Abstract
The risk of TB reactivation among infected children increases as they reach the age of adolescence. BCG vaccination history seems to increase the risk of TB disease reactivation among adults exposed to Mycobacterium tuberculosis. http://bit.ly/2YReXej
To the Editor:
The risk of a person progressing to tuberculosis (TB) disease after infection with Mycobacterium tuberculosis remains poorly understood, with some contacts developing TB in the early period following exposure, while others take many years to progress or never do so [1, 2]. We described profiles and patterns of contacts' progression to TB disease following exposure by linking a large, prospectively collected contact investigation dataset from Victoria, Australia to data on subsequent cases of active TB disease, after obtaining ethical approval from the Monash University, Human Research Ethics Committee. Unlike many past studies, this approach offers the opportunity to disaggregate by various characteristics of both index patient and exposed contact.
The main outcomes of interest were early progression (development of active TB within 6 months of exposure) and late reactivation (development of TB thereafter). Our definition for early progression is supported by previous findings that the risk of TB is highest in the first 3–9 months after infection [3, 4]. We used a survival analysis approach to describe patterns of TB disease reactivation and used logistic and Cox regression to quantify risks associated with early and late reactivation respectively, incorporating variables pertaining to both contact and index patients.
17 740 contacts of pulmonary TB patients were included in the analysis, of which 82.0% were aged 15 years and above. Over more than 7 years of median follow-up, 224 (1.3%) contacts were diagnosed with TB disease, resulting in an overall incidence rate of 189 cases per 100 000 person-years. Most cases (62.5%) accrued during the first 6 months following exposure and so were defined as early progression episodes. We found that male contacts had a slightly greater rate of TB than their female counterparts (log-rank p=0.03), while children aged <5 years at exposure had the greatest rate of active TB development, followed by children aged 5–14 years (log-rank p<0.001). Tuberculin skin test (TST) positivity (≥10 mm) was associated with a greater rate of TB (4.6%, 148/3213) than for TST-negative contacts (0.2%, 29/12 233), despite TST positives being more likely to receive preventive therapy. Younger age of contact, close contact, high-burden country of birth, absence of BCG vaccination and recurrent TB in the index patient were associated with early progression, while close contact, high-burden country of birth, BCG vaccination history and BCG status not stated were associated with contacts' higher hazard of late reactivation (table 1).
Every TB case observed among children aged <5 years occurred in the first 18 months of follow-up, of which 44/51 (86.3%) were early progressions. Although a similar proportion (49/60, 81.6%) of early progressions was found among child TB contacts aged 5–14 years, there were seven TB activation episodes that occurred after more than 2 years of follow-up in this age group. These episodes predominantly occurred as these children aged into adolescence and reached approximately 15 years of age. This pattern has also been observed in historical works, which have reported that when a first exposure occurred in children (aged 1–14 years), TB typically did not develop until after the age of 15 years [3]. A prospective study from Hong Kong provided similar evidence that infected children were at greater risk of TB beyond the age of 15 years after an initial low-risk window period [4]. The predilection of TB to affect young adults has been recognised since Hippocrates, although there are multiple possible explanations for the considerable burden of TB in adolescents, particularly assortative mixing by age leading into the years in which active TB is most infectious [5–7]. Given our very low transmission setting and the fact that all of these children were born in low-burden settings, our findings provide considerable evidence for the explanation that infected children enter a higher risk phase as they progress towards adulthood, which has major implications for clinical care and public health responses. The absence of any cases occurring in adolescence among those aged 0–4 years at the time of exposure likely reflects insufficient follow-up duration, as only the oldest such contacts entering during the earliest period of the study would have reached adolescence, and could also reflect more consistent use of preventive therapy in this group.
Consistent with previous studies, we found that BCG vaccination was highly protective against early progression in contacts aged 0–14 years at exposure, which is a similar effect size to those reported by a previous meta-analysis [8]. However, unexpectedly, risk of late reactivation was greater among BCG vaccinated and BCG not stated adult contacts than BCG unvaccinated counterparts. Previous studies have consistently reported that BCG is effective in preventing TB disease among children and that its effectiveness is reduced when all ages are considered [9–11]. We note similar findings from the largest ever trial estimating the efficacy of BCG, which reported nonsignificantly higher rates of TB among every age group of vaccinated adults than in the unvaccinated over 15 years of follow-up [12]. One possible interpretation of this evidence is that BCG is effective in reducing TB-related child morbidity and mortality, but may increase late reactivation rates in adults. These phenomena may lead to perverse effects on the overall epidemic because the adult forms of TB are typically most infectious, and they are critical when considering BCG revaccination programmes and interpreting recent trials of novel vaccines whose outcomes consider immunological responses to M. tuberculosis rather than incidence of TB disease [13]. Although we found no evidence of an association between BCG vaccination and absence of preventive therapy, this or other unmeasured factors could confound these findings. Furthermore, adults with unknown BCG status were also associated with higher rates of late reactivation, such that ensuring comprehensive collection of BCG status will be a future focus of our research.
Although the development of active TB in contacts after infection is largely dependent on contacts' endogenous factors, we also found associations with index patients' characteristics. Specifically, index patients with recurrent TB appeared more infectious, with their contacts at a considerably greater odds of early progression than for new TB cases. This may be explained by patients with recurrent TB having longer infectious periods or more frequently being smear-positive TB than those with a first episode, although this effect was not seen for late reactivation.
Important limitations of our study include that transmission was inferred based on epidemiological links due to the absence of genotyping data and we did not have universal information on the provision of preventive therapy that could have significantly influenced our findings. However, because preventive therapy is recommended universally for all contacts of patients with active TB in our setting, we would generally expect this to reduce the absolute TB risk across the cohort, but be less likely to systematically bias the trends reported in this manuscript.
In conclusion, child contacts were at highest risk of early progression, but high rates of late reactivation among infected children as they reached adolescence imply that children should be strongly considered for preventive interventions even after their initial high-risk period has elapsed. BCG seems to have an important effect on TB epidemiology in preventing early activation but increasing rates of late reactivation in adults. This effect requires further investigation with more comprehensive data on contacts' receipt of preventive therapy.
Shareable PDF
Supplementary Material
This one-page PDF can be shared freely online.
Shareable PDF ERJ-00353-2019.Shareable
Footnotes
Conflict of interest: Y.A. Melsew has nothing to disclose.
Conflict of interest: A.C. Cheng has nothing to disclose.
Conflict of interest: E.S. McBryde has nothing to disclose.
Conflict of interest: J.T. Denholm has nothing to disclose.
Conflict of interest: E. Tay has nothing to disclose.
Conflict of interest: R. Ragonnet has nothing to disclose.
Conflict of interest: J.M. Trauer has nothing to disclose.
- Received November 29, 2018.
- Accepted May 25, 2019.
- Copyright ©ERS 2019