnp.bmj.com/ (1964) like us, found it to be 2 : 1. Jacobsen and Various aetiological factors have been impli- Lester's report (1969) was unusual in that the cated in the development of spinal arachnoiditis. ratio was 6: 1. The relative importance of each has changed with Spinal arachnoiditis may be localised or the passage of time. Most of our patients had pre- generalised. Though in 1951 Elkington wrote that existing intraspinal disease. there was still a slight preponderance of cases Infection was the first condition to be associated situated in the cervical region, he had previously with spinal arachnoiditis. Schwarz's (1897) case was found (1936) that 29% were restricted to the secondary to syphilis, as were some of Horsley's on September 25, 2021 by cervical, 59% to the thoracic (including the (1909). The latter author attributed the remaining thoracolumbar junction), and 12% to the lumbo- cases to the gonococcus on indirect and, therefore, sacral spine. This is in agreement with other inconclusive evidence. Since then many authors reports (Horsley, 1909; Lombardi et al., 1962) in have described the condition after bacterial, in- which the majority of cases were thoracic. French cluding tuberculous, infections (Elkington, 1936; (1946) was the first to report the swing to a high Rocovich, 1947; Lombardi et al., 1962). This was incidence in the lumbar region (46%). This trend felt to be the most common cause, and has been J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.41.2.97 on 1 February 1978. Downloaded from

The changing pattern of spinal arachnoiditis 103 compounded by the use of intrathecal therapy- (Kaufmann and Jeans, 1976) than other patients. for example, penicillin (Erickson et al., 1946; Whether the patients in our retrospective series Millikan and Sahs, 1947), and sulphonamide drugs were allergic individuals cannot be certain, but it (Lombardi et al., 1962). Lymphocytic meningitis would seem unlikely as only one of the 61 patients has also been described as a predisposing condition undergoing myelography developed a temperature (Barker and Ford, 1937). after the myelogram (as compared with 6% of 60 Blood in the cerebrospinal fluid produces an random myelograms), and none developed any inflammatory reaction in the leptomeninges other allergic phenomena such as bronchospasm (Bagley, 1928a,b; Hammes, 1944; Alpers and or skin rashes. A single patient was known to have Forster, 1945). Spinal arachnoiditis after subarach- a drug sensitivity, namely to PAS. noid haemorrhage though described (Lombardi Removal of the iophendylate immediately upon et al., 1962) is uncommon (1% in the present completion of the investigation does not neces- series). However, this may account for the rela- sarily protect the patient from the risk of develop- tively high incidence of spinal trauma, with or ing arachnoiditis (Peacher and Robertson, 1945). without a fracture (Elkington, 1936; Elsberg, The water-soluble contrast media are not free 1942; Lombardi et al., 1962) in patients with from this complication (Autio et al., 1972; arachnoiditis. Spinal anaesthetics were cited as the Ahlgren, 1973; Halaburt and Lester, 1973; Rad- commonest aetiological factor by Winkelman et al. berg and Wennberg, 1973; Irstam et al., 1974; (1953) and Austin (1972). Rosenbleum et al. (1952) Irstam and Rosencrantz, 1973, 1974; Liliequist described six cases after 1272 spinal anaesthetics, and Lundstrom, 1974; Slatis et al., 1974; Malmros, while Kennedy et al. (1945) estimated the risk of 1976). Though various authors (Ahlgren, 1973; grave neurological deficit after spinal anaesthesia Halaburt and Lester, 1973; Irstam et al., 1974; guest. Protected by copyright. to be 0.5%, of which approximately half were Irstam and Rosencrantz, 1974) have claimed that due to spinal arachnoiditis (Kennedy et al., 1950). the incidence varies between the different con- Particles of haemosiderin found in a biopsy speci- trast media, Malmros's composite figures suggest men (Weiss et al., 1962) supported the suggestion little difference between methiodal (Kontrast U), made by Austin (1972) that faulty technique was methylglucamine iothalamate (Dimer-X), and the cause. However, Greenfield (quoted by Meyer, meglumine iothalamate (Conray 60). An increas- 1963) thought that chemical contamination was ing amount of contrast medium, and hence con- the causative factor. Kennedy et