Abstract
Rationale: Interferon-γ (IFN-γ) release assays are widely used to diagnose latent infection with Mycobacterium tuberculosis in adults, but their performance in children remains incompletely evaluated to date.
Objectives: To investigate factors influencing results of IFN-γ release assays in children using a large European data set.
Methods: The Pediatric Tuberculosis Network European Trials group pooled and analyzed data from five sites across Europe comprising 1,128 children who were all investigated for latent tuberculosis infection by tuberculin skin test and at least one IFN-γ release assay. Multivariate analyses examined age, bacillus Calmette-Guérin (BCG) vaccination status, and sex as predictor variables of results. Subgroup analyses included children who were household contacts.
Measurements and Main Results: A total of 1,093 children had a QuantiFERON-TB Gold In-Tube assay and 382 had a T-SPOT.TB IFN-γ release assay. Age was positively correlated with a positive blood result (QuantiFERON-TB Gold In-Tube: odds ratio [OR], 1.08 per year increasing age [P < 0.0001]; T-SPOT.TB: OR, 1.14 per year increasing age [P < 0.001]). A positive QuantiFERON-TB Gold In-Tube result was shown by 5.5% of children with a tuberculin skin test result less than 5 mm, by 14.8% if less than 10 mm, and by 20.2% if less than 15 mm. Prior BCG vaccination was associated with a negative IFN-γ release assay result (QuantiFERON-TB Gold In-Tube: OR, 0.41 [P < 0.001]; T-SPOT.TB: OR, 0.41 [P < 0.001]). Young age was a predictor of indeterminate IFN-γ release assay results, but indeterminate rates were low (3.6% in children < 5 yr, 1% in children > 5 yr).
Conclusions: Our data show that BCG vaccination may be effective in protecting children against Mycobacterium tuberculosis infection. To restrict use of IFN-γ release assays to children with positive skin tests risks underestimating latent infection.
Keywords: interferon-γ release tests, latent tuberculosis, BCG vaccine, pediatrics
At a Glance Commentary
Scientific Knowledge on the Subject
Although interferon-γ (IFN-γ) release assays are a powerful tool, their performance in children, especially under 5 years old, remains under evaluation, and data from large cohorts are still lacking. The protective effect of bacillus Calmette-Guérin vaccination in children has been attributed to prevention of disseminated tuberculosis (TB) disease.
What This Study Adds to the Field
This large European multicenter pediatric study indicates a previously underappreciated role of bacillus Calmette-Guérin vaccination in preventing latent TB infection in children. Stepwise use of IFN-γ release assays only in tuberculin skin test–positive children may underestimate latent TB infection.
New cases of childhood tuberculosis (TB) usually arise after exposure to a smear-positive adult index case (1, 2). Data from the prechemotherapy era showed that in the absence of intervention, up to 40% of children less than 2 years of age could progress from infection to disease, usually within 12 months (3). Appropriate treatment of latent TB infection (LTBI) provides a golden opportunity to prevent childhood TB disease or later reactivation of disease in adult life. On this basis, the correct identification and preventive treatment of latently infected individuals represents an important tool for TB control worldwide (4).
The 100-year-old tuberculin skin test (TST) has been the mainstay of diagnosing LTBI; it relies on a delayed-type hypersensitivity response to purified protein derivative (PPD) of Mycobacterium tuberculosis (MTB) (5). Its advantages include low cost and no dependence on phlebotomy or laboratories. However, shared antigens between MTB, nontuberculous mycobacteria, and bacillus Calmette-Guérin (BCG) vaccine compromise its specificity. Furthermore, it is reliant on correct administration and interpretation at 48–72 hours, with associated loss to follow-up (6).
Newer blood-based interferon-γ (IFN-γ) release assays (IGRAs) detect IFN-γ released by sensitized T cells when re-exposed to MTB-specific antigens in vitro (7). The use of IGRAs in diagnosing LTBI has already been the subject of many research studies, systematic reviews, and meta-analyses (1, 6–13). In summary, in adults, IGRAs are perceived to be as sensitive as the TST, more specific, and to correlate better than TST to TB exposure (6). However, comprehensive data sets in children are still lacking, and there have been particular concerns over the performance of IGRAs in children under 5 years old, including a weak response to the positive control (14, 15) and higher proportions of indeterminate results (11, 16–19).
The benefits of BCG vaccination in pediatric populations have been attributed to the prevention of severe forms of TB (20, 21). In the past, there appeared to be little evidence that BCG protected against infection, because studies aiming to examine this question used TST results as the measure of MTB sensitization status. The absence of a gold standard test for LTBI complicates investigation of the effectiveness of BCG vaccination against infection. IGRAs are proving to be a useful tool in re-examining this question, but despite careful epidemiological studies examining correlation of IGRA results with risk of MTB infection in children, some studies show a protective effect (22), whereas others demonstrate no impact (23).
The Pediatric Tuberculosis Network European Trials group (pTB-NET), a multicenter European initiative of clinicians, scientists, and epidemiologists investigating childhood TB, set out to examine age, sex, and BCG vaccination status as predictors of TST and IGRA outcome among children evaluated for LTBI. We performed a subanalysis of household contacts as an indicator of exposure risk and thus a proxy for MTB infection. We also investigated contributory factors to indeterminate IGRA results and the likelihood of a positive IGRA in the context of a negative TST.
Methods
Data Collection
Each country investigator conducted a retrospective notes review of all children referred for evaluation of LTBI who had received both TST and IGRA as part of the TB screening process in their setting. All children who had the minimal mandatory data points collected between 2006 and 2009 were included. These were age, sex, reason for testing, and the results of TST and IGRA. BCG vaccination as determined by the presence of a scar and/or vaccination record was also documented. Anonymized data were then submitted to the pTB-NET, where it was pooled into a single database (Excel; Microsoft, Redmond, WA).
