ATBI during pregnancy remains associated with poor maternal and fetal outcomes globally, including a 3 times increase in maternal morbidity, 9 times increase in miscarriage, 2 times increase in preterm birth and low birth weight, and 6 times increase in perinatal death ^[1]. Meanwhile, severe immune reconstitution inflammatory syndrome may occur in the postnatal period ^[2]. However, the diagnosis of ATBI during pregnancy can be difficult because of the atypical clinical presentation ^[3]. At the same time, LTBI is easy to be underestimated and its incidence varies by race and economic status. According to epidemiological findings, LTBI occurred in 14–48% of pregnant women in the United States ^[4]. LTBI during pregnancy and the puerperium may develop into ATBI due to a physiological suppression of the T-helper 1 (Th1) pro-inflammatory response ^[5]. A high incidence of ATBI was reported within 180 days postpartum, and the postpartum diagnosis reflects the onset of active prenatal disease ^[4]. Therefore, it is important to monitor M. tuberculosis infection during pregnancy. However, the suppression of Th1 cells significantly affects the performance of IGRAs and the TST.

The efficacy of IGRAs compared with the TST for diagnosing LTBI during pregnancy varied in different studies. In both low and high prevalence areas, IGRAs are more specific than the TST due to there being no cross-reaction with BCG ^[4]. In terms of sensitivity, most studies showed little difference between the two, but IGRAs were less sensitive under certain conditions. The reason may be that as time from M. tuberculosis exposure increases, the sensitivity of IGRAs decreases ^[6]. A meta-analysis that involved 22 studies showed 77–91% concordance between IGRAs and the TST ^[4]. Recent evidence suggested that IGRAs were a good predictive tool for ATBI and that a positive IGRA result or a recently elevated IGRA strongly predicted the risk of reactivation, even among strongly immunosuppressed women ^[7][8]. A study showed that IGRAs also had a higher sensitivity (84.5–100.0%) and specificity (37.6–96.4%) than the TST for ATBI ^[9], though they have not been approved for the diagnosis of ATBI in this population

However, the use of IGRAs has some issues. First, a higher proportion of indeterminate results during pregnancy may be due to a low mitogen response, which was reported to be around 4–6% ^[10]. Fortunately, this is often temporary, and repeated testing after 1–2 months is essential ^[11]. Second, the cut-off value for this population is not clear, which needs further exploring ^[12]. Third, it was reported that IFN-γ released during pregnancy and the postpartum period varied and the conversion rate of IGRAs was relatively high ^[13][14]. Therefore, the optimal time for IGRAs remains an unresolved issue. Despite its lower efficacy compared with non-pregnant populations, it is still the first recommendation in both high and low prevalence areas and is highly cost-effective. The TST is even more influenced than IGRAs due to skin anergy ^[14][15]. In addition, IGRAs are an important tool for screening LTBI in low-income areas where follow-up rates are extremely low. Some women have only one medical visit for their whole life (IGRAs require one visit only, while the TST requires two visits). A new generation of IGRAs (QFT-Plus) promises to improve the detection of LTBI during pregnancy, which detects IFN-γ secretion from both CD4+ and CD8 + T-cells ^[12]. This is based on the reduction in CD4 + T-cells without altering CD8 + T-cells in pregnant women ^[16].

The elderly are an easily overlooked group that is vulnerable to M. tuberculosis infection. Epidemiological surveys over the past century showed they had the highest proportion of LTBI, possibly related to exposure to M. tuberculosis early in their lives ^[17]. The risk of progression to ATBI is 1.5 times higher than adults ^[18] due to a dysfunctional immune system marked by inflammation and oxidative stress ^[19]. ATBI in the elderly is often the result of reactivation rather than primary infection. They are prone to have poor treatment response and high mortality (up to 20–30%), which is 6 times higher than that of adults ^[20].

Due to the lack of a gold standard for the diagnosis of LTBI, it is hard to evaluate the sensitivity and specificity of IGRAs and the TST. No significant decrease in the positive rate of IGRAs was found in the elderly, while the positive rate of the TST gradually decreased due to the decline in skin immunity ^[18]. IGRAs from older people had a poorer concordance compared with adults ^[18]. Higher indeterminate rates (5–40%) in older people were found when compared with younger adults. The high proportion of indeterminate results is related not only to a lower mitogen response but also to a higher IFN-γ background (high baseline inflammatory status) ^[18]. Despite indeterminate or borderline results on initial IGRAs, a second attempt with another IGRA seems to resolve the problem ^[18]. LTBI in the elderly is highly heterogeneous and may arise from recent or very distant M. tuberculosis infection. Unfortunately, neither the TST nor IGRAs can distinguish between recent and remote infections ^[21][22].

