Abstract
Background
Latent tuberculosis infection (LTBI) is a huge reservoir for the deadlier TB disease. Accurate identification of LTBI is a key strategy to eliminate TB. Therefore, a systematic review and meta-analysis approach was used to assess diagnostic potential of IL-2 for LTBI.
Methods
PubMed, Web of Science, the Cochrane Library and Embase were searched. The pooled sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR), area under the summary receiver operating characteristic curve (AUROC) and hierarchical summary receiver operating characteristic curve (HSROC) were estimated by bivariate and HSROC models.
Results
Twenty-seven studies including 1404 participants and 1986 samples met the inclusion criteria. The pooled sensitivity, specificity, PLR, NLR, DOR and AUROC of IL-2 were separately as 87%, 98%, 34.78, 0.14, 256.41 and 0.98, indicating a very powerful differentiating ability of IL-2 for LTBI from non-TB controls. For differentiating ATB from LTBI, the pooled sensitivity, specificity, PLR, NLR, DOR and AUROC of IL-2 were 83%, 76%, 3.41, 0.22, 15.47 and 0.87, respectively, suggesting a good differentiating ability of IL-2.
Conclusions
These findings showed that IL-2 is a powerful marker for differentiating LTBI from non-TB controls and a good marker for differentiating ATB from LTBI individuals.
Keywords: Interleukin-2, latent tuberculosis infection, non-TB controls, active tuberculosis, differentiation
Introduction
The global disease burden of tuberculosis (TB), caused by Mycobacterium tuberculosis (Mtb), is immense [1]. Approximately, 1.3 million individuals died of TB and more than one-fourth of these cases were coinfected with human immunodeficiency virus (HIV) [2]. Latent tuberculosis infection (LTBI), an asymptomatic state of TB, affects approximately 1.7 billion people worldwide [2,3]. Healthcare workers, HIV-infected patients, and household contacts of ATB cases are especially easy to develop LTBI [4–6]. 5%–15% of individuals with LTBI are at risk of developing active TB (ATB). Therefore, these LTBI individuals are a huge reservoir for the deadlier ATB disease [7,8]. Accurate identification of LTBI is a key strategy for eliminating TB by 2050, a new target established by the World Health Organization (WHO) [9,10].
In defect of clinical symptoms and a recognized optimal method, differentiation of cases with LTBI from non-TB controls and ATB patients are challenges for clinicians [11,12]. At present, the tuberculin skin test (TST) and interferon gamma release assay (IGRA) are routinely used to identify LTBI [13,14]. TST, a traditional immunodiagnostic method to detect LTBI, is globally used. However, false-positive TST results (21.2–41.8%) could be seen in population with Bacille Calmette-Guerin (BCG) vaccination [15]. IGRA, including QuantiFERON-TB Gold In-Tube (QFT-IT) assay, T-SPOT.TB assay, and the newly developed QFT-plus assay, could replace TST in many settings, such as BCG-vaccinated and HIV-coinfected individuals [9,16,17]. However, IGRA frequently has indeterminate results in children, especially in those less than 5 years old [18]. Clinically, positive Mtb culture is a gold standard for diagnosing ATB [19]. However, culture is time wasting, which takes around six weeks for truly negative detection [20]. Considering the limitations of TST, IGRA and Mtb culture, additional diagnostic tools for differentiating LTBI from non-TB controls and ATB patients are urgently needed.
Interleukin-2 (IL-2) is a cytokine mainly produced by antigen-activated T-cells [21,22]. It can promote cellular immunity especially T-cell replication after Mtb infection [23]. Furthermore, it is notable that the specific IL-2 cannot be detected in Mtb-uninfected contacts [23]. Recent studies have evaluated the auxiliary differentiating value of IL-2 for LTBI [23–40]. However, the results vary (LTBI vs. non-TB controls: 42.9–100% of sensitivity, 91.1–100% of specificity; ATB vs. LTBI: 63.1–100% of sensitivity, 24.14–100% of specificity). We therefore performed a systematic review and meta-analysis to assess the sensitivity and specificity of IL-2 for differentiating LTBI from non-TB controls and ATB patients.
Material and methods
Literature search
We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Diagnostic Test Accuracy (PRISMA-DTA) criteria published in 2018 [41]. The study has been registered (PROSPERO CRD42019140124). PubMed, Web of Science, the Cochrane Library, and Embase were searched for relevant English language citations from database inception through January 2020. Our search terms were “latent tuberculosis infection”, “latent tuberculosis”, “LTBI”, “interleukin-2”,“T cell growth factor”, “IL-2”, and “TCGF”. A comprehensive search strategy was determined based on the combination of “MeSH” and “all fields” for the PubMed database. In addition, the relevant research letters and reference lists of pertinent articles were reviewed to find other potentially relevant studies.
Literature selection
Studies reporting IL-2 for the differentiation of LTBI individuals from non-TB controls and/or LTBI from ATB patients were included according to the following criteria: (1) reporting on LTBI individuals and non-TB controls and/or ATB patients; (2) Mtb-specific antigen-stimulated IL-2 in blood as the index test;(3)IGRA (QFT-IT, T-SPOT.TB, QFT-plus) and/or TST as the gold standard for LTBI, positive microbiological findings (e.g. Mtb culture, microscopy) as the reference standard for ATB; (4) diagnostic performance (sensitivity and specificity) of IL-2 as the primary outcome; (5) the study design being randomized controlled trials, cohort, or cross-sectional studies; (6) more than five participants in each study. The exclusion criteria included articles published in languages other than English, case–control researches, veterinary experiments, patents, case reports, and conference abstracts. Two investigators independently screened the titles, abstracts, and full-text, and followed the determined inclusion eligibility. Disagreements were resolved by discussing and reaching a consensus.
