VALUE OF VISUAL EVOKED RESPONSE AND OLIGOCLONAL BANDS IN CEREBROSPINAL FLUID IN DIAGNOSIS OF SPINAL MULTIPLE SCLEROSIS

769

VALUE OF VISUAL EVOKED RESPONSE AND OLIGOCLONAL BANDS IN CEREBROSPINAL FLUID IN DIAGNOSIS OF SPINAL MULTIPLE SCLEROSIS

TABLE I-SPINAL CORD LESIONS IN COHORT

(42 PATIENTS) USED

TO ESTIMATE FALSE-POSITIVE RATES FOR VISUAL EVOKED

RESPONSE AND OLIGOCLONAL BANDS IN CSF

R. IANSEK P. M. DENNIS

P. A. KEMPSTER

J. I. BALLA B. BIEGLER

Departments of Neurology, Medicine, and Chemical Pathology, Prince Henry’s Hospital, Melbourne, Australia To assess the predictive value of prolongation of visual evoked response (VER) plus the presence of oligoclonal bands in the cerebrospinal fluid (CSF) in the diagnosis of multiple sclerosis (MS) in patients with isolated cord lesions, the false-positive rate for the two tests combined was determined by a prospective analysis of 42 patients with a structural spinal cord lesion. The false-positive rate for the

Summary

combination of VER and CSF abnormalities was zero, but the individual false-positive rates were 10% for VER and 12% for CSF protein oligoclonal bands. According to Bayes’ theorem, and taking into account the false-positive rates of the tests in the population studied and the prevalence of MS and the true-positive rates for the tests as derived from published reports, abnormalities of VER and CSF protein together gave a probability of MS of 100%. However, if only either the VER or the CSF were abnormal, the probability of MS was 44% or 49%, respectively. An abnormal result for both tests thus indicates a sufficiently high probability of MS to exclude myelography, but a positive result in only one test warrants the procedure. Introduction HAVING to differentiate between cord compression and sclerosis (MS) in a young or early middle aged adult who presents with an isolated spinal cord lesion is a common problem in neurology. Myelography is usually done to exclude cord compression but the procedure is invasive and carries a risk of adverse effects.1 An alternative approach is to look for the presence of subclinical disease in other parts of the central nervous system (CNS) by the use of evoked potentials or to demonstrate the presence of abnormal proteins in the cerebrospinal fluid (CSF) in the form of oligoclonal bands. Mastaglia et aP have suggested that prolongation of evoked potentials is sufficient evidence of demyelination to exclude the need for myelography. Their method of assessment is not suitable for assessing whether the combination of oligoclonal bands in the CSF plus abnormalities in visual evoked response (VER) strengthens the evidence for demyelination, because the degree to which the two tests overlap in their positivity, both in MS and in other forms of spinal cord pathology, is not known. In our report3 of the use of clinical decision analysis to assess the probability of MS in the face of an abnormal VER and/or oligoclonal bands in the CSF, we had to estimate the false-positive rates for the two tests since the information was unavailable. We now report on a prospective study designed to give the false-positive rates of the VER and CSF oligoclonal bands, both singly and in combination, in patients with structural disorders of the spinal cord other than MS.

multiple

-

Subjects and Methods All patients with clinical features of a spinal cord syndrome and a demonstrable lesion on myelography who were assessed by the

neurology and neurosurgery units of a teaching hospital over a period of 2 years were included in the study. A limited number of patients who also had clinical features of a spinal cord syndrome but normal myelographic findings and who were classified by the McAlpine criteria’ as probably or possibly having MS, were also included. CSF was taken at the time of myelography and was analysed for oligoclonal banding by isoelectric focusing with silver staining on polyacrylamide geLS Serum was also examined to exclude the possibility that oligoclonal bands in the CSF could be due to contamination by abnormal bands in the blood. The CSF was deemed to be positive for oligoclonal bands if such bands were present in the serum and if the CSF contained one or more discrete bands, as detected by silver staining, in the gamma globulin region. The sensitivity of this method of detecting such bands in CSF is 2 mg protein/1 or 0-04 Ilg in a 20 µg sample. The VER was measured by monocular full-field pattern reversal. The mean +2-5 SD VER in 24 normal subjects was taken as the normal range. Prolongation of the PI 00 latency for either eye or prolonged inter-eye difference was regarded as abnormal. not

