Answers are grounded in this paper and cite their sources.
(See the Major Article by Lu et al on pages 2006–15.) Built on a previous study of tuberculosis (TB) testing of 5405 persons in 2 villages in eastern China, Wei Lu and colleagues provide 6 years of follow-up on disease incidence by result of tuberculin skin test (TST) and QuantiFERON-TB Gold In-Tube (QFT) test [1]. Follow-up was by active surveillance for the first 2 years and passive record matching for the final 4 years. Although there have been many comparisons of TST and both commercial interferon gamma release assay (IGRA), this study is large with follow-up that is long enough to capture most reactivation in a high-burden setting. One unfortunate deficiency is the lack of quantitative QFT data, whereas TST data are compared as 5-, 10-, and 15-mm cutoffs. Reporting a QFT test as positive at or above 0.35 IU and negative at or below 0.34 IU inflates the significance of small reactions and loses the considerable progressive certainty of infection conveyed by larger QFT responses. However, the reason for this editorial commentary is not to critique the methods used or the conclusions reached on factors that might predict discordance between TST and QFT. It is to question the underlying infection-latency-progression model that is its foundation. The assumption of the study is that sustained infection occurs in high-prevalence settings over a lifetime and persists unless it progresses to disease early or late or is effectively treated by preventive chemotherapy. Within this model, the discordance between TST and QFT is problematic since the goal is to evaluate patients for treatment of latent infection as part of the refocused World Health Organization Global Plan for TB Elimination. Which test do you believe, and why do the tests differ? The authors analyze factors such as age and BCG vaccination that correlate with discordance. However, while the infection–latency–disease model of pathogenesis is more likely to be true in low-burden settings where transmission is low and reinfection rare, it may not be true for many patients in the study and in most high-burden settings. At one time, TB specialists were unclear on how primary TB in childhood related to postprimary, chronic TB seen in adults. William Stead, working in the United States, introduced the “unitary theory” of TB pathogenesis, that adult TB resulted from reactivation of foci seeded at the time of primary infection [2]. “Unitary” implied that it was all one process. Exogenous reinfection, a second infection, was thought to be unusual due to adaptive immunity and low transmission risk in the United States and other low-prevalence settings. This was the prevailing thinking when, together with coauthors, I published an article on an outbreak of isoniazid and streptomycin-resistant TB cases in a homeless shelter, suggesting transmission, including reinfection of patients known to have had previous TB disease or infection [3]. Identical bacteriophage types confirmed that exogenous reinfection occurred in the high-transmission setting of a homeless shelter. No patients had human immunodeficiency virus coinfection. When as an assistant professor I shared the data with Dr Stead, he was gracious and admitted an obvious exception to his unitary rule. More recently, molecular epidemiology has shed light on reinfection in those uncommon cases where fingerprints of an earlier and contemporary isolate occur. In an earlier article in Shanghai, Gao and colleagues documented a different genotype in nearly 42% of cases of recurrence, indicating new infection rather than relapse [4]. In another study in Shanghai, Gao and colleagues showed that transmission and reinfection accounted for 32% of incident multidrug-resistant TB cases [5]. Although the current study is about infection and progression to, presumably, mostly drug-susceptible TB, I raise this local data as evidence of high levels of TB transmission and reinfection of adults in China. A sizable fraction of TB disease results from recent transmission, including reinfection, not from reactivation of latent infection identified by testing with either test. Behr et al recently reviewed the natural history of reactivation of initial infection in the prechemotherapy era [6]. Their conclusion is that most progression of new infection occurs in the first year or so and that reactivation potential decreases thereafter, even in immunosuppressed patients. This is not to say that late reactivation does not occur—it does—but how often relative to progression of recent infection in high-burden settings? This is fundamental to the question of testing, interpretation, and treatment of latent infection. Further, how much of what appears to be reactivation is actually reinfection? Our own studies of a naturally infected guinea pig model indicate that many 6–14 mm TST reactions in guinea pigs revert to negative upon repeat testing, with no demonstrable risk of TB disease, even with immunosuppression. Reinfection is also seen after reversion [7]. Transient TB infection is also seen among contacts of human infectious TB cases [8]. However, we are so indoctrinated in the unitary theory, “once positive, always positive,” that we rarely retest positive tests before treating to be sure that the responses are sustained. When untreated positive healthcare workers were retested by Friedman and colleagues using TST or QFT, the majority were negative [9]. As expressed in the introduction to this editorial commentary, its purpose is to better explain discordance between the 2 common tests used to predict risk of reactivation for the purpose of preventive treatment. Indeed, many studies indicate that treating patients for latent TB infection can be highly effective at the population level. The study suggests that countries that use BCG vaccination should switch to an IGRA test to avoid false-positives and unnecessary treatment. Not part of this study, but as important perhaps, quantitative IGRA results are useful. Many small IGRA-positive reactions may not remain positive with repeat testing. As usual, the treatment focus should be on children and contacts likely to be recently infected. In patients with underlying risk factors for TB, the benefits of treatment likely outweigh the risks, even if the reactivation risk has diminished over time. In high-transmission settings, the possibility of reinfection may justify treatment regardless of test results; however, treatment does not prevent further reinfection despite adaptive immunity. Potential conflicts of interest. The author: No reported conflicts of interest. The author has submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest.
No takes yet. Share an insight, caveat, or question.
Edward A. Nardell (2020) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: