Drug-resistant tuberculosis (TB) is a worldwide public health problem.1 Many studies have focused on multidrug-resistant TB (resistant to isoniazid [H] and rifampicin [R]), but little attention has been given to isoniazid-resistant, rifampicin-sensitive (HRRS) TB.2 HRRS is the most common form of resistance (8.5% of cases1,3,4) and is a risk factor for poor clinical progress if treatment is not appropriate.5 Resistance rates to H in Spain are estimated at 5% among adults and 9.6% among children.6 There are few pediatric studies of HRRS TB due to the difficulty of isolating the Bacilli, and those that are available are generally retrospective.7,8

During the period 2014–2016, 17 children were diagnosed with TB in our hospital, of whom 7 (41.2%) were considered HRRS (Table 1). The study was prospective, using the REDCAP on-line database created by the pTBred working group, and the children were treated by a single pediatrician. Clinical samples were collected before starting treatment, and sensitivity studies were conducted in the local laboratory and national reference laboratory.

The 7 children with probable HRRS TB were between 3 and 6 years of age, of Spanish nationality, and not BCG-vaccinated. They were all detected following the diagnosis and contact tracing of 4 adults with active TB who were subsequently found to have H-resistant disease: 2 had single resistance to H and 2 were also streptomycin-resistant, and the mutation study detected 2 cases with the inhA and 2 with the katG gene mutation (Table 1). The Bacilli was isolated in only 1 child, confirming the resistance pattern of the source case. In the remaining 6 children the diagnosis of HRRS was based on the resistance of the source case and the absence of a different sensitivity pattern in the TB patients detected in the contact tracing study. Four of these 6 children had more than 6h daily contact with the resistant adult and 1 was the brother of the child with the positive culture.

Changes were observed on the chest X-ray of 5 children at the time of diagnosis. Patient 7, whose X-ray was normal, began treatment for latent tuberculosis infection, but 1 month later presented persistent cough, doubtful X ray and pathological CT. Chest CT was performed in another 2 cases: 1 had clinical symptoms with normal X-ray and the CT was abnormal and the other because of a radiological deterioration during treatment. Fiberoptic bronchoscopy was performed in this patient and in another case, both of whom had endobronchial involvement and corticosteroids were prescribed.

All patients started treatment with HRZE and continued for 1-2 months, until resistance was determined, after which they continued with RZE up to a total of 7–12 months The reasons for administering longer treatment regimens to 2 children were a failure of treatment compliance and a slow-to-resolve endobronchial TB.

Ophthalmological monitoring due to the administration of ethambutol was normal and 4 patients had mild side effects, not requiring treatment discontinuation. Patients were followed up for at least 12 months after completing treatment; at discharge, 6 cases had a normal X-ray, and 1 showed a small residual lymphadenopathy.

TB is usually classified as monoresistant, polyresistant, multidrug-resistant or extesively drug resistant TB,9 but the term HRRS TB is important because it is the most frequent form of drug resistant TB and its treatment must be appropriate.5,8 In our study, diagnosis of the HRRS pattern by culture took 1–2 months, as only rapid resistance testing for rifampicin was performed (GeneXpert®). It would also be advisable to perform a rapid diagnostic test such as MTBDRplus® (Hain Lifescience, Nehren, Germany), which identifies most cases resistant to R and/or H.10 However, the children were appropriately treated, since they started the treatment with 4 drugs because the sensitivity pattern of the source case was initially unknown and the rate of resistance to H in Spain is greater than 4% .11,12 Furthermore, our hospital area has an unusually high prevalence of HRRS TB, mainly due to strains with a unique pattern suggestive of TB reactivations. After HRRS TB was diagnosed, RZE was continued for a total of 7–12 months11,13 with good tolerance, and cure was confirmed without sequelae or reactivation in long-term follow-up.

The guidelines for initial treatment (2HRZE+4HR) and retreatment (2SHRZE+1HRZE+5HRZ) in cases of HRRS can lead to treatment failures, relapse, and acquired resistance, and the 6–9RZE combination has greater hepatotoxicity in adults.5 A meta-analysis comparing different treatments for HRRS TB concluded that the addition of a fluoroquinolone improved treatment success, although it only included 37 children.2,14 A retrospective study in children with HRRS TB, mainly with severe disease and administration of fluoroquinolones in 75% of cases, reported a treatment failure rate of only 4%.8 In our study, patients did not have advanced disease and the 6–12RZE regimen was effective and showed no side effects.

It could be argued that some of the 6 children without bacteriological confirmation might not have been HRRS. The definition of probable resistant disease in children has been recommended when the source case is resistant.9 This study was prospective and all the children were diagnosed following the contact tracing of an adult that caused an outbreak of TB with many secondary cases of TB infection and TB disease ( with the same strain and sensitivity pattern) and having 6 of the children high probability of contagion from the source case (proximity, frequency of exposure and infectivity).15 Moreover, the children were young, with little exposure to other TB patients outside of the family setting. In any case, the low rate of isolates in children means that the diagnosis of resistance is often made from the drug suscetibility testing of the source case. This, however, is a barrier to publishing pediatric studies of drug-resistant TB and proving the effectiveness of the different treatment regimens.

We conclude that it is important to determine resistance to H by performing a rapid drug sensitivity testing in the child and in their source case. Our patients responded favorably to 7–12RZE without side effects, sequelae or relapse.