Results 42 patients aged 19-71 years (mean 55) had abnormal myelograms, the commonest lesion demonstrated being cervical spondylosis. Spondylotic bars were considered significant if the spinal cord was focally flattened and there was complete or partial obstruction to the passage of contrast material at this level. Some patients underwent surgical decompression and in most of these the CSF protein

concentration was raised. The other lesions were extradural malignant compression, neurofibroma, and syringomyelia

(table I). The VER was abnormal in 4 (10%) patients and oligoclonal bands were detected in the CSF in 5 patients (table II). No patient had abnormalities of both tests. 13 patients aged 26 to 56 years (mean 42) had normal myelograms. They were presumed to have MS, since other causes of non-structural spinal cord pathology, such as vitamin B 12 or folic acid deficiency, syphilis, and hereditary systems disease, were excluded. 6 (46%) had oligoclonal bands in the CSF, 4 (39%) had abnormal VERs, and 3 (23%) had abnormal results for both tests (table III). Decision

Analysis

The value of any diagnostic test in predicting the presence of disease can be determined with the knowledge of three TABLE II—FALSE-POSITIVE RATES FOR COMBINATION OF VISUAL

EVOKED RESPONSE AND OLIGOCLONAL BANDS IN CSF

770 TABLE III—TRUE-POSITIVE RATES FOR COMBINATION OF VISUAL EVOKED RESPONSE AND OLIGOCLONAL BANDS IN CSF

TABLE IV-SENSITIVITY ANALYSIS OF PREVALENCE OF MS IN POPULATION AND FALSE-POSITIVE RATE FOR COMBINATION OF VER ABNORMALITIES AND

PRESENCE OF OLIGOCLONAL BANDS IN CSF

True +ve for combination=23%; for CSF bands=23%; for VER- 16%.

factors-prior probability of disease (PD), true-positive rate, and false-positive rate for the test-in question. Prior probability of disease refers to the likelihood of finding a patient with MS in a group of young to middle aged patients with otherwise unselected spinal cord syndromes. The true positive rate refers to the probability of a patient with MS having an abnormal VER or CSF. The false-positive rate indicates the probability of one of these abnormal results being found in a patient without MS. Bayes’ theorem was used to revise the probability of disease in the light test results. Bayes’ theorem can be expressed as: Probability of disease given a positive test PD x True positive rate of test [PD x True positive rate] + [(1—PD) x False positive rate] Here we are concerned with young or early middle aged adults with undiagnosed isolated spinal cord lesions, and on the basis of previous long-term clinical and pathological studieswe assume a prior probability of MS of 33%. We use our 42 patients with abnormal myelograms to provide the data for false-positive rates. We also make use of our true-positive data, even though the number of patients is .

small, since the results we obtained were very similar to the average in many published series.7-10 We prefer to use our figures since they are derived from a prospective study that examined the positivity of both tests simultaneously in the same patients, which was not the case for previously published figures. We first consider a patient with oligoclonal bands in the CSF but with a normal VER. The probability of MS is calculated from tables II and III as follows:

positive is below 50% for either test. This small likelihood of MS implies that myelography becomes mandatory when only one test is positive. An important factor in such an analysis is the prior probability of disease. It may be argued that in a younger age group, the prevalence could be different from that used in our analysis. However, when the combination of tests has a false-positive rate of zero, the prevalence of disease is irrelevant; when both tests are abnormal then the probability of MS is 100%, no matter what the prior probability. Nevertheless, our cohort may have been too small for the detection of a low false-positive rate for the two combined, and a more realistic figure could well be 1-2%. To test the robustness of our conclusions, we substituted plausible values (for prevalence and false-positive rates) for the ones used in our original equation. This method of sensitivity analysis (table IV) showed that, with a prevalence of MS of greater than 15% and with a false-positive rate of less than 2%, the probability of MS is still high (>70%) when both tests are abnormal. The robustness of our fmding leads us to recommend an algorithm for the management of patients with spinal cord lesions suspected of being due to MS (see accompanying figure). This algorithm indicates that one lumbar puncture is necessary to test for oligoclonal bands in the CSF and, should these not be found, a second would be necessary for myelography. Such would probably be the approach if old water-soluble contrast agents are used for myelography, since the high (68%)’1 frequency with which they produce side-effects would make a lumbar puncture preferable to myelography. However, the lower (26%)" frequency with which the newer agents produce side-effects might mean that a myelogram could be done at the outset if the newer water-soluble contrast agents are available. We would, tests

Discussion Decision analysis of our results indicates that a patient with a clinically isolated spinal cord syndrome should be diagnosed as having MS if the VER is abnormal and there are oligoclonal bands in the CSF. The probability of MS is high enough for us to recommend that under these circumstances myelography should be withheld in favour of observation. The probability of MS if only one test is

Algorithm for sclerosis.

the management of

patient with suspected spinal

771

however, favour the two-step approach since not all patients will need a myelogram and in our experience myelography is more distressing than a lumbar puncture to the patient, but the final decision rests with the individual doctor. What is clear, though, is that an abnormal VER is not sufficient reason for withholding myelography. The major policy implication of our recommendations would be to reduce the number of invasive investigations and associated hospital costs. VER can be done as an outpatient procedure and lumbar punctures would require only a brief hospital stay. In some cases a repeat admission may be needed for myelography, but this is preferable to and more cost-effective than doing myelograms on all patients. Our findings also demonstrate how decision analysis allows a rational appraisal of the contributions made by investigations in the management of common clinical problems. This procedure also pinpoints areas where data are inadequate, so that the necessary information can then be collected. We no longer need to assess the value of newer investigations in everyday patient management intuitively when rational methods are available.

HEPATIC GLUTATHIONE S-TRANSFERASE RELEASE AFTER HALOTHANE ANAESTHESIA: OPEN RANDOMISED COMPARISON WITH ISOFLURANE AMANDA J. HUSSEY LAUREN G. ALLAN GEOFFREY J. BECKETT JANE HOWIE ALISTAIR F. SMITH JOHN D. HAYES GORDON B. DRUMMOND

University Departments of Anaesthetics and Clinical Chemistry, Royal Infirmary, Edinburgh

Plasma

of concentrations hepatic S-transferase (GST) are a glutathione more sensitive measure of acute hepatic damage than aminotransferase activity. Plasma GST concentrations have been measured by radioimmunoassay in an open randomised study after halothane or isoflurane anaesthesia. The concentration of GST was significantly increased after anaesthesia in patients who received halothane in 30% oxygen/70% nitrous oxide (n=37) and in patients who received halothane in 100% oxygen (n = 17). The frequency of abnormal GST concentrations, defined as 4µg/l or more, was 35% and 24%, respectively. GST concentrations usually reached a peak 3-6 h after the end of anaesthesia. In 17 patients who received isoflurane in 30% oxygen/70% nitrous oxide, there was no significant rise in GST concentration and no patient had a concentration above 4 µg/l. No patient in any of the groups had a significant increase in alanine aminotransferase. In clinically identical situations, anaesthesia with halothane but not isoflurane leads to demonstrable impairment of hepatocellular

Summary

integrity. Introduction THE frequency of unexplained hepatitis after halothane administration has not been established. The incidence of fulminant hepatic failure may be as high as 1 in 7000 with a 20-50% mortality rate.2 First described in 1958,3 the nature,

incidence, mechanism,

and

predisposing

factors for this

Correspondence should be addressed to R. I., Department of Neurology, Hospital, St Kilda Road, Melbourne, Australia 3004.

