Article Text

Download PDFPDF
BCG protects toddlers during a tuberculosis outbreak
  1. J T Gaensbauer,
  2. M Vandaleur,
  3. M O’Neil,
  4. A Altaf,
  5. M Ní Chróinín
  1. Division of Paediatrics, Cork University Hospital, Wilton Road, Wilton, Cork, Ireland
  1. James T Gaensbauer, Cork University Hospital, Division of Paediatrics, Wilton Road, Wilton, Cork, Ireland; jgaens{at}


In 2007, an outbreak of tuberculosis occurred in a toddler population attending two child care centres in Cork, Ireland. Of 268 children exposed, 18 were eventually diagnosed with active tuberculosis. We present the initial clinical and radiographic characteristics of the active disease group. Mantoux testing was positive in only 66% of cases. All cases were either pulmonary or involved hilar adenopathy on chest radiograph; there were no cases of disseminated disease or meningitis. 24% of the exposed children had been previously vaccinated with BCG, and no case of active disease was found in this group (p = 0.016), suggesting a profound protective effect of BCG in this population. Our experience provides evidence supporting a protective effect of BCG against pulmonary disease in young children.

Statistics from

Request Permissions

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.

In the past decade, Ireland has seen significant regional variation in the use of BCG vaccination.1 2 In recent years BCG was not routinely administered to infants in Cork but given on parental request, resulting in variable vaccine coverage throughout the region. In March 2007 an outbreak of tuberculosis (TB) occurred in a toddler population attending two child care centres in Cork. Children attending the centres were exposed over a period of 4 months to two child care workers with smear-positive pulmonary TB, one of whom had cavitating disease on chest radiograph. The two source cases were both Irish born and without a history of travel to TB endemic countries. Cultures from each revealed fully sensitive Mycobacterium tuberculosis. In addition to providing valuable information about the clinical presentation of tuberculosis disease in an otherwise healthy pre-school age population, this outbreak has provided a unique opportunity to examine the protective efficacy of BCG, since a known proportion of the exposed children had received BCG at birth.


Cork University Hospital (CUH) is the primary paediatric referral centre for County Cork and contains the only paediatric pulmonology service in Ireland’s southern region. Approval to utilise clinical and demographic patient data was obtained from the hospital’s ethics committee.

Basic demographic data and the medical histories of the entire exposed population were gathered at initial interview performed by the local department of public health. All exposed children underwent a detailed evaluation process which included serial Mantoux testing, clinical review and chest radiography. Patients with abnormalities in any of these primary areas were referred for further review at CUH.

As the yield from microbiological testing in children is low, we established a clinical case definition for the diagnosis of active tuberculous disease in our population based on a triad of (i) known exposure to TB, (ii) abnormalities on chest radiograph or physical examination consistent with TB, and (iii) a Mantoux test measuring >5 mm. All of the children had a known high-risk exposure. Diagnostic clinical abnormalities on physical examination considered to be consistent with TB included chest findings, lymphadenopathy and erythema nodosum. Hilar adenopathy, lobar consolidation, miliary opacification or pleural effusion were considered typical radiographic signs of pulmonary TB disease. All x rays were interpreted by a single experienced radiologist who was informed of clinical findings but was unaware of BCG status. A positive Mantoux test was defined as >5 mm induration, based on American Thoracic Society guidelines.3 In the context of a child with an abnormal radiograph highly suggestive of TB, a positive Mantoux test was not required to meet our case definition as anergy is a well-described cause of a false-negative result in young children with active disease. A subset of patients also underwent early morning gastric aspiration for mycobacterial culture. A positive result on any sample was considered unequivocally diagnostic.

For patients eventually diagnosed with TB, we obtained further detailed demographic and clinical data, including vaccine status, past history, and presenting symptoms and signs. Results of Mantoux testing and chest radiography were prospectively collected and the entire dataset was entered into an Excel (Microsoft, Redmond, USA) database.


During the outbreak period, a total of 268 children from the two crèches were exposed to TB. There were 20 cases of latent TB infection based on positive Mantoux testing in the absence of clinical signs or symptoms of disease. Eighteen children (median age 2.8; range 2.25–4.5 years) were diagnosed with active tuberculosis disease. No children had disseminated or meningeal TB; all cases were pulmonary or had hilar adenopathy on chest radiograph. Mantoux testing was negative in 6/18 (33%) children diagnosed with active TB. Presenting symptoms and signs among cases are shown in table 1. Chest radiographs were abnormal in all 18 cases and the spectrum of x ray findings is presented in table 2. Gastric washings were performed in 11 children and five of these were culture positive for M tuberculosis.

Table 1 Frequency of symptoms and clinical signs among 18 TB cases
Table 2 Frequency of radiographic findings at diagnosis among 18 TB cases

All children were treated with a standard three-drug protocol (isoniazid/rifampicin/pyrazinamide for 2 months, followed by isoniazid and rifampicin for 4 months) according to the antibiotic sensitivities of the index case. Directly observed therapy was not deemed necessary. Only one of the 18 developed complications (bronchiectasis) requiring further investigation and prolonged treatment. The remaining cases were well at completion of therapy. Therapy itself was well tolerated. There were no cases of significant liver or other toxicity, although a number of parents described increased hyperactivity in their child during the treatment course. No further cases of TB were detected from the original exposed cohort 1 year following the outbreak.

The rate of BCG immunisation in the exposed population was 24%. All cases of active infection occurred in patients who had not been vaccinated (18/204 vs 0/64, p = 0.0162; Fisher’s exact test). As there were zero cases in the vaccinated group, calculation of relative risk is not mathematically possible, but based on an absolute risk reduction, the number needed to treat to prevent one case of active TB disease in our population is 11.3.


The spectrum of clinical findings among our patient population provides valuable information regarding the approach to tuberculosis outbreaks in similar settings. Notably, Mantoux testing as an exclusive screening tool would have delayed diagnosis in up to one third of cases who were anergic despite a diagnosis of active TB. Utilisation of clinical and radiographic screening in all exposed children in addition to Mantoux testing may in part explain the absence of more invasive or advanced forms of tuberculosis disease in this outbreak as these may have been prevented by early diagnosis. Our resources allowed us to perform clinical review, chest x ray and Mantoux testing for each exposed child. In resource-limited settings, symptom-based screening has been proposed, and in our population 17 of 18 actively infected patients would have been identified based on their symptoms alone.4 The lack of severe forms of TB may also be in part attributed to a highly organised and well-coordinated management plan involving public health, paediatric and radiological services.

Our findings also suggest a pronounced protective effect of BCG against active TB disease in this setting. The relative effectiveness of BCG in the prevention of severe childhood tuberculosis disease, in particular the miliary and meningeal forms, has been reliably established.5 Studies examining protection against all forms or less invasive forms of TB disease have been much less consistent, with efficacy rates ranging from 0% to 80%. These studies have relied primarily on data obtained from large populations in settings of high TB prevalence. The outbreak in Cork city resulted from a prolonged high-risk exposure in a clearly defined population consisting of a group of children who were previously healthy, at no socioeconomic disadvantage, of a defined high-risk age range and for whom BCG status was readily available. Furthermore, the vaccinated group and the unvaccinated group were equivalent in terms of their exposure and risk of TB. Therefore, the statistical limitations of relatively small numbers are counterbalanced by a lack of confounding factors. While the number needed to treat to prevent one case of active TB disease of 11.3 will not be directly applicable outside of a high-risk smear-positive exposure of children in the toddler age group, it still provides a measure of BCG effectiveness in what could be considered an ideal test situation.

The International Union against Tuberculosis and Lung Disease has suggested criteria for cessation of BCG immunisation programs in more developed countries which have lower rates of TB infection and adequate health surveillance systems to identify and control isolated outbreaks.6 We believe that the data from our recent outbreak suggest that any such transition should be undertaken with caution. Two very similar outbreaks involving large numbers of vulnerable children occurred in Western Europe prior to that reported here7 8 and other outbreaks are likely in the future, particularly as migration from TB endemic areas of Europe and the developing world is increasingly facilitated. Additionally, the protective effect of BCG demonstrated here adds to the evidence supporting widespread BCG vaccination in developing countries with high TB prevalence.



  • Competing interests: None.

  • Ethics approval: Approval was obtained from the hospital’s ethics committee.

Linked Articles

  • Atoms
    Howard Bauchner