© 2006 American Thoracic Society
Tuberculosis Diagnostic Tests: Sensitivity, Specificity, and Comparing Apples with ApplesFrom the Author:The assessment of available evidence in my update on the diagnosis of tuberculosis infection was based on objective and critical appraisal of published or in-press papers (1). Radford and colleagues consider that the results reported in the 12 studies published by Lalvani's group with ELISpot are not directly comparable to those by other investigators using T-SPOT.TB. However, the actual differences between the Lalvani ELISpot test and the regulatory authority–approved commercially available T-SPOT.TB test are few. In the case of the Lalvani ELISpot, all published studies, except one (2), used 15-mer peptides and/or recombinant antigens. From ESAT-6, the same comprehensive set of 17 peptides, and from CFP-10, the same comprehensive set of 18 peptides were always used. These constitute the same peptides used in T-SPOT.TB, which is based on the Lalvani ELISpot. The main difference between the tests is that the Lalvani ELISpot generally places the 35 peptides across 6 pairs of duplicate wells, whereas T-SPOT.TB places all the peptides in two single wells. Radford and coworkers' statement about 35 combinations of peptides thus suggests a misunderstanding of both assay formats. Contrary to Radford's assertion, Lalvani's papers report the standardized level of purity of his peptides and also describe the purity of recombinant antigen, when used (3). The Lalvani ELISpot has always used a uniform predefined cutoff of 5 spots per well for peptides and 10 spots for recombinant antigens, and concordance between peptide and antigen results is generally very high (4). In the one study where concordance between antigen and peptide results was low, the possible reasons for this were discussed in the paper (3). Moreover, the Lalvani ELISpot and T-SPOT.TB use the same ELISpot plates and developing systems. It is surprising that Radford and colleagues cite the study by Goletti and coworkers (5), as this used a type of ELISpot assay that was very different from both the Lalvani ELISpot and T-SPOT.TB. Their assay used only 2 peptides from ESAT-6, 3 peptides from CFP-10, 40 h incubation (instead of 16 h), and a cutoff of 34 spots per million peripheral blood mononuclear cells. Also, the version of QuantiFERON used in that paper was not QuantiFERON-TB Gold but a research product known as the QuantiFERON-CMI kit. Radford and coworkers may also have misinterpreted the paper by Lee and coworkers (6) and the recent editorial by Davies and Drobniewski (7): both state that, although T-SPOT.TB has higher diagnostic sensitivity than QuantiFERON-TB Gold, there is no statistically significant difference in specificity. In conclusion, I agree that the evidence base using the Lalvani ELISpot (i.e., pre–T-SPOT.TB) is much larger than that with the commercial test; however, to date the performance of T-SPOT.TB is consistent with the original Lalvani test, on which it is based.
University of Modena and Reggio Emilia Modena, Italy FOOTNOTES
Conflict of Interest Statement: L.R. received REFERENCES
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1,900 in 2005 as a consultant in an Oxford Immunotec (manufacturer of T-SPOT.TB) advisory board meeting. His institution (the University of Modena and Reggio Emilla) received