Rational application of adenosine deaminase activity in cerebrospinal fluid for the diagnosis of tuberculous meningitis

Infection. 2015 Oct;43(5):531-5. doi: 10.1007/s15010-015-0777-7. Epub 2015 Apr 14.

Abstract

Purpose: Tuberculous meningitis (TBM) is one of the most serious and difficult to diagnose manifestations of TB. An ADA value >9.5 IU/L has great sensitivity and specificity. However, all available studies have been conducted in areas of high endemicity, so we sought to determine the accuracy of ADA in a low endemicity area.

Methods: This retrospective study included 190 patients (105 men) who had ADA tested in CSF for some reason. Patients were classified as probable/certain TBM or non-TBM based on clinical and Thwaite's criteria. Optimal ADA cutoff was established by ROC curves and a predictive algorithm based on ADA and other CSF biochemical parameters was generated.

Results: Eleven patients were classified as probable/certain TBM. In a low endemicity area, the best ADA cutoff was 11.5 IU/L with 91 % sensitivity and 77.7 % specificity. We also developed a predictive algorithm based on the combination of ADA (>11.5 IU/L), glucose (<65 mg/dL) and leukocytes (≥13.5 cell/mm(3)) with increased accuracy (Se: 91 % Sp: 88 %).

Conclusions: Optimal ADA cutoff value in areas of low TB endemicity is higher than previously reported. Our algorithm is more accurate than ADA activity alone with better sensitivity and specificity than previously reported algorithms.

Keywords: Algorithm; Central system fluid; Meningitis; Tuberculosis.

Publication types

  • Evaluation Study

MeSH terms

  • Adenosine Deaminase / cerebrospinal fluid*
  • Adolescent
  • Adult
  • Aged
  • Aged, 80 and over
  • Case-Control Studies
  • Cerebrospinal Fluid / chemistry*
  • Female
  • Humans
  • Male
  • Middle Aged
  • ROC Curve
  • Retrospective Studies
  • Sensitivity and Specificity
  • Tuberculosis, Meningeal / diagnosis*
  • Young Adult

Substances

  • Adenosine Deaminase