ATBI in older people is often clinically and radiographically atypical because of multi-underline diseases. Furthermore, difficulties in sputum production make it difficult to obtain microbiological evidence. Therefore, immunological tests become extremely important ^[20]. In the diagnosis of ATBI, researchers found a significant decrease in the sensitivity of the TST and IGRAs ^[23]. The decline in the TST is more significant. A Japanese study on people over 80 years old showed that the sensitivity of IGRAs (67%) was higher than that of the TST (27%). High proportions of the indeterminate results (up to 33.3%) were also found in ATBI ^[23]. It is impractical to use an IGRA alone to diagnose ATBI in older people and a combination of tests is required. QFT-Plus was reported to perform well in clinical trials in the elderly, with significantly improved sensitivity ^[24].

Every year, over 7.6 million children under 15 years old worldwide are infected with M. tuberculosis and 1 million of them develop ATBI, with nearly 140,000 deaths ^[25]. Young children are at high risk of developing fatal, disseminated tuberculosis, accounting for nearly 20% of all tuberculosis-related deaths in high-burden countries ^[26]. Infants, young children, and adolescents have a greater risk of developing ATBI rather than primary school children. The risk of progression in infants is 40–50%, with 15% developing meningitis or disseminated disease; the risk of children aged 1–2 years is reduced to 25%, with 5% developing meningitis or disseminated disease ^[27][28]. Due to their difficulty in expectoration and concerns about invasive operation, the diagnosis of children often lacks a microbiological basis. This means that they lack the absolute evidence to evaluate the accuracy of the TST and IGRAs.

Among children more than 5 years old, the diagnostic value of IGRAs is positive. Though IGRAs do not have enough sensitivity and specificity to diagnose or exclude ATBI ^[29], it is an important auxiliary diagnostic technology. The sensitivity of IGRAs has no obvious advantage over the TST, but the specificity of IGRAs is higher ^[30][31]. Sensitivities were 78–93% for ELISA, 58–93% for ELISPOT, and 82–100% for the TST, while the specificities for IGRAs were 98–100% ^[32]. In unvaccinated children, concordance between TSTs and IGRAs was high (κ = 0.802), whereas discordance was mostly seen in the BCG-vaccinated children ^[33]. With anti-tuberculosis therapy, IGRAs were seen to switch from positive to negative in most patients, but it is not possible to rely on an IGRA monitoring treatment response for individuals ^[34][35][36]. In terms of LTBI, the concordance between the two is fairly poor, and the results of TST+/IGRA− often appear ^[37]. When risk factors are present, children with TST+/IGRA− may develop into ATBI. When children lack risk factors, those with TST+/IGRA− may be falsely positive due to BCG. In countries with high BCG coverage and high rates of NTMs infection, IGRAs are more reliable than the TST ^[38].

Using IGRAs for the diagnosis of M. tuberculosis infection in children under 5 is controversial. First, there is concern about the reduced sensitivity of IGRAs. A study from the United States showed that the sensitivity of IGRAs was low in the diagnosis of laboratory-confirmed ATBI among children under 5 years of age, with 58.7% positivity in the 2–4-years-old group and 51.9% positivity in the under-2-years-old group ^[39]. Despite this, there is a high level of concordance between IGRAs and the TST in the diagnosis of ATBI ^[33]. Different studies had different conclusions regarding the value of IGRAs and the TST for diagnosing LTBI. A cohort study of BCG-vaccinated children younger than 5 years showed that there was a high discordance (12.3%) between IGRAs and the TST, mostly TST+/IGRA− ^[40]; another study showed that for high-risk children aged 2–5 years old, IGRAs’ positivity was lower than that of the TST ^[41]. None of the TST+/IGRA− children developed into ATBI during the follow-up even without prophylactic treatment, demonstrating the false positivity of the TST that came from the cross-reaction with BCG ^[40][41][42]. Using the TST as a standard, a meta-analysis showed that IGRAs had higher sensitivity and specificity, with the pooled sensitivity and specificity of ELISA being 84.1 and 89.5%, respectively, and the pooled sensitivity and specificity of ELISPOT were 93.1 and 76.7%, respectively ^[43]. An IFN-γ-inducible protein 10 (IP-10) assay compared with current IGRAs will improve the diagnosis of LTBI in patients younger than 5 years ^[44]. Another reason to hesitate to use IGRAs is that younger age will significantly increase indeterminate results due to the low mitogen response caused by immunologic immaturity ^[45][46]. Indeterminate results in children under 2 years old can reach up to 40%. In addition, iron deficiency anemia, helminths, and malaria infection, which are common in children under 5 years old, may also increase the indeterminate rates ^[45]. However, the largest study found less than 1% of results were indeterminate, where the researchers included more than 5000 IGRA results in children under 2 years old ^[47].

At present, an increasing number of guidelines approve the application of IGRAs in the diagnosis of LTBI in children. When high specificity is required (e.g., BCG-vaccinated children in low-risk areas), IGRAs are a better choice. When sensitivity is the first consideration (e.g., children about to receive immunomodulatory biologics), a TST and IGRA need to be performed simultaneously, and any positive result should be regarded as infection. However, for the diagnosis of ATBI in children, IGRAs are thought to be a complementary diagnostic technology ^[48][49].