MAJOR ARTICLE

Culture-Conrmed Multidrug-Resistant Tuberculosis in Children: Clinical Features, Treatment, and Outcome

James A. Seddon,1,2 Anneke C. Hesseling,1 Marianne Willemse,3 Peter R. Donald,1 and H. Simon Schaaf1,4
1Desmond Tutu Tuberculosis Centre, Faculty of Health Sciences, Stellenbosch University, South Africa; 2Department of Clinical Research, Faculty of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine, United Kingdom; 3Brooklyn Hospital for Chest Diseases, Cape Town, and 4Tygerberg Children's Hospital, South Africa

Downloaded from http://cid.oxfordjournals.org/ at HINARI Bolivia Administrative Account on May 3, 2012

Background. Multidrug-resistant (MDR) tuberculosis in children is frequently associated with delayed diagnosis and treatment. There is limited evidence regarding the management and outcome of children with MDR-tuberculosis. Methods. All children ,15 years of age with a diagnosis of culture-conrmed MDR-tuberculosis were included in this retrospective cohort study from 1 January 2003 to 31 December 2008, with follow-up documented until 31 May 2011. We identied children from Brooklyn Hospital for Chest Diseases and Tygerberg Childrens Hospital, Western Cape Province, South Africa. Treatment outcomes were dened as 2-month sputum-culture conversion, treatment episode outcome, and survival. Results. A total of 111 children (median age, 50 months) were included. The diagnosis was delayed in children who had no identied MDR-tuberculosis index case (median delay, 123 vs 58 days; P , .001). Sixty-two percent of patients (53 of 85) were sputum-smear positive, and 43% of patients (43 of 100) were human immunodeciency virus (HIV) infected. Overall, 82% had favorable treatment outcomes; total mortality was 12%. Malnutrition was associated with failure to culture-convert at 2 months (odds ratio [OR], 4.49 [95% condence interval {CI}, 1.3215.2]; P 5 .02) and death (OR, 15.0 [95% CI, 1.17192.5]; P 5 .04) in multivariate analysis. HIV coinfection (OR, 24.7 [95% CI, 1.79341.1]; P 5 .02) and the presence of extrapulmonary tuberculosis (OR, 37.8 [95% CI, 2.78513.4]; P 5 .006) predicted death. Conclusions. Despite advanced disease at presentation and a high prevalence of human immunodeciency virus coinfection, children with MDR-tuberculosis can be treated successfully, using individualized treatment under routine conditions.

Multidrug-resistant (MDR) tuberculosis is caused by Mycobacterium tuberculosis resistant to isoniazid and rifampin. The World Health Organization (WHO) estimated that there were 440 000 new MDR-tuberculosis cases worldwide during 2008 [1]. Treatment outcomes are generally poor in adults, with favorable outcomes

Received 12 June 2011; accepted 15 September 2011; electronically published 3 November 2011. Correspondence: James Seddon, MBBS, MA, MRCPCH, DTM&H, Department of Paediatrics and Child Health, Desmond Tutu Tuberculosis Centre, Clinical Building, Rm 0085, Faculty of Health Sciences, Stellenbosch University, PO Box 19063, Tygerberg, South Africa (jseddon@sun.ac.za). Clinical Infectious Diseases 2012;54(2):15766 The Author 2011. Published by Oxford University Press on behalf of the Infectious Diseases Society of America. All rights reserved. For Permissions, please e-mail: journals.permissions@oup.com. DOI: 10.1093/cid/cir772

reported in only 60% of those receiving treatment [2]. Even though childhood tuberculosis makes up 15%20% of the global tuberculosis burden [3], MDR-tuberculosis is poorly studied in children, the literature including mainly case reports or small case series [416]. The diagnosis of tuberculosis in young children is challenging and often delayed [17]. Symptoms and signs may be nonspecic, especially in children ,3 years of age and in children infected with human immunodeciency virus (HIV) [18]. Because of the paucibacillary nature of childhood tuberculosis, a microbiological diagnosis is typically made in only 20%40% of cases with radiological evidence of intrathoracic disease [19]. Because drug susceptibility testing (DST) is only possible following bacteriological conrmation, conrmed MDR-tuberculosis in children is infrequent. In
MDR-Tuberculosis in Children CID 2012:54 (15 January) 157

the absence of a known MDR-tuberculosis source case, children are often initially treated for drug-susceptible tuberculosis, with MDR-tuberculosis treatment started only once treatment is failing, microbiological and DST results become available, or an MDR-tuberculosis source case is identied. Treating children with MDR-tuberculosis is complex. Few of the multiple drugs routinely used to treat MDR-tuberculosis have been studied in children, and guidance on drug regimens, dosages, appropriate monitoring, and duration of therapy is frequently extrapolated from adult data. As young children metabolize drugs more rapidly than adults and generally have paucibacillary disease [20], this may not always be appropriate. We describe the clinical presentation, treatment, and outcome for a large cohort of children with conrmed MDR-tuberculosis and evaluate factors inuencing treatment response. METHODS
Setting and Population

were conrmed through polymerase chain reaction (PCR) [23]. Phenotypic DST was performed using 2 different assays that have been shown to yield highly concordant results, as previously described [24].
Definition of Tuberculosis Episodes and Treatment Delay

