Comparison of Tuberculin Skin Test and QuantiFERON TB Gold In-tube Assay for the diagnosis of tuberculosis infection and disease in young children; study conducted as part of the development of a TB vaccine site in Southern India
MetadataShow full item record
Background: An estimated 35% of tuberculosis (TB) cases worldwide are accounted by India and China alone. Pediatric TB accounts for 11% of TB cases worldwide. The most common type of TB in children is smear negative pulmonary TB. Data from the Revised National Tuberculosis Control Programme (RNTCP) in India indicate that in 2009, the percentage of new smear positive cases was 72% of the total case detection rate of 132/100,000. Of all the new cases detected, 7% were pediatric cases. Forty per cent of the population in India is thought to be infected with M. tuberculosis and the population has an estimated annual risk of infection of 1.5%. Due to difficulty in obtaining a sputum sample in children, M. tuberculosis very often goes undetected in smear and culture. In children <5 years, the tuberculin skin test (TST) is widely used as an aid in the diagnosis of TB infection. A decade ago, the QuantiFERON test (QFT), an Interferon Gamma Release Assay (IGRA) was approved by the Food and Drug administration (FDA) and European Medicines Agency (EMEA) and is now used as an aid in the diagnosis of TB infection. Although the TST is known for its limitations related to variations in dose, inter-observer errors related to administration and reading, potential lack of specificity related to non tuberculous mycobacteria (NTM) and Bacillus Calmette-Guerin (BCG) vaccination, it may have a higher sensitivity than the QFT. However, the available literature shows that QFT has a higher specificity and potentially higher predictive value than the TST. Diagnosis of childhood TB remains a challenge. Although, QFT has shown promising results in few studies, there is limited data on its diagnostic performance in children. One of the limitations of the QFT in TB-endemic areas is that it will not differentiate between latent and active TB. However, in children – especially young children – all TB infection can be considered to be of recent origin and given the shorter duration of progression from infection to disease in these children, they are therefore at ‘high risk’ of developing TB. The available studies suggest that QFT may therefore prove clinically useful in the evaluation of TB in children and feasible to perform. There is a need to perform more studies in children in India to evaluate different tests in relation to TB infection and disease with a larger sample size in order to overcome the diagnostic challenge that exists in the field of childhood TB. The overall aim of this study is to contribute to the diagnosis of TB infection and disease in young children in India, a high TB burden country. Methodology: The present study is nested within a prospective cohort study. BCG vaccinated infants <14 days of age and born to mothers from Palamaner taluk in Andhra Pradesh were enrolled. Informed consent was obtained and information such as age, weight of the infant, parents’ demographic characteristics, and socio economic data were collected from clinical records and by interview. Infants were randomly assigned to an active or passive follow up group and followed for 2 years. Infants suspected of having TB based on symptoms, failure to thrive (FTT) and a recent history of contact with TB were referred to the case verification ward (CVW) for diagnostic evaluation of TB. A TST and chest X-Ray was performed. One sample each of a gastric aspirate and induced sputum were collected on two consecutive days which were used for smear microscopy and culture. Mycobacterial cases were confirmed by speciation using Genotype MTBC and CM kit (HAIN) or COBAS TaqMan MTB test kit (Roche). Blood was collected, plasma isolated and QuantiFERON gold In tube assay (QFTGIT) performed. Results: There were 4,382 children enrolled in the study of which 2,215 were included in the active surveillance group and 2,167 in the passive surveillance group. Seven hundred and forty six participants were referred to the CVW of which 53 came for a second visit, three for a third visit and one for a fourth visit during the two year period of follow up. In 709 participants both TST and QFT-GIT were performed. The percentage of children who were TST positive was 10.2% (n=72/709) while 5.8% were QFT-GIT positive (n=41/709).There were a total of four cases with definite (bacteriologically confirmed) TB; three of which were positive for TST and QFT. The 4th case was negative on culture, TST and QFT-GIT but was positive on smear microscopy. A direct polymerase chain reaction (PCR) on the induced sputum sample was positive for M. tuberculosis. There were 9 cases with ‘probable’ TB, of which one was positive for TST but negative for QFT-GIT and the rest were negative for both. Children with a recent history of contact with a TB case had increased odds of being TST positive (OR 2.11[95% CI 1.07; 4.17], p=0.03), QFT-GIT positive (OR 3.46 [95% CI 1.50; 7.81], p=<0.01) and combined TST and QFT-GIT positive (OR 7.08[95% CI 2.51; 19.65], p=<0.01) in a univariate analysis. In a multivariate analysis, the association between TST positivity and recent history of contact with TB was not statistically significant (OR 1.58 [95% CI 0.69; 3.63], p=0.27). However, for QFT-GIT positivity the association with contact remained statistically significant in a multivariate analysis (OR 4.36 [95% CI 1.79; 10.54], p=0.01). Children with wasting based on the weight for height Z score (<-2SD) had reduced odds of being TST positive in a univariate analysis (OR 0.31 [95% CI 0.17; 0.56], p=<0.01) as also those with FTT (OR 0.19 [95% CI 0.12; 0.33], p= <0.01]. There was no association between QFT-GIT positivity and nutritional status. The agreement between TST (cut off ≥10mm) and QFT-GIT was fair [κ 0.31 (95% CI 0.19; 0.40), p=<0.01] and TST (cut off ≥5mm) and QFT-GIT was poor [κ 0.09 (95% CI 0.04; 0.11), p=<0.01] for all children with both TST and QFT-GIT tests performed and referred to the CVW for the first time (n=709). The percentage of indeterminate QFT-GIT results was 3.6%. Conclusion: The incidence of TST positivity was higher than that of QFT-GIT positivity. Due to a lack of a gold standard for the diagnosis of TB infection in children and the low number of bacteriologically confirmed cases, sensitivity and specificity of the tests in relation to TB infection and disease could not be calculated. A recent history of contact with TB was associated with QFT-GIT positivity, but not TST positivity in a multivariate analysis and this association was stronger when both TST and QFT-GIT positivity was considered together. Additional studies to confirm the utility of a two-step approach i.e. QFT-GIT test on those who are TST positive would be relevant. This approach would be strengthened if the likelihood of progression to TB disease was higher for those infants with combined TST / QFT-GIT positives rather than either test alone. TST and not QFT-GIT positivity was reduced in undernourished participants; this is significant given the high prevalence of under nutrition in children of this age group in India. The data suggest that QFT-GIT may be better for the diagnosis of a recent infection as compared to the TST in a setting where under nutrition is high. The overall agreement between TST and QFT-GIT was fair at a cut off of ≥10mm for TST. The number of indeterminate test results in the QFT-GIT was low.