Improved Diagnostic Evaluation of Suspected Tuberculosis

  1. Davinder P.S. Dosanjh, DPhil;
  2. Timothy S.C. Hinks, MD;
  3. John A. Innes, MD;
  4. Jonathan J. Deeks, PhD;
  5. Geoffrey Pasvol, DPhil;
  6. Sarah Hackforth, RGN;
  7. Hansa Varia, RGN;
  8. Kerry A. Millington, DPhil;
  9. Rubamalar Gunatheesan, MD;
  10. Valerie Guyot-Revol, PhD; and
  11. Ajit Lalvani, DM
  1. From Tuberculosis Immunology Group, Imperial College London, London; University of Oxford, Oxford; Birmingham Heartlands Hospital and University of Birmingham, Birmingham; and Northwick Park Hospital, Harrow, United Kingdom.
    1. Figure 1.
      View larger version:
      Figure 1. Study flow diagram.

      ELISpot = enzyme-linked immunospot assay incorporating early secretory antigenic target-6 and culture filtrate protein-10; ELISpotPLUS = enzyme-linked immunospot assay incorporating early secretory antigenic target-6, culture filtrate protein-10, and Rv3879c; TB = tuberculosis; TST= tuberculin skin testing. *Results were indeterminate because of no achievement of positive control (11 patients), high background (1 patient), peptide contamination (1 patient), inconclusive assay (1 patient), or defective ELISpot plate (1 patient). †Results were not available because of history of TB (clinically contraindicated) (45 patients), patient did not return for reading (8 patients), result not recorded (8 patients), reason unknown (3 patients), death (1 patient), test performed elsewhere (1 patient), or patient declined consent (1 patient). ‡Tuberculin skin test results were based on a 15-mm cutoff point and considered positive if induration was ≥15 mm on the Mantoux test or grade 3 to 4 on the Heaf test regardless of bacille Calmette–Guérin vaccination status.

    2. Appendix Figure 1.
      View larger version:
      Appendix Figure 1. Study flow diagram, stratified by final diagnosis and then by test result.

      These are the same data shown in Figure 1 but displayed in a different manner to allow scrutiny of the entire raw data set. ELISpot = enzyme-linked immunospot assay incorporating early secretory antigenic target-6 and culture filtrate protein-10; ELISpotPLUS = enzyme-linked immunospot assay incorporating early secretory antigenic target-6, culture filtrate protein-10, and Rv3879c; TB = tuberculosis; TST= tuberculin skin testing. *Results were not available because of history of TB (clinically contraindicated) (45 patients), patient did not return for reading (8 patients), result not recorded (8 patients), reason unknown (3 patients), death (1 patient), test performed elsewhere (1 patient), or patient declined consent (1 patient). †Tuberculin skin test results were based on a 15-mm cutoff point and considered positive if induration was ≥15 mm on the Mantoux test or grade 3 to 4 on the Heaf test regardless of bacille Calmette–Guérin vaccination status. ‡Results were indeterminate because of no achievement of positive control (11 patients), high background (1 patient), peptide contamination (1 patient), inconclusive assay (1 patient), or defective ELISpot plate (1 patient).

    3. Figure 2.
      View larger version:
      Figure 2. Likelihood ratios, sensitivities, and specificities of tests used in combination, using ELISpot or ELISpotPLUS first.

      Data are for patients in whom results on both tests were available (n = 265). Except where stated, values are likelihood ratios with 95% CIs. Tuberculin skin test thresholds for positivity were induration ≥15 mm on the Mantoux test or grade 3 to 4 on the Heaf test. ELISpot = enzyme-linked immunospot incorporating early secretory antigenic target-6 and culture filtrate protein-10; ELISpotPLUS = enzyme-linked immunospot incorporating early secretory antigenic target-6, culture filtrate protein-10, and Rv3879c; TB = tuberculosis; TST = tuberculin skin testing. Top. ELISpot followed by TST. *Combined sensitivity of 1 or more positive results from tests used in combination, 97% (CI, 93% to 99%). †Combined specificity for a double-negative result from tests used in combination, 62% (CI, 53% to 71%). Bottom. ELISpotPLUS followed by TST. ‡Combined sensitivity of 1 or more positive results from tests used in combination, 99% (CI, 95% to 100%). §Combined specificity for a double-negative result from tests used in combination, 61% (CI, 52% to 70%).

    4. Appendix Figure 2.
      View larger version:
      Appendix Figure 2. Likelihood ratios, sensitivities, and specificities of tests used in combination for diagnostic evaluation, using tuberculin skin testing (TST) as the first test.

      Data are for patients in whom results on both tests were available (n = 265). Except where stated, numbers are likelihood ratios with 95% CIs. Tuberculin skin test thresholds for positivity were induration ≥15 mm on the Mantoux test or grade 3 to 4 on the Heaf test. ELISpot = enzyme-linked immunospot incorporating early secretory antigenic target-6 and culture filtrate protein-10; ELISpotPLUS = enzyme-linked immunospot incorporating early secretory antigenic target-6, culture filtrate protein-10, and Rv3879c; TB = tuberculosis. Top. TST followed by ELISpot. *Combined sensitivity of 1 or more positive results from tests used in combination, 97% (CI, 93% to 99%). †Combined specificity for a double-negative result from tests used in combination, 62% (CI, 53% to 71%). Bottom. TST followed by ELISpotPLUS. ‡Combined sensitivity of 1 or more positive results from tests used in combination, 99% (CI, 95% to 100%). §Combined specificity for a double-negative result from tests used in combination, 61% (CI, 52% to 70%).

    Responses to this article

    « Previous | Next Article »Table of Contents