This retrospective cohort study was conducted in Western Cape Province, South Africa, where the tuberculosis notication rate was 978 per 100 000 population in 2009 [21]. Among all children with routinely diagnosed culture-conrmed tuberculosis at Tygerberg Childrens Hospital (TCH) during 20052007, tuberculosis in 6.7% was identied as MDR [22]. In this setting, children initially present to a variety of regional primary (ie, community-based) health clinics and referral hospitals, while TCH serves as the main referral center for children with drugresistant tuberculosis exposure or disease. Once children are stable during treatment, they are managed at the Brooklyn Hospital for Chest Diseases, a specialist tuberculosis hospital. After discharge, children are routinely followed as outpatients at TCH.
Eligibility Criteria

A previous tuberculosis episode was dened as standard tuberculosis treatment (ie, isoniazid, rifampin, and pyrazinamide, with or without ethambutol) for at least 1 month, followed by a symptom-free period of $6 months (determined on the basis of reports by the parent/caregiver and the absence of presentations to any healthcare providers) before the start of the current MDR-tuberculosis episode; this is a commonly used programmatic denition of a separate episode [25]. An MDRtuberculosis episode was dened as beginning (if MDRtuberculosis was subsequently conrmed) at the childs initial documented presentation to the healthcare system, when the specimen was obtained conrming MDR-tuberculosis, or when tuberculosis treatment was commenced. Treatment delay was dened as the time from the start of the MDR-tuberculosis episode to the initiation of MDR-tuberculosis treatment. Treatment delay could not be determined for children who died or were lost to follow-up before the start of MDR-tuberculosis treatment or for those treated inadvertently with rst-line drugs only.
Clinical Data and Standard of Care

Downloaded from http://cid.oxfordjournals.org/ at HINARI Bolivia Administrative Account on May 3, 2012

Children ,15 years of age were included if they received a diagnosis of conrmed MDR-tuberculosis between 1 January 2003 and 31 December 2008. DST for all children with cultureconrmed tuberculosis was routine during the study period. Follow-up was documented until 31 May 2011.
Mycobacterial Culture and Drug Susceptibility Testing

Mycobacterial culture was completed at the accredited National Health Systems Reference Laboratory following a standard protocol to prevent mycobacterial cross-contamination. Primary mycobacterial cultures were established by inoculation of routine clinical samples into Middlebrook 7H9 broth-based Mycobacterial Growth Indicator Tubes (MGIT960; Becton Dickinson) following a standard protocol for decontamination, while lymph node aspirates, pleural uid, cerebrospinal uid, and other bodily uids were directly inoculated. M. tuberculosis complex isolates
158
d

MDR-tuberculosis treatment was based on the local standard of care, informed by international recommendations and available literature [2633]. High-dose isoniazid (1520 mg/ kg) was used in the majority of cases, as there is evidence that isolates with an inhA promoter region mutation usually have a low minimum inhibitory concentration [34]. An injectable agent, most frequently amikacin (1530 mg/kg), was typically used for 6 months; capreomycin (1530 mg/kg) was substituted if resistance to amikacin was detected. Ooxacin (1520 mg/kg), the most effective uoroquinolone available in the South African National Tuberculosis Control Programme, was used. Further drugs were added to result in at least 4 effective drugs. These included ethionamide (1520 mg/kg), para-aminosalicylic acid (150 mg/kg), terizidone (1020 mg/kg), amoxicillin/clavulanic acid (1015 mg/kg every 8 hours), clarithromycin (7.515 mg/kg every 12 hours), and linezolid (10 mg/kg every 12 hours). All anti-tuberculosis drugs were given as directly observed therapy for the full treatment duration. Most children remained hospitalized during the intensive phase, when an injectable agent was given. Thereafter, children were treated at their local tuberculosis clinic, with hospital outpatient follow-up every 2 months. Children with extensively drug-resistant tuberculosis (XDR-tuberculosis; dened as MDR-tuberculosis with resistance to a uoroquinolone and a second-line injectable agent) were

CID 2012:54 (15 January)

