Sudanese Journal of Paediatrics 1984; 3: 10-19

Review Article

LABORATORY DIAGNOSIS OF MENINGITIS: AN OVERVIEW INRELATION TO THE SUDAN,

Mustafa AbdallaH Salih,MB, BS, MPCH,MDDepartment of Paediatrics and Child Health,Faculty of Medicine, University of Khartoum

INTRODUCTION

Sudan lies within the meningitis belt and merringo-coccal disease has been one of its major problemsthroughout the twentieth century. Over 12,000 and 55,000cases were recorded in the years 1935 and 1950 respec-tively; and during the period 1968-1980, over fortythousands (40,513) were inflicted with the disease ofwhom 1,220 have died1, The magnitude of the problem inthe paediatric age group could easily be imagined in acountry where almost half of its population (= 20 mil-lions) are below 16 years of age. During the first 6months of the 1980 outbreak, 1,643 children (age group1=14 years) had meningitis from a total of 3,8001• Over60,% of those cases occurred in Khartoum Province 'Whichhas the largest crowded residential areas in the country.

The establishment of the diagnosis early in thecourse of the disease is not only important from auepidemiological point of view ~ through highlighting theneed for specific immunization for prevention as it isthe case in meningococcal meningitis - but also mattersfor every patient in such a life-threatening infection.The clinical signs and symptoms of meningeal disease ..though of great help are nonspecific. Fever, bulgingfont.ane Ll , nucka I rigidity, Kerrring+s sign and B'rudz.-niski's sign only suggest an intracranial pathology. Theycan result from tuberculous, fungal and aseptic meningi-tis and can be the manifestation of a brain abscess or atumour2. In the tropics, other diseases like malaria may

11 NAN Salih

closely simulate the disease. In a study of 117 patientswith meningitis admitted to the medical wards of Muhim-bili Medical Centre at Dar es Salaam, the disease wasconsidered initially in only 40 (34%) patients. Twenty-nine (25%) cases were labelled as malaria while theremaining patients either had no referral diagnosis orwere thought to have other problems. The impact of such adelay in diagnosis is even more harmful in children withmeningitis. It was found to influence the frequency ofneurological sequelae in Ethiopian children and thepresence of neurological abnormality at the time ofdiagnosis was found to increase the case fatality ratesignificantly4.

LABORATORY DIAGNOSIS

The diagnosis of meningitis depends upon carefulexamination of cerebrospinal fluid (SCF) obtained bylumbar puncture. The SCF specimen is probably one of themost critical specimens processed by a laboratory sinceit is obtained throught a potentially - though negligibly- hazardous procedure, requires skill to undertake andmight lead the way to the correct management of a life-~threatening disease. So the chance should be taken toextract maximum information when undertaking a lumbarpuncture.

CSF pressure, proteins and glucoseMeasurement of CSF pressure is an important compo-

nent to examine since it reflects the degree of intra-cranial pressure. When found to be considerably high,care should be taken not to draw too much of the fluid.The colour of CSF is also a helpful guide, whether tur-bulent as it is in the usual case of pyogenic meningitisor xanthochromic following subarachnoid haemorrhage orwhen pus is present in considerable amounts.

CS;F proteins are moderately Increased in bacterialmeningitiS usually to a level below 0.2 g/lOO ml. How-ever, such an increase may accompany other inflammatorydiseases of the nervous system such as encephalitis andpoliomyelitis. On· the other hand, remarkable depressionof CSF glucose and the ratio of CSF to blood glucose

Laboratory diagnosis of meningitis 12

levels (normally about 66%) is highly indicative ofbacterial meningitis. CSF glucose might even be absent insome cases.