Data extraction
For each eligible article, one investigator extracted the following data: the first author, year of publication, country, number of participants (ATB patients, LTBI individuals and non-TB controls), types of LTBI populations, diagnostic reference standard (ATB and LTBI), cut-off and antigen condition for the index test (IL-2), and the status of HIV-coinfection. Then, the data with regard to TB burden (regulated by WHO) [2], study design type, the subjects’ age, the method for the index test (IL-2), sensitivity, specificity, true positive (TP), false positive (FP), false negative (FN), and true negative (TN) were extracted. A second investigator examined the accuracy and completeness of the data.
Quality assessment
According to the Quality Assessment of Diagnostic Accuracy Studies tool-2 (QUADAS-2) recommended by the Cochrane Collaboration, two investigators independently reviewed the methodological quality of eligible articles [42]. Disagreements were resolved by consensus. Revman (version 5.3) software was used to perform the quality assessment.
Data analysis
According to the data inclusion criteria, data with TP, FP, FN, TN of IL-2 for differentiating LTBI from non-TB controls and ATB from LTBI individuals were included. According to the recommendations of the Cochrane Collaboration, the hierarchical summary receiver operating characteristic (HSROC) model and bivariate model were utilised [43,44]. The HSROC curve was computed with the “metandi” command. All analyses were performed using Stata (version 14.0, Stata Corporation).
The main outcomes evaluated were the immunodiagnostic performance of IL-2 for differentiating LTBI individuals from non-TB controls and ATB patients, as evaluated by the summary estimates of sensitivity, specificity, positive likelihood ratio (PLR), negativelikelihood ratio (NLR), diagnostic odds ratio (DOR), and area under the summary receiver operating characteristic curve (AUROC). The AUROC is a measure of the diagnostic accuracy of a test [45]. An AUROC less than 0.75 is “not accurate”, between 0.75 and 0.93 is “good”, and more than 0.93 is “excellent”. Stata 14.0 was used to perform the detailed command “midastpfpfntn, id(author year) ms(0.75) ford forsbfor(dss)” to obtain the sensitivity and specificity.
We used subgroup and meta-regression analyses to explore potential sources of heterogeneity with the “midas” command. Four subgroups were created: TB burden of the country (high or low), study design type (cohort or not), age (adults or not), and the IL-2 method (multiplex cytokines assay or not). For publication bias assessment, the Deeks test was used with the “midas” command [46].
Results
Search results
Totally, 945 records were identified from the literature searches. Three hundred and sixty-eight duplicates (the same articles) were removed, and 522 records were excluded after reading the titles and abstracts, including a total of 36 records were due to patient selection criteria (psoriasis, rheumatoid arthritis, Crohn’s disease, etc.), 333 records were due to the intervention standards (tumour necrosis factor-alpha, new TB vaccines, tocilizumab, etc.), 34 records were due to animal experiments (mice, cynomolgus macaques, guinea pigs, etc.), 62 records were due to reviews, abstracts, innovations, and case reports, 29 records were focussed on TB mechanisms (signalling pathways, apoptosis, autophagy, etc.), 18 records were focussed on TB detecting methods (novel electrochemiluminescence-sensing platform, multicolour flow cytometric, doublet-discriminator strategy, etc.), six records were repeated (different articles included the same participants), and four records were not in English (Chinese, French, Russian, etc.). Thus, we reviewed the full texts of 55 articles. Ultimately, 18 articles were included in our analysis [23–40] (Figure 1).
Figure 1.
Flow chart of the process of the identified articles regarding IL-2 and TB infection.
Characteristics of the included studies
The main characteristics of the 18 articlesare listed in Table 1 [23–40]. In total, 1404 participants (ATB patients: 373; LTBI individuals: 542; non-TB controls: 489) were involved. These studies, came from 10 countries, ranged from 2008 to 2018. The LTBI populations mostly included household contacts, healthcare workers and close contacts. Positive Mtb culture and/or sputum smear microscopy were the gold standards for ATB. The IGRA and/or TST were as the reference standards for LTBI individuals. The cut-offs of IL-2 ranged from 10.7 to 976.3 pg/ml or 12.5 to 275 spot-forming cells per million peripheral blood mononuclear cells. IL-2 were reported by Mtb-specific antigen stimulated, and the specific antigens included early secreted antigenic target-6 (ESAT-6), culture filtrate protein-10 (CFP-10), TB7.7, L-alanine dehydrogenase (Ala-DH) and purified protein derivative (PPD). Only one article reported that the populations were co-infected with HIV. Other baseline characteristics (gender, age and migrants) are shown in Supplementary Table 1.
Table 1.
The main characteristics of included studies regarding IL-2 and TB infection.