Seddon et al

treated for longer periods, typically 24 months. If an HIVinfected child was not yet receiving combination antiretroviral therapy (cART), cART was started after the initiation of MDR-tuberculosis treatment, consistent with national guidelines.
Clinical Data Collection

Data were collected through chart review. HIV testing was performed after written informed consent was obtained from the parent or legal guardian, with pretest and posttest counseling; an enzyme-linked immunosorbent assay was used for children aged .18 months, and a DNA PCR test was used for younger and breastfed children. Immunological staging in HIV-infected children used the WHO classication [35]. Weight was recorded and plotted on the National Center for Health Statistics weight-for-age percentile chart. Malnutrition was classied as a weight of less than the third percentile for age. Two tuberculin units were injected intradermally (puried protein derivative RT23; Statens Serum Institut) for tuberculin skin testing. Results were read at 4872 hours, with transverse diameters of $10 mm and $5 mm considered positive in HIV-uninfected children and HIV-infected children, respectively. Tuberculosis severity was dened using chest radiographic features following review by a single expert reader, who read radiographs systematically with standardized recording [36]. Disease was classied as pulmonary if there were any chest radiographic changes attributable to tuberculosis or if any thoracic samples were positive for M. tuberculosis. Extrapulmonary disease was classied if any imaging (ie, ultrasonography, plain lm radiology, or computed tomography) demonstrated extrathoracic tuberculosis or if a microbiological sample conrmed tuberculosis at a site other than the lungs. Intrathoracic radiological features were classied as nonsevere (ie, normal, hilar lymphadenopathy, airway compression, lobar/segmental collapse/opacication, or pleural effusions) and severe (ie, cavities, miliary opacication, or a widespread bronchopneumonic picture).
Outcome

negative result, in the presence of treatment completion. Treatment outcomes were further classied as favorable (ie, cured and treatment completed) or unfavorable (ie, died, lost to follow-up, treatment failure, or transferred out). For MDRtuberculosis episodes with positive results of an initial sputum culture, culture conversion was dened as the time from initiation of therapy until the rst culture with negative results, provided there were no subsequent cultures with positive results and at least 2 further cultures with negative results. We describe 2-month culture conversion, as it is a frequently used surrogate marker for nal treatment outcome in adult treatment trials [38]. We further describe the total proportion of children who died.
Downloaded from http://cid.oxfordjournals.org/ at HINARI Bolivia Administrative Account on May 3, 2012

Statistical Analysis

Data were analyzed using Stata statistical software, version 11 (Stata). All identier details were dissociated from clinical data by unique study numbers. Missing data were excluded from analysis. Continuous variables were used for age and delay. Associations between clinical characteristics at presentation were assessed using the v2 test (or Fisher exact test) when comparing categorical data; effect estimates (ie, odds ratios [ORs]) and 95% condence intervals (CIs) were calculated. The MannWhitney U test was used to assess the effect of age and delay, given the nonnormal distribution of the data. Risk factors for treatment outcomes (ie, 2-month culture conversion, nal treatment outcome, and death) were assessed in univariate analysis. Multivariate logistic models were used to analyze the relationship between presenting characteristics and outcomes if the univariate relationships showed statistical signicance (dened as P , .05) or when variables were thought to be clinically or epidemiologically relevant. Variables classied as collinear were not used in combination in the model. This study was approved by the Committee for Human Research, Stellenbosch University, and the London School of Hygiene and Tropical Medicine (a waiver of informed consent was obtained). RESULTS During the study period, 111 children with MDR-tuberculosis were identied, with a median age of 50 months (interquartile range [IQR], 19108 months); all were included. Forty-two samples underwent DST to second-line drugs, which identied 3 MDR-tuberculosis cases resistant to amikacin, 4 resistant to ooxacin, and 5 resistant to both ooxacin and amikacin (XDR-tuberculosis). An overview of the cohort with treatment outcomes is provided in Figure 1. Demographic and clinical characteristics at the start of MDR-tuberculosis treatment are described in Table 1. The median time to MDR-tuberculosis treatment initiation (n 5 102) was shorter in the presence
MDR-Tuberculosis in Children
d

Respiratory samples in children with pulmonary involvement were obtained monthly to monitor response to therapy. Standard international denitions of nal treatment outcome were used, with minor adaption for children [37]. The denition of cure for adults with MDR-tuberculosis is completion of treatment with 5 cultures with negative results in the nal 12 months of treatment [26]. For children with drug-susceptible tuberculosis, cure is dened as being sputum culture or smear negative in the last month of treatment and on $1 previous occasions [27]. Cure for this study was dened as negative results of 3 consecutive respiratory cultures obtained at least 1 month apart, with no positive culture results after the rst

CID 2012:54 (15 January)

159

Downloaded from http://cid.oxfordjournals.org/ at HINARI Bolivia Administrative Account on May 3, 2012

Figure 1.