Sites
Data were gathered from sites in Athens, Greece; Barcelona, Spain; Padua, Italy; Sofia, Bulgaria; and London, United Kingdom. The study was approved by the ethics committees at each of the sites where IGRAs formed part of research studies and informed consent was obtained. Where IGRAs were in routine clinical practice, data were collected as part of clinical audit. Participating sites have a low prevalence of pediatric HIV and no severe malnutrition.
Participants
Children were 16 years of age or younger. At all sites, they were included if referred as household contacts of an index case treated for TB or if referred for evaluation of LTBI having originated from a high TB prevalence area. At the sites in Greece and Spain, children were also included if they had a positive TST result in their national universal childhood TB screening programs and were then referred to the clinical site for further evaluation. Children were excluded from analysis if there was clinical, radiological, or microbiological suspicion of active TB.
Tuberculin Skin Testing
All children were investigated with a TST using 2 units of PPD RT 23 (Statens Serum Institut, Copenhagen, Denmark), placed by trained nursing staff experienced in the use of TST as part of routine clinical care. World Health Organization (WHO, Geneva, Switzerland) criteria were used to interpret TST results, where in the absence of HIV infection or severe malnutrition, a TST induration at least 10 mm in diameter is regarded as positive (4). In TST-negative subgroup analysis, cutoffs of 5, 10, and 15 mm were used to maximize relevance to international guidelines, which differ in their cutoffs when diagnosing LTBI in children and recommending prophylactic treatment.
Laboratory Assays
IGRAs were performed simultaneously or at latest within 7 days of the TST result. The choice of IGRA depended on availability at the sites. The QuantiFERON-TB Gold In-Tube test (QFT-GIT; Cellestis, Carnegie, Australia) (24) and T-SPOT.TB test (Oxford Immunotec, Oxford, UK) (25) are described in detail elsewhere. Tests were performed and the results interpreted at country level according to the manufacturer’s instructions.
Statistical Analysis
The primary outcome measures were odds ratios (ORs) for a positive result in the TST, QFT-GIT, and T-SPOT.TB in multivariate logistic regression analyses. The covariates were age, sex, and BCG vaccination status, adjusting for site. Ninety-five percent confidence intervals (CIs) were calculated for each odds ratio, and P values less than 0.05 were considered significant. The same analysis was performed for children who were household contacts of adults treated for pulmonary TB, children who had a negative TST and were tested with an IGRA, and for those with an indeterminate IGRA result. Where children were evaluated with both IGRAs, agreement was evaluated with the kappa statistic (κ), using the Landis and Koch interpretation (26).
Spearman’s correlation was used to investigate the relationship between TST induration and BCG vaccination status, and between age and TST induration.
Means and standard deviations (SDs) or medians and interquartile ranges (IQRs) were used for parametric and nonparametric data, respectively.
Data were analyzed with Stata 9.0 (Stata Statistical Software Release 9, 2005; StataCorp, College Station, TX).
Results
Study Participant Characteristics
Data from 1,128 children were included in the study. The baseline demographic and relevant clinical data are summarized in Table 1. The majority of the cases came from the Greek (n = 491, 43.5%) and Spanish (n = 459, 40.7%) sites. There was an even sex distribution with a male-to-female ratio of 1:0.98. The mean (SD) age was 8.2 (4.4) years. There were 268 children less than 5 years of age (23.8%), and 99 children less than 2 years of age (8.8%). HIV status was known in 42 (3.7%) of the cases, all of whom were negative. BCG vaccination status was known in 1,108 (98.2%) of the cases, of whom 684 (61.7%) were vaccinated. Indications for investigation of LTBI are shown in Table 2. The reason for IGRA testing was being a household contact for 43.5% of the children.
TABLE 1.
BASELINE DEMOGRAPHIC DATA INCLUDING BACILLUS CALMETTE-GUÉRIN VACCINATION STATUS
| n (%)* | |
| Country | |
| Bulgaria | 26/1,128 (2.3) |
| Greece | 491/1,128 (43.5) |
| Italy | 42/1,128 (3.7) |
| Spain | 459/1,128 (40.7) |
| UK | 110/1,128 (9.8) |
| Males | 569/1,128 (50.4) |
| Age, yr: mean (SD) | 8.2 (4.4) |
| <2 yr | 99/1,128 (8.8) |
| ≥2 yr | 1,029/1,128 (91.2) |
| <5 yr | 268/1,128 (23.8) |
| ≥5 yr | 860/1,128 (76.2) |
| BCG+ | 684/1,108 (61.7) |
Definition of abbreviation: BCG = bacillus Calmette-Guérin.
Unless otherwise stated.
TABLE 2.
REASONS FOR IFN-γ RELEASE ASSAY SCREENING IN STUDY PARTICIPANTS
| Reason for IGRA Screening | n (%) |
| Household contact | 491/1,128 (43.5) |
| Universal screening policy | 405/1,128 (35.9) |
| Origin from high-prevalence area | 170/1,128 (15.1) |
| School outbreak | 55/1,128 (4.9) |
| Hospital exposure | 6/1,128 (0.5) |
| Maternal TB during pregnancy | 1/1,128 (0.1) |
Definition of abbreviations: IGRA = IFN-γ release assay; TB = tuberculosis.
Summary of Test Results for All Children
The basic characteristics of the investigations and their results are displayed in Table 3. TST results were available for all children, with a median (IQR) induration across all groups of 12 (8–16) mm. There were 682 (60.5%) positive and 446 (39.5%) negative TST results as defined by WHO criteria (cutoff, 10 mm) (4). The QFT-GIT assay was used in 1,093 cases (96.9%), of whom 331 (30.3%) had positive results, 742 (67.9%) had negative results, and 20 (1.8%) had indeterminate results. T-SPOT.TB results were available for 382 cases (33.9%). Of these, 145 (37.9%) had positive results, 231 (60.5%) had negative results, and 6 (1.6%) had indeterminate results. Both IGRAs were conducted in 333 children simultaneously.
