TB is the most frequent opportunistic infection and a leading cause of death in people living with HIV. IPT has been shown to reduce the risk of TB and is now universally recommended in HIV-infected individuals with LTBI or at high risk of having LTBI. However, the optimal test for identifying HIV-infected individuals who would benefit most from IPT remains uncertain. Consequently, the most recent guidelines differ in their recommended LTBI screening strategies, ranging from screening with TST if feasible3
to dual testing with TST and IGRAs.50, 51
Our systematic review addressed two questions relevant for determining whether IGRAs should replace TST as a screening test for LTBI in people living with HIV: (1) Are IGRAs better than TST at predicting which HIV-infected individuals are at highest risk of progression to active TB and (2) Are IGRAs more sensitive than TST for diagnosis of MTB infection, particularly in HIV-infected individuals with advanced immunosuppression? For both questions, we found insufficient evidence to conclude that either test is superior to the other.
It is well established that the majority of persons latently infected with MTB, including persons co-infected with HIV, do not develop active TB.52
The clinical utility of any diagnostic test for LTBI is therefore dependent on its ability to identify which persons are truly at increased risk for progression to active TB. We identified 3studies of the predictive value of IGRAs in HIV-infected individuals, each of which showed that IGRAs have poor positive predictive value but high negative predictive value for active TB. While these results suggest that a negative IGRA result is reassuring (no person with a negative IGRA result developed culture-positive TB), the studies had serious limitations, including small sample sizes with short-duration of follow-up 23, 24
and differential evaluation and/or follow-up of persons with positive and negative IGRA results.22–24
These limitations would be expected to result in an underestimation of active TB in persons with negative IGRA results. In contrast, randomized controlled trials in HIV-infected persons demonstrate that IPT confers a 20–60% reduction in the risk of active TB among persons with positive TST results.2
In addition, large prospective cohort studies have established that persons with a positive TST have a 1.4 to 1.7-fold higher rate of active TB within one year compared to persons with a negative TST result.1, 53
Unfortunately, similar high quality data on the clinical impact and predictive value of IGRA testing are currently lacking. Though a recent meta-analysis that included studies of HIV-infected and -uninfected patients found that IGRA results are more strongly associated with progression to active TB than TST results, the vast majority (>85%)of IGRA-positive individuals did not progress to active TB.54
In spite of the limited data on important outcomes, it has been suggested that IGRAs may have a role for identifying MTB infection in HIV-infected individuals given the sub-optimal performance of TST in immunosuppressed individuals.55
In support of this role, data from high-income countries suggest that TSPOT performance may be less affected by advanced immunosuppression, possibly because the testing platform ensures that an adequate number of peripheral blood mononuclear cells are available despite overall low CD4+ cell counts in whole blood.47
However, the point estimates and lower limits of the 95% CI for the difference in the proportion of positive test results in HIV-infected individuals with and without advanced immunosuppression were similar for TSPOT, QFT-GIT and TST. Moreover, in low-and middle-income countries, 2 of 5 studies of TSPOT and 1 of 2 studies of QFT-GIT found a large (range 23–31%) and statistically significant absolute reduction in the proportion of positive test results in HIV-infected individuals with advanced immunosuppression. Reasons for the stronger impact of immunosuppression on IGRA performance in low/middle-income vs. high-income settings are unclear but may be related to disease severity and anti-retroviral treatment status. However, overall, the available data suggest, but do not clearly confirm, that IGRAs are less affected by HIV-related immunosuppression than TST.
The major limitation of this review, and studies of IGRAs in general, is the lack of an adequate reference standard for diagnosis of LTBI. Though we developed a pre-specified hierarchy that could support the use of IGRAs, there were no data on whether IGRAs identify HIV-infected individuals who would benefit from preventive therapy and minimal data on the predictive value of IGRAs for active TB – the 2 strongest outcomes in the hierarchy. In addition, the majority of studies were small (<150 patients in 22of 37studies), only 6studies performed a head-to-head comparison of IGRA and TST results to a reference standard, and there were insufficient studies to perform meta-analysis in many sub-groups. Secondly, though IGRAs have potential operational advantages relative to TST, we did not find any studies that evaluated the impact of implementing IGRAs in LTBI screening programs targeting HIV-infected populations. Lastly, since we only included studies that evaluated IGRAs, we did not review historic data on the effect of HIV-related immunosuppression on TST results.
Given that both TST and IGRAs have only modest predictive value and sub optimal sensitivity, it would be relevant to evaluate outcomes when both tests are used, either simultaneously or sequentially, for diagnosing LTBI in HIV-infected persons. Though we did not find any studies of a dual testing approach, the most recently updated US national guidelines endorse such an approach. While routine use of dual-testing is not recommended, the 2010US Centers for Disease Control and Prevention (CDC) guidelines indicate that the results from both tests (IGRA and TST) may be useful in HIV-infected individuals when the initial test is negative.50
Similarly, the 2010 Canadian Tuberculosis Committee guideline recommends starting with TST but performing an IGRA if the TST is negative and there is a strong clinical suspicion for LTBI in immunocompromised individuals.51
While dual testing approaches will surely increase the number of HIV-infected individuals with positive test results, clinicians should balance this potential benefit against the lack of evidence supporting the efficacy of IPT in TST-negative but IGRA-positive individuals. Other strategies should also be considered, including modifying the definition for a positive IGRA result in HIV-infected individuals, monitoring trends in IGRA results in individual patients (i.e., serial testing), measuring levels of other biomarkers, and developing risk prediction models.
Current evidence suggests that IGRAs perform similarly to the TST at identifying HIV-infected individuals who could benefit from LTBI treatment. Important questions remain unanswered despite the substantial body of literature on IGRAs. HIV-infected individuals with a negative IGRA result may have a low risk of progression to active TB, but this result should be confirmed in larger studies that simultaneously perform TST and include a longer duration of follow-up. IGRAs (particularly TSPOT) may be more sensitive than TST in HIV-infected individuals and less affected by advanced immunosuppression. However, these results have not been observed consistently in head-to-head comparisons. Clinical trials evaluating outcomes in HIV-infected individuals randomized to different LTBI screening strategies (IGRA vs. TST vs. dual-testing) are needed to more definitively determine whether IGRAs could improve the identification of people living with HIV who could benefit from IPT. Until such data are available, the decision to use IGRA or TST (or both) will depend on national guidelines as well as resource and logistical considerations.