Contact investigation (CI) and migrant screening are key components of tuberculosis (TB) control in low endemic settings. Conducted in a comprehensive and timely manner, they can identify infected persons before disease develops, or while it is asymptomatic. Of 134 children diagnosed with TB at our hospital from 2001 to 2015, 43 (32%) were identified at CI and 44 (33%) at migrant screening. Moreover, among 102 children with clinically presenting TB in 1971 to 2015, 44 might have been averted by preventive treatment of TB infection, had they been screened per present-day standards.1
Using a tuberculin skin test (TST), the rates of TB infection were 30% and of TB disease 1.4% among close contacts in high income settings.2 When interferon-gamma release assays (IGRAs) were available, studies showed them to be more specific than TSTs in Calmette–Guérin bacillus (BCG) vaccinated persons.3–5 Test positivity correlated better with the degree of exposure to the index case when IGRAs were used, and rates of progression to TB disease were higher in untreated persons with positive IGRAs (IGRA+) than in those with TST+.6
Initially, IGRAs were recommended primarily for the confirmation of positive TSTs in individuals with a low risk of infection.3 More recently, we and other pediatric TB clinics have used IGRAs as the main test for TB screening.7
In Stockholm, all CI involving patients <18 years of age is performed at 2 pediatric TB clinics.8 Exposed persons are required by law to participate in CI. The aim of this study was to compare the results using TSTs and IGRAs at the northern clinic from 2000 to August 2018.
MATERIALS AND METHODS
R.B. entered the information about referred children <18 years of age and their index cases into a Microsoft Access database in a prospective manner as previously described and made all diagnostic decisions.9 Missing data were retrieved by retrospective chart review.
Index cases with positive Mycobacterium tuberculosis culture of respiratory secretions in sputum, bronchoalveolar or gastric lavage fluid were considered infectious. We performed CI also for contacts to noninfectious index cases, under the presumption that they might have been exposed to other contagious persons. Some children were not tested because detailed interview made it clear that they were not sufficiently exposed.
The child’s closeness to the index case was grouped into “household contacts” and “other.” We included first-degree relatives, other persons living in the household, day-care personnel and children, and classmates among household contacts.
In children exposed at school or a day-care center, initial TSTs were sometimes performed locally and only those with a positive result or those needing a repeat test referred to the TB clinic. These cases were not included in calculations of test positivity rates.
TSTs were performed using Mantoux technique with purified protein derivative SSI RT23 2 TU, read after 72 hours by the TB clinic’s specialized nurse. In unvaccinated or immunosuppressed children, TSTs were considered positive if the induration was ≥6 mm, otherwise if ≥10 mm. From 2005, the IGRAs T-SPOT.TB (Oxford Immunotec, Oxfordshire, UK), at our clinic used mainly in children <3 years of age, and QuantiFERON-TB Gold (QFT; Qiagen Pty Ltd, VIC, Australia) gradually replaced the TST (Table, Supplemental Digital Content 1, http://links.lww.com/INF/D398). Blood for IGRAs was drawn at the pediatric laboratory at our hospital and reached the analyzing laboratory within 4 hours. In children with household exposure to infectious patients, we obtained chest radiograph at the first visit, otherwise only if they were symptomatic or tested positive.
When we used both tests in a child, we usually relied on the IGRA result for interpretation, after consideration of clinical factors such as age, intensity of exposure, TST reaction size and test results in other contacts of the index case.
If a negative test was obtained less than 2 months after the last exposure to an infectious index case, a repeat test was ordered.
In children with TB disease, we assumed it to be caused by the strain isolated from the index case. However, material for culture was obtained if the child had severe disease, if the index case had a multi-drug resistant M. tuberculosis strain or was noninfectious, or if there were other possible infectious sources, for example, during time spent in high-endemic countries.
We used the online facilities at http://www.quantitativeskills.com/sisa/statistics/twoby2.htm. and https://www.graphpad.com/quickcalcs/kappa1/, accessed on September 29, 2018, for statistical analysis of 2 × 2 tables. Age difference was analyzed with Statistica v. 13.2 (Dell Inc., Tulsa, OK).
If a patient later progressed to TB disease, we would be alerted of this through regular TB control conferences held at county level, or through follow-up at the national Swedish TB registry.
The ethics committee at the Karolinska University Hospital approved the study (2010/1686-31/1).
The characteristics of the 877 index cases are shown in Table, Supplemental Digital Content 2, http://links.lww.com/INF/D399.
We evaluated 2136 children, of which 2047 were tested with TST, IGRA or both (Table, Supplemental Digital Content 1, http://links.lww.com/INF/D398 and Figure, Supplemental Digital Content 3, http://links.lww.com/INF/D401). Of the 89 not tested, 47 were considered unexposed, 12 were previously treated for TB infection or disease and 30 were lost to follow-up.
A total of 216 children were selectively referred from institutional outbreaks after testing. Of these, 148 (69%) had a positive test result and 20 (9%) had TB disease.
The test results of the remaining 1831 children are shown in Table 1. The IGRA tested children were older than the TST tested, because initially we did not use IGRAs in young children. Of 309 household contacts to microscopy positive index cases, 102/194 (53%) were TST+ and 35/115 (30%) IGRA+ (P < 0.001).
In 254 cases, we had both TST and IGRA results, and BCG vaccination status was known in 201. The 2 tests were concordant in 30% of vaccinated and in 87% of unvaccinated children (P < 0.001; Table, Supplemental Digital Content 4, http://links.lww.com/INF/D400).