al. (1945) felt centration in the cerebrospinal fluid, results in an that pre-existing central nervous system disease increased incidence of severe arachnoiditis (Hala- was a contraindication to the use of spinal burt and Lester, 1973). The resultant increase in anaesthesia, because this combination led to an osmolarity of the cerebrospinal fluid could be the increased risk of serious neurological deficit. precipitating factor (Slatis et al., 1974), but Hala- Chemical may be the mechanism of burt and Lester (1973) also found that the risk was production of arachnoiditis in the two cases who higher if the cerebrospinal fluid protein was above had dermoid tumours removed in the present the normal range. In this series, of 61 patients who series, though the condition has been described had a lumbar puncture at the time of the previous after surgery for other types of benign intraspinal spinal complaint, only six had a cerebrospinal tumours (Davidoff et al., 1947). fluid protein content greater than 0.5 g/l, and Positive contrast myelography has been impli- only one of these underwent myelography: the http://jnnp.bmj.com/ cated in the aetiology of spinal arachnoiditis, and other five patients had all suffered infections. The experimentally (Bergeron et al., 1971), at least striking feature in the present series is the very using iophendylate, blood potentiates the effect. high percentage (39%) of myelograms which Howland and Curry (1966) demonstrated that showed evidence of a technical difficulty, even intrathecal steroids introduced simultaneously though very few cases had bloody fluid even on with the blood and iophendylate greatly reduce the microscopy. This is significantly greater than the risk of developing a severe arachnoiditis. However, incidence of 8% found in a consecutive series of Bernat et al. (1976) implicated intrathecal "Depo 50 myelograms (x2=13.1507 P<0.001). It would, on September 25, 2021 by Medrol" in the production of a severe stenotic therefore, seem important that the lumbar meningeal reaction in patients with multiple puncture, necessary for myelography, should be sclerosis, which they thought might be due to the carried out by someone who has considerable vehicle rather than to the steroid itself. Patients and frequent experience of the procedure in order suffering from multiple sclerosis seem to be more to reduce the chances of an extra- or subdural reactive to intrathecal injections (Bernat, 1977) injection of contrast medium. If water-soluble and, in particular, to iophendylate myelography myelography is followed by surgery, a much J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.41.2.97 on 1 February 1978. Downloaded from

104 M. D. M. Shaw, J. A. Russell, and K. W. Grossart higher incidence of spinal arachnoiditis has been the arachnoiditis varies. Half developed symptoms noted-for example, Liliequist and Lindstrom in less than one year, but in 16% it was greater (1974) found 16% in the non-operation group as than 10 years. Patients usually complained of against 58% in an operation group, and Irstam bilateral sciatica. Only one patient did not and he and Rosencrantz (1974) found 24% in the non- suffered back pain only, although his arachnoiditis operation group and 69% in the operation group. involved the lumbar region. Most patients show These high incidences may be due to the less objective evidence of multiple root lesions, but rigid criteria adopted for the diagnosis of the frequently it is the severe pain, which is not condition as compared with the present series. always strictly radicular, that is the predominant Furthermore, there were only three patients who complaint. had an uncomplicated myelogram, no surgery, The role of direct surgery in the treatment of and no previous spinal illness, suggesting that the spinal arachnoiditis remains in question. Elsberg incidence of spinal arachnoiditis in this situation (1942) obtained relief for 25% of his patients and is very low. Since 1971 water soluble myelograms a further 25% were improved. The improvement have been in regular use in the Institute, con- was related to a short history. Brain and Walton stituting 44% of the new investigations. In view (1969) state that surgery improves some 30% and of this, together with the facts that only one of results from the relief of the cystic component of the patients in this series had been investigated the arachnoiditis. Elkington (1951) agreed that if using water-soluble contrast medium and that half cystic changes were present surgery was of help, the patients with the condition developed their but in the majority of cases there was, as earlier symptoms within one year, it would seem that the writers had found, an inexorable progression which risk of developing symptomatic spinal arachnoiditis at the extreme would result in paraplegia due to guest. Protected by copyright. is considerably less with water-soluble contrast intramedullary cavitation (Schwarz, 1897; Mackay, medium, than after iophendylate. 