Overview of treatment outcomes in 111 children with multidrug-resistant tuberculosis (MDR-tuberculosis).

of a known adult MDR-tuberculosis index case (median time to MDR-tuberculosis treatment initiation, 58 days [IQR, 25120 days] vs 123 days [IQR, 67231 days]; P , .001). Fiftythree (62%) of 85 children with a sputum result were smear positive; a positive sputum smear was more common in older children (median age, 85 months [IQR, 25132 months] vs 24 months [IQR, 1559 months]; P , .001) and in children with more-severe chest radiographic changes (OR, 9.95 [95% CI, 2.9833.3]; P , .001). Of 100 children (90.1%) tested for HIV, 43 (43%) were positive; 27 (64%) had severe immunosuppression. Nineteen children were receiving cART before initiation of MDR-tuberculosis treatment; cART for 21 was started afterward (median time from start of MDR-tuberculosis treatment to cART initiation, 75 days [IQR, 18123 days]). Fifty of 109 children (46%) had severe chest radiographic changes at diagnosis; children with severe chest radiographic
160
d

changes were older (median age, 84 months [IQR, 27121 months] vs 28 months [IQR, 1568 months]; P 5 .002) (Table 1). Of the 111 patients, 91 (82%) had a favorable treatment outcome. Of these, 3 were treated successfully with only rstline drugs: 2 had cervical lymph node disease and the other only hilar lymphadenopathy. Four patients were transferred to another hospital, and 3 were lost to follow-up. One adolescent, who received a diagnosis of pulmonary XDR-tuberculosis, was a repeat defaulter, and her sputum never converted. She was declared to have failed treatment after 2 years of intermittent treatment and eventually was transferred into adult care. One patient was still receiving treatment at the end of the study period, having been initially treated for MDR-tuberculosis with additional resistance to amikacin, according to his bacteriological data, and then for XDR-tuberculosis, based on his mothers bacteriological data. The overall mortality was 12% (13 deaths;

CID 2012:54 (15 January)

Seddon et al

Table 1. Clinical Characteristics at Initial Presentation in Children With Bacteriologically Confirmed Multidrug-Resistant Tuberculosis
Characteristic Age, median months (IQR) Male sex Treatment delay, median days (IQR) (n 5 105) Index case None identied Index case identied with no evidence of MDR-tuberculosis MDR-tuberculosis index case indentied Index case defaulter, treatment failure, or died Previous treatment episode Positive Mantoux skin test (n 5 89) Weight ,3rd percentile for age 3rd10th percentile for age .10th percentile for age Tuberculosis type PTB only EPTB only Both PTB and EPTB Site of EPTB (.1 possible; n 5 38) Miliary Tuberculous meningitis Pericardial effusion Pleural effusion Abdominal area Peripheral lymph node Bone/joint/spine Ear Sputum smear positive (n 5 85) Drug susceptibility test pattern (n 5 42) MDR MDR with resistance to amikacin MDR with resistance to ooxacin XDR CR features (.1 present in the majority; n 5 109) Normal (all had EPTB) Hilar lymphadenopathy or airway compression Lobar/segmental collapse/opacication Large pleural effusion Cavities Miliary opacication Widespread bronchopneumonic changes CR severity (n 5 109) Nonsevere Severe Time to sputum conversion, median months (IQR) (n 5 74) HIV-infected (n 5 100 tested) 59 (54.1) 50 (45.9) 2 (13) 43 (43.0) 11 (10.1) 52 (47.7) 76 (69.7) 7 (6.4) 38 (34.9) 7 (6.4) 21 (19.2) 7 (18.4) 6 (15.8) 2 (5.3) 7 (18.4) 8 (21.1) 16 (42.1) 9 (23.7) 5 (13.2) 53 (62.3) 30 (71.4) 3 (7.1) 4 (9.5) 5 (11.9) 73 (65.8) 12 (10.8) 26 (23.4) 41 (36.9) 28 (25.2) 42 (37.8) 33 (29.7) 17 (15.3) 45 (40.5) 16 (14.4) 28 (25.2) 63 (70.8) No. (%) 50 (19108) 46 (41.4) 91 (51166)