| Author | Year | Country | Participants (N) |
Types of LTBI populations | Reference standard |
Index test (IL-2) |
HIV-coinfection | ||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| ATB patients | LTBI individuals | TB uninfected controls | ATB | LTBI | Cut-off | Antigen condition | |||||
| Wang S | 2018 | China | / | 34 | 26 | Household contacts | / | IGRA | 14.3 pg/ml | TBAg (ESAT-6 + CFP-10) | No |
| Balcells ME | 2018 | Chile | 72 | 39 | 33 | Household contacts | Sputum smear microscopy, culture and chest X-ray | IGRA | >14 pg/ml (LTBI vs non-TB), <35.3 pg/ml (ATB vs LTBI) | TBAg-Nil (ESAT-6 + CFP-10) | No |
| La Manna MP | 2018 | Italy | / | 32 | 20 | Household or equivalent close contacts (work) | / | TST + IGRA | 42.13 pg/ml | TBAg-Nil (ESAT-6 + CFP-10 + TB7.7) | No |
| Nausch N | 2017 | Germany | / | 15 | 12 | Children contacts | / | TST +/- IGRA | 28 pg/ml | TBAg (ESAT-6 + CFP-10 + TB7.7) | No |
| Kamakia R | 2017 | Kenya | 19 | 16 | 8 | Household contacts | Sputum smear microscopy and chest X-ray | IGRA | 210.3 pg/ml (LTBI vs non-TB), 448 pg/ml (ATB vs LTBI) | TBAg (ESAT-6 + CFP-10 + TB7.7) | No |
| Wergeland I | 2016 | Norway | / | 36/11 | 16 | Known exposure of TB or origin from a high TB endemic country | / | IGRA (>0.70 IU/ml/0.35-0.70 IU/ml) | 37 pg/ml/unclear | TBAg-Nil (ESAT-6 + CFP-10 + TB7.7) | No |
| Sauzullo I | 2014 | Italy | / | 17 | 50 | Healthcare workers | / | TST + IGRA | 117 pg/ml | TBAg (ESAT-6 + CFP-10 + TB7.7) | No |
| Wang S | 2012 | China | / | 28 | 45 | Household and close contacts | / | IGRA | / | TBAg (ESAT-6 + CFP-10 + TB7.7) | No |
| Wang S | 2012 | China | / | 13 | 63 | Students and staff | / | IGRA | / | TBAg (ESAT-6 + CFP-10 + TB7.7) | No |
| Rubbo PA | 2012 | France | / | 8 | 17 | Healthcare workers | / | IGRA | 66 pg/ml | TBAg (ESAT-6 + CFP-10 + TB7.7) | No |
| Krummel B | 2010 | Germany | / | 7 | 151 | Contacts | / | IGRA | / | TBAg (ESAT-6 + CFP-10) | No |
| Ruhwald M | 2008 | Nigeria | / | 53 | 48 | Children contacts | / | IGRA | / | TBAg (ESAT-6 + CFP-10 + TB7.7) | / |
| Della Bella C | 2018 | Italy | 73 | 88 | / | BCG-vaccinated subjects or not | culture and chest X-ray | TST + IGRA | 12.5/71.25/231.25 SFC per million PBMCs | TBAg (Ala-DH/ESAT-6/CFP-10) | No |
| Movahedi B | 2017 | Iran | 33 | 33 | / | Household contacts | sputum smear microscopy and/or culture | TST | 235/275 SFC per million PBMCs | TBAg (ESAT-6 + CFP-10/Ala-DH) | No |
| Wu J | 2017 | China | 25 | 36 | / | Household contacts | Sputum smear microscopy, culture and chest X-ray | IGRA | 976.3 pg/ml | TBAg (PPD) | / |
| Mamishi S | 2016 | Iran | 30 | 30 | / | Unknown | Culture | TST + IGRA | 11.6/10.7 pg/ml | TBAg (ESAT-6/CFP-10) | No |
| Suzukawa M | 2016 | Japan | 31 | 29 | / | Hospital workers | Sputum smear microscopy, TRC, culture and chest X-ray | IGRA | 333.2 pg/ml | TBAg-Nil (ESAT-6 + CFP-10 + TB7.7) | No |
| Sun Q | 2016 | China | 65 | 43 | / | Household contacts | Sputum smear microscopy, culture and chest X-ray | IGRA | 266.3/329.1 pg/ml | TBAg (ESAT-6 + CFP-10/PPD) | 2 in ATB patients |
| Chiappini E | 2012 | Italy | 25 | 21 | / | Close contacts | Sputum smear microscopy, polymerase chain reaction, culture and chest CT | TST + IGRA | 12.5 SFC per million PBMCs | TBAg (Ala-DH) | No |
Ala-DH: L-alanine dehydrogenase; ATB: active tuberculosis; CT: computed tomography; CFP-10: culture filtrate protein-10; ESAT-6: early secreted antigenic target-6; HIV: human immunodeficiency virus; IGRA: interferon-gamma release assay; IL-2: interleukin-2; LTBI: latent tuberculosis infection; PBMCs: peripheral blood mononuclear cells; PPD: purified protein derivative; SFC: spot forming cells; TB: tuberculosis; TBAg-Nil: TB-specific antigen-stimulated minus Nil; TBAg: TB-specific antigen-stimulated; TRC: transcription reverse transcription concerted amplification; TST: tuberculin skin test.
TB burden, study design type, subjects’ age, the method for detecting IL-2, and the sensitivity, specificity, TP, FP, FN, and TN of IL-2 for 13 trials detecting LTBI from non-TB controls and 14 trials discriminating ATB from LTBI individuals are shown in Table 2.
Table 2.