1939; Lubin, 1940). Furthermore, surgery may That arachnoiditis might complicate lumbar result in more nerve root damage (French, 1946). disc disease was recognised by French (1946), In the present series there were marginally more Smolik and Nash (1951), and Ransford and patients in the operation group who were im- Harries (1972). No particular feature leading to proved than in the non-operation group (P<0.05), an increased incidence of this complication was but this was not related to the duration of the identified. The number of intervertebral disc symptomatology. However, it is important to levels explored was not markedly different in our exclude additional pathology as the cause of the series as compared with the control group, but pain, for example a prolapsed intervertebral disc there were over eight times as many bilateral or the superior facet syndrome, because there are explorations in the patients with arachnoiditis. A some patients with proven asymptomatic arach- possible explanation is that in unilateral ex- noiditis. There are four such patients in this series plorations the dura mater is displaced in only and Malmros (1976) has described a further two one direction, but in bilateral explorations there cases. There are insufficient cases in this series is less space and circumferential leptomeningeal to decide whether leaving open the dura mater damage may be inflicted as a result of the confers any added benefit. retraction. Feder and Smith (1962) reported that radio-

In a few cases (five in the present series), no therapy resulted in an excellent clinical response http://jnnp.bmj.com/ preceding reason for the arachnoiditis was found in 23.5% of patients, a limited response in 23.5%, and in particular no history of allergy was ob- and arrest of progression in 29%; 23.5% con- tained. No single site was preferentially involved. tinued to progress. Of the excellent responders Duke and Hashimoto (1974) described six cases in 50% had also undergone surgery, and of those one family which they felt was genetically deter- who responded in a limited way the majority (all mined as an autosomal dominant factor, but in except one) would have been considered as static none of our cases was there a familial tendency. in this present series. Thus the percentage of

Various factors are, therefore, involved in the patients in the static and progressive groups do not on September 25, 2021 by production of spinal arachnoiditis. Not infre- differ from those in the present series. We, there- quently these are multiple-for example, two fore, agree with French (1946) that radiotherapy separate diseases affecting the spinal column or its is of no benefit. contents, or a single condition with technical Though steroids may have some protective difficulties. value, if given at the time of the precipitating The time interval between the precipitating event (Howland and Curry, 1966), they are un- factor and the onset of symptoms arising from likely to be of value in the established disease. The J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.41.2.97 on 1 February 1978. Downloaded from

The changing pattern of spinal arachnoiditis 105 patients given steroids in this series were few and have any clear-cut role in the management of this did not gain any additional benefit. chronic condition. The difference between Elkington's findings (1936, 1951) and ours concerning progression of We would like to acknowledge the kindness of the the clinical state to paraplegia may well reflect consultant staff of the Institute of Neurological the alteration in predominance of involvement of Sciences, Glasgow, for allowing us access to their the various sites in the . Major vascular case material, and to thank Miss H. Wilson, Mrs supply to the cord from one of the