Statistics from Altmetric.com
You are asked to review a 7-month-old infant born in Australia who was given BCG the previous day in preparation for travel to India. Her parents have noticed a 6 mm diameter swelling at the injection site. You are aware that an accelerated response at a BCG injection site may be associated with underlying tuberculosis (TB) infection. The infant is otherwise completely well with a normal examination. You wonder whether you should investigate further.
Structured clinical question
Is an accelerated local reaction following BCG vaccination in a child indicative of underlying TB (latent TB infection (LTBI) or active TB disease)?
Search strategy and outcome
Medline (1946 to current), PubMed and Embase databases were searched in September 2021 with the search terms detailed in the online supplemental search strategy. Our systematic search identified 139 articles. Of these, 14 fulfilled the inclusion criteria of English-language publications addressing accelerated local reaction to intradermal BCG for diagnosing TB in children under 18 years of age. All the studies reported sensitivity but only seven included controls enabling specificity to be calculated.
BCG is one of the most extensively used vaccines worldwide, given to children to protect against TB since 1921. It has received more recent attention for its beneficial off-target effects on the immune system, associated with protection against other diseases.1 The normal reaction to intradermal BCG vaccination is the appearance of a small, red papule or swelling at the injection site within 2–3 weeks. Usually, the papule softens, resulting in a small ulcer, healing over several weeks to months into a small flat scar.2 3 In some individuals, an ‘accelerated local reaction’ occurs, commonly defined as the early development of an induration or swelling of 5 mm or more at the injection site, beginning within 24–72 hours of vaccination.4–6 An accelerated BCG reaction has also been called a ‘positive BCG test’ and was first described as a potential diagnostic tool for TB in 1949.7 The criteria for a positive BCG test vary and are sometimes subclassified into ‘hyperaccelerated’ reactions occurring within 6–12 hours,8 and ‘delayed’ occurring between 72 hours and a week. Regardless, the accelerated onset is commonly followed by accelerated healing, characterised by pustule formation within 5–7 days and healing within 10–15 days.2
The 14 studies (11 cross-sectional and 3 case–control) involved 2617 children (table 1). The sensitivity of accelerated BCG reactions for the diagnosis of active TB disease (confirmed or probable) was reported in all 14 studies, ranging from 71% to 100%. This was higher than the tuberculin skin test (TST) sensitivity (range 19%–67%), although a lower strength of purified protein derivative (1 tuberculin unit (TU) rather than the current standard strength of 5 TU) was used in most of these studies.9 Seven of these studies also reported specificity, and these were included in a meta-analysis, in which the combined sensitivity of an accelerated BCG reaction was 87% (95% CI: 81% to 91%) and specificity 90% (95% CI: 85% to 94%) (figure 1 and online supplemental figure 1).
As BCG is live-attenuated Mycobacterium bovis, an accelerated BCG reaction in an individual with TB has been compared with a phenomenon described by Koch in 1891 (online supplemental figure 2).3 10 In studies in adults, an accelerated BCG reaction was also more sensitive than TST in diagnosing TB.11 However, its suboptimal specificity led to it being used only as a rule-out test in sputum-negative cases of suspected pulmonary TB.12 The use of the ‘BCG test’ as a diagnostic tool for TB has also been questioned in children, as accelerated reactions have also been shown to occur with non-tuberculous mycobacterial infections in this age group,13 and false positives can occur with repeated use.14
In asymptomatic children, although an accelerated reaction might be indicative of underlying TB infection, it is also influenced by previous BCG vaccination. Among the six studies5 14–18 reporting accelerated BCG reactions in asymptomatic children, five included BCG-naïve children14–18 and four included BCG-revaccinated children.5 14 15 17
In BCG-naïve children (>1 month of age), between 0% (0 of 20) and 18% (9 of 50) developed an accelerated BCG reaction. Of those who developed an accelerated reaction and had TST administered concurrently, between 24% and 100% had evidence of TB infection, such as a positive TST ≥10 mm14 17 or active TB disease.18
In contrast, in BCG-revaccinated children, between 10% (10 of 100) and 78% (39 of 50) developed an accelerated reaction. In three studies,14 15 17 accelerated reactions were more common in BCG-revaccinated compared with BCG-naïve asymptomatic children (29% vs 17%, p=0.0514; 10% vs 6%, p=0.317; 78% vs 18%, p<0.0000115). In the latter study, all children were TST negative at baseline, suggesting an alternative reason for their accelerated reaction, as well as the influence of prior BCG.15 Furthermore, in a large study of over 1000 children aged less than 5 years who all had prior neonatal BCG, younger infants were more likely to develop an accelerated reaction compared with older children.5 It is unclear whether younger age or a shorter time interval between prior BCG and revaccination increased the likelihood of developing an accelerated reaction.
Limitations of studies in this review include the paucity of studies describing the proportion of TST-positive results in asymptomatic children with accelerated compared with non-accelerated reactions to BCG. This would be helpful to investigate the potential association between accelerated BCG reaction and LTBI. However, the diagnosis of LTBI in children is challenging, due to the absence of a recognised gold standard.19–24
Other limitations include spectrum bias through case–control design, and potential interpretation bias due to lack of blinding of investigators measuring BCG and TST reactions. Another limitation is that most children with TB in this review had ‘probable TB’ as microbiological confirmation (the reference standard for active TB disease) is seldom achieved in children due to the paucibacillary nature of their disease.25 Moreover, the majority of included studies were done prior to the year 2000 (when TB diagnostic armamentarium was even more limited than currently).
Several studies did not specify BCG strain, dose or vaccination history, all of which can influence the local reaction.26–28 Lastly, all of the studies were done in countries with high TB prevalence meaning it is uncertain how this applies to other settings.29 A lower disease prevalence would decrease the positive predictive value and increase the negative predictive value of an accelerated BCG reaction.
However, as the available evidence suggests that accelerated BCG reactions might reflect current or prior mycobacterial exposure, to err on the side of caution, thorough further evaluation of children with accelerated reactions seems warranted, as early detection and treatment of LTBI to prevent progression to TB, particularly in young children, is important.
In conclusion, an accelerated BCG reaction should prompt investigation for potential LTBI or active TB disease. The index of suspicion should be higher in children with no BCG history and those living in high-TB prevalence settings. Further research reporting outcomes of children with accelerated reactions in low-TB prevalence settings would be beneficial.
Clinical bottom line
An accelerated BCG reaction in symptomatic children has both a high sensitivity and specificity for tuberculosis (TB) infection in high-TB prevalence setting and warrants investigation for latent TB infection (LTBI) or active TB disease. (GRADE B)
An accelerated BCG reaction might also reflect previous mycobacterial exposure, including non-tuberculous mycobacteria or BCG (as it is more common in BCG-revaccinated than BCG-naïve children). (GRADE B)
As no studies have addressed accelerated BCG reactions in low-TB prevalence countries, the safest approach for children in such settings is to investigate for LTBI or active TB disease, particularly if there are any risk factors (such as travel to an endemic area or parents from an endemic area).
Patient consent for publication
This study does not involve human participants.
Twitter @Paola_Villa14, @PittetLaure, @nigeltwitt
Contributors PV drafted the initial manuscript. LFP, CN and NC critically revised the manuscript and all authors approved the final version as submitted.
Funding PV is supported by the Australian Government Research Training Programme Scholarship provided by the Australian Commonwealth Government and the University of Melbourne, and a Murdoch Children's Research Institute (MCRI) PhD Top-Up Scholarship.
Competing interests None declared.
Provenance and peer review Not commissioned; internally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.