TABLE 3.
SUMMARY OF ALL TUBERCULIN SKIN TEST AND IFN-γ RELEASE ASSAY RESULTS FOR ALL CHILDREN
| Assay | All Ages | Age, <2 yr | Age, 2–5 yr | Age, >5 yr |
| TST, n = 1,128 | ||||
| Median induration, mm (IQR) | 12 (8–16) | 9 (0–13) | 10 (0–15) | 13 (10–17) |
| Positive, no. (%) | 682 (60.5) | 32/99 (32.3) | 96/215 (44.7) | 554/814 (68.1) |
| Negative, no. (%) | 446 (39.5) | 67/99 (67.7) | 119/215 (55.3) | 260/814 (31.9) |
| QFT-GIT, n = 1,093 | ||||
| Positive, no. (%) | 331 (30.3) | 22/96 (22.9) | 50/206 (24.3) | 259/791 (32.7) |
| Negative, no. (%) | 742 (67.9) | 71/96 (74.0) | 148/206 (71.8) | 523/791 (66.1) |
| Indeterminate, no. (%) | 20 (1.8) | 3/96 (3.1) | 8/206 (3.9) | 9/791 (1.1) |
| T-SPOT.TB, n = 382 | ||||
| Positive, no. (%) | 145 (37.9) | 9/36 (25.0) | 19/73 (26.0) | 117/273 (42.9) |
| Negative, no. (%) | 231 (60.5) | 26/36 (72.2) | 51/73 (69.9) | 154/273 (56.4) |
| Indeterminate, no. (%) | 6 (1.6) | 1/36 (2.8) | 3/73 (4.1) | 2/273 (0.7) |
Definition of abbreviations: IQR = interquartile range; QFT-GIT = QuantiFERON-TB Gold In-Tube; TST = tuberculin skin test.
Influence of Age
Age was positively correlated with TST induration (in millimeters) with Spearman’s rank correlation (r = 0.29, P < 0.0001). For the multivariate logistic regression analysis (Table 4), we examined age as a continuous variable and also in two age groups (older than 2 yr and older than 5 yr). Age as a continuous variable showed a significant association with a positive TST result ≥ 10 mm (OR, 1.09 per year increasing age; CI, 1.06–1.13; P < 0.001). Similarly, children more than 2 years of age (OR, 2.75; CI, 1.74–4.36; P < 0.001) or more than 5 years of age (OR, 2.47; CI, 1.82–3.34; P < 0.001) were significantly more likely to have a positive TST result compared with younger children.
TABLE 4.
MULTIVARIATE LOGISTIC REGRESSION ANALYSIS FOR ASSOCIATION BETWEEN POSITIVE IFN-γ RELEASE ASSAY OR TUBERCULIN SKIN TEST RESULT GREATER THAN 10 MILLIMETERS, AS OUTCOME VARIABLE AND PREDICTOR VARIABLES OF AGE, SEX, AND BACILLUS CALMETTE-GUÉRIN VACCINATION STATUS, ADJUSTING FOR STUDY SITE
| Predictor Variable | Descriptor | Outcome Variable (Positive Result) | OR | P Value (95% CI) |
| Age | Continuous | TST | 1.09* | <0.001 (1.06–1.13) |
| QFT-GIT | 1.08* | <0.001 (1.05–1.12) | ||
| T-SPOT.TB | 1.14* | <0.001 (1.08–1.20) | ||
| >2 yr | TST | 2.75 | <0.001 (1.74–4.36) | |
| QFT-GIT | 1.72 | 0.04 (1.03–2.85) | ||
| T-SPOT.TB | 1.77 | 0.16 (0.80–3.96) | ||
| >5 yr | TST | 2.47 | <0.001 (1.82–3.34) | |
| QFT-GIT | 1.91 | <0.001 (1.35–2.70) | ||
| T-SPOT.TB | 2.27 | 0.003 (1.33–3.87) | ||
| BCG | Vaccinated | TST | 3.22 | <0.001 (2.41–4.32) |
| QFT-GIT | 0.41 | <0.001 (0.30–0.55) | ||
| T-SPOT.TB | 0.41 | <0.001 (0.25–0.66) | ||
| Sex | Male | TST | 1.04 | 0.75 (0.81–1.35) |
| QFT-GIT | 1.49 | 0.004 (1.14–1.96) | ||
| T-SPOT.TB | 1.13 | 0.60 (0.72–1.76) |
Definition of abbreviations: BCG = bacillus Calmette-Guérin; CI = confidence interval; OR = odds ratio; QFT-GIT = QuantiFERON-TB Gold In-Tube; TST = tuberculin skin test.
OR per year of increasing age.
Both QFT-GIT and T-SPOT.TB were also significantly more likely to be positive when age was examined as a continuous variable (QFT-GIT: OR, 1.08 per year increasing age; CI, 1.05–1.12 [P < 0.0001]; T-SPOT.TB: OR, 1.14 per year increasing age; CI, 1.08–1.20 [P < 0.001]). Both IGRAs were more likely to be positive in children more than 5 years old (QFT-GIT: OR, 1.91; CI, 1.35–2.70 [P < 0.001]; T-SPOT.TB: OR, 2.27; CI, 1.33–3.87 [P = 0.003]). Children more than 2 years old were again more likely to have a positive QFT-GIT, although this did not reach statistical significance for T-SPOT.TB (QFT-GIT: OR, 1.72; CI, 1.03–2.85 [P = 0.04]; T-SPOT.TB: OR, 1.77; CI, 0.80–3.96 [P = 0.16]).