CSF lactate

The increase in CSF lactate concentration inpatients with bacterial meningitis was noticed as earlyas in the 1920' 85 and was found later to be associatedwith anaerobic glycolysis of cerebral tissue6 possiblydue to decreased cerebral blood flow and cerebralhypoxia7. The value of this investigation was emphasizedin earlier studies and was reported to be helpful in theearly and differential diagnosis of bacterial meningi-tisS, 9. However~ subsequent investigations observed theoccasional occurrence of normal lactate level in thepresence of septic meningitis10 whereas othersl1 com-pletely doubted its use. In a study of 185 specimens ofCSF including those of 29 patients with culture-provenbacterial meningitis, Rutledge et al11 reported thepredicitive value of increased CSF lactate concentration(Le. greater than 3 mMol/l) to be 26.7%, the remainderbeing false positives. They also observed in an exper1-mental model that only after 3 days of meningitis did theCSF lactate concentration increase above 3 mMol/l sug..gesting that the lactate measurement has no role in earlydiagnosis of bacterial meningitis.

Cytology

The normal CSF contains a small number of cells,usually not exceeding 5 lymphocytes. To examine for thenumber and identify the different types of cells thatreflect the pathological state, the fluid should beexamined immediately. After centrifugation, a counterchamber is used to define the total nwnber of white bloodcells (WBCs) and a differential cell count is performedon a Wright-stained smear of the sediment. Pyogenicorganisms excite mainly polymorphonuclear leucocytosis.Infections with neurotropic viruses present with mono-nuclear pleocytosis although polymorphonuclear cellsmight be observed during the acute stage of the latterdisease. The number of polymorphonuclear cells present inpyogenic meningitis are usually very large; but this isnot always the case. Children admitted rather early in

13 MAM Salih

the course of .the disease may have a normal CSF initial~Iy , which if repeated later due to deterioration of thepatient wb1l.1d reveal meningitis 12, Moreover, cases ofbacterial meningitis have been reported who showed nomeningeal signs and had relatively low BCY pleocytosis.Sixteen (1.5%) of the 1~064 cases of bacterial meningitisbeyond the neonatal period = reviewed by Gieseler andNelson 13 - had no meningeal signs during their entirehospitalization. In nine of those (56%) J first lumbarpuncture disclosed fewer than 1,000 \¥'"BCs~tJhereas in seven(44%) the CSF contained· ( WBCs/mm3,

Direct microsc~py

Direct micros£opy of stained CSF, although notspecific, is a. sinsitive method that guides treatment andcould give a rapid answer. The yield obtained by Gramstain in cases of pyogenic meningitis was reported to beabout 75%11;14. However ~ its interpretation might provedifficult in mixed infections15,

Cultural methods

The culture technique in meningitis is a well-established diagnostic tool. A specimen of CSF fluidshould be cultured On a blood agar plate, a chocolateagar plate, on Fildes or Levent.ha1 media and in broth.However, pretreatment tends to render sterile the CSFfluid; and this was mostly found In pneumococcal andmeningococcal disease2. This presents a significantproblem in the Sudan; since inadvertent intake of anti-biotics by the patients before reporting to the hospitalwas found to be a major handicap in identifying theorganisms during epidemics of infections diseases inSudanese children. During an outbreak of diphtheria inKhartoum that involved 107 children~ only 41 cases(38.3%) could be bacteriologically proven. Cultures werenegative in the remaining 66 patinets of whom 43 (40.3%)had received antibiotics before reporting to hospital16•In the study of Sippel et all of the 1980 cerebrospinalmeningitis (CSM) outbreak Khartoum Province, only 25%of 114 CSF specimens of patients with probable CSM gave apositive yield.

MAN Salih· 14

Specific se~ologic testsRecently, several rapid and specific serological

methods were found to be of great use in demonstratingbacteria in the CSF. These were immunofluorescence(IFL)17, immunoelectroosmophoresis (IEOP)'8 and radio-immunoassayl? IFL detects cell-bound antigens in CSFwhereas IEOP demonstrates soluble antigens. The formertechnique was found to have higher sensitivity whencumpared to the other serological methods and gave acorrect diagnosis in 84% of 75 patients with meningitisin one of Olcen's series20. In this study, IEOP yieldedcorrect results in 55% of ca~es and was comparable to theresults reported by others2/1• However, the latter threemethods became less popular because they require high-powered experience for interpretation, are laborious todo and necessitate the use of comparatively expensiveequipment.