Baseline data of included studies regarding IL-2 and TB infection.
| Author | Year | TB burden | Study design | Age | IL-2 method | Sensitivity (%) | Specificity (%) | TP | FP | FN | TN |
|---|---|---|---|---|---|---|---|---|---|---|---|
| LTBI vs. non-TB | |||||||||||
| Wang S | 2018 | High | Cohort | Children + adults | Multiplex cytokines assay | 85.3 | 96.2 | 29 | 1 | 5 | 25 |
| Balcells ME | 2018 | Low | Cohort | Children + adults | Multiplex cytokines assay | 68.2 | 92.6 | 15 | 2 | 7 | 25 |
| La Manna MP | 2018 | Low | Cohort | Children + adults | Multiplex cytokines assay | 74.19 | 95 | 23 | 1 | 8 | 19 |
| Nausch N | 2017 | Low | Cohort | Children | Multiplex cytokines assay | 93.3 | 100 | 14 | 0 | 1 | 12 |
| Kamakia R | 2017 | High | Cross-sectional | Adults | Multiplex cytokines assay | 100 | 91.7 | 16 | 1 | 0 | 7 |
| Wergeland I | 2016 | Low | Cross-sectional | Children + adults | Multiplex cytokines assay | 82 | 94 | 9 | 1 | 2 | 15 |
| Wergeland I | 2016 | Low | Cross-sectional | Children + adults | Multiplex cytokines assay | 89 | 94 | 32 | 1 | 4 | 15 |
| Sauzullo I | 2014 | Low | Cohort | Adults | ELISA | 94.12 | 100 | 16 | 0 | 1 | 50 |
| Wang S | 2012 | High | Cross-sectional | Adults | ELISA | 89.3 | 91.1 | 25 | 4 | 3 | 41 |
| Wang S | 2012 | High | Cross-sectional | Adults | ELISA | 84.6 | 100 | 11 | 0 | 2 | 63 |
| Rubbo PA | 2012 | Low | Cohort | Adults | Multiplex cytokines assay | 100 | 100 | 8 | 0 | 0 | 17 |
| Krummel B | 2010 | Low | Cohort | Children + adults | ELISpot | 42.9 | 98 | 3 | 3 | 4 | 148 |
| Ruhwald M | 2008 | High | Cohort | Children | Multiplex cytokines assay | 92.5 | 100 | 49 | 0 | 4 | 48 |
| ATB vs. LTBI | |||||||||||
| Della Bella C | 2018 | Low | Cohort | Adults | ELISpot | 96 | 100 | 70 | 0 | 3 | 88 |
| Della Bella C | 2018 | Low | Cohort | Adults | ELISpot | 86 | 36 | 63 | 56 | 10 | 32 |
| Della Bella C | 2018 | Low | Cohort | Adults | ELISpot | 80 | 54 | 58 | 37 | 15 | 51 |
| Balcells ME | 2018 | Low | Cohort | children + adults | Multiplex cytokines assay | 91.2 | 59.1 | 31 | 9 | 3 | 13 |
| Movahedi B | 2017 | Low | Cross-sectional | Adults | ELISpot | 69.7 | 42.4 | 23 | 19 | 10 | 14 |
| Movahedi B | 2017 | Low | Cross-sectional | Adults | ELISpot | 75.8 | 78.8 | 25 | 7 | 8 | 26 |
| Wu J | 2017 | High | Cohort | Children + adults | Multiplex cytokines assay | 84 | 58.3 | 21 | 15 | 4 | 21 |
| Kamakia R | 2017 | High | Cross-sectional | Adults | Multiplex cytokines assay | 75 | 91 | 12 | 1 | 4 | 10 |
| Mamishi S | 2016 | Low | Cross-sectional | children + adults | ELISA | 72 | 79 | 21 | 6 | 8 | 23 |
| Mamishi S | 2016 | Low | Cross-sectional | children + adults | ELISA | 75 | 79 | 21 | 6 | 7 | 23 |
| Suzukawa M | 2016 | Low | Cross sectional | Adults | Multiplex cytokines assay | 96.77 | 24.14 | 30 | 22 | 1 | 7 |
| Sun Q | 2016 | High | Cohort | Adults | ELISA | 70.8 | 76.7 | 46 | 10 | 19 | 33 |
| Sun Q | 2016 | High | Cohort | Adults | ELISA | 63.1 | 97.7 | 41 | 1 | 24 | 42 |
| Chiappini E | 2012 | Low | Cohort | Children | ELISpot | 100 | 81 | 25 | 4 | 0 | 17 |
ATB: active tuberculosis; FP: false positive; FN: false negative; ELISA: enzyme-linked immuno sorbent assay; ELISpot: enzyme-linked immunospot assay; IL-2: interleukin-2; LTBI: latent tuberculosis infection; TB: tuberculosis; TP: true positive; TN: true negative.
Quality of the included studies
The methodological quality of the included articles was determined by QUADAS-2 (Figure 2). Patient selection bias was unclear for seven articles (7/18), which used cross-sectional design in six articles and did not report the time of patient selection in one article [25,27,28,30,35,37,38]. In addition, the index tests bias was unclear in 13 articles (13/18), because the results of IL-2 cannot be obtained in blinded conditions [26–34,36–39]. The reference standard bias was unclear for one article (1/18), which had intermediate results 33. Flow and timing bias were unclear for six articles (6/18), in which patients were lost during the analysis of the diagnostic potential of IL-2 [23,25,28,31,33,37]. The bias of applicability concerns was generally low. The overall quality was good.
Figure 2.
Methodological quality regarding IL-2 and TB infection.
Summary statistics
For differentiating LTBI from non-TB controls, a total of 790samples were included in this analysis. The overall sensitivity and specificity for IL-2 were respectively 0.87(95%CI: 0.80–0.92) and 0.98(95%CI: 0.95–0.99). The pooled PLR was 34.78 (95%CI: 15.89–76.11), and the pooled NLR was 0.14 (95%CI: 0.09–0.21). The pooled DOR was 256.41 (95%CI: 95.02–691.90), suggesting that the differentiating effect of IL-2 for LTBI and non-TB controls was greatly credible. The AUROC was 0.98(95%CI: 0.96–0.99), indicating that IL-2 had an excellent diagnostic accuracy. The HSROC curves of IL-2 are shown in Figure 3.
Figure 3.