Influence of BCG Vaccination Status
BCG vaccination waspositively correlated with TST induration in millimeters (r = 0.28, and P < 0.0001). Vaccinated children had significantly larger TST indurations (median, 14 mm; interquartile range, 10–17 mm) than unvaccinated children (median, 10 mm; interquartile range, 0–14 mm; P < 0.001). In the multivariate logistic regression (Table 4) BCG vaccination was strongly associated with a positive TST result (≥10 mm) (OR, 3.22; CI, 2.41–4.32; P < 0.001). The opposite pattern is shown for both IGRAs, where BCG vaccination is significantly associated with a negative IGRA result (QFT-GIT: OR, 0.41; CI, 0.30–0.55 [P < 0.001]; T-SPOT.TB: OR, 0.41; CI, 0.25–0.66 [P < 0.001]). These patterns also hold true in the subgroup analysis of household contacts (Table 5). BCG vaccination was again associated with a positive TST result (OR, 2.08; CI, 1.36–3.20; P = 0.001) but a negative IGRA (QFT-GIT: OR, 0.52; CI, 0.33–0.81 [P = 0.003]; T-SPOT.TB: OR, 0.31; CI, 0.16–0.58 [P < 0.001]).
TABLE 5.
MULTIVARIATE LOGISTIC REGRESSION ANALYSIS FOR HOUSEHOLD CONTACTS ONLY: ASSOCIATION BETWEEN POSITIVE IFN-γ RELEASE ASSAY OR TUBERCULIN SKIN TEST RESULT GREATER THAN 10 mm, AS OUTCOME VARIABLE AND PREDICTOR VARIABLES OF AGE, SEX, AND BACILLUS CALMETTE-GUÉRIN VACCINATION STATUS, ADJUSTING FOR SITE
| Predictor Variable | Descriptor | Outcome Variable (Positive Result) | OR | P Value (95% CI) |
| Age | Continuous | TST | 1.14* | <0.001 (1.09–1.19) |
| QFT-GIT | 1.10* | <0.001 (1.05–1.15) | ||
| T-SPOT.TB | 1.14* | <0.001 (1.07–1.21) | ||
| >2 yr | TST | 2.54 | 0.005 (1.33–4.86) | |
| QFT-GIT | 2.14 | 0.04 (1.05–4.36) | ||
| T-SPOT.TB | 3.12 | 0.03 (1.10–8.85) | ||
| >5 yr | TST | 3.57 | <0.001 (2.27–5.61) | |
| QFT-GIT | 2.93 | <0.001 (1.78–4.84) | ||
| T-SPOT.TB | 3.64 | <0.001 (1.87–7.11) | ||
| BCG | Vaccinated | TST | 2.08 | 0.001 (1.36–3.20) |
| QFT-GIT | 0.52 | 0.003 (0.33–0.81) | ||
| T-SPOT.TB | 0.31 | <0.001 (0.16–0.58) | ||
| Sex | Male | TST | 1.30 | 0.19 (0.88–1.93) |
| QFT-GIT | 1.52 | 0.04 (1.02–2.27) | ||
| T-SPOT.TB | 1.20 | 0.53 (0.68–2.10) |
Definition of abbreviations: BCG = bacillus Calmette-Guérin; CI = confidence interval; OR = odds ratio; QFT-GIT = QuantiFERON-TB Gold In-Tube; TST = tuberculin skin test.
OR per year of increasing age.
Influence of Sex
In the multivariate logistic regression for all children (Table 4) and for household contacts only (Table 5), male sex was significantly associated with a positive QFT-GIT result.
Concordance between TST, T-SPOT.TB, and QFT-GIT Tests
In the 1,073 children whose QFT-GIT test gave a definitive result (see Table E1A in the online supplement), concordance with a TST result of at least 10 mm was fair (κ = 0.23, standard error [SE] 0.03). In the 376 children whose T-SPOT.TB test gave a definitive result (Table E1B), concordance with a TST result of at least 10 mm was fair (κ = 0.35, SE 0.05). There were 333 children who were tested with a TST, and both IGRAs (Table E1C). In 290 of the 333 children who had both IGRAs, there was substantial agreement between QFT-GIT and T-SPOT.TB (κ = 0.73, SE 0.06). Concordance between the IGRAs and TST for these 290 children was fair (κ = 0.37, SE 0.05) (Table E1D).
IGRA in “TST-Negative” Children
A subgroup analysis was performed to describe the rates of positivity of IGRA in the context of a “negative” TST, as defined by the three different cutoffs most commonly used in international studies (Table 6). The basic demographic and clinical details of these subgroups are summarized in Table E2. There was an even sex distribution of children within this subgroup. For the 200 children with a TST result less than 5 mm, 3.5% of those tested with T-SPOT.TB (2 of 57), to 5.5% of those tested with QFT-GIT (11 of 199), had a positive IGRA result. Using a cutoff of less than 10 mm for a positive TST, 12.8% of those children with a negative TST tested with T-SPOT.TB (18 of 141) to 14.8% of those tested with QFT-GIT (65 of 439) had a positive IGRA result. Of the 702 children with a TST result less than 15 mm, 20.2% (140 of 692) of those tested with QFT-GIT to 23.9% (50 of 209) of those tested with T-SPOT.TB had a positive IGRA result (Table 6).
TABLE 6.
