[Possible presence of a booster effect in people repeatedly exposed to Mycobacterium tuberculosis demonstrated in a Mycobacterium tuberculosis-specific interferon-gamma release assay].

INTRODUCTION

Recently, several reports have described a waning response of Mycobacterium tuberculosis (TB)-specific interferon-gamma (IFN-gamma) release assays (IGRAs) performed years after the initial TB infection. Since a considerable part of elderly people in Japan was infected with TB in their youth, it was hypothesized that if the general population was screened with IGRAs, the mean IFN-gamma value of positive subjects (POSITIVES) would be reduced as the age of the subjects increased. Conversely, in subjects who were recently exposed to TB (CONTACTS), it was assumed that the POSITIVES would include subjects with new TB infections who show a high IFN-gamma value and those with old TB infections who show a low IFN-gamma value. Since the presence of subjects with old infections among the POSITIVES cannot be avoided during examinations, it is hypothesized that the mean IFN-gamma value of the POSITIVES decreases as the age of the CONTACTS increases.

MATERIALS AND METHODS

To test this hypothesis for CONTACTS, data acquired during the contact examinations at Adachi Public Health Center in Tokyo, Japan were analyzed. Since it is thought that the grade of exposure of the TB index case (INDEX CASE) influences the IFN-gamma value, its relationship with the INDEX CASE and the infectivity of the INDEX CASE were also investigated. In the CONTACTS, as only a few healthcare workers are < 20 years old or > 70 years old, the subjects were chosen from individuals aged in the range of 20-69 years who had been in contact with a TB patient within the past 3 months. The IFN-gamma value of the 188 POSITIVES in the 1145 CONTACTS who underwent a contact examination using the IGRA QuantiFERON-TB Gold In-Tube (QFT-G-IT) assay was investigated. The POSITIVES were divided according to their age class, and their IFN-gamma values were compared. In addition, after dividing all POSITIVES into groups to assess their relationship with the INDEX CASE (i.e., household, relative, healthcare worker, and others), the IFN-gamma values of the 20-49- and 50-69-year-old classes were compared in each group.

RESULTS

There was no significant difference in the IFN-gamma values between the age classes, (Kruskal-Wallis test, P = 0.598). When the IFN-gamma values of the POSITIVES in the 20-49- and 50-69-year-old classes were compared for each relationship group, the mean IFN-gamma value of the POSITIVES increased in proportion to the rise in age only in the healthcare worker group. This occurred even though the mean IFN-gamma value of the POSITIVES decreased in many groups as their age increased. A significant difference was confirmed in the IFN-gamma values between the age classes in the healthcare worker group (Wilcoxon rank-sum test, P < 0.001). No significant difference was observed in the infectivity of the INDEX CASE between the age classes in the healthcare worker group.

CONCLUSION

Initially, it seemed that the high infectivity of the INDEX CASE affected the IFN-gamma values of the POSITIVES in the oldest age class; in other words, the factor expected to decrease actually increased. However, comparison of the IFN-gamma value by age class in each relationship group revealed that, in the healthcare worker group only, the IFN-gamma value of POSITIVES increased in proportion to the rise in age regardless of infectivity of the INDEX CASE. Since it has been hypothesized that healthcare workers have an increased chance of contact with TB patients than other relationships, this outcome suggests the existence of a booster effect on people who are repeatedly exposed to TB, as assessed using IGRAs, as one of the several possibilities. If this booster effect seen with IGRAs is proven, predicting the development of symptoms and presuming the infection time by using IFN-gamma values will be difficult. To validate the present results, animal experiments that can be adjusted for various biases and confounding factors are necessary.