The hierarchical summary receiver operating characteristic (HSROC) curve for assessing the overall diagnostic performance of IL-2 differentiating LTBI from non-TB controls.
For differentiating ATB from LTBI individuals, a total of 1196 samples were included. The sensitivity ranged from 63.1% to 100% (overall sensitivity: 0.83, 95%CI: 0.76–0.89). The specificity ranged from 24.14% to 100% (overall specificity: 0.76, 95%CI: 0.58–0.88). The pooled PLR and NLR were, respectively 3.41 (95%CI: 1.83–6.35) and 0.22 (95%CI: 0.14–0.34). The pooled DOR was 15.47 (95%CI: 6.12–39.10). The AUROC was 0.87 (95%CI: 0.84–0.90), indicating that IL-2 had a good differentiating value. The HSROC curves are shown in Figure 4.
Figure 4.
The hierarchical summary receiver operating characteristic (HSROC) curve for assessing the overall diagnostic performance of IL-2 differentiating ATB from LTBI individuals.
Heterogeneity
As shown in Table 3, heterogeneity was assessed by subgroup and meta-regression analyses. Comparing LTBI with non-TB controls, heterogeneity was not present regarding the high TB burden versus low TB burden countries (p = .25), the cohort versus cross-sectional designs (p = .10), adults versus children (p = .11), the test methods of multiplex cytokine assays versus enzyme-linked immunosorbent assay (ELISA) and enzyme-linked immune absorbent spot (ELISpot) (p = .62). Comparing ATB with LTBI individuals, heterogeneity was not present regarding to the above aspects as in comparing LTBI with non-TB controls (p = .11, 0.45, 0.75, 0.23).
Table 3.
Heterogeneity assessment regarding IL-2 and TB infection.
| Covariate | Studies | Sensitivity (95%) | Specificity (95%) | p Value | |
|---|---|---|---|---|---|
| LTBI vs. non-TB | |||||
| TB burden | High | 5 | 0.91 (0.85-0.96) | 0.97 (0.94-1.00) | .25 |
| Low | 8 | 0.83 (0.75-0.90) | 0.98 (0.96-1.00) | ||
| Design type | Cohort | 8 | 0.85 (0.77-0.92) | 0.99 (0.97-1.00) | .1 |
| Cross-sectional | 5 | 0.90 (0.84-0.97) | 0.96 (0.91-1.00) | ||
| Age | Adults | 5 | 0.94 (0.88-0.99) | 0.98 (0.95-1.00) | .11 |
| Children and adults/children | 8 | 0.83 (0.76-0.90) | 0.98 (0.95-1.00) | ||
| IL-2 method | Multiplex cytokines assays | 9 | 0.88 (0.81-0.94) | 0.97 (0.94-1.00) | .62 |
| ELISA | 4 | 0.84 (0.72-0.97) | 0.98 (0.96-1.00) | ||
| ATB vs. LTBI | |||||
| TB burden | High | 4 | 0.73 (0.59-0.87) | 0.86 (0.68-1.00) | .11 |
| Low | 10 | 0.86 (0.80-0.92) | 0.70 (0.51-0.89) | ||
| Design type | Cohort | 8 | 0.86 (0.78-0.93) | 0.80 (0.63-0.97) | .45 |
| Cross-sectional | 6 | 0.79 (0.68-0.90) | 0.69 (0.43-0.95) | ||
| Age | Adults | 9 | 0.82 (0.73-0.90) | 0.77 (0.59-0.95) | .75 |
| Children and adults/children | 5 | 0.86 (0.77-0.96) | 0.73 (0.47-0.99) | ||
| IL-2 method | Multiplex cytokines assays | 4 | 0.89 (0.80-0.98) | 0.59 (0.24-0.94) | .23 |
| ELISA/ | 10 | 0.81 (0.73-0.88) | 0.80 (0.66-0.95) | ||
ATB: active tuberculosis; IL-2: interleukin-2; LTBI: latent tuberculosis infection; TB: tuberculosis.
Publication bias
Deeks’ funnel plot showed no statistically significant differences in differentiating LTBI from non-TB controls (p = .23) and ATB from LTBI individuals (p = .54). Therefore, there was no publication bias in this study.
Discussion
Currently, diagnostic methods for differentiating LTBI from non-TB controls and ATB patients are not optimal and restricted by age, BCG-vaccination status, and immune status etc [15,47]. LTBI affects one-third population worldwide [6]. Searching for new markers for auxiliary diagnosis of LTBI is necessary. This systematic review and meta-analysis assessed 27 original studies to evaluate the overall diagnostic performance of IL-2 for differentiating LTBI individuals from non-TB controls and ATB patients.
The pooled sensitivity and specificity were 87% and 98% in differentiating LTBI from non-TB controls. This means a relatively low possibility of missed diagnoses (13%) and an extremely low rate of misdiagnoses (2%), indicating a sufficient level of diagnostic accuracy of IL-2. The PLR was 34.78(>10), indicating that the probability of IL-2 in correctly diagnosing LTBI was large. The NLR of 0.14 (0.1–0.2) suggesting that the probability of IL-2 in incorrectly diagnosing non-TB controls was moderate. The DOR and AUROC were respectively 256.41 and 0.98, suggesting an excellent diagnostic performance of IL-2 in differentiating LTBI from non-TB controls. Therefore, IL-2 is an accurate marker which can be used as an auxiliary method to differentiate LTBI from non-TB controls. For differentiating ATB from LTBI individuals, the sensitivity and specificity were separately as 83% and 76%, which suggesting a moderate to large differentiating ability of IL-2. The PLR and NLR were separately as 3.41 (2-5) and 0.22 (0.2–0.5), suggesting the probability of IL-2 in differentiating ATB from LTBI was small. The DOR and AUROC of IL-2 were separately as 15.47 and 0.87, suggesting the overall performance was good. Therefore, we thought that IL-2 could also be a good marker to differentiate ATB from LTBI individuals.