IFN-γ RELEASE ASSAY RESULTS FOR CHILDREN WITH NEGATIVE TUBERCULIN SKIN TEST RESULTS, AS DEFINED BY THREE DIFFERENT CUTOFFS
| TST Results | IGRA Results | All Ages | Age, <2 yr | Age, 2–5 yr | Age, <5 yr | Age, ≥5 yr |
| <5 mm | QFT-GIT, n (%) | |||||
| Positive | 11/199 (5.5) | 2/38 (5.3) | 3/75 (4.0) | 5/103 (4.9) | 6/96 (6.3) | |
| Negative | 180/199 (90.5) | 33/38 (86.8) | 69/75 (92.0) | 92/103 (89.3) | 88/96 (91.7) | |
| Indeterminate | 8/199 (4.0) | 3/38 (7.9) | 3/75 (4.0) | 6/103 (5.8) | 2/96 (2.1) | |
| T-SPOT.TB, n (%) | ||||||
| Positive | 2/57 (3.5) | 0/8 (0.0) | 2/22 (9.1) | 2/29 (6.9) | 0/28 (0.0) | |
| Negative | 54/57 (94.7) | 8/8 (100.0) | 19/22 (86.4) | 26/29 (89.7) | 28/28 (100.0) | |
| Indeterminate | 1/57 (1.8) | 0/8 (0.0) | 1/22 (4.6) | 1/29 (3.5) | 0/28 (0.0) | |
| <10 mm | QFT-GIT, n (%) | |||||
| Positive | 65/439 (14.8) | 8/54 (14.8) | 4/102 (3.9) | 19/163 (11.7) | 46/276 (16.7) | |
| Negative | 363/439 (82.7) | 43/54 (79.6) | 93/102 (91.2) | 136/163 (83.4) | 227/276 (82.3) | |
| Indeterminate | 11/439 (2.5) | 3/54 (5.6) | 5/102 (4.9) | 8/163 (4.9) | 3/276 (1.1) | |
| T-SPOT.TB, n (%) | ||||||
| Positive | 18/141 (12.8) | 1/14 (7.1) | 2/32 (6.3) | 4/52 (7.7) | 14/89 (15.7) | |
| Negative | 120/141 (85.1) | 12/14 (85.7) | 29/32 (90.6) | 46/52 (88.5) | 74/89 (83.2) | |
| Indeterminate | 3/141 (2.1) | 1/14 (7.1) | 1/32 (3.1) | 2/52 (3.9) | 1/89 (1.1) | |
| <15 mm | QFT-GIT, n (%) | |||||
| Positive | 140/692 (20.2) | 13/76 (17.1) | 27/156 (17.3) | 34/203 (16.8) | 106/489 (21.7) | |
| Negative | 539/692 (77.9) | 60/76 (78.9) | 124/156 (79.5) | 161/203 (79.3) | 378/489 (77.3) | |
| Indeterminate | 13/692 (1.9) | 3/76 (4.0) | 5/156 (3.2) | 8/203 (3.9) | 5/489 (1.0) | |
| T-SPOT.TB, n (%) | ||||||
| Positive | 50/209 (23.9) | 3/23 (13.0) | 9/51 (17.7) | 10/66 (15.2) | 40/143 (28.0) | |
| Negative | 154/209 (73.7) | 19/23 (82.6) | 40/51 (78.4) | 54/66 (81.8) | 100/143 (69.9) | |
| Indeterminate | 5/209 (2.4) | 1/23 (4.4) | 2/51 (3.9) | 2/66 (3.0) | 3/143 (2.1) |
Definition of abbreviations: IGRA = IFN-γ release assay; QFT-GIT = QuantiFERON-TB Gold In-Tube; TST = tuberculin skin test.
Multivariate logistic regression analyses of these children with “negative” TST results with positive QFT-GIT results as outcome variables were likely influenced by lack of statistical power (data not shown).
Indeterminate IGRA Results
The overall proportion of indeterminate results was low, and there was no significant difference between the two IGRAs (20 of 1,093, 1.8% for QFT-GIT vs. 6 of 382, 1.57% for T-SPOT.TB; P = 0.77). There were higher indeterminate rates in younger children (Table 3), up to 4.1%. In multivariate logistic regression analysis using age as a continuous variable (Table 7), older children were less likely to have an indeterminate QFT-GIT result (OR, 0.83 per year of increasing age; CI, 0.73–0.94; P = 0.004). This was also true when age was investigated as a binary variable with children more than 5 years old (OR, 0.25; CI, 0.09–0.67; P = 0.006).
TABLE 7.
MULTIVARIATE LOGISTIC REGRESSION ANALYSIS FOR CHILDREN WITH AN INDETERMINATE IFN-γ RELEASE ASSAY RESULT: ASSOCIATION BETWEEN INDETERMINATE IFN-γ RELEASE ASSAY RESULT, AS OUTCOME VARIABLE AND PREDICTOR VARIABLES OF AGE, SEX, AND BACILLUS CALMETTE-GUÉRIN VACCINATION STATUS ADJUSTED FOR STUDY SITE
| Predictor Variable | Descriptor | Outcome Variable (Indeterminate Result) | OR | P Value (95% CI) |
| Age | Continuous | QFT-GIT | 0.83 | 0.004 (0.73–0.94) |
| T-SPOT.TB | 0.90 | 0.26 (0.74–1.09) | ||
| ≥2 yr | QFT-GIT | 0.53 | 0.34 (0.15–1.94) | |
| T-SPOT.TB | 0.50 | 0.54 (0.06–4.52) | ||
| ≥5 yr | QFT-GIT | 0.25 | 0.006 (0.09–0.67) | |
| T-SPOT.TB | 0.34 | 0.19 (0.07–1.71) | ||
| BCG | Vaccinated | QFT-GIT | 1.52 | 0.44 (0.52– 4.43) |
| T-SPOT.TB | 2.57 | 0.41 (0.27–24.83) | ||
| Sex | Male | QFT-GIT | 0.93 | 0.88 (0.37–2.33) |
| T-SPOT.TB | 0.46 | 0.37 (0.08–2.55) |
Definition of abbreviations: BCG = bacillus Calmette-Guérin; CI = confidence interval; OR = odds ratio; QFT-GIT = QuantiFERON-TB Gold In-Tube.