Table 1 continued.
Characteristic HIV immunological stage (n 5 42) Not signicant Mild Moderate Severe Time of cART initiation (n 5 43) Never (all died) Receiving at start of MDR-tuberculosis episode Initiated between start of episode and start of MDR-tuberculosis treatment Initiated after start of MDR-tuberculosis treatment Time from start of MDR-tuberculosis treatment to cART initiation, median days (IQR) (n 5 21)
Data are no. (%) of 111 children, unless otherwise indicated. Abbreviations: cART, combination antiretroviral therapy; CR, chest radiograph; EPTB, extrapulmonary tuberculosis; HIV, human immunodeciency virus; IQR, interquartile range; MDR-tuberculosis, multidrug-resistant tuberculosis; PTB, pulmonary tuberculosis; XDR, extensively drug resistant. For nonsignicant, CD4 cell values (age) are .35% (#11 mo), .30% (1235 mo), .25% (3659 mo), and .500 cells/mm3 ($5 y); for mild, 30%35% (#11 mo), 25%30% (1235 mo), 20%25% (3659 mo), and 350499 cells/mm3 ($5 y); for moderate, 25%29% (#11 mo), 20%24% (1235 mo), 15%19% (3659 mo), and 200349 cells/mm3 ($5 y); and for severe, ,25% (#11 mo), ,20% (1235 mo), ,15% (3659 mo), and ,200 cells/mm3 or ,15% ($5 y).
a

No. (%)
a

7 (16.7) 5 (11.9) 3 (7.1) 27 (64.3) 3 (7.0) 9 (20.9) 10 (23.3) 21 (48.8) 75 (18123)

Downloaded from http://cid.oxfordjournals.org/ at HINARI Bolivia Administrative Account on May 3, 2012

Table 2), regardless of treatment initiation. Eleven children died during their MDR-tuberculosis episode, 1 was cured of tuberculosis but died in the year following the end of treatment, and 1 died following treatment failure. Of the 88 patients successfully treated with MDR-tuberculosis drugs, 79 (89.8%) were alive and well 12 months after completion of treatment. Of the remaining 9 patients, 3 had been transferred to another institution, 5 had been lost to follow-up, and 1 had died. Those successfully treated were treated for a median duration of 18 months, including a median duration of 6 months with an injectable agent, with a median of 7 drugs over the total course of treatment (Table 3). Univariate analysis of clinical features and their association with outcome are shown in Table 4. Malnutrition and severe chest radiographic changes were associated with a failure to culture-convert by month 2, unsuccessful treatment outcome, and death. HIV infection and extrapulmonary disease were associated with death. Children with positive smears at diagnosis were less likely to have culture-converted by month 2. Findings of multivariable analysis are shown in Table 5. After adjustment for age and smear status (or chest radiographic severity), malnutrition at diagnosis predicted failure to culture-convert by 2 months (OR, 4.49 [95% CI, 1.3215.2]; P 5 .02). Malnutrition (OR, 15.0 [95% CI, 1.17192.5]; P 5 .04), HIV infection (OR, 24.7 [95% CI, 1.79341.1]; P 5 .02), and extrapulmonary disease
MDR-Tuberculosis in Children
d

CID 2012:54 (15 January)

161

Table 2. Description of 13 Children With Culture-Confirmed Multidrug-Resistant Tuberculosis Who Died

CD4 cell value, count (%) .

Age at death 11 years

Sex M

HIV status Unknown

Tuberculosis site Spine

History Presented with severe disease; underwent surgery, with biopsy sample taken; started rst-line treatment but died 1 month later, before diagnosis was made Died after 1 month of rst-line therapy; MDR-tuberculosis diagnosed posthumously Died after 12 months of MDR-tuberculosis treatment

Attributed cause of death MDR-tuberculosis of the spine

6 months

Pos

1652 (36%) Pulmonary

10 years

Pos

37 (8%)

Disseminated (miliary)

Hepatotoxicity, HIV infection, MDR-tuberculosis, Down syndrome, severe cardiac defect, marasmic, bacterial pneumonia Developed disseminated tuberculosis, including tuberculosis meningitis, while receiving full MDR-tuberculosis treatment in hospital; possible XDR-tuberculosis Disseminated MDR-tuberculosis

Downloaded from http://cid.oxfordjournals.org/ at HINARI Bolivia Administrative Account on May 3, 2012

6 years

Pos

33 (2%)

Pulmonary and abdominal

9 months

Pos

825 (19%)

Pulmonary, miliary, and lymph node

Treated for 10 months with rst-line drugs before sample was sent for culture; MDR-tuberculosis diagnosed posthumously Died after 3 months of MDR-tuberculosis treatment

Multisystem failure, HIV infection, extensive disseminated tuberculosis at presentation, sepsis, heart failure, thrombocytopenia Severe tuberculosis meningitis (stage 3) Severe HIV infection, systemic candidal infection, osteomyelitis, respiratory failure, pneumonia, S. aureus sepsis Disseminated MDR-tuberculosis meningitis