In 2014, Mamishi et al. reported a systematic review that IL-2 is a valid marker (sensitivity: 0.81, specificity: 0.95, AUROC: 0.96) for diagnosing LTBI from non-TB individuals [48]. Mamishi et al. included 5 trails, and we included four trials of this systematic review and one trial was excluded without the specificity of IL-2. And, different data were extracted in two trials. Eventually, we selected 13 trials with a repetition of two trials, and thus, a more stable result of IL-2 for immunodiagnostic testing of LTBI from non-TB could be used. Besides, we only included Mtb-specific antigen stimulated IL-2, which avoided influence by other bacterial and viral infections [49,50].
TST and IGRA are recommended for the diagnosis of LTBI by the WHO [13,14]. The TST could show cross-reactivity in BCG-vaccinated individuals. However, IL-2 is less influenced by BCG vaccination. Ruhwald et al. reported the performance of IGRA and IL-2-based tests among children at high risk for LTBI who were exposed to sputum smear-positive adults [33]. It was observed that IGRA was positive in 42/59 cases (71%) and IL-2 was positive in 44/59 cases (75%), suggesting that IL-2 was an alternative biomarker to IGRA. Wang S et al. indicated the indeterminate rates of IGRA and IL-2 for ATB, household contacts and healthy controls were 2.3% and 1.9%, respectively [24].
Subsequently, the TB-burden, study design, age and IL-2 detection method were studied using bivariate analyses. And they were not significant source of heterogeneity (p > .05). The IL-2 result was independent of age, which was consistent with the report by Fisher et al. [51]. Multiplex cytokine assays and ELISA were comparable with respect to reliability and reproducibility [52]. Considering the cost, ELISA was preferable to multiplex cytokine assays.
Certainly, this systematic review and meta-analysis had several limitations. First, IL-2 is usually detected in combination with other immune markers such as IL-6 and IL-8. In this situation, other markers will decrease the diagnostic performance of IL-2 alone. The reliability and incremental benefits of IL-2 alone are could not be judged properly. Second, although the TB burden, study design, age, and IL-2 detection method were not significant sources of heterogeneity, the heterogeneity cannot be ignored since the various cut-offs of IL-2 could increase heterogeneity. Furthermore, we cannot perform a bivariate analysis using cut-offs since its range are too wide. Despite of no publication bias, it was still a concern. The concern was that we only included the studies published in English journals. We just analysed publication bias by the Deek’s plot for these English publications. However, we did not analyse the publication bias for articles in other languages. Therefore, a concern for publication bias could not be excluded.
Original studies just for detecting IL-2 alone without combination with other markers might resolve the reliability in judging the performance of IL-2. It will be very helpful if the cut-off standard of IL-2 is unified. Furthermore, if we could include all of the publications in non-English journals, which may decrease the publication bias.
In conclusion, this systematic review and meta-analysis shows that IL-2 is a powerful marker for differentiating LTBI from non-TB controls and a good marker for differentiating ATB from LTBI individuals. Further large-scale, multi-centre, prospective studies are warranted to support our findings.
Supplementary Material
Funding Statement
This work was supported by the National Science Foundation of China [81630038, 81971433, 81971428, 81771634, 81842011, 81330016], the National Key R&D Programme of China [2017YFA 0104200], the grants from the Ministry of Education of China [IRT0935], the grants from the Science and Technology Bureau of Sichuan Province [2016TD0002], and the grant from the clinical discipline programme (Neonatology) from the Ministry of Health of China [1311200003303].
Disclosure statement
The authors report no conflict of interest.
Data availability statement
All data generated within this study are available from the corresponding author on request.
References
- 1.Kawahara JY, Irvine EB, Alter G.. A case for antibodies as mechanistic correlates of immunity in tuberculosis. Front Immunol. 2019;10:996. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Committee. WGAbtGR. WHO guidelines on tuberculosis infection prevention and control: 2019 update. Geneva: World Health Organization; 2019. [PubMed] [Google Scholar]
- 3.Lee MR, Huang YP, Kuo YT, et al. Diabetes mellitus and latent tuberculosis infection: a systematic review and metaanalysis. Clin Infect Dis. 2017;64(6):719–727. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Restrepo BI, Schlesinger LS.. Impact of diabetes on the natural history of tuberculosis. Diabetes Res Clin Pract. 2014;106(2):191–199. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Cattamanchi A, Smith R, Steingart KR, et al. Interferon-gamma release assays for the diagnosis of latent tuberculosis infection in HIV-infected individuals: a systematic review and meta-analysis. J Acquir Immune Defic Syndr. 2011;56(3):230–238. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Getahun H, Matteelli A, Chaisson RE, et al. Latent mycobacterium tuberculosis infection. N Engl J Med. 2015;372(22):2127–2135. [DOI] [PubMed] [Google Scholar]