Discussion
To our knowledge, this is the largest data set available for IGRAs in children investigated for LTBI in low TB prevalence areas, with 1,128 participants from five sites across Europe. It illustrates the opportunities provided by collaborative research within research networks, such as the pTB-NET.
As previously shown, the association between BCG vaccination and a positive TST result was confirmed by our data. More importantly, however, we also showed a significant association with BCG vaccination and a negative IGRA for both QFT-GIT and T-SPOT.TB, which was also observed when the analysis was restricted to household contacts only. This supports the hypothesis that BCG vaccine not only protects against disseminated disease in childhood, but may also protect against infection (22, 27, 28). Our findings are in contrast to some studies that showed no association with either a positive or negative IGRA result (23, 29–31), but are supported by data from the United Kingdom (27, 28) and Turkey (22), where—like in our analysis—both TST and IGRA were used in conjunction. Although there is no gold standard test for LTBI and IGRA results were used as a surrogate, our findings could have implications for BCG vaccination policy worldwide, including in low-prevalence areas, and for the evaluation of the performance of novel anti-TB vaccines.
Our results highlight the challenges of using TST and IGRAs in pediatric practice. Using WHO criteria for a positive TST, both IGRAs suggested a lower proportion of children with LTBI (30.3% positive, QFT-GIT; and 37.9% positive, T-SPOT.TB) than the TST (60.5%). There was substantial agreement between the two types of IGRA in children when tested with both. However, the proportion of children with a negative TST and a positive IGRA varied from 3.5 to 23.9% depending on the IGRA and TST cutoff. This suggests that stepwise use of IGRAs primarily in TST-positive individuals as recommended in the U.K. National Institute for Health and Clinical Excellence (NICE) guidelines (32) might encounter the risk of underestimating and therefore undertreating childhood LTBI.
Our data show that younger children were significantly less likely to have either a TST result greater than 10 mm, or a positive IGRA. Whether this reflects the reality and is indeed due to absence of exposure or, rather, is indicative of false negative results of these tests can be debated. In addition to a maturing immune response reflected in the mechanism of these tests, the particularly marked age effect on TST may represent increasing exposure to nontuberculous mycobacteria. Furthermore, although we report relatively low frequencies of indeterminate IGRA results, younger age was a significant factor for indeterminate QFT-GIT results (12, 16–19). Both the NICE guidelines (32) and American Academy of Pediatrics guidelines (33) highlight the difficulty of interpreting a negative IGRA result in children, the latter stating that IGRA “cannot be recommended routinely for use in children younger than 5 years of age because of a lack of published data about their utility.” Our data support these observations in principle—although the data set for children under 5 years of age was limited by its size, indeterminate test results were rare and the dilemma of distinguishing between a “true negative” due to lack of infection and a “false negative” due to a suboptimal immune response in young children remains. A large Croatian study of children less than 5 years old showed no evidence of age influencing QFT-GIT results (34), as did a much smaller French study (35). Further studies of the negative predictive value of IGRA including young children are currently ongoing.
Interestingly, boys were substantially more likely to have a positive QFT-GIT result than girls, a finding that is well recognized in adult populations (36) but not in pediatric populations (37). Sex-based differences in TB (37), inflammatory responses in childhood (38), severe sepsis childhood presentations in the developed world (39), and vaccine-related morbidity in the developing world (40) are all well-recognized phenomena. It is unclear whether our results are due to an underlying biological difference, higher risk of exposure, or reflect a sociocultural characteristic, but further investigation would be warranted.
Our study has several limitations: although we were able to analyze a large data set through the combined efforts of several European investigators, the number of children younger than 2 and 5 years of age was still relatively small. The data collection was retrospective in nature and cross-sectional, and involved pooling of data from several sites across Europe. The same strain of BCG is not used in the various countries, and therefore controlling for site in the analysis was of particular importance. There was an uneven case distribution across the sites and as a consequence of local policy, practical, and commercial factors more children were tested with QFT-GIT than T-SPOT.TB. Although we performed a subanalysis of household contacts as a proxy of risk of MTB infection, further quantification of the degree of exposure to TB cases including proximity to the index case, number of index cases in the household, and infectivity of the index case was not possible. The nature of the data collection precluded evaluation of any potential boosting of IGRA results through the TST, and this is unlikely to play a significant role (41).
In summary, this multicenter pTB-NET project highlights several key areas when investigating children for LTBI that are important for TB control programs:
The role of BCG vaccination should be reappraised in the light of our findings and other publications. Our data suggest that BCG vaccination may reduce the risk of LTBI in children by greater than 50% and that its decline in use in the developed world may have been premature.
In our study, young age was a predictor of indeterminate IGRA results. Only an improved understanding of the immune responses in young children is likely to allow the development of more reliable tests, which can measure host responses in this age group.
The high proportions of “TST-negative,” IGRA-positive children in our study highlight the importance of evaluating the existing clinical guidelines, many of which recommend that an IGRA be used only in TST-positive individuals.
The sex differences in our data highlight yet unexplored areas for biological and sociological research within the field of TB.
Supplementary Material
Footnotes
Supported by the Wellcome Trust (B.K.), MRC (B.K.), NIHR fellowships (B.K. and R.B.R.), and Fondo de Investigaciones Sanitarias Proyecto no. PI081738 (N.A.-G.).
Author Contributions: Conception and design: B.K., G.S., N.A.-G., M.T., E.R., S.V., and the pTB-NET (Pediatric Tuberculosis Network European Trials group); data contribution and collation: R.B.R., B.K., N.A.-G., M.T., E.R., and S.V.; statistical analysis: G.S.; drafting the manuscript for important intellectual content: R.B.R. and B.K.; editing and review of the manuscript: R.B.R., B.K., G.S., N.A.-G., M.T., E.R., and S.V.
This article has an online supplement, which is available from this issue’s table of contents at www.atsjournals.org
Originally Published in Press as DOI: 10.1164/rccm.201201-0026OC on June 14, 2012
Author disclosures are available with the text of this article at www.atsjournals.org.