lower lumbar W. McDermid, and Mrs J. Rubython for typing vessels is unusual, and hence arachnoiditis at this the manuscript. site manifests itself as a multiple root problem, whereas if the cervical or thoracic regions are References involved intramedullary cavitation may occur (Mackay, 1939; Lubin, 1940) due either to Ahlgren, P. (1973). Long term side effects after vascular deprivation or compression of the spinal myelography with water soluble contrast media: cord from which paraplegia results. It is, therefore, Conturex Conray meglumine 282 and Dimer-X. of interest that none of our cases progressing to Neuroradiology, 6, 206-211. paraplegia had arachnoiditis restricted to the Alpers, B. J., and Forster, F. M. (1945). The repara- lumbar region. Indeed, in five of the seven tive process in subarachnoid hemorrhage. Journal patients, the third to sixth thoracic segments, of Neuropathology and Experimental Pathology, 4, which represents the vascular watershed, were 262-268. in- Austin, G. (1972). The Spinal Cord. Basic Aspects involved. In the remaining two patients the and Surgical Considerations. Second edition. Charles was of the tenth and eleventh thoracic volvement C. Thomas: Springfield, Illinois. guest. Protected by copyright. segments and this suggests that the artery of Autio. E., Suolanen, J., Norrback, S., and Slatis, P. Adamkiewicz in these patients entered the spinal (1972). Adhesive arachnoiditis after lumbar my- column at or above this level. It is difficult to elography with meglumine iothalamate. Acta explain why arachnoiditis which was distal to the Radiologica Diagnosis, 12, 17-24. sixth thoracic segment should result in a quadri- Bagley, C. (1928a). Blood in the cerebrospinal fluid: paresis in one patient, but the three patients who resultant functional and organic alterations in the had spastic paraparesis had involvement of the central nervous system. A. Experimental data. In the remaining cases Archives of Surgery, 17, 18-38. thoracolumbar junction. Bagley, C. (1928b). Blood in the cerebrospinal fluid: the condition was restricted to the lumbar region. resultant functional and organic alterations in the central nervous system. B. Clinical data. Archives Conclusion of Surgery, 17, 39-81. Barker, L. F., and Ford, F. R. (1937). Chronic arach- The patterns of production of spinal arachnoiditis noiditis obliterating the spinal subarachnoid space. and hence the sites involved and the degree of Journal of the American Medical Association, 109, neurological progression have changed. Usually 785-787. several factors are involved, and rarely does con- Bergeron, R. T., Rumbaugh. C. L., Fang, H. and trast medium alone produce spinal arachnoiditis. Cravioto, H. (1971). Experimental pantopaque on the spine such as arachnoidlitis in the monkey. Radiology. 99, 95- It is essential that procedures 101. lumbar punctures should be undertaken if possible Bernat, J. L. (1977). lophendylate myelography in http://jnnp.bmj.com/ by experienced personnel. Furthermore, if there multiple sclerosis. Lancet, 1, 48-49. are technical difficulties, or if the patient has had Bernat, J. L., Sadowsky, C. H., Vincent, F. M., a previous spinal illness, contrast media or other Nordgren, R. E., and Margolis, G. (1976). Sclerosing agents should not be inserted into the spinal theca spinal pachymeningitis: a complication of intra- if at all avoidable. thecal administration of Depo-Medrol for multiple The most commonly associated previous spinal sclerosis. Journal of Neurology, Neurosurgery, and condition is now intervertebral disc disease, and Psychiatry, 39, 1124-1128. Brain, W. R., and Walton, J. N. (1969). Brain's this accounts for the change in outcome in the on September 25, 2021 by to have radicular Diseases of the Nervous System. Seventh edition. majority of cases who continue Oxford University Press: London. manifestations, in particular pain, and rarely, if Christensen. E. (1942). Chronic adhesive spinal arach- ever, progress to paraplegia unless the thoracic or noiditis. Acta Psychiatrica Scandinavica, 17, 23-38. cervical regions are involved. Davidoff, L. M., Gass, H., and Grossman, J. (1947). The outcome is not related to the presence of a Post operative spinal adhesive arachnoiditis and myelographic block or a high cerebrospinal fluid recurrent spinal tumors. Journal of Neurosurgery, protein content. Direct surgery does not appear to 4, 451-464.