- 7.Rangaka MX, Wilkinson KA, Glynn JR, et al. Predictive value of interferon-gamma release assays for incident active tuberculosis: a systematic review and meta-analysis. Lancet Infect Dis. 2012;12(1):45–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Jung YEG, Schluger NW.. Advances in the diagnosis and treatment of latent tuberculosis infection. Curr Opin Infect Dis. 2020;33(2):166–172. [DOI] [PubMed] [Google Scholar]
- 9.Committee. WGAbtGR. Latent tuberculosis infection: updated and consolidated guidelines for programmatic management [Internet]. Geneva: World Health Organization; 2018. [PubMed] [Google Scholar]
- 10.Uplekar M, Weil D, Lonnroth K, et al. ; for WHO's Global TB Programme . WHO's new end TB strategy. Lancet. 2015;385(9979):1799–1801. [DOI] [PubMed] [Google Scholar]
- 11.Getahun H, Matteelli A, Abubakar I, et al. Management of latent Mycobacterium tuberculosis infection: WHO guidelines for low tuberculosis burden countries. Eur Respir J. 2015;46(6):1563–1576. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Won EJ, Choi JH, Cho YN, et al. Biomarkers for discrimination between latent tuberculosis infection and active tuberculosis disease. J Infect. 2017;74(3):281–293. [DOI] [PubMed] [Google Scholar]
- 13.Qiu X, Tang Y, Yue Y, et al. Accuracy of interferon-γ-induced protein 10 for diagnosing latent tuberculosis infection: a systematic review and meta-analysis. Clin Microbiol Infect. 2019;25(6):667–672. [DOI] [PubMed] [Google Scholar]
- 14.Kahwati LC, Feltner C, Halpern M, et al. Primary care screening and treatment for latent tuberculosis infection in adults: evidence report and systematic review for the US preventive services task force. JAMA. 2016;316(9):970–983. [DOI] [PubMed] [Google Scholar]
- 15.Farhat M, Greenaway C, Pai M, et al. False-positive tuberculin skin tests: what is the absolute effect of BCG and non-tuberculous mycobacteria? Int J Tuberc Lung Dis. 2006;10(11):1192–1204. [PubMed] [Google Scholar]
- 16.Barcellini L, Borroni E, Brown J, et al. First evaluation of QuantiFERON-TB Gold Plus performance in contact screening. Eur Respir J. 2016;48(5):1411–1419. [DOI] [PubMed] [Google Scholar]
- 17.Meier NR, Jacobsen M, Ottenhoff THM, et al. A systematic review on novel mycobacterium tuberculosis antigens and their discriminatory potential for the diagnosis of latent and active tuberculosis. Front Immunol. 2018;9:2476. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lucas M, Nicol P, McKinnon E, et al. A prospective large-scale study of methods for the detection of latent Mycobacterium tuberculosis infection in refugee children. Thorax. 2010;65(5):442–448. [DOI] [PubMed] [Google Scholar]
- 19.Leon-Janampa N, Zimic M, Shinkaruk S, et al. Synthesis, characterization and bio-functionalization of magnetic nanoparticles to improve the diagnosis of tuberculosis. Nanotechnology. 2020;31(17):175101. [DOI] [PubMed] [Google Scholar]
- 20.Warsinske H, Vashisht R, Khatri P.. Host-response-based gene signatures for tuberculosis diagnosis: A systematic comparison of 16 signatures. PLoS Med. 2019;16(4):e1002786. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Connell TG, Tebruegge M, Ritz N, et al. The potential danger of a solely interferon-gamma release assay-based approach to testing for latent Mycobacterium tuberculosis infection in children. Thorax. 2011;66(3):263–264. [DOI] [PubMed] [Google Scholar]
- 22.Chegou NN, Heyckendorf J, Walzl G, et al. Beyond the IFN-γ horizon: biomarkers for immunodiagnosis of infection with Mycobacterium tuberculosis. Eur Respir J. 2014;43(5):1472–1486. [DOI] [PubMed] [Google Scholar]
- 23.Balcells ME, Ruiz-Tagle C, Tiznado C, et al. Diagnostic performance of GM-CSF and IL-2 in response to long-term specific-antigen cell stimulation in patients with active and latent tuberculosis infection. Tuberculosis (Edinb)). 2018;112:110–119. [DOI] [PubMed] [Google Scholar]
- 24.Wang S, Li Y, Shen Y, et al. Screening and identification of a six-cytokine biosignature for detecting TB infection and discriminating active from latent TB. J Transl Med. 2018;16(1):206. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.La Manna MP, Orlando V, Li Donni P, et al. Identification of plasma biomarkers for discrimination between tuberculosis infection/disease and pulmonary non tuberculosis disease. PLoS One. 2018;13(3):e0192664. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Nausch N, Lundtoft C, Schulz G, et al. Multiple cytokines for the detection of Mycobacterium tuberculosis infection in children with tuberculosis. Int J Tuberc Lung Dis. 2017;21(3):270–277. [DOI] [PubMed] [Google Scholar]
- 27.Kamakia R, Kiazyk S, Waruk J, et al. Potential biomarkers associated with discrimination between latent and active pulmonary tuberculosis. Int J Tuberc Lung Dis. 2017;21(3):278–285. [DOI] [PubMed] [Google Scholar]
- 28.Wergeland I, Assmus J, Dyrhol-Riise AM.. Cytokine patterns in tuberculosis infection; IL-1ra, IL-2 and IP-10 differentiate borderline QuantiFERON-TB samples from uninfected controls. PLoS One. 2016;11(9):e0163848. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Sauzullo I, Mastroianni CM, Mengoni F, et al. Long-term IFN-γ and IL-2 response for detection of latent tuberculosis infection in healthcare workers with discordant immunologic results. J Immunol Methods. 2014;414:51–57. [DOI] [PubMed] [Google Scholar]