References
- 1.Jones C, Whittaker E, Bamford A, Kampmann B. Immunology and pathogenesis of childhood TB. Paediatr Respir Rev 2011;12:3–8 [DOI] [PubMed] [Google Scholar]
- 2.Schaaf H, Michaelis I, Richardson M. Adult-to-child transmission of tuberculosis: household or community contact? Int J Tuberc Lung Dis 2003;7:426–431 [PubMed] [Google Scholar]
- 3.Marais BJ, Gie RP, Schaaf HS, Hesseling AC, Obihara CC, Starke JJ, Enarson DA, Donald PR, Beyers N. The natural history of childhood intra-thoracic tuberculosis: a critical review of literature from the pre-chemotherapy era. Int J Tuberc Lung Dis 2004;8:392–402 [PubMed] [Google Scholar]
- 4.World Health Organization Stop TB Partnership Childhood TB Subgroup Guidance for national tuberculosis programmes on the management of tuberculosis in children. Chapter 1. Introduction and diagnosis of tuberculosis in children. Int J Tuberc Lung Dis 2006;10:1091–1097 [PubMed] [Google Scholar]
- 5.Daniel T. The history of tuberculosis. Respir Med 2006;100:1862–1870 [DOI] [PubMed] [Google Scholar]
- 6.Dheda K, van Zyl Smit R, Badri M, Pai M. T-cell interferon-γ release assays for the rapid immunodiagnosis of tuberculosis: clinical utility in high-burden vs. low-burden settings. Curr Opin Pulm Med 2009;15:188–200 [DOI] [PubMed] [Google Scholar]
- 7.Diel R, Goletti D, Ferrara G, Bothamley G, Cirillo D, Kampmann B, Lange C, Losi M, Markova R, Migliori GB, et al. Interferon γ release assays for the diagnosis of latent M. tuberculosis infection: a systematic review and meta-analysis. Eur Respir J 2011;37:88–99 [DOI] [PubMed] [Google Scholar]
- 8.Ahmad S. Pathogenesis, immunology, and diagnosis of latent Mycobacterium tuberculosis infection. Clin Dev Immunol 2011;2011:814943. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Lalvani AMK. T cell–based diagnosis of childhood tuberculosis infection. Curr Opin Infect Dis 2007;20:264–271 [DOI] [PubMed] [Google Scholar]
- 10.Ling D, Zwerling A, Steingart K, Pai M. Immune-based diagnostics for TB in children: what is the evidence? Paediatr Respir Rev 2011;12:9–15 [DOI] [PubMed] [Google Scholar]
- 11.Machingaidze S, Wiysonge CS, Gonzalez-Angulo Y, Hatherill M, Moyo S, Hanekom W, Mahomed H. The utility of an interferon γ release assay for diagnosis of latent tuberculosis infection and disease in children: a systematic review and meta-analysis. Pediatr Infect Dis J 2011;30:694–700 [DOI] [PubMed] [Google Scholar]
- 12.Mandalakas A, Detjen A, Hesseling A, Benedetti A, Menzies D. Interferon-γ release assays and childhood tuberculosis: systematic review and meta-analysis. Int J Tuberc Lung Dis 2011;15:1018–1032 [DOI] [PubMed] [Google Scholar]
- 13.Pai M, Zwerling A, Menzies D. Systematic review: T-cell–based assays for the diagnosis of latent tuberculosis infection: an update. Ann Intern Med 2008;149:177–184 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Connell TG, Curtis N, Ranganathan SC, Buttery JP. Performance of a whole blood interferon γ assay for detecting latent infection with Mycobacterium tuberculosis in children. Thorax 2006;61:616–620 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Kampmann B, Tena-Coki G, Anderson S. Blood tests for diagnosis of tuberculosis. Lancet 2006;368:282. [DOI] [PubMed] [Google Scholar]
- 16.Moyo S, Isaacs F, Gelderbloem S, Verver S, Hawkridge AJ, Hatherill M, Tameris M, Geldenhuys H, Workman L, Pai M, et al. Tuberculin skin test and QuantiFERON assay in young children investigated for tuberculosis in South Africa. Int J Tuberc Lung Dis 2011;15:1176–1181 [DOI] [PubMed] [Google Scholar]
- 17.Bergamini BM, Losi M, Vaienti F, D’Amico R, Meccugni B, Meacci M, De Giovanni D, Rumpianesi F, Fabbri LM, Balli F, et al. Performance of commercial blood tests for the diagnosis of latent tuberculosis infection in children and adolescents. Pediatrics 2009;123:e419–e424 [DOI] [PubMed] [Google Scholar]
- 18.Grare M, Derelle J, Dailloux M, Laurain C. QuantiFERON-TB Gold In-Tube as help for the diagnosis of tuberculosis in a French pediatric hospital. Diagn Microbiol Infect Dis 2010;66:366–372 [DOI] [PubMed] [Google Scholar]
- 19.Haustein T, Ridout DA, Hartley JC, Thaker U, Shingadia D, Klein NJ, Novelli V, Dixon GLJ. The likelihood of an indeterminate test result from a whole-blood interferon-γ release assay for the diagnosis of Mycobacterium tuberculosis infection in children correlates with age and immune status. Pediatr Infect Dis J 2009;28:669–673 [DOI] [PubMed] [Google Scholar]
- 20.Kaufmann SHE, Hussey G, Lambert P-H. New vaccines for tuberculosis. Lancet 2010;375:2110–2119 [DOI] [PubMed] [Google Scholar]
- 21.Trunz BB, Fine P, Dye C. Effect of BCG vaccination on childhood tuberculous meningitis and miliary tuberculosis worldwide: a meta-analysis and assessment of cost-effectiveness. Lancet 2006;367:1173–1180 [DOI] [PubMed] [Google Scholar]
