Role of the BACTEC radiometric method in the evaluation of patients with clinically probable tuberculous meningitis

Background: Confirmation of tuberculous meningitis (TBM) and initiation of treatment are often delayed due to limitations in isolating Mycobacterium tuberculosis from cerebrospinal fluid (CSF). Objectives: To evaluate the role of the BACTEC radiometric method in a clinical setting for the early diagnosis of TBM. Materials and Methods: Patients meeting criteria for clinically probable TBM over a 3 year period were included. Clinical features, results of CSF investigations (protein, glucose, cell count, Ziehl-Neelsen staining, culture in Löwenstein-Jensen (LJ) medium and BACTEC) and brain CT imaging were reviewed. Drug sensitivity was tested using BACTEC. Patients were started on standard treatment and functional outcome, and response at discharge and follow-up were assessed. Patients were divided according to whether or not M. tuberculosis was isolated by BACTEC and the clinical, radiological, and laboratory features compared. Results: Sixty patients were evaluated. The mean age was 30 years ± 11.7 years. Headache and fever were the most common symptoms and the mean duration was 26 days. CT findings were hydrocephalus (n=21), basal exudates (n=16), and tuberculoma (n=14). In 40 patients, M. tuberculosis was isolated by BACTEC and average 15 days was required for detection, whereas it was 30 days in LJ medium. Results of drug-sensitivity testing (n=32) were obtained average 7 days after isolation. Patients from whom M. tuberculosis had been isolated by BACTEC more often had tuberculomas in CT imaging (P=0.018). Conclusion: Use of the BACTEC method allows early confirmation in patients with clinically probable TBM. It can guide clinicians in the rational use of anti-tuberculosis treatment by confirming diagnosis and identifying drug- sensitivity.


Introduction
Tuberculosis is a serious public health problem. [1] Though tuberculous meningitis (TBM) accounts for only around 10% of cases, it is the most severe form of extrapulmonary tuberculosis and most common manifestation of central nervous system disease. [2] It is a common cause for chronic meningitis and is associated with signifi cant morbidity and mortality. Early recognition and initiation of treatment are the most important factors predicting favorable outcome in TBM. [2] Though the diagnosis may be suspected based upon clinical features, [3] isolation of Mycobacterium tuberculosis from cerebrospinal fl uid (CSF) is the gold standard for confi rming the diagnosis. The standard tests used include microscopy by Ziehl-Neelsen (ZN) staining, culture in Löwenstein-Jensen (LJ) medium, and polymerase chain reaction. However, limitations in isolating M. tuberculosis from CSF using these standard laboratory techniques delay the confi rmation of the diagnosis. [2,4,5] The BACTEC radiometric method has been used for several years for rapid isolation of bacteria. [6] It is based on the

Role of the BACTEC radiometric method in the evaluation of patients with clinically probable tuberculous meningitis
computed tomography (CT) study showing basal meningeal enhancement, hydrocephalus, infarcts, or (3) evidence of extraneural tuberculosis. [3] Patients were evaluated by history taking and standard physical and neurological examination. Severity of TBM was staged according to Medical Research Council (MRC) criteria. [9]

Microbiological Evaluation
Clinical specimens 10-15 mL CSF was obtained by lumbar puncture and protein, glucose and cell count were assayed. ZN technique was used to stain for acid-fast bacilli (AFB). Slides were examined under the oil immersion lens for about 30 min, with care taken to view at least 300-500 high power fi elds and re-examined by an independent examiner.
0.5-1 mL of CSF was inoculated in BACTEC 12B vials, incubated at 37°C, and evaluated for 14 CO 2 using the BACTEC 460TB system (Becton Dickinson, USA). As recommended, 0.1 mL of PANTA supplement (polymyxin B 50 µg/mL, amphotericin B 5 µg/mL, nalidixic acid 20 µg/mL, trimethoprim 5 µg/mL and azlocillin 10 µg/mL, Becton Dickinson) was added to each vial prior to inoculation. Culture media were incubated at 37°C for 6 weeks and tested for growth twice weekly for the fi rst 3 weeks and weekly for the subsequent 3 weeks. 14 CO 2 production was expressed in terms of the growth Index (GI) and change in GI was proportional to mycobacterial growth. Vials with GI greater than a value of 10 were evaluated daily, and when the GI surpassed 50, specimens were stained with ZN and Gram's stain. BACTEC NAP (p-nitro-α-acetylaminoβ-hydroxypropiophenone) test was used to differentiate M. tuberculosis from Mycobacteria other than tubercle (MOTT) bacilli. [6] Cultures were negative for growth if there was no signifi cant change in GI aft er 6 weeks of incubation.
Drug-sensitivity for streptomycin, isoniazid, rifampicin, and ethambutol was tested using the BACTEC system by the modifi ed proportion method. [6] LJ culture About 0.5 ml of CSF was inoculated on LJ slopes and incubated at 37°C for 8 weeks. The LJ slants were inspected weekly and growth on LJ slants resembling mycobacterial colonies, i.e. rough, tough and buff -colored, were stained with ZN and Gram's stain to confi rm the presence of AFB and rule out contamination.

Chronic meningitis evaluation
Relevant investigations were performed to exclude other causes for chronic meningitis such as complete blood count, erythrocyte sedimentation rate, infection screen, CSF bacterial and fungal culture, and cytology of cytospin centrifuged samples (Shandon Cytospin, UK).

Computed tomography
Computed tomography was performed during the acute stage of illness using third-generation scanner and contrast-enhanced images were reviewed.