In 1973, Dr Goran Kronvall22 from the University ofLund, Sweden, introduced the term "co-agglutination" as anew serological method for typing pneumococci. The prin-ciple of this new technique was based on the phenomenonthat when adding a specific antipneumococcal antibodiesto a stabilised suspension of Cowan 1 strain of Staphylo·coccus aureus, the antibodies bind with the Fe parts toprotein A on the cell wall, leaving their Fab parts freefor reactions with the corresponding antigen. The re-sulting staphylococcal-bound antibodies readily co-ag-glutinate when coming in contact with a specific pneumo-coccal capsular antigen; and the result of the reactionis readily visible. This technique has been used sincethen for identification of other bacteria includingstreptococci and meningococci and in the latter case itwas found to detect both cell-bound and soluble bacterialantigens which are present in the CSF2b. In 1974,Pharmacia Diagnostics in Sweden obtained the commercialrights for the co-agglutination technique and producedlater Phadebact CSF Test which contains 4 differentreagents for the identification of the commonest orga-nisms causing ,bacterial meningitis i.e. H influenzae typeb , S Pnewnoniae and N men inqi.t-idi.e . The reagents arecapable of detecting all clinically relevant serotypes ofS Pnewnoniae and can detect groups A, B, C, Y and W-135of N meninqi.tridie- Before performing the test, the CSF

15 MAM Salih

should be heated for 5 minutes at 80°C to destroy theproteins which were found to cause autoagglutinationyielding a noninterpretable result. Prior centrifugationof the specimen may also be required in cases of thepresence of excessive amounts of protein or blood. Thetest only needs one drop of CSF to be added to one dropof each reagent and the result is read within 2 minutes.The sensitivity and specificity of the test is well-established now. In a study by Drow et a123 where a totalof 2,817 individual tests were performed on 577 CSFspecimens, 74% of the positive results could be detectedusing Phadebact CSF Test whereas counterimmunoelectro-phoresis detected 65%. No falso-positive results wereobtained when using the co-agglutination method.

On the other hand, latex agglutination technique issimilar to co-agglutination, the reagent being in thiscase latex particles sensitised with specific antisera.However, the antibodies are randomly adsorbed onto latexparticles and some of the antibodies will be attached viatheir Fab portion rendering them unable to react withantigen. The test is now available commercially in a kitform (Slidex meningite-Kit; bio Merieux). It requires onedrop of CSF (preferably using the supernatant aftercentrifugation for 10 minutes) to be added to one drop ofeach of latex suspensions and the reaction (aggluti-nation) is read after 2 minutes. Its 4 reagents candetect N meningitidisgroup A and C, H influenzae type band S pneumoniae (83 serotypes). Comparing it to theconventional bacterial cultures and other immunologictechniques in 718 cases of purulent meningitis, Denis eta124 found it to have the same sensitivity as positivedirect examination of CSF (82% and 80.4%, respectively)and it yielded slightly more positive results than theculture technique (73.1%). On the other hand counter-immunoelectrophoresis was positive in 90% of cases.

