Thesis (MScMedSc (Pathology. Medical Microbiology))--University of Stellenbosch, 2010. / ENGLISH ABSTRACT: INTRODUCTION Mycobacterium tuberculosis (MTB) and Cryptococcus neoformans are the most common causes of chronic meningitis in South Africa. Conventional microbiology has limited utility in diagnosing these pathogens due to the paucibacillary nature of cerebrospinal fluid (CSF) and the diagnostic delay associated with culturing methods. This study aimed to evaluate the utility of an in-house polymerase chain reaction (PCR) method for the detection of the etiological agent of chronic meningitis. METHODS CSF samples (where volume exceeded 5ml) were submitted to the Medical Microbiology diagnostic laboratory of the Tygerberg Hospital from patients with suspected tuberculosis meningitis (TBM). Following routine bacteriology, the sample was used to inoculate two mycobacterial growth indicator tubes (MGIT A and B) and subsequently incubated in the BACTEC 960 automated system. MGIT A followed standard operating procedures and the time to culture positivity was noted. Weekly aliquots (up to 6 weeks) were removed from MGIT B. These samples were boiled to inactivate the bacteria and then the DNA was extracted using the Promega Wizard SV Genomic DNA kit. The DNA was then speciated by PCR and high-resolution melting analysis (HRM) by using primers specific to either the RD9 region of MTB complex or primers specific to the partial internal transcribed spacer 1 (ITS1), 5.8S rRNA gene and partial ITS2 sequence of C. neoformans. RESULTS Routine CSF microscopy indicated that 14 of the 78 patients (17.9%) had typical CSF findings of TBM (lymphocytes predominant, increased protein levels and decreased glucose levels). IV Ziehl-Neelsen (ZN) stains were positive for 12 (15.4%) samples, and MTB was cultured from 19 samples (24.4%). Our optimized PCR and HRM method was able to detect M. tuberculosis in 17 of the 19 culture positive specimens with a sensitivity of 89.5% and a specificity of 62.7%. The sensitivity of this method was higher than that of direct microscopy. In all of the PCR positive samples, the time to detection, compared to culture, could be shortened by 1 to 2 weeks. Only one sample was positive for Cryptococcus culture and another sample was positive with a Cryptococcus latex test. PCR for Cryptococcus was positive in 2 cases (n=78), sensitivities and specificities could not be reported due to the low number of positive cases. CONCLUSION We demonstrated that a short culture period and the use of commercial DNA extraction kit on CSF samples increases the sensitivity of molecular tests to diagnose tuberculosis. Furthermore, the molecular techniques could significantly reduce the time to positivity of results, when compared to culture. Due to the low occurrence of Cryptococcus in the samples included in our study, we could not comment on the diagnostic utility of PCR in the diagnosis of Cryptococcal meningitis, when compared to the conventional methods. / AFRIKAANSE OPSOMMING: INLEIDING Mycobacterium tuberculosis (MTB) en Cryptococcus neoformans is die mees algemeenste oorsake van kroniese meningitis in Suid-Afrika. Routine mikroskopie dra beperkte waarde in die diagnose van hierdie patogene as gevolg van die klein hoeveelhede organismes wat in die SSV (serobrospinale vog) voorkom en die lang tyd wat dit benodig om hierdie organisms te kweek. Hierdie studie beoog om die diagnostiese waarde van ‘n polymerase ketting reaksie (PKR) metode wat intern ontwerp is te evalueer vir die identifikasie van patogene verantwoordelik vir kroniese meningitis. METODES SSV monsters (waarvan die volume 5ml oorskry) en waar daar ‘n kliniese vermoede van tuberkulose meningitis (TBM) was, is na die diagnostiese Mediese Mikrobiologie laboratorium van Tygerberg hospitaal gestuur vir roetine bakteriologiese ontleding. Die oorblywende monsters is gebruik om twee mikobakteriële groei-indikasiebuise (MGIT A en B) te innokuleer en hulle is geïnkubeer in ‘n BACTEC 960 geautomatiseerde sisteem. MGIT A is volgens roetine diagnostiese metodes geanaliseer en die tyd tot ‘n positiewe resultaat is aangeteken Weeklikse monsters (tot en met week 6) is uit MGIT B verwyder en die monsters is gekook om sodoende die bakterië te inaktiveer. Die Promega Wizard SV Genomiese DNS ekstraksiemetode is gebruik om die DNS te versuiwer. Spesiëring van die DNS is deur middel van ‘n intern ontwerpte PKR en hoëresolusiesmeltingsmetode (HRS) gedoen met inleiers wat spesifiek is tot die RD9 gedeelte van die MTB kompleks en inleiers spesifiek tot die gedeeltelike interne getranskribeerde spasieerder 1 (ITS1), 5.8S rRNS geen en die gedeeltelike ITS2 DNS volgorde van C. neoformans. VI RESULTATE Roetine SSV mikroskopie het aangedui dat 14 uit 78 (17.9%) pasiënte tipiese SSV bevindings van TBM (oorwegend limfosiete, verhoogde proteïene en verlaagde glukose) gehad het. Ziehl- Neelsen (ZN) kleurings was positief vir 12 (15.4%) monsters, en MTB is gekweek in 19 (24.4%) van hierdie monsters. Ons geoptimaliseerde PKR en HRS metode het daarin geslaag om M. tuberculosis in 17 van die 19 kultuurpositiewe monsters aan te toon met ‘n sensitiviteit van 89.5% en ‘n spesifisitiet van 62.7%. Die sensitiwiteit van die direkte PKR was hoër in vergelyking met mikroskopie. In al die PKR positiewe monsters was die tyd tot aantoning, in vergelyking met kultuur, verkort met 1 tot 2 weke. Slegs een monster het C. neoformans gekweek en ‘n ander monster was positief met die kriptokokkale latekstoets. PKR vir C. neoformans was positief in 2 gevalle (n=78). Die sensitiwiteit en spesifisiteit van die C. neoformans PKR kon nie bepaal word nie weens te min gevalle. GEVOLGTREKKINGS Ons het aangetoon dat ‘n verkorte inkubasieperiode en die gebruik van ‘n kommersiële DNS ekstraksiemetode op SSV monsters die sensitiwiteit van die molekulêre tegniek vir die diagnose van tuberkulose verhoog en dat hierdie metode die tyd na positiwiteit aansienlik verkort in vergelyking met kultuur. Weens die lae getalle van kriptokokkale meningitis in ons studie kon ons nie kommentaar lewer op die akkuraatheid van PKR in die diagnose van kriptokokkale meningitis, in vergelyking met meer konvensionele metodes, nie.