B J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.41.2.97 on 1 February 1978. Downloaded from

106 M. D. M. Shaw, J. A. Russell, and K. W. Grossart Duke, R. J., and Hashimoto, S. A. (1974). Familial anaesthesia. Surgery, Gynaecology and Obstetrics, spinal arachnoiditis. A new entity. Archives of 91, 385-398. Neurology (Chicago), 30, 300-303. Kennedy, F., Somberg, H. M., and Goldberg, B. B. Elkington, J. St. C. (1936). Meningitis serosa circum- (1945). Arachnoiditis and paralysis following spinal scripta spinalis (spinal arachnoiditis). Brain, 59, anaesthesia. Journal of the American Medical 18 1-203. Association, 129, 664-667. Elkington, J. St. C. (1951). Arachnoiditis. In Modern Kulowski, J., and Scott, W. (1934). Localised ad- Trends in Neurology. 5. Edited by A. Fielding. hesive spinal arachnoiditis: an obscure cause of Butterworth: London. radiating low back pain. Journal of Bone and Joint Elsberg, C. A. (1942). Surgical Diseases of the Spinal Surgery, 16, 699-703. Cord Membranes and Nerve Roots: Symptoms, Liliequist, B., and Lundstrom. B. (1974). Lumbar Diagnosis and Treatment. Lewis: London. myelography and arachnoiditis. Neuroradiology, 7, Erickson, T. C., Masten, M. G., and Suckle, H. M. 91-94. (1946). Complications of intrathecal use of penicil- Lombardi, G., and Passerini, A. (1964). Spinal Cord lin. Journal of the American Medical Association, Diseases: a radiological and myelographic analysis. 132, 561-565. Williams and Wilkins: Baltimore. Feder, B. H., and Smith, J. L. (1962). Roentgen Lombardi, G., Passerini, A., and Migliavacca, F. therapy in chronic spinal arachnoiditis. Radiology, (1962). Spinal arachnoiditis. British Journal of 78, 192-197. Radiology. 35, 314-320. French, J. D. (1946). Clinical manifestations of lumbar Lubin, A. J. (1940). Adhesive spinal arachnoiditis spinal arachnoiditis: a report of thirteen cases. as a cause of intramedullary cavitation: Comparison Surgery, 20, 718-729. with . Archives of Neurology and Halaburt, H., and Lester, J. (1973). Leptomeningeal Psychiatry (Chicago), 44, 409-420. changes following lumbar myelography with water Mackay, R. P. (1939). Chronic adhesive spinal soluble contrast media (meglumine iothalamate and arachnoiditis: a clinical and pathological study. guest. Protected by copyright. methiodal sodium). Neuroradiology, 5, 70-76. Journal of the American Medical Association, 112, Hammes, E. M. (1944). Reaction of the to 802-808. blood. Archives of Neurology and Psychiatry Malmros. R. (1976). Spinal arachnoiditis after myelog- (Chicago), 52, 505-514. raphy with water soluble contrast media. Presented Horsley, V. (1909). Chronic spinal meningitis: its at the combined Meeting of Nederlandse Verenig- differential diagnosis and surgical treatment. British ing Van Neurochirurgen and the Society of British Medical Journal, 1, 513-517. Neurological Surgeons, Amsterdam, Netherlands. Pantopaque Mendel, K., and Adler, S. (1908). Cited by Elkington Howland, W. J., and Curry, J. L. (1966). (1936). arachnoiditis: experimental study of blood as a: In Greenfield's potentiating agent and cortico steroids as an Meyer, A. (1963). Intoxications. 