4 years

Pos Pos

Not tested Meninges 117 (6%) Pulmonary

Died after 3 months of MDR-tuberculosis treatment Died after 5 months of MDR-tuberculosis treatment

12 months F

2 years

Neg

15 years

Unknown

Meninges, Died after 1 month of rst-line pulmonary, treatment; MDR-tuberculosis abdominal, and diagnosed posthumously lymph node Pulmonary Died after 1 month of MDR-tuberculosis treatment; pre-XDR result returned after death Given 24 days of rst-line treatment for stage 3 tuberculosis meningitis before MDR-tuberculosis diagnosis; died after 1 month of MDR-tuberculosis treatment

Congenital myopathy, aplastic anemia, requiring multiple transfusions, extensive pre-XDR tuberculosis MDR-tuberculosis meningitis, HIV infection

2 years

Pos

1065 (26%) Meninges

8 years

Unknown

Pulmonary

12 years

Neg

Pulmonary

Spastic quadriplegia from previous Severe MDR-tuberculosis, preexisting illness; died 6 d after starting neurological condition, making rst-line treatment; MDRdiagnosis challenging tuberculosis diagnosed after death Adolescent repeated defaulter with Died of pre-XDR tuberculosis under the pre-XDR-tuberculosis; never sputum care of adult physician culture-converted and declared a treatment failure after 12 months of MDR-tuberculosis treatment; transferred to adult care Registered as cured of MDR-tuberculosis with 20 months of therapy; signicant lung damage as a result of tuberculosis Developed bronchiectasis and chronic lung abscess; died within 1 year of nishing MDR-tuberculosis treatment with a bacterial infection

8 years

Neg

Pulmonary

A total of 111 children were included in the study. Abbreviations: F, female; HIV, human immunodeciency virus; M, male; MDR-tuberculosis, multidrug-resistant tuberculosis; Neg, negative; Pos, positive; S. aureus, Staphylococcus aureus; XDR, extensively drug resistant.

162

CID 2012:54 (15 January)

Seddon et al

Table 3. Treatment Characteristics for 88 Children Successfully Treated With a Multidrug-Resistant Tuberculosis Regimen
Variable Drugs used at any point during treatment, median no. (range) Duration of intensive phase therapy, median months (range) Total duration of therapy, median months (range) Anti-tuberculosis drug used Isoniazid (high dose) Rifampin Pyrazinamide Ethambutol Streptomycin Amikacin Capreomycinb Ooxacin Ethionamide Terizidone or cycloserine Para-aminosalicylic acidb Clarithromycin Augmentin Linezolidb
Data are no. (%) of children, unless otherwise indicated.
a b

Value 7 (413) 6 (018)a 18 (826) 83 (94.3) 14 (15.9) 81 (92.0) 82 (93.2) 1 (1.1) 80 (90.9) 6 (6.8) 86 (97.7) 86 (97.7) 57 (64.8) 7 (8.0) 7 (8.0) 6 (6.8) 2 (2.3)

No injectable was used in 6 cases. Available only from 2007.

(OR, 37.8 [95% CI, 2.78513.4]; P 5 .006) all independently predicted mortality in a model that also adjusted for age. DISCUSSION We document the clinical presentation, treatment, and outcomes of children with culture-conrmed MDR-tuberculosis under routine clinical conditions in a setting where the tuberculosis burden is high. To our knowledge, this is the largest cohort of children with culture-conrmed MDR-tuberculosis described to date. Our data indicate that children with cultureproven MDR-tuberculosis tend to be young and malnourished, are frequently HIV-infected, and often present with radiological features of advanced disease. Furthermore, the absence of a known MDR-tuberculosis source case led to considerable delay in initiation of appropriate therapy. Treatment regimens were long (median duration, 18 months), with 6 months including an injectable agent. Of key importance is that, despite advanced disease and the presence of extrapulmonary disease in .30% of children, the majority were treated successfully, with .80% having favorable outcomes. We identified important risk factors for clinically and programmatically relevant treatment outcomes, including mortality. HIV infection, malnutrition, and extrapulmonary