- 30.Wang S, Diao N, Lu C, et al. Evaluation of the diagnostic potential of IP-10 and IL-2 as biomarkers for the diagnosis of active and latent tuberculosis in a BCG-vaccinated population. PLoS One. 2012;7(12):e51338. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Rubbo PA, Nagot N, Le Moing V, et al. Multicytokine detection improves latent tuberculosis diagnosis in health care workers. J Clin Microbiol. 2012;50(5):1711–1717. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Krummel B, Strassburg A, Ernst M, et al. Potential role for IL-2 ELISpot in differentiating recent and remote infection in tuberculosis contact tracing. PLoS One. 2010;5(7):e11670. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Ruhwald M, Petersen J, Kofoed K, et al. Improving T-cell assays for the diagnosis of latent TB infection: potential of a diagnostic test based on IP-10. PLoS One. 2008;3(8):e2858. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Della Bella C, Spinicci M, Grassi A, et al. Novel M. tuberculosis specific IL-2 ELISpot assay discriminates adult patients with active or latent tuberculosis. PLoS One. 2018;13(6):e0197825. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Movahedi B, Mokarram P, Hemmati M, et al. IFN-gamma and IL-2 Responses to Recombinant AlaDH against ESAT-6/CFP-10 fusion antigens in the diagnosis of latent versus active tuberculosis infection. Iran J Med Sci. 2017;42(3):275–283. [PMC free article] [PubMed] [Google Scholar]
- 36.Wu J, Wang S, Lu C, et al. Multiple cytokine responses in discriminating between active tuberculosis and latent tuberculosis infection. Tuberculosis (Edinb). 2017;102:68–75. [DOI] [PubMed] [Google Scholar]
- 37.Mamishi S, Pourakbari B, Shams H, et al. Improving T-cell assays for diagnosis of latent TB infection: Confirmation of the potential role of testing Interleukin-2 release in Iranian patients. Allergol Immunopathol (Madr)). 2016;44(4):314–321. [DOI] [PubMed] [Google Scholar]
- 38.Suzukawa M, Akashi S, Nagai H, et al. Combined analysis of IFN-gamma, IL-2, IL-5, IL-10, IL-1RA and MCP-1 in QFT supernatant is useful for distinguishing active tuberculosis from latent infection. PLoS One. 2016;11(4):e0152483. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Sun Q, Wei W, Sha W.. Potential role for mycobacterium tuberculosis specific il-2 and ifn-γ responses in discriminating between latent infection and active disease after long-term stimulation. PLoS One. 2016;11(12):e0166501. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Chiappini E, Della Bella C, Bonsignori F, et al. Potential role of M. tuberculosis specific IFN-gamma and IL-2 ELISPOT assays in discriminating children with active or latent tuberculosis. PLoS One. 2012;7(9):e46041. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.McInnes MDF, Moher D, Thombs BD, et al. ; and the PRISMA-DTA Group . Preferred reporting items for a systematic review and meta-analysis of diagnostic test accuracy studies: the PRISMA-DTA statement. JAMA. 2018;319(4):388–396. [DOI] [PubMed] [Google Scholar]
- 42.Whiting PF, Rutjes AW, Westwood ME, et al. ; QUADAS-2 Group . QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med. 2011;155(8):529–536. [DOI] [PubMed] [Google Scholar]
- 43.Rutter CM, Gatsonis CA.. A hierarchical regression approach to meta-analysis of diagnostic test accuracy evaluations. Stat Med. 2001;20(19):2865–2884. [DOI] [PubMed] [Google Scholar]
- 44.Chappell FM, Raab GM, Wardlaw JM.. When are summary ROC curves appropriate for diagnostic meta-analyses? Stat Med. 2009;28(21):2653–2668. [DOI] [PubMed] [Google Scholar]
- 45.Eguchi H, Horita N, Ushio R, et al. Diagnostic test accuracy of antigenaemia assay for PCR-proven cytomegalovirus infection-systematic review and meta-analysis. Clin Microbiol Infect. 2017;23(12):907–915. [DOI] [PubMed] [Google Scholar]
- 46.Deeks JJ, Macaskill P, Irwig L.. The performance of tests of publication bias and other sample size effects in systematic reviews of diagnostic test accuracy was assessed. J Clin Epidemiol. 2005;58(9):882–893. [DOI] [PubMed] [Google Scholar]
- 47.Mandalakas AM, van Wyk S, Kirchner HL, et al. Detecting tuberculosis infection in HIV-infected children: a study of diagnostic accuracy, confounding and interaction. Pediatr Infect Dis J. 2013;32(3):e111–8–e118. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Mamishi S, Pourakbari B, Teymuri M, et al. Diagnostic accuracy of IL-2 for the diagnosis of latent tuberculosis: a systematic review and meta-analysis. Eur J Clin Microbiol Infect Dis. 2014;33(12):2111–2119. [DOI] [PubMed] [Google Scholar]
- 49.Maertzdorf J, Kaufmann SH, Weiner J. 3rd.. Toward a unified biosignature for tuberculosis. Cold Spring Harb Perspect Med. 2014;5(1):a018531. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Walzl G, Haks MC, Joosten SA, et al. Clinical immunology and multiplex biomarkers of human tuberculosis. Cold Spring Harb Perspect Med. 2015;5(4):a018515–a018515. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Lighter-Fisher J, Peng CH, Tse DB.. Cytokine responses to QuantiFERON(R) peptides, purified protein derivative and recombinant ESAT-6 in children with tuberculosis. Int J Tuberc Lung Dis. 2010;14(12):1548–1555. [PubMed] [Google Scholar]
- 52.Luo J, Zhang M, Yan B, et al. Diagnostic performance of plasma cytokine biosignature combination and MCP-1 as individual biomarkers for differentiating stages Mycobacterium tuberculosis infection. J Infect. 2019;78(4):281–291. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All data generated within this study are available from the corresponding author on request.