In comparative studies using reagents sensitisedwith the same antisera, both co-agglutination and latextests were found to be about equally sensitive regardingtheir specificity and sensitivity25, 26. Both tests werealso found to be of special value in areas with limitedlaboratory facilities. In a study by Whittle et a127 inZaria (Nigeria) comparing latex test, counter-current

MAM Salih 16

immunoelectrophoresis, eIE and bacteriological tests in229 CSF samples from patients with pyogenic meningitislatex agglutination was reported to a higher yieldthan routine bacteriology (Gram stain or culture) andidentified respectively 86%" 82% and 94% cases ofmeningococcal 9 pneumococcal and H inftuenzae meningitis 0

It showed about equal sensitivity to eIE in detectingmeningococcal and H inftuenzae but was less sensitive inrevealing pneumococcal meiningitis (respectively 82%versus 98%). The latex reagents were found to be stableand not adversely affected if left inadvertently at ahigh room rempe rat.ure for a few days. On the other hand ~in a recent study by Sanborn and Toun:!28, co+agg l.ut i>nation technique proved its adaptability for use in ruralAfrica. Designing simple, easily operated diagnostic kitthey found that West African medical at.t endat.s 1 afteronly 6 teaching hours, could the tests indepen-dently and with high degree of accuracy, The co+agg.l ut i ~nation reagents which had been prepared in Californiashipped to Upper Volta showed no 1013.13 of reactivity whenexposed to a temperature of 30-40oC at the field for 8weeks. The authors drew the attention to the fact thatthe co-ag91utination can eosily belocally using the laboratorysuccess of implementing this method in ct.ne rcount.rri.es woul.d mean a boost to primarycare in rural areas where of population

I am indebted to Professor Yngve Hofvande r of theInternational Child Health at the University Hos~pitaI of v Upps.a.l.a for his deep interest and supportthroughout the period of this study; and to Dr GunnarAhlsten of the Paediatric Department for placing facili-ties at my disposal which formed the background to thiswork. My sincere thanks are due to Dr Per O'lcen and DrDan Dani.e Lsson from the Departments of InfectiousDiseases and Clinical Batterology at 6rebro MedicalCentre Hospital for their stimulating discussionswritings which helped me laying the foundations for thisarticle. I am also grateful to Iva Kulhanek for preparingthe manuscript. This work is supported by the SwedishCommission for Technical Co-operation and the MedicalResearch Council.

17 MAM Sa1ih

REFERENCES

1. Sippel JE, Sid Ahmed H, Abu Ahmed H, Mikhail IA.Cerebrospinal meningitis in the Sudan in 1980.

Sudan Med J 1982; 18: 39-46.2. Feigin DF. Bacterial meningitis beyond the neonatal

period. In: Feigin RD, Cherry JD, eds, Textbook ofPaediatric Infectious Diseases. Philadelphia: WESaunders, 1981:851-7.

3. Mc Larty DG, Maseu GNK. Pitfalls in clinical diag-nosis of meningitis in adults.Tropical Doctor 1983;13:29-30.

4. Hailemeskel H, Tafari N. Bacterial meningitis inchildhood in an African city., Factors influencingaetiology and outcome. Acta Paediatr Scand1978;67:725-730.

5. Nishimua K. The lactic acid content of blood andspinal fluid. Proc Sac Exp Biol Med 1925;22:322-324.

6. Menkes JR. The causes for low spinal fluid sugar inbacterial meningitis: another look.Pediatrics 1969;44:1-3.

7. Paulson OB, Hansen EL, Kristensen HS, Brodersen P.Cerebral blood flow, cerebral metabolic rate ofoxygen and CSF acid-base parameters in patients withacute pyogenic meningitis and with acute encephali-tis.Acta Neural Scand 1972;48 (suppl 51):407-408.

8. Controni G,~Rodriquez WJ, Hicks 3M, Ricke M, Ross S,Friedman G, Kahn W. Cerebrospinal fluid lactate acidlevels in meningitis. J Pediatr 1977;91:379-384.

9. Lauwers S. Lactic-acid concentration in cerebro-spinal fluid and differential diagnosis of meningi-tis. Letter. Lancet 1978;2:163.

10. Donald PR, Malan C, Dykman JR. Determination oflactate concentrations in cerebrospinal fluid forrapid diagnosis early in bacterial meningitis.Letter. J Infect Dis 1982;145:187-188.