involvement were independent risk factors for death in adjusted analyses. Five of the 13 deaths occurred before MDR-tuberculosis was conrmed and appropriate treatment started, indicating the importance of early diagnosis. Although severe disease based on chest radiographic ndings was associated with all outcome measures in univariate analyses, this association was less pronounced in adjusted analyses. Our ndings suggest that, once MDR-tuberculosis is identied and treated appropriately with individualized therapy based on available DST, children with the disease have a good prognosis, even in areas with a high prevalence of HIV coinfection. Other reports of MDR-tuberculosis in children include a previous study from Cape Town of 39 children with cultureconrmed disease; similar to the present study, treatment delay was common if an MDR index case of tuberculosis was not identied [4]. A study from Peru described 38 children treated for MDR-tuberculosis, 28 of whom had conrmed disease, and also found considerable delay in the initiation of appropriate therapy [5]. A study from New York of 20 children treated for MDR-tuberculosis (6 of whom had culture-conrmed disease), demonstrated good outcomes and minimal toxicity [6]. A recent case series of 13 children with culture-conrmed MDRtuberculosis from Johannesburg (54% of whom were HIVinfected) indicated high mortality of 30% [12]. Other case reports and small series have been reported from other settings [711, 1316]. Despite the delay in diagnosis, severe disease at presentation, and the need for second-line medications, these studies all describe good outcomes, in dramatic contrast to adult MDR-tuberculosis data [2]. Reasons for this contrast are unclear. Children may have less severe and paucibacillary disease, may tolerate and adhere to medications better, may have a lower prevalence of HIV-infection, or may have a lower prevalence of concomitant pathology. Further research is required in this eld. The absence of an identied adult MDR-tuberculosis index case was strongly associated with delay in initiation of appropriate treatment in children. Of note, 16 (14.4%) children had index cases who were documented to have died, failed treatment, or undergone retreatment, indicating a high risk of MDR-tuberculosis exposure. These factors highlight the importance of careful history taking from both the childs caregiver and from health services regarding contact with adults with known MDR-tuberculosis or with known risk factors for MDR-tuberculosis. Since we only report on children with culture-conrmed disease, our data are not representative of all children with MDR-tuberculosis, many of whom are treated on the basis of MDR-tuberculosis contact history or poor clinical response to therapy. As bacteriological yield is associated with radiological extent of disease [19], we likely report children with more advanced disease. Since we report on children who received
MDR-Tuberculosis in Children
d

Downloaded from http://cid.oxfordjournals.org/ at HINARI Bolivia Administrative Account on May 3, 2012

CID 2012:54 (15 January)

163

Downloaded from http://cid.oxfordjournals.org/ at HINARI Bolivia Administrative Account on May 3, 2012

164
d

Table 4.

Univariate Analysis of Clinical Features, 2-Month Culture Conversion, Treatment Outcome, and Death in Children With Multidrug-Resistant Tuberculosis
Failure to culture-convert by month 2 Proportion No. failing analyzed to convert 74/74 46/74 28/74 53/74 21/74 42/71 29/71 26 13 13 14 12 14 11 1.22 (.453.31) .69 3.71 (1.2211.3) .01 2.20 (.806.01) .11 Unfavorable treatment outcome Proportion analyzed No. died 111/111 65/111 1.53 (.574.07) .39 46/111 70/111 3.21 (1.158.96) .02 41/111 57/100 2.16 (.746.36) .15 43/100 11 6 5 2 9 1 7 10.9 (1.17101.0) .008 9.56 (1.7951.2) .001 1.20 (.344.22) .78 Death

CID 2012:54 (15 January)

d

Characteristic Age Sex Female Male Nutrition $3rd percentile ,3rd percentile HIV infection Negative Positive Timing of cART initiation Before MDR-tuberculosis treatment After MDR-tuberculosis treatment Mantoux skin test Negative Positive MDR-tuberculosis contact No contact Contact Treatment delay Site of tuberculosis No EPTB EPTB Smear status Negative Positive CR severity Nonsevere Severe

OR (95% CI) .

No. with Proportion unfavorable P value analyzed outcome .06 111/111 65/111 46/111 70/111 41/111 57/100 43/100 20 10 10 8 12 7 10

OR (95% CI) .

P value .80

OR (95% CI) .

P value .97

Seddon et al

14/29 15/29

4 7 2.19 (.4410.8) .32

22/43 21/43

5 5 1.06 (.254.45) .93

22/43 21/43

4 3 0.75 (.143.92) .73

16/62 46/62 35/74 39/74 74 60/74 14/74 27/68 41/68 39/74 35/74

2 28 13 13 26 20 6 4 20 9 17 3.15 (1.118.92 .022 5.48 (1.4720.4) .004 1.50 (.454.97) .50 0.85 (.322.21) .73 .25 4.50 (.8523.7) .05

26/89 63/89 66/111 45/111 103 73/111 38/111 32/85 53/85 59/109 50/109

7 8 15 5 15 10 10 4 10 5 13 3.79 (1.2012.0) .014 1.63 (.465.77) .443 2.25 (.836.11) .10 0.43 (.141.29) .12 .36 0.39 (.121.27) .10

26/89 63/89 66/111 45/111 103 73/111 38/111 32/85 53/85 59/109 50/109

3 5 8 3 8 3 8 2 5 2 8 5.43 (1.0528.2) .02 1.56 (.288.68) .61 6.22 (1.4526.6) .005 0.52 (.132.09) .35 .18 0.66 (.143.03) .59

Abbreviations: cART, combination antiretroviral therapy; CI, condence interval; CR, chest radiograph; EPTB, extrapulmonary tuberculosis; HIV, human immunodeciency virus; MDR-tuberculosis, multidrug-resistant tuberculosis; OR, odds ratio.