Treatment and follow up
Patients were started on standard four drug anti-tuberculosis treatment (isoniazid 5-10 mg/kg/day, rifampicin 10 mg/kg/ day, pyrazinamide 35 mg/kg/day and ethambutol 20 mg/kg/ day or streptomycin 15 mg/kg/day) and prednisolone 1 mg/kg/ day, if not contraindicated. Once clinically stabilized, patients were discharged to home and continued on anti-tuberculosis treatment for 3 months and thereaft er changed to two-drug regimen with isoniazid and rifampicin. Functional outcome at discharge was graded using the modifi ed Rankin Scale (mRS). This is a seven point scale from 0 (no symptoms) to 6 (death) and patients' outcome was classifi ed as good (mRS ≤ 3) or poor (mRS 4 or 5) according to the evaluation at discharge. [10] Patients were regularly reviewed at the out-patient clinic once at 6 weeks and thereaft er every 3-6 months depending upon clinical condition. The functional outcome was reassessed at 6 months.
Patients were divided according to whether or not M. tuberculosis was isolated by BACTEC, and clinical, radiological, and laboratory features were compared between the two groups. Statistical analysis was performed using SPSS for Windows version 11.0 (SPSS Inc. Chicago, IL, USA). Chi-square test and Student's t test were used to compare clinical, CSF and radiological fi ndings. P-value of <0.05 was considered signifi cant.

Results
Sixty patients meeting the criteria for clinically probable TBM were evaluated (41 males, 19 females). Clinical features, CSF parameters, and radiological fi ndings are shown in Table 1.
In 40 patients, M. tuberculosis was isolated by BACTEC method and average 15 days was required to detect growth. M. tuberculosis was isolated in LJ medium in 15 patients [ Table 2] and average 30 days were required to detect growth. Results of drug-sensitivity testing using BACTEC (n=32) were obtained within a mean of 7 days (range 5 to 12 days) aft er isolating M. tuberculosis. Three patt erns were observed: no resistance to drugs (n=20), resistance to isoniazid alone (n=11), and resistance to both isoniazid and streptomycin (n=1).
The mean duration of hospital stay was 13.8 to 14.5 days. Six patients expired. Patients were discharged on anti-tuberculosis treatment and were regularly followed up during out-patient visits. The mean follow-up was 8.9 months. 82% follow-up was obtained and all patients showed favorable response to anti-tuberculosis treatment, without recurrence of meningitis symptoms [ Table 1].
Patients in whom M. tuberculosis had been isolated from CSF by BACTEC more oft en had tuberculomas in CT imaging (P=0.018) [ Table 1]. There were no diff erences in clinical features, functional outcome, or response to treatment at follow up.
to slow multiplication of M. tuberculosis or the strong host immunological response. [2,5,11] The duration of meningitis symptoms as well as CSF parameters such as glucose levels and neutrophil counts may infl uence bacterial isolation from CSF. [12] Previous laboratory based-studies that compared isolation of M. tuberculosis from CSF samples demonstrated increased sensitivity of the BACTEC radiometric method over culture in LJ medium. [7,8] The present study correlated clinical and radiological features of patients with probable TBM with the results of BACTEC and evaluated the role of this investigation in a clinical sett ing. The time required for isolating M. tuberculosis was less when using the BACTEC method (average 15 days) compared to the standard culture in LJ medium (average 30 days), resulting in earlier confi rmation of TBM. Use of the BACTEC method also enabled more rapid assessment of drug-sensitivity, and results were obtained in an average of 7 days aft er isolation. This is in contrast to the usual 1-2 months required when using LJ medium. [2] Drug resistance was seen in 37% of Patients with poor outcome at discharge (n)*

Discussion
Establishing that the cause for chronic meningitis is due to tuberculosis depends upon microbiological tests. Demonstration of acid-fast bacilli in CSF, either in centrifuged smears or aft er isolation in culture, is essential for confi rming the diagnosis of TBM. [2,5] However, TBM is a paucibacillary condition and the yield from standard investigations is low and results are oft en delayed. This may be partly due isolates and isoniazid resistance was most common (n=11). Primary drug resistance was more likely, as nine of these patients were never exposed to anti-tuberculosis drugs previously. In ten patients with single drug resistance, second line anti-tuberculosis treatment could not be started due to fi nancial constraints. Incidentally, these patients continued to show good clinical response.
The observation that patients from whom M. tuberculosis was isolated in BACTEC more oft en had tuberculomas requires to be confi rmed in a larger series of patients. Necrosis is associated with higher bacillary load and the possibility that some of the lesions had additional necrosis could not be excluded. [13] The lack of association between isolation in BACTEC and LJ culture may refl ect diff erences between the two media. Several factors are known to contribute to the increased isolation rate in BACTEC including addition of CO 2 which promotes mycobacterial growth, and antimicrobial agents (PANTA) which reduce contamination. [2,6] However in two patients, M. tuberculosis was isolated in LJ medium but not in BACTEC. This may have been due to inherent differences between solid (LJ medium) and liquid media (BACTEC) with regard to the presence of specifi c growth promoting factors. [7,14] This observation emphasizes the importance of using both types of media when att empting to isolate M. tuberculosis from the CSF of a patient with clinically probable TBM.
Comparison between microbiological techniques was beyond the scope of this clinical study. Drawbacks associated with the BACTEC method include involvement of laborintensive procedures, potential exposure to radio-labeled substrates and expense for set-up and maintenance. [6] These factors would restrict the use of this investigation to tertiary referral centers.
Notwithstanding these drawbacks, our prospective study demonstrates numerous advantages of the BACTEC method that make it a useful investigation for clinicians to confi rm the clinical diagnosis of TBM. Patients may have an early confi rmation of the diagnosis during the acute stage of the illness. Reports of culture and drug-sensitivity would be available within 3-4 weeks, while a negative report would be available within 6 weeks.