5, Neuropathology. Edited by W. Blackwood, W. H. ameliorating agent. Acta Radiologica Diagnosis, McMenemy, A. Meyer, R. M. Norman, and D. S. 1032-1041. Russell. Second edition. Edward Arnold: London. Irstam, L., and Rosencrantz, M. (1973). Water soluble Adhesive contrast media and adhesive arachnoiditis: 1, re- Millikan, C. H., and Sahs, A. L. (1947). Radio- arachnoiditis following the intrathecal administra- investigation of non operated cases. A cta tion of penicillin. Journal of Iowa State Medical logica Diagnosis, 14, 497-506. Society, 37, 198-201. Irstam, L. and Rosencrantz, M. (1974). Water soluble Peacher, W. G., and Robertson, R. C. L. (1945). contrast media and adhesive arachnoiditis: 2, re- Pantopaque myelography: results, comparisons of investigation of operated cases. A cta Radiologica contrast media and spinal fluid reaction. Journal Diagnosis, 15, 1-15. of Neurosurgery, 2, 220-231. http://jnnp.bmj.com/ Irstam, L., Sundstrom, R., and Sigstedt, B. (1974). Radberg, C., and Wennberg, E. (1973). Late sequelae Lumbar myelography and adhesive arachnoiditis. following lumbar myelography with water soluble Acta Radiologica Diagnosis, 15, 356-368. contrast media. A cta Radiologica Diagnosis, 14, Jacobsen, H. H., and Lester, J. (1969). Diagnosis of 507-512. spinal arachnoiditis by air myelography. Acta Ransford, A. C., and Harries, B. J. (1972). Localised Neurologica Scandinavica, 45, 376-379. arachnoiditis complicating lumbar disc lesions. Jennett, B., and Shaw, M. D. M. (1976). The natural Journal of Bone and Joint Surgery, 54, 656-665. history of lumbar disc disease. Presented at the Rocovich, P. M. (1947). Adhesive spinal arachnoiditis: combined meeting of Nederlanclse Vereniging Van a short review. Bulletin of Los Angeles Neurological on September 25, 2021 by Neurochirurgen and the Society of British Neuro- Society, 12, 69-77. logical Surgeons, Amsterdam, Netherlands. Rosenbleum, H. E., Long, F. B., Hinchley, T. R., and Kaufmann, P., and Jeans, W. D. (1976). Reactions to Trufant, S. A. (1952). Paralysis with saddle block iphendylate in relation to multiple sclerosis; Lancet, anesthesia in obstetrics. Archives of Neurology 2, 1000-1001. and Psychiatry (Chicago), 68, 783-790. Kennedy, F., Effron, A. S., and Perry, G. (1950). Schwarz, E. (1897). Praparate von einem falle The grave spinal cord paralyses caused by spinal syphilitischer myelomeningitis mit h6hlenbildung J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.41.2.97 on 1 February 1978. Downloaded from

The changing pattern of spinal arachnoiditis 107

im ruchenmarke und besonderen degenerativ en Spiller, W. G., Musser, J. H., and Martin, E. (1903). veranderungen der neurologia. Wiener Klinische A case of intradural spinal cyst with operation and Wochenschrift, 10, 177-179. recovery. University of Pennsylvania Medical Seaman. W. B., Marder, S. N., and Rosenbaum, H. E. Bulletin, 16, 27-31. (1953). The myelographic appearance of adhesive Stookey, B. (1927). Adhesive spinal arachnoiditis spinal arachnoiditis. Journal of Neurosurgery, 10, simulating spinal cord tumor. Archives of Neur- 145-153. ology and Psychiatry (Chicago), 17, 151-178. Slatis, P., Autio, E., Suolanen, J., and Norrback, S. Weiss, R. M.. Sweeney, L., and Dreyfuss, M. (1962). (1974). Hyperosmolarity of the cerebrospinal fluid Circumscribed adhesive spinal arachnoiditis. as a cause of adhesive arachnoiditis in lumbar Journal of Neurosurgery, 19, 435-438. myelography. Acta Radiologica Diagnosis, 15, 619- Winkelman, N. W., Gotten, N., and Scheibert, D. 629. (1953). Localised adhesive spinal arachnoiditis: a Smolik. E. A., and Nash, F. D. (1951). Lumbar spinal study of twenty-five cases with reference to etiology. arachnoiditis: a complication of the intervertebral Transactions of the A merican Neurological disc operation. A nnals of Surgery, 133, 490-495. Association, 78, 15-18. guest. Protected by copyright. http://jnnp.bmj.com/ on September 25, 2021 by