As far as we are aware this is the largest study so far to evaluate ELISPOT reversion after anti-TB treatment. A significant proportion of patients who successfully completed a course of standard TB treatment changed from a positive to a negative EC ELISPOT 12 months after their initial diagnosis. There was also a significant reduction in mean ESAT-6, CFP-10 and PPD ELISPOT counts. Seven per cent of cases had an increase in their ELISPOT counts for either ESAT-6 or CFP-10, compared to 21% for PPD.
The data reported here cannot be explained by the inherent variability of the ELISPOT readout and are consistent with those from other, albeit small, studies from other settings. Two previous studies of tuberculosis patients in the United Kingdom (UK; n = 4) and Italy (n = 18) reported a decline in quantitative ESAT-6 ELISPOT counts after drug treatment [11, 12], However, in one UK study, ESAT-6 ELISPOT counts in 10 patients who had completed TB therapy were marginally but not significantly higher than the counts in 8 different patients early in their TB treatment course .
The ex vivo ELISPOT response is thought to identify recently activated lymphocytes with immediate effector memory that persist only a limited time in circulation once antigen is cleared . It follows therefore that antigen withdrawal after anti-TB therapy would be accompanied by a decline in T cell frequencies. There are several possible explanations why even 'specific' antigen (ESAT-6 and CFP-10) counts did not universally drop below the pre-defined, and relatively conservative, cut-off for positivity (10 SFU) in The Gambia. Firstly, some patients may have been re-exposed to M. tuberculosis after the completion of treatment and re-aquired latent M. tuberculosis infection. Secondly, they may have been exposed to environmental mycobacteria, commonly encountered in the tropics, which share the esat-6 and cfp-10 genes (eg. M. marinum). Thirdly, and most intriguingly, there is evidence that in some individuals there is persistence of a population of activated T cells in the absence of direct mycobacterial antigen stimulation, even for several years after completing therapy .
Having determined that, even in this TB-endemic tropical setting, ELISPOT counts decrease significantly in TB cases after treatment, we will explore the decay kinetics of specific antigen T-cell responses during the course of treatment. It has been suggested that the decay rate of antigen-specific T-cells is approximately 5.5% of the total count per week,  which would lead to an approximately 90% decline in counts over 40 weeks. In a small study in South Africa, Nicol et al showed that 10 children with TB had an initial rise, then a decline in ELISPOT counts over a 6-month treatment course . Vekemans et al, in The Gambia, showed that TB patients had relatively low T cell responses to purified mycobacterial antigens, compared to healthy controls with evidence of M. tuberculosis infection . An obvious explanation for these findings is that patients at diagnosis have suppressed T cell responses because they are unwell. M. tuberculosis can itself interfere with IFN-γ signalling , providing another possible explanation. Other factors may also be important, such as increased early antigen presentation as a result of pathogen killing at the onset of antibiotic therapy. Thus the implied trajectory of cellular kinetics is of T-cell responses that increase early in treatment, then decline as the M. tuberculosis antigen load diminishes. A possible sampling frame would be to conduct repeat ELISPOT tests at 1 month, 2 months (to compare with sputum conversion in smear positive cases), 4 months and 6 months. We also aim to follow those who are persistently ELISPOT positive as a cohort over time.
While it is of interest that all four of the treatment failures (3 HIV-negative, 1 HIV-positive) in our study had positive EC and PPD ELISPOT tests at 12 months, larger numbers of similar cases will be required to identify the true sensitivity of ELISPOT persistence with respect to treatment failure. That a few cured cases had a rise in ELISPOT counts for ESAT-6 and/or CFP-10 suggests that the positive predictive value of this finding for treatment failure is not 100% in The Gambia. However, it is quite possible that 'cured patients' may relapse at a later date.