Abstract

To the Editors: New guidelines for the management of latent infection with Mycobacterium tuberculosis (LTBI) have been released by the World Health Organization subsequent to the announcement of their novel strategy for the elimination of tuberculosis (TB) in countries of low TB incidence. According to this guideline, there is a strong recommendation for high-income countries for systematic testing and treatment of LTBI in people living with HIV, adult and child contacts of pulmonary TB cases, patients initiating tumor necrosis factor antagonist treatment, patients receiving dialysis, patients preparing for organ or hematologic transplantation, and patients with silicosis. Either interferon-g release assays (IGRAs) or Mantoux tuberculin skin test (TST) should be used to test for LTBI. Recently, we performed a multinational observational cohort study in Western Europe to evaluate the ability of the TST and the currently available IGRAs (the QuantiFERON-TB Gold intube assay and the T-SPOT.TB assay) to predict the progression to TB in patients living with HIV and in other immunocompromised patients in countries of low TB incidence. It was found that the frequency of positive test results by TST or IGRAs and the risk for the progression to TB differed among patients from different groups of immunocompromised individuals, and that patients with HIV infection clearly had the highest risk for the development of TB when compared to patients with chronic renal failure, rheumatoid arthritis, solid organ, or stem-cell transplant recipients in the absence of preventive chemotherapy. However, the overall risk of progression to TB in HIV-infected patients in Western-Europe was less than 3.5 per 100 person-years over a 2-year period and was thus not substantially higher than in recent contacts of TB patients in Western Europe where the risk for the progression to TB was also found to be very low. Physicians in many European countries, in the United States, and in Canada are reluctant to strictly test all HIVinfected patients and provide preventive chemotherapy in case of a positive TST or IGRA test result. This emphasizes the need to identify additional risk factors for targeted TB prevention in persons living with HIV in low TB incidence countries. As ongoing viral replication is a strong risk factor of opportunistic infections in chronic HIV infection, we reanalyzed the data from our study stratified by detection of viral replication in patients with HIV infection. Among 635 patients, 338 (53%) had suppression of viral replication (,50 copies/mL) on antiretroviral therapies at the time of screening for LTBI. Among the remaining 297 patients with detectable HIV load, 114 (38%) were on antiretroviral therapy. Positive TST, enzyme-linked immunespot assay (ELISPOT) (T-SPOT.TB), and enzyme-linked immunosorbent assay (ELISA) (QuantiFERON-TB Gold intube) results were observed in 8.1%, 10.4%, and 8.4%, respectively, of patients with detectable viral replication; these percentages were not significantly different from those of patients with undetectable HIV viral load (5.3%, 9.8%, and 7.7% for TST, ELISPOT, and ELISA, respectively). Within a 2-year follow-up period, none of the patients with viral suppression at the time of LTBI screening developed TB (0/338), regardless of the result of the screening test. In contrast, 6/297 patients with ongoing viral replication developed TB. The incidence of TB in patients with ongoing viral replication was therefore significantly higher compared with patients without viral replication (rate ratio, 13.4; 95% confidence interval: 2.7 to 66.3). The rate of TB in patients with viral replication was between 6.6 and 13.5 times higher for those with a positive LTBI test than for those without a positive LTBI test (Table 1). Comparing patients by antiretroviral therapy status rather than viral suppression did not result in such a clear distinction with regard to risk of TB. Two of the 6 patients who developed TB during follow-up were on antiretroviral therapy at the time of screening. One of these had a positive LTBI test, whereas the other had not. In contrast to the finding of viral suppression, we did not find an association between numbers of circulating CD4+ T cells and the risk for the development of TB in this cohort. The study may have been underpowered to show such an association that was observed elsewhere. To estimate the effect of different testing and treatment scenarios in patients with HIV, we calculated the numbers needed to treat with preventive chemotherapy to prevent one case of TB based on the absolute risk reduction (1/absolute risk reduction). When targeting preventive chemotherapy to patients with detectable viral replication and a positive immunodiagnostic test, the numbers needed to treat to prevent one case of TB was as low as 8, 13, and 10 with the TST, ELISA, and ELISPOT, respectively. In a public health approach of screening only patients with detectable viral replication, the numbers needed to test and treat when positive to prevent one case of TB was 79, 85, and 59 with the TST, ELISA, and ELISPOT, respectively. This increased number is due to the inability to prevent TB in screen negatives. This approach is more than twice as efficient compared with testing all patients with HIV and treating those positive, where the numbers needed to test and treat if positive would be 166, 199, and 137 with the TST, ELISA, and ELISPOT, respectively. Thus, the presence of viral replication identifies HIV-infected patients at risk for progression to TB substantially better than immunodiagnostic testing C.L. reports payments for lectures from AbbVie, Gilead, and Chiesi outside the submitted work; F.v.L. has nothing to disclose; M.S. reports payments for lectures from Qiagen and from Oxford Immunotech, and research grants from Pfizer and Novartis outside the submitted work; in addition, M.S. has a patent pending on “In vitro process for the quick determination of a patient’s status relating to infection with Mycobacterium tuberculosis” (international patent number WO2011113953/A1).