651

DETECTION OF MYCOBACTERIAL ANTIGENS step procedurewas added, and a 1 mg/ml solution of

IN CEREBROSPINAL FLUID OF PATIENTS WITH p-nitrophenylphosphate was used as substrate. Absorbance at 405
TUBERCULOUS MENINGITIS BY ENZYME- nm was read in an ELISA reader (’Titertek’; Dynatech, Alexandria,

LINKED IMMUNOSORBENT ASSAY Virginia); BCG protein standards of 3 ng/ml, 7 ng/ml, and 14 ng/ml
gave absorbance values of of 0 -122,0 - 213, and 0 - 325, respectively.
In samples values ≥0. 20 were considered positive for the presence
EDUARDO SADA GUILLERMO M. RUIZ-PALACIOS of M tuberculosis antigens. Each sample was run in duplicate, and
YOLANDA LÓPEZ-VIDAL SERGIO PONCE DE LEÓN commercial BCG protein was used as positive control. Statistical
Department of Infectious Diseases and Teaching Division, analysis was done by the Mann Whitney U test for two samples.
Instituto Nacional de la Nutrición, Mexico
Results
Summary Culture and identification of Mycobacterium The mean absorbance of the sixteen samples from patients
tuberculosis in cerebrospinal fluid (CSF) of with tuberculous meningitis (group I) was 0.619±SD0.676
patients with tuberculous meningitis is slow, and prognosis (see figure). In all four patients with confirmed diagnosis
depends on early diagnosis and tratament. An enzyme-linked (group IA) the absorbance was above 0-20 (mean
immunosorbent assay with commercial BCG and anti-BCG 0.511±0.112). Of the twelve patients with probable
was designed to detect mycobacterial antigens in human body tuberculous meningitis (group IB), absorbance was above
fluids. Absorbance at 405 nm was measured and readings 0’ 20 in nine (mean for all twlve 0 -655±0.783). In group II,
≽0·20 were considered positive. The readings were above all samples but one had an absorbance below 0 - 20 (mean
0·20 for four patients with confirmed tuberculous meningitis 0.098±0.071); the exception was the patient with
and nine of twelve patients with probable tuberculous cryptococcal meningitis, whose sample had an absorbance of
meningitis. In all but one of ten patients with meningitis of 0’ 293. In all eleven control subjects (group III) readings were
other causes and in eleven control subjects the readings were below 0’ 20 (mean 0 -.086±0.037). The mean absorbance in
below 0·20. The specificity of the test was 95% and its group I was significantly higher than that in groups II and III
sensitivity 81&middot; 25% (U value 4-701, p<0. 001). The specificity of the test was
95% (one false-positive out of twenty-two negative samples
Introduction
tested). There were three false-negative results among sixteen
TUBERCULOUS meningitis remains a common disease in CSF samples from patients with tuberculous meningitis,
developing countries, and the prognosis is closely related to therefore the sensitivity was 8125%.
the stage of the illness at the initiation of treatment.’Culture
and identification of Mycobacterium tuberculosis in
cerebrospinal fluid (CSF) from patients with tuberculous
meningitis is slow and tedious. Results are not available for
2-4 weeks with routine methods,2 or for 1 week with other
assays such as selected ion monitoring,3 and a rapid and
reliable test to detect mycobacterial antigens in body fluids is
therefore needed. We now report the development of an
enzyme-linked immunosorbent assay (ELISA) to detect
soluble antigens in CSF of patients with tuberculous
meningitis. Chaparas and Hendricks have shown that
M tuberculosis and BCG have the same surface antigens, as
tested by a reference antiserum;4 we therefore used antibody
against BCG in the assay.
Patients and Methods
CSF samples were obtained from thirty-eight subjects. Group I,
consisting of sixteen patients with tuberculous meningitis, was
divided into group lA, four patients with bacteriologically Study groups
confirmed disease, and group IB, twelve patients with the Detection of mycobacterial antigens in 38 CSF samples by ELISA in
presumptive diagnosis of tuberculous meningitis, supported by a patients with tuberculous meningitis and controls.
CSF glucose level of <40 mg/dl, protein >40 mg/dl, pleocytosis
with lymphocytosis, negative culture for other bacteria and fungi, Discussion
plus two or more of the following-clinical picture of tuberculous The diagnosis of tuberculous meningitis is usally made on
meningitis, positive tuberculin skin test, evidence of tuberculosis in
other sites, or a clinical response to treatment. Group II consisted of the basis of a history of exposure, clinical picture,
five patients with bacterial meningitis other than tuberculous characteristic CSF, positive tuberculin skin test, and the
meningitis, two with viral encephalitis, two with cysticercosis, one culture of M tuberculosis in CSF.7-9 However, cultures are
patient with meningeal cryptococcosis, and one with meningeal positive in only 55-75% of patients.7,8 Other methods have
tumoral infiltration. Group III consisted of eleven subjects without been used in the diagnosis of tuberculous meningitis, such as
neurological disorders who had lumbar punctures for epidural radioactive bromide partition, reactivity of adenosine
anaesthesia.
The ELISA method used was essentially that of Engvall and
deaminase,1I or detection of mycobacterial metabolites such
as 3- (2-ketohexyl) indoline.12 However, all of these, though
Perlmann.5 Microtitre plates (’Immulon II’; Dynatech, Alexandria,
Virginia) were coated with 100 &micro;l of a 1/800 dilution of rabbit IgG helpful, are cumbersome and require sophisticated
anti-BCG (Dako Laboratories, New York). After overnight technology and are therefore not available in most
incubation at37&deg;C, 100 &micro;l samples of undiluted anda 1/2 dilution of laboratories. In contrast, ELISA is a simple and rapid method
CSF were added and incubated at 37&deg;C for 2 h. A 1/1000 dilution of and is useful in antigen detection of several other infectious
rabbit IgG anti-BCG conjugated to alkaline phosphatase by the two- diseases.13 Our results show that it is possible to detect
652

soluble antigens of mycobacteria in CSF of patients with

tuberculous meningitis. The high sensitivity and specificity
of the test suggest that it may become a useful tool in clinical Preliminary Communication
practice.
The false-positive result in the patient with cryptococcosis ALBENDAZOLE AS A POTENTIAL TREATMENT
could be due to the antibody used in our test cross-reacting FOR HUMAN HYDATIDOSIS
with the cryptococcal polysaccharide or other structures.
Other microorganisms, such as other mycobacteria, A. G. SAIMOT A. MEULEMANS
A. C. CREMIEUX M. D. GIOVANANGELI
corynebacteria, nocardia, and even more common bacteria, B. DELAITRE
share antigens withM tuberculosis. 14,15 The explanation of the J. M. HAY
J. P. COULAUD
false-negative results may be that at the time of the lumbar
puncture some patients were already receiving treatment, Institut de M&eacute;decine et d’Epid&eacute;miologie Tropicales, H&ocirc;pital Claude
which might have lowered the soluble-antigen concentration Bernard, Paris; Laboratoire de Biophysique, UER Xavier Bichat et
in the CSF. One way to improve the test is to use monoclonal Unit&eacute; INSERM U 13, Paris; H&ocirc;pital Louis Mourier, Colombes;
antibodies against M tuberculosis16 or antibodies directed H&ocirc;pital Cochin, Paris, France
against antigens 5 and 6 of M tuberculosis, which are specific
for this bacillus. 14,15 Summary The pharmacokinetics of albendazole was
The ELISA method has been used to detect mycobacterial evaluated in 11 patients with hydatid disease
who underwent surgery 12 h after the last dose of drug.
antigens in animal tissues 17 but, to our knowledge, this is the
first time that it has been used to demonstrate these antigens Albendazole and its main metabolite, albendazole-
in body fluids in human tuberculosis. This test could- have sulphoxide, were assayed in the serum from peripheral and
several clinical applications. It is possible that mycobacterial portal blood, the liver, bile, lungs, and hydatid cyst walls and
fluid. After a 10-14 mg/kg daily oral dose of albendazole,
antigens could be detected in other body fluids, simplifying concentrations of 1844&plusmn;904 ng/g wet tissue of albendazole-
the diagnosis of tuberculosis. The serial determination of
antigen concentration in the fluids of patients during sulphoxide were found in the liver and 749&plusmn;34 ng/g wet
treatment could be useful in follow-up and in assessing tissue in the lungs. The level in hydatid cyst fluid was
921&plusmn;314 ng/ml. The drug was also excreted through the bile.
prognosis.
The same daily oral dose produced very stable blood levels
We thank Dr Javier BAez-Villaselior, Dr Agustin Ocejo, Dr Judith
after 2 to 4 days of treatment (600-1000 ng/ml). In another
Dominguez, and the personnel of the Hospital de Infectologia Centro Medico
La Raza, IMSS, Mexico, for their collaboration in obtaining samples; Dr part of the study 3 patients with liver cysts, 2 with peritoneal
Beatriz Remus de Ruiz-Palacios for reviewing the manuscript; and Ms cysts, and 5 with bone cysts received 7 mg/kg twice a day
Guadalupe Perez R. for typing. This work was supported by a research prophylactically and/or therapeutically for 30 days. In all
training scholarship granted to E. S. from Secretaria de Salubridad y cases the treatment was repeated several times after intervals
Asistencia, Mexico.
of 2 weeks. The 3 patients with liver cysts were cured
Correspondence should be addressed to G. M. R.-P., Department of
Infectious Diseases, Instituto Nacional de la Nutricion, Vasco de Quiroga 15, (probably due to treatment); there was no recurrence in the 2
Tlalpan, Mexico 14000 DF, Mexico. patients with peritoneal cysts and very slight improvement in
the 5 cases with bone cysts. The drug was clinically and
REFERENCES
biologically extremely well tolerated.
1. Kennedy DH, Fallon RJ. Tuberculous meningitis. JAMA 1979; 241: 264-68.