In 325 cases, we used repeat TSTs and in 144, repeat IGRAs. In total, 57 (18%) had TST conversion and 5 (3.5%) IGRA conversion (P < 0.001). Of 37 children with TST conversion and a subsequent IGRA test, 33 (89%) were IGRA−.
We used early prophylaxis with isoniazid (n = 72), rifampin (n = 18) or both (n = 6) in 96 young children with a median age of 1.7 years [interquartile range (IQR): 0.5–4.7] before a decision on infection status could be made. Eventually, we considered 10 of them infected.
In all, there were 658 children with at least 1 positive test. Of these, 425 were diagnosed with TB infection and 55 with TB disease. Seven had previously been treated for TB infection or disease. We considered the remaining 171 TST+ children uninfected, 123 because of TST+/IGRA− discordance and 48 because they were BCG vaccinated, and detailed history showed them to be unexposed.
Preventive therapy was completed by 358/425 (84%) children with TB infection.
All 55 children with TB disease were TST+ or IGRA+. Twenty belonged to day-care center or classroom outbreaks.10 The remaining 35 made up 6.1% of the TST+ and 13% of the IGRA+ children (P < 0.05) and had a distribution of age and sex similar to the entire material. There were 53 cases with intrathoracic disease whereof 28 pulmonary, 23 with lymphadenitis and 2 with pleuritis. Two with extrathoracic disease, comprising 1 with lymphadenitis and 1 with multisite disease. Material for culture was obtained from 25 children, with growth of M. tuberculosis in 10. Five strains were isoniazid mono-resistant and 2 were multi-drug resistant. In 8 cases where strains from both index case and contact were available, they were identical by molecular genotyping. We successfully treated 54 children and 1 was lost to follow-up.
This study shows lower rates of positive test results in contacts using IGRAs compared with TSTs. When both tests were used in BCG-vaccinated children, agreement was poor, also in those with TST conversion. We believe the differences between the tests to be because of a low specificity of TSTs in BCG vaccinated children, which is supported by the lower rate of active TB in TST+ children.
In Toronto children, 34% (31/90) of household contacts and 36% (26/72) of contacts to microscopy positive index cases were QFT+, higher than our rates of 18% and 19%, respectively.5 The Toronto material was, however, a convenience sample, and the higher rate of TB disease (5%) among their children compared with 1.6% in ours is compatible with a higher rate of infection. When we counted only the most heavily exposed children, we also had an IGRA+ rate of 30%.
Some experts warn against relying on IGRA results in children with positive TSTs. Although such a routine is not suitable everywhere, our experience from a low burden setting is reassuring, also when it comes to later progression to TB (manuscript submitted). Still, the sensitivity of the IGRAs remains uncertain, as illustrated by the 12% rate of TST+/IGRA− discordance in our unvaccinated subjects (Table, Supplemental Digital Content 4, http://links.lww.com/INF/D400). Therefore, when the risk of infection or severe disease is high in an unvaccinated child, we also recommend the confirmation of a negative IGRA with a TST. In vaccinated children, the use of TSTs is compromised by their low specificity. Clinical judgement remains essential, with attention to known risk factors.
1. Bennet R, Eriksson M. Paediatric tuberculosis
cases increased in Stockholm from 1971 to 2015 following the rising number of children
with immigrant backgrounds. Acta Paediatr. 2016;105:1480–1486.
2. Fox GJ, Barry SE, Britton WJ, et al. Contact investigation
: a systematic review and meta-analysis. Eur Respir J. 2013;41:140–156.
3. Kang YA, Lee HW, Yoon HI, et al. Discrepancy between the tuberculin skin test
and the whole-blood interferon gamma assay for the diagnosis of latent tuberculosis
infection in an intermediate tuberculosis
-burden country. JAMA. 2005;293:2756–2761.
4. Zellweger JP, Sotgiu G, Block M, et al; TBNET. Risk Assessment of Tuberculosis
in Contacts by IFN-γ Release Assays. A Tuberculosis
Network European Trials Group Study. Am J Respir Crit Care Med. 2015;191:1176–1184.
5. Rose W, Read SE, Bitnun A, et al. Relating Tuberculosis
(TB) Contact Characteristics to QuantiFERON-TB-Gold and Tuberculin Skin Test
Results in the Toronto Pediatric TB Clinic. J Pediatric Infect Dis Soc. 2015;4:96–103.
6. Diel R, Goletti D, Ferrara G, et al. Interferon-γ release assays for the diagnosis of latent Mycobacterium tuberculosis
infection: a systematic review and meta-analysis. Eur Respir J. 2011;37:88–99.
7. Nejat S, Bennet R. Interferon-gamma release assays can effectively screen migrants for the tuberculosis
infection, but urgent, active cases need clinical recognition. Acta Paediatr. 2016;105:671–675.
8. Nejat S, Buxbaum C, Eriksson M, et al. Pediatric tuberculosis
in Stockholm: a mirror to the world. Pediatr Infect Dis J. 2012;31:224–227.
9. Bennet R, Eriksson M. Tuberculosis
infection and disease in the 2015 cohort of unaccompanied minors seeking asylum in Northern Stockholm, Sweden. Infect Dis (Lond). 2017;49:501–506.
10. Gillman A, Berggren I, Bergström SE, et al. Primary tuberculosis
infection in 35 children
at a Swedish day care center. Pediatr Infect Dis J. 2008;27:1078–1082.