- 22.Soysal A, Millington K, Bakir M, Dosanjh D, Aslan Y, Deeks J, Efe S, Staveley I, Ewer K, Lalvani A. Effect of BCG vaccination on risk of Mycobacterium tuberculosis infection in children with household tuberculosis contact: a prospective community-based study. Lancet 2005;366:1443–1451 [DOI] [PubMed] [Google Scholar]
- 23.Hill PC, Brookes RH, Adetifa IMO, Fox A, Jackson-Sillah D, Lugos MD, Donkor SA, Marshall RJ, Howie SRC, Corrah T, et al. Comparison of enzyme-linked immunospot assay and tuberculin skin test in healthy children exposed to Mycobacterium tuberculosis. Pediatrics 2006;117:1542–1548 [DOI] [PubMed] [Google Scholar]
- 24.Cellestis.com. QuantiFERON-TB Gold: the whole blood IFN-γ test measuring responses to ESAT-6, CFP-10 & TB7.7(P.4) peptide antigens [package insert] [accessed May 1, 2012]. Available from: http://www.cellestis.com/irm/content/pdf/39111_pi_05990301g_aus_july%202011_ttf.Pdf
- 25.Oxford Immunotec. T-SPOT.TB: An aid in the diagnosis of tuberculosis infection [package insert] [accessed May 1, 2012]. Available from: http://www.oxfordimmunotec.com/8-uk
- 26.Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159. [PubMed] [Google Scholar]
- 27.Eisenhut M, Paranjothy S, Abubakar I, Bracebridge S, Lilley M, Mulla R, Lack K, Chalkley D, McEvoy M. BCG vaccination reduces risk of infection with Mycobacterium tuberculosis as detected by γ interferon release assay. Vaccine 2009;27:6116–6120 [DOI] [PubMed] [Google Scholar]
- 28.Eisenhut M. Significance of reduced reactivity in ELISPOT assay to RD antigens of Mycobacterium tuberculosis in BCG vaccinated individuals exposed to a patient with tuberculosis. Virulence 2010;1:105–106 [DOI] [PubMed] [Google Scholar]
- 29.Whalen CC, Zalwango S, Chiunda A, Malone L, Eisenach K, Joloba M, Boom WH, Mugerwa R. Secondary attack rate of tuberculosis in urban households in Kampala, Uganda. PLoS ONE 2011;6:e16137. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Lewinsohn D, Zalwango S, Stein C, Mayanja-Kizza H, Okwera A, Boom W, Mugerwa R, Whalen C. Whole blood interferon-γ responses to Mycobacterium tuberculosis antigens in young household contacts of persons with tuberculosis in Uganda. PLoS ONE 2008;3:e3407. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Lucas M, Nicol P, McKinnon E, Whidborne R, Lucas A, Thambiran A, Burgner D, Waring J, French M. A prospective large-scale study of methods for the detection of latent Mycobacterium tuberculosis infection in refugee children. Thorax 2010;65:442–448 [DOI] [PubMed] [Google Scholar]
- 32.National Institute for Health and Clinical Excellence CG117 tuberculosis: clinical diagnosis and management of tuberculosis, and measures for its prevention and control. London: National Institute for Health and Clinical Excellence; 2011 [PubMed] [Google Scholar]
- 33.Committee on Infectious Diseases, American Academy of Pediatrics The Red Book—report of the Committee on Infectious Diseases. Elk Grove Village, IL: American Academy of Pediatrics; 2009 [Google Scholar]
- 34.Pavic I, Topic RZ, Raos M, Aberle N, Dodig S. Interferon-γ release assay for the diagnosis of latent tuberculosis in children younger than 5 years of age. Pediatr Infect Dis J 2011;30:866–870 [DOI] [PubMed] [Google Scholar]
- 35.Debord C, De Lauzanne A, Gourgouillon N, Guerin-El Khourouj V, Pedron B, Gaudelus J, Faye A, Sterkers G. Interferon-γ release assay performance for diagnosing tuberculosis disease in 0- to 5-year-old children. Pediatr Infect Dis J 2011;30:995–997 [DOI] [PubMed] [Google Scholar]
- 36.Mahan CS, Johnson DF, Curley C, van der Kuyp F. Concordance of a positive tuberculin skin test and an interferon γ release assay in bacille Calmette-Guérin vaccinated persons. Int J Tuberc Lung Dis 2011;15:174–178 [PubMed] [Google Scholar]
- 37.Neyrolles O, Quintana-Murci L. Sexual inequality in tuberculosis. PLoS Med 2009;6:e1000199. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Casimir G, Heldenbergh F, Hanssens L, Mulier S, Heinrichs C, Lefevre N, Désir J, Corazza F, Duchateau J. Gender differences and inflammation: an in vitro model of blood cells stimulation in prepubescent children. J Inflamm 2010;7:1–7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Watson R, Carcillo J, Linde-Zwirble W, Clermont G, Lidicker J, Angus D. The epidemiology of severe sepsis in children in the United States. Am J Respir Crit Care Med 2003;167:695–701 [DOI] [PubMed] [Google Scholar]
- 40.Agergaard J, Nante E, Poulstrup G, Nielsen J, Flanagan KL, Østergaard L, Benn CS, Aaby P. Diphtheria–tetanus–pertussis vaccine administered simultaneously with measles vaccine is associated with increased morbidity and poor growth in girls: a randomised trial from Guinea-Bissau. Vaccine 2011;29:487–500 [DOI] [PubMed] [Google Scholar]
- 41.van Zyl-Smit RN, Zwerling A, Dheda K, Pai M. Within-subject variability of interferon-γ assay results for tuberculosis and boosting effect of tuberculin skin testing: a systematic review. PLoS One 2009;4:e8517. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.