2. Runyon EH, Karlson AG, Kubican GP, Wayne LG. Mycobacterium. In. Lennette INTRODUCTION
EH, Balows A, Hausler WJ, Truant JP. Manual of clinical microbiology. 3rd ed.
Washington DC: American Society for Microbiology, 1980: 151-79. HYDATID disease is the commonest and one of the most
3. Odham G, Lennart L, Mardh P. Demonstration of tuberculostearic acid in sputum
from patients with pulmonary tuberculosis by selected ion monitoring. J Clin Invest severe human larval cestodiases; it is also one of the rare
1979; 63: 813-19 helminth infections which has not benefited from the general
4. Chaparas SD, Hendricks SR. Comparison of strains of BCG. I. Antigenic analysis and
tuberculin reactivity. Infect Immun 1973; 7: 777-80. progress in chemotherapy. Surgery is not always effective or
5. Engvall E, Perlman P. Quantitation of specific antibodies by enzyme-labeled anti- possible and under some circumstances it might even increase
immunoglobulin in antigen-coated tubes. J Immunol 1972; 109: 129-35. the risks of dissemination.
6. Avrameas S. Coupling of enzymes to proteins with glutaraldehyde. The use of the
conjugate for the detection of antigens and antibodies. Immunochemistry 1969; 6: Research on therapy with benzimidazole-carbamates is in
43-52.
progress. So far the clinical trials of mebendazole and
7. Bell WE, McCormick WF. Neurological infections in children. Philadelphia: WB
Saunders, 1975: 73-89. flubendazole have not been well coordinated. The results are
8. Sumaya CV, Simek J, Smith MHD, et al. Tuberculous meningitis in children during inconclusive and should be reassessed by well-designed and
the isoniazid era. J Pediatr 1975; 87: 43-49.
9. Meyer BR. Tuberculous meningitis Med Clin North Am 1982; 66: 755-62. accurately controlled studies.
10. Mandal BK, Evans DI, Ironside AG, Puliam BR. Radioactive bromide partition test in Since another benzimidazole-carbamate, albendazole, has
differential diagnosis of tuberculous meningitis. Br Med J 1972; iv: 413-15.
been effective against hepatico-peritoneal cysticercosis in
11. Mann MD, Macfarlane CM, Verburg CJ, Wiggelinkhuizen J. The bromide partition
test and CSF adenosine deaminase activity in the diagnosis of tuberculosis rabbitsu2 and more recently against Echinococcus
meningitis in children. S Afr Med J 1982; 62: 431-33. multilocularis in rats (Euz&eacute;by J, unpublished), we decided to
12. Brooks JB, Choundhary G, Craven RB, et al. Electron capture gas chromatography
detection and mass spectrum identification of 3- (2-ketohexyl) indoline in spinal investigate this drug in three stages-the pharmacokinetics of
fluid of patients with tuberculous meningitis. J Clin Microbiol 1977; 5: 625-28. albendazole in healthy volunteers;3,4 the pharmacokinetics
13. Yolken RH. Enzyme immunoassay for the detection of infectious antigens in body
fluids: current limitations and future prospects. Rev Infect Dis 1982; 4: 35-68.
in patients with hydatid disease; and the therapeutic effect of
14. Daniel TM, Janicki BW Mycobacterial antigens: a review of their isolation, chemistry, the drug on various sites of infection in patients with hydatid
and immunological properties. Microbiol Rev 1982; 42: 84&mdash;113
15. Goren MB. Immunoreactive substances of mycobacteria. Am Rev Resp Dis 1982; 125: cysts. We report here the results of the second and third
50-69. stages.
16. Coates ARM, Hewitt J, Allen BW, Ivanyi J, Mitchison DA. Antigenic diversity of
Mycobacterium tuberculosis and Mycobacterium bovis detected by means of
monoclonal antibody. Lancet 1981; ii: 167-69. PATIENTS AND METHODS
17 Thoen CO, Malstrom C, Himes EM, Mills K. Use of enzyme-linked immunosorbent
assay for detecting mycobacterial antigens in tissues of Mycobacterium bovis infected
The 21 subjects consisted of 18 adults (15 men, 3 women, age-
cattle. Am J Vet Res 1981; 42: 1814-15. range 23-56) and 3 children (2 boys, 1 girl, age-range 10-12). 13