Prince Henry’s

REFERENCES

RA, Hillman BJ, McLennan JE, Strand RD, Kaufinan SM. Sequelae of metrizamide myelography in 200 examinations. AJR 1978, 130: 499-502. 2. Mastaglia FL, Black JL, Cala LA, Collins DWK. Electrophysiology and avoidance of invasive neuroradiology in multiple sclerosis. Lancet 1980; i: 144. 3. Iansek R, Balla JI. Decision analytical approach to the role of VER and CSF abnormalities in the management of singular spinal sclerosis. Clin Exp Neurol 1985; 21: 249-55. 4. McAlpine D, Lumsden CE, Acheson ED. Multiple sclerosis: an appraisal. Edinburgh: Churchill Livingston, 1972. 5. Merrill CR, Goldman D, Van Keuren ML. Simplified silver protein detection and image enhancement methods in polyacrilamide gels. Electrophoresis 1982, 3: 17-23. 6. Marshall J. Spastic paraplegia of middle age. Lancet 1955; i: 643-46 7. Bartel DR, Markand ON, Kolar OT. The diagnosis and classification of multiple sclerosis: evoked responses and spinal electrophoresis. Neurology 1983; 33: 611-17. 8. Paty DW, Blume WT, Brown WF, Taaloul N, Kerstesz A, McInnis W. Chronic progressive myelopathy: investigation with CSF electrophoresis, evoked potentials and CAT scan. Trans Am Neurol Assoc 1978, 103: 110-12. 9. Assalman P, Chadwick DW, Marsden CD. Visual evoked responses in the diagnosis and management of patients suspected of multiple sclerosis. Brain 1975; 98: 1. Baker

261-82. 10. Blumhardt LD, Barrett G, Halliday AM. Clinical application of evoked potentials in neurology. In: Courjon J, ed. The pattern evoked potential in the clinical assessment of undiagnosed spinal cord disease. New York: Raven Press, 1980. 11. Latchaw RE, Hirsch WL, Horton JA, Bissonette D, Shaw DD, Iohexiol VS. Metrizamide: Study of efficacy and morbidity in cervical myelography AJNR 1985; 6: 931-33.

remain unknown.4,5 Medical evidence and legal decisions that have incriminated halothane, mainly by temporal association of anaesthesia with jaundice, have caused a dramatic decline in halothane use for adults in the USA. Elsewhere, the rarity of unexplained hepatitis after halothane and the failure to find irrefutable evidence against halothane have ensured this agent’s continued popularity and sustained the controversy. 1,7 Diagnosis of mild halothane-induced hepatitis is difficult whether made on clinical, biochemical, or histological criteria. Plasma or serum aminotransferase activity8 is generally regarded as a sensitive measure of acute hepatic damage.9 However, aminotransferase activity is not specific to the liver, because these enzymes are released in other conditions,8 and may correlate poorly with hepatic histology.lO Alanine and aspartate aminotransferase (ALT and AST) are commonly used to assess hepatic injury after general anaesthesia but have yielded conflicting results and, if used alone, may be misleading.4 The measurement of hepatic glutathione S-transferase liver

damage

(GST) by radioimmunoassay (RIA)11 offers potential advantages over the aminotransferases in the investigation of hepatic damage. GST has a molecular mass of 45 000 to 50 000 daltons and is readily and rapidly released into blood after hepatic damage. In contrast to the periportal location of aminotransferases, GST is primarily distributed in centrilobular hepatocytes.12 This location may be more relevant to the study of unexplained hepatitis after halothane, which classically causes centrilobular necrosis.13 Plasma GST concentrations provide a more sensitive index of acute hepatocellular damage than aminotransferase activity. 14,15 Also, when the active phase of hepatic damage is over, plasma GST concentrations rapidly revert to normal, a feature of the short plasma half-life (under 90 min), whereas aminotransferase activity may be abnormal for much longer. 14 Plasma GST concentrations correlate better than aminotransferases with hepatic histology, 10 an important factor when inferring hepatotoxicity by suspected agents.9 These features suggest that GST measurement would have specific advantages over conventional hepatic enzymes in the investigation of acute hepatic damage.