11. Rutledge J, Benjamin D, Hoos L, Smith A. Is the CSFlactate measurement useful in the management ofchildren with suspected bacterial meningitis?J Pediatr 1981;98:20-24.

12. Onorato 1M, Wormser GP, Nicholas P. Normal CSF inbacterial meningitis. JAMA 1980 ;244: 1469-1471.

MAM Salih 18

13. Geiseler PJ, Nelson KE. Bacterial meningitis withoutclinical signs of meningeal irritation.SouthMed J 1982;75:448-450.

14. Olcen P. Serological methods for rapid diagnosis ofHaemophilus influenzae, Neisseria meningitidis andStreptococcus pneumoniae in cerebrospinal fluid: Acomparison of co-agglutination, immunofluorescenceand immunoelectroosmophoresis.!Scand J Infect Dis

1978;10:283-289.15. Herweg JC, Middelkamp IN, Hartmann AF. Simultaneous

mixed bacterial meningitis in children. J Pediatr1963;63:76-83.

16. Salih HAM, Suliman GI, Hassan HS. Complications ofdiphtheria seen during the 1978 outbreak in Khartoum.

Ann Trop Paedat:r 1981;1:97-10l.17. Biegeleisen JZ, Mitchell MS, Marcus BB, Rhoden DL,

Blumberg RW. Immunofluorescence technique fordemonstrating bacterial pathogens associated withcerebrospinal meningitis. I. Clinical evaluation ofconjugates on smears prepared directly from cerebro-spinal fluid sediments.IJ Lab Clin Med1965;65:976-989.

18. Feiging RD, Wong M, Shachelford PG, Stechenberg BW,Dunkle LM, Kaplan S. Countercurrent immunoelectro-phoresis of urine as well as of CSF and blood fordiagnosis of bacterial meningitis. J Pediatr 1976;89:773-775.

19. Kayhty H, Makela PH, Ruoslahti E. Radioimmunoassayof capsular polysaccharide antigens of group A and Cmeningococci and Haemophilus influenzae typ b incerebrospinal fluid. J Cl.in Path 1977;30:831-833.

20. Olcen P. Infections with Neisseria meningitis:laboratory aspects, epidemiological and clinicalimplications. MD Thesis: Goteborg, Sweden, 1978.

21. Colding H, Lind I. Counterimmunoelectrophoresis inthe diagnosis of bacterial meningiti~ Clin Microbiol

1977;5:405-409.22. Kronvall G. A rapid slide-agglutination method for

typing pneumococci by means of specific antibodyadsorbed to protein A-containing staphylococci.

Med Microbiol 1973;6:187-190.

19 MAN Salih

23. Draw DL, Welch DF, Hensel D, Eisenach K, Long E,Slifkin M. Evaluation of the Phadebact CSF test fordetection of the four most co~~on causes of bacte-rial meningitis J GUn Microbial 1983; 18: 1358-1361.

24. Denis F 1 Saulnier M, Chiron SP. Rapid etiologicdiagnosis of purulent meningitis by reversed passiveagglutination of latex particles and by counter-immunoelectrophoresis: experience and prospects.

BulZ- rVHO 19S1;59:143-154.25. Onododi JK, Wauters G. Capsular typing of Kleb-

s co-agglutination and latex agglutination.26. Thirumoorthi Me, Dajani AS. Comparison of staphylo-

coccal agglutination, latex agglutination, andcount.e r.immuneLect rcpho res Ls for bacterial antigendetection. J" GUn Microbial 1979; 9: 28-32.

27. Whittle He, Tugwell P, Egler LJ, Greenwood Eli. Rapidbacteriological diagnosis of pyogenic meningitis bylatex agglutination. Lancet "1974;ii:619-621.

28. Sanborn WR, Toure IM. A simple kit system for rapiddiagnosis of cerebrospinal meningitis in rural areasof developing cout r i es . BuZZ WHO 1984,-62: 293-297.