Improving the microbiological diagnosis of tuberculous meningitis: A prospective, international, multicentre comparison of conventional and modified Ziehl-Neelsen stain, GeneXpert, and culture of cerebrospinal fluid

Abstract

Objectives: Tuberculous meningitis (TBM) is the severest form of tuberculosis, but current diagnostic tests are insensitive. Recent reports suggest simple modifications to conventional cerebrospinal fluid (CSF) Ziehl-Neelsen (ZN) staining may greatly improve sensitivity. We sought to define the performance of modified and conventional ZN stain for TBM diagnosis.

Methods: In hospitals in Vietnam, South Africa and Indonesia we conducted a prospective study of modified ZN with or without cytospin, conventional ZN smear, GeneXpert, and culture on CSF in adults with suspected TBM.

Results: A total of 618 individuals were enrolled across 3 sites. Compared with the TBM clinical diagnostic gold standard for research (definite probable or possible TBM), sensitivity of conventional ZN and modified ZN with cytospin were 33.9% and 34.5% respectively (p = 1.0 for the difference between tests), compared with culture 31.8% and Xpert 25.1%. Using culture as a reference, sensitivities of conventional ZN, modified ZN with cytospin, and Xpert were 66.4%, 67.5%, and 72.3%, respectively. Higher CSF volume and lactate, and lower CSF:blood glucose ratio were independently associated with microbiologically confirmed TBM.

Conclusions: Modified ZN stain does not improve diagnosis of TBM. Currently available tests are insensitive, but testing large CSF volumes improves performance. New diagnostic tests for TBM are urgently required.

Keywords: Cytospin; Diagnosis; Tuberculous meningitis; Xpert MTB/RIF; Ziehl–Neelsen stain.

Fig. 1

Flow chart of patients screened and recruited to study. ZN; Ziehl–Neelsen smear, MZN; modified Ziehl–Neelsen smear, CZN; conventional ZN smear, culture; mycobacterial culture by MGIT (Becton Dickinson, USA) (Vietnam and South Africa) or MODS (Indonesia), Xpert; Xpert MTB/RIF (Cepheid), TBM; tuberculous meningitis. ¹Reasons for exclusion were as follows; entry criteria not met (n = 18), at least one exclusion criteria met (n = 50), consent not given by patient (n = 9), consent could not be obtained (n = 1).

Fig. 2

Venn diagram of mycobacterial tests (conventional ZN, modified ZN with cytospin, Xpert and culture). CZN; conventional Ziehl–Neelsen smear, MZN; modified Ziehl–Neelsen smear, Xpert; Xpert MTB/RIF (Cepheid), culture; mycobacterial culture by MGIT (Becton Dickinson, USA) (Vietnam and South Africa) or MODS (Indonesia). For each test, the number of positive results is shown divided by the number of tests performed. Test denominators (CZN 174, MZN with cytospin 172, Xpert 174, culture 172) refer to the number of definite cases of TBM for which each test was performed.