Article Text

PDF

False negative polymerase chain reaction on cerebrospinal fluid samples in tuberculous meningitis established by culture
  1. M MELZER,
  2. T J BROWN
  1. Department of Microbiology, St Thomas’ Hospital, Lambeth Palace Road, London SE1 7EH, UK
  2. King George Hospital, Barley Lane, Goodmayes, Essex IG3 8YB, UK
  1. Dr M Melzer, Department of Microbiology, St Thomas’ Hospital, Lambeth Palace Road, London SE1 7EH, UK.
  1. J FLOOD,
  2. S LACEY,
  3. L R BAGG
  1. Department of Microbiology, St Thomas’ Hospital, Lambeth Palace Road, London SE1 7EH, UK
  2. King George Hospital, Barley Lane, Goodmayes, Essex IG3 8YB, UK
  1. Dr M Melzer, Department of Microbiology, St Thomas’ Hospital, Lambeth Palace Road, London SE1 7EH, UK.

Statistics from Altmetric.com

The polymerase chain reaction (PCR) has been reported to be of diagnostic value when performed on CSF samples in tuberculous meningitis.1-4 Rapid amplification ofMycobacteruim tuberculosis specific DNA enables results to be available within 48 hours and can influence treatment decisions.

Recently two patients presented to our hospital with symptoms and signs suggestive of tuberculous meningitis. Examination of CSF disclosed a lymphocytic exudate. Repeated samples were sent to a British referral laboratory where CSF PCR for M tuberculosiswas reported negative. Despite this, antituberculous treatment was continued for 12 months and both patients responded clinically. Several weeks after the negative PCR result, M tuberculosis was cultured on Lowenstein-Jensen slopes from CSF taken from both patients. False negative CSF PCR in tuberculous meningitis established by culture has rarely been reported. The two patients are described to emphasise the dangers of overreliance on PCR in cases of suspected tuberculous meningitis. Premature cessation of treatment would have had tragic consequences for the two patients concerned.

The first patient was a 28 year old Asian man, last in India 8 years previously. He was sent from a clinic to hospital for incision and drainage of two deep seated Staphylococcus aureus abscesses. While an inpatient he complained of headaches and nausea and developed a low grade pyrexia and meningism. Brain CT was normal. Lumbar puncture disclosed a high opening pressure (19 cm CSF), 133 white blood cells/μl, predominately lymphocytes, a raised protein (1.61g/l), and a low CSF/blood glucose ratio (1.7/6.1). A sample of 0.5 ml CSF was sent to a British referral laboratory and PCR for M tuberculosis was negative. Twenty four hours later, because of increasing confusion and agitation, treatment with intravenous acyclovir, antituberculous chemotherapy (600 mg rifampicin, 300 mg isoniazid, 2 g pyrazinamide, and 10 mg pyridoxine daily), and dexamethasone was commenced. Clinically he showed signs of improvement and was discharged home 2 weeks later on the above treatment. A repeat lumbar puncture 4 weeks later showed similar results. A CSF PCR for M tuberculosis was again negative although a fully sensitive M tuberculosis grew 12 weeks later from the first sample on Lowenstein-Jensen slopes.

The second patient was a 21 year old Kenyan woman living in the united Kingdom for 3 years. She presented with a 3 month history of photophobia and occipital headaches. She had no other systemic symptoms. She had had peritoneal tuberculosis diagnosed at the age of 6 years during laparotomy for an appendicectomy and had received antituberculous medication for 1 month only. On examination she had mild neck stiffness and a partial left third cranial nerve palsy. Brain CT was normal. Lumbar puncture results showed a high opening pressure (15cm CSF), 90 white blood cells/μl, predominantly lymphocytes, a raised protein concentration (1.62 g/l), and a low CSF/blood glucose ratio. At the same referral laboratory CSF PCR forM tuberculosis was negative but culture after 8 weeks grew a fully sensitive organism. Despite the negative PCR antituberculous therapy was started empirically. After 2 months of treatment her symptoms had resolved although a partial third nerve palsy remains.

Adequate volumes of both patients’ CSF (0.5 ml) were sent to our referral laboratory where the samples were spun and PCR performed using three primer sets and appropriate controls.5-7 The assay included primers for the target IS6110, an insertion sequence normally present in multiple copies in the M tuberculosis genome, which has been used successfully for the detection of M tuberculosis in CSF.2 4 Multiple primer sets were used as this is thought to increase the probability of detecting target DNA within a specimen.

Recent studies suggest that CSF PCR for M tuberculosis is more sensitive than culture in cases of clinically suspected tuberculous meningitis that responded to empirical treatment.2-4 Some authors have even suggested the usefulness of serial CSF PCR in assessing the efficacy of treatment.4 8 False negatives and positives are rarely reported in the literature and unless these results are critically reviewed patients could, tragically, have treatment prematurely stopped or be started on prolonged antituberculous chemotherapy. False negatives occurred in two studies, in which reported CSF PCR sensitivities were 32% and 85%.2 3 In one study 6.1% of CSF specimens received from patients with no evidence of tuberculous meningitis were falsely PCR positive.3 These results also show that sensitivity and specificity can vary when different assays and laboratories are used. Claims that PCR can detect 1–10M tuberculosis organisms “in vitro” seems not to be the case in clinical samples such as CSF.

In the two patients presented above adequate volumes and repeated samples of CSF were assayed using suitable primers and appropriate controls at a British referral laboratory. Results for these two patients show the dangers of overreliance on CSF PCR when tuberculous meningitis is clinically suspected.

Acknowledgments

We are grateful to Dr Deborah Binzi-Gascoigne of the Leeds mycobacterium laboratory, where the PCR tests were performed and who provided additional information for the manuscript .

References

View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.