In the 1980s, the emergence of HIV/AIDS led to a dramatic increase in the incidence of tuberculosis, primarily in countries with fewer resources, initially sub-Saharan Africa, followed several years later by south-east Asia. In 1993, the World Health Organization (WHO) declared tuberculosis to be a global health emergency.1

Epidemics of multi-drug resistant tuberculosis (MDR-TB), defined as Mycobacterium tuberculosis infection resistant to isoniazid and rifampicin, were first described in the United States and Europe at the beginning of the 90s. MDR-TB occurs mainly in HIV-infected patients, the transmission pattern is explosive, attack rates are high and incubation periods are short due to the immunosuppressed status of the host and the lack of adequate preventive measures against tuberculosis transmission in hospitals and penitentiary institutions affected by these outbreaks. Healthcare personnel are not exempt from risk.2–4

In 1994, the WHO, together with the International Union against Tuberculosis and Lung Disease, set up their Surveillance of Drug Resistance in Tuberculosis program. Until then, drug resistance had been a minor problem, usually related with poor treatment adherence, and primary multi-drug resistance was very rare.

Poverty, overpopulation, lack of economic resources for obtaining second-line drugs, the absence of effective programs for the management of tuberculosis, the HIV epidemic, and moving populations; all are reasons for the appearance and international dissemination of multi-resistant strains of M. tuberculosis. This has become a worldwide health problem, but one that is unequally distributed – countries such as Brazil, China, India, the Russian Federation, and South Africa are much more severely affected.5

Confirmation of diagnosis in a patient with suspected MDR-TB and the considerable delay in starting treatment was an additional problem, since cultures and sensitivity studies had to be performed. This particularly affected the start of second-line treatment.6 The situation has improved, thanks to new techniques that detect M. tuberculosis DNA in respiratory specimens, allowing the diagnosis and detection of isoniazid and/or rifampicin resistance in a matter of hours.7,8

Tuberculosis epidemics in schools are not uncommon, and when they occur, contact tracing is essential to diagnose secondary cases, to initiate prompt treatment of patients and carriers who require it, and to identify the real index case.9

This is a description of an outbreak of MDR-TB in a school setting and the difficulties we encountered in managing both patients with disease and carriers.

We conducted an observational, prospective study of an MDR-TB outbreak detected in November 2008 in 2 schools in Onda (O) and Nules (N) in the province of Castellón (Spain). A microepidemic or epidemic outbreak was defined as the appearance of 3 or more cases of tuberculosis related in space and time, or the diagnosis of at least 2 patients generated by the same index case.9,10

The index case was the first patient diagnosed, and the source case, or real index case, was the patient who was most likely to be the origin of the outbreak. A contact was any person who had shared enclosed spaces with a tuberculosis carrier.

After the index case was detected, contact tracing was performed, in accordance with previously published guidelines,9,10 using the concentric circles or “stone-in-the-water” model.11

The Castellón Public Health Center worked with the health centers in O and N to coordinate contact tracing in the 2 schools in which the index case was employed as a teacher, and among family members. They were also responsible for providing information to the parents of pupils at the school.

Tuberculin testing was performed using the Mantoux technique with 2TU PPD RT 23.

Latent tuberculosis infection was defined as a positive Mantoux test with a 5mm induration at 72h. An initial Mantoux test was performed and repeated 2 months later in cases which were initially negative. Individuals aged over 50 years with a negative result repeated the test 7 days later, to determine the booster effect.

All individuals with a positive Mantoux test had a chest X-ray.

If mycobacteria were cultured, the intraspecific differentiation was identified using restriction fragment length polymorphism (RFLP) analysis of the IS6110 sequence.12

A tuberculosis outbreak in a school setting is a situation of particular concern for public health,13–15,19,20 but outbreaks have also been described in churches,16 bars,17 buses,18 and even in healthcare centers.2–4 In this article, we report an outbreak in 2 schools, which was complicated by the fact that it was caused by a multiresistant strain of M. tuberculosis. Schools are an environment in which outbreaks are likely to occur: contact is continuous, classrooms are poorly ventilated, and diagnosis is delayed.19–21 In countries with high prevalence of MDR-TB, outbreaks in schools are common.21

Case 4 deserves particular attention, since she was diagnosed 1 year before the school outbreak; she presented no risk factors for MDR-TB,22 having never traveled to countries with high rates of multiresistant strains. This was the first case in our province,23 and there was a considerable delay (4 months) in diagnosis before definitive results were received from the reference laboratory. The patient had had contact with case 1, but case 1 could not attend the contact tracing visit. This underlines the importance of well-designed anti-tuberculosis programs,9,10 since an earlier diagnosis of case 1 may have prevented the outbreak. The index case had done voluntary work in Colombia for 2 summers, and we believe that he may have been infected there. Case 1 infected case 4, and subsequently triggered the school outbreak, so case 1 is considered as the source or real index case. Diagnosis of case 1 was delayed by 2 months, which meant that the schoolchildren and teaching staff started chemoprophylaxis treatment with isoniazid. When drug sensitivity study results were received, isoniazid had to be discontinued. A meeting was held with the parents, and we explained that, in multiresistant cases, no effective drug was available to prevent the infection from developing to tuberculous disease.10 This generated great anxiety and concern, so after several meetings coordinated by the Public Health Center, we decided to continue clinical and radiological monitoring for at least 2 years in Mantoux-positive cases, as indicated in the guidelines.10 We offered optional isoniazid, being the less toxic choice in children, in view of the possibility that individuals may be infected with sensitive strains. Perhaps the follow-up period should have been extended, since case 5 was diagnosed after 3 years. Very few studies have been published that confirm the efficacy of treating latent MDR-TB infection, but adverse events do seem to be an issue.24 Attamma et al. followed up 476 contacts of patients with MDR-TB over a mean duration of 6 years. Patients received isoniazid or a combination of ciprofloxacin and pyrazinamide, and 387 received no treatment. They did not detect any case of MDR-TB, thus supporting the clinical monitoring approach.25 The situation has changed since then, since effective chemoprophylaxis with fluoroquinolones has been reported in studies following MDR-TB cases for periods of up to 12 months.26 However, reluctance to use fluoroquinolones in children and teenagers persists, in view of their adverse effects on growing cartilage.

The rate of infection among exposed cases was 14.4%. This is relatively low, compared to other reports of 30%–100% in school outbreaks caused by sensitive M. tuberculosis.14,15 The same is true of the 0.9% rate of disease, which is much lower than other reported MDR-TB outbreaks that were highly explosive, probably due to co-infection with HIV.2–4 Kritski et al.27 also reported a rate of 8% in a contact tracing study in MDR-TB, and higher rates have also been reported in school outbreaks of M. tuberculosis.14,15 This finding concurs with the observation that multiresistant strains cause less secondary cases than sensitive strains,28 but for epidemiological purposes, we cannot assume that multiresistant strains are less infectious or less virulent than sensitive strains.29

In our study, the BCG vaccine conferred some protective effect among teachers, as previously noted by Kritski et al.27

After diagnosis, treatment was started and cases were closely monitored in monthly clinical check-ups with blood tests and sputum studies, since patients on second-line drugs often encounter toxicity and need encouragement to comply.30,31 All patients achieved a negative sputum culture within 4 months of starting treatment, and no relapses have been seen to date. This is probably because our patients were immunocompetent, non-HIV carriers, and had a good prognosis for treatment.21,32

A very important aspect of the study of outbreaks is genotyping of the causative strains. This information may help determine transmission mechanisms and determine if the disease is due to endogenous reactivation or exogenous reinfection, if it is found to be caused by more than 1 strain of M. tuberculosis. Genotyping of a strain is also a starting point for retrospective contact tracing.2,14,33–35

One limitation of our study is that we could not characterize latent tuberculous infection with interferon gamma release assays (IGRAS), which would have provided greater specificity to Mantoux reactions and tuberculin conversions.

To conclude, we report the first outbreak of MDR-TB in a school setting in Spain. A meticulous study of cases and contacts is crucial, and these individuals must be followed up closely and clearly informed, as the situation can generate considerable anxiety. As globalization continues, we can expect to encounter more tuberculosis caused by multiresistant strains. New technologies must be put to work to provide rapid diagnoses and to guide the right therapeutic intervention.

Dr. Miravet contributed to the follow-up of cases, data acquisition, analysis and interpretation, manuscript preparation and the review of the final version submitted for evaluation; Dr. García participated in the follow-up of cases and the review of the final version submitted for evaluation; Dr. Arnedo participated in contact tracing, data acquisition and interpretation, manuscript preparation and the review of the final version submitted for evaluation; Dr. Bellido participated in contact tracing, data analysis and interpretation, and the review of the final manuscript submitted for evaluation; Dr. Romeu contributed to contact tracing and the review of the final manuscript submitted for evaluation; Dr. Gil contributed to the microbiological study and the review of the final manuscript submitted for evaluation; Dr. Cortés participated in the follow-up of cases, and contributed to the review of the final manuscript submitted for evaluation.

The authors declare that they have no conflict of interests.