Table 5. Multivariate Analysis of Clinical and Bacteriological Features at Diagnosis, 2-Month Culture Conversion, Treatment Outcome, and Death in Children With Multidrug-Resistant Tuberculosis
No. in analysis

Characteristic Failure to culture-convert by 2 moa Smear positivity Malnutrition Age Poor treatment outcomeb Severe CR changes Malnutrition HIV positivity Age Death EPTB Malnutrition HIV positivity Age

OR (95% CI)

P value

68 68 68 99 99 99 99 99 99 99 99

3.24 (.8212.8) 4.49 (1.3215.2)

.10 .02 .15

2.50 (.689.19) 1.87 (.586.07) 1.46 (.464.63)

.17 .30 .52 .51

37.8 (2.78513.4) 15.0 (1.17192.5) 24.7 (1.79341.1)

.006 .04 .02 .18

Abbreviations: CI, condence interval; CR, chest radiograph; EPTB, extrapulmonary tuberculosis; HIV, human immunodeciency virus; OR, odds ratio.
a

a greater risk for tuberculosis transmission than previously thought. Infection control should be an important consideration in the management of children with MDR-tuberculosis. This retrospective study has limitations, including reliance on routine data. We did not have systematic data regarding adverse effects, and the tolerability of multiple medications, frequently unpalatable, was not systematically assessed. Although all samples were conrmed to be MDR-tuberculosis, DST for second-line drugs was not routinely completed during the study period. We have reported the second-line DST results that were available, but because of inconsistent testing and signicant bias in completion of DST, we are unable to draw meaningful conclusions. Finally, although treatment outcomes were good, we do not comment on morbidity as a result of MDR-tuberculosis disease and treatment. All of these aspects need assessment in future prospective studies. In conclusion, although South African children with conrmed MDR-tuberculosis often present with severe disease and have a high frequency of HIV infection, excellent treatment outcomes can be achieved in high-burden settings with individualized clinical care under standard programmatic conditions.
Notes
Acknowledgments. We thank Mrs R. Rabie, Mrs W. Brittle, Mrs K. Zimri, and Miss Z. Mramba, for assistance with clinical and laboratory data collection, and Mr J. Harvey, for statistical advice. Financial support. This work was supported by grants from TREAT TB (GHN-A-00-08-00004-00), the US Agency for International Development (to J. A. S. and H. S. S.), the Sir Halley Steward Trust (to J. A. S.), the South African Medical Research Council (to H. S. S.), and the National Research Foundation of South Africa (to P. R. D. and H. S. S.). Potential conicts of interest. All authors: No reported conicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conicts of Interest. Conicts that the editors consider relevant to the content of the manuscript have been disclosed.

Downloaded from http://cid.oxfordjournals.org/ at HINARI Bolivia Administrative Account on May 3, 2012

In an alternative model, in which all variables were held constant but smear positivity was replaced with severe CR changes, ndings were similar (n 5 74; severe CR changes: OR, 1.88 [95% CI, .615.78]; P 5 .270; malnutrition: OR, 3.51 [95% CI, 1.1211.0]; P 5 .031; age: P 5 .148).

In an alternative model, in which all variables were held constant but severe CR changes was replaced with EPTB, ndings were similar (n 5 100; EPTB: OR, 2.59 [95% CI, .867.75]; P 5 .90; malnutrition: OR, 2.43 [95% CI, .807.40]; P 5 .115; HIV positivity: OR, 2.43 [95% CI, .655.98]; P 5 .232; age: P 5 .679).

a diagnosis of MDR-tuberculosis by use of combined sources of surveillance (ie, laboratory and tuberculosis register sources), we do not report on only those who started therapy (which is the traditional tuberculosis treatment cohort approach). Given this conservative approach, we probably underestimate treatment success in our cohort. The long duration of treatment (a median of 6 months with a second-line injectable medication and 18 months overall) could possibly be reduced in children with less severe disease in the future, if adequate evidence from clinical studies becomes available. In our study, many children had severe disease at initiation of treatment. There was a high proportion with a positive smear, and nearly half had cavities, severe bronchopneumonic changes, or a miliary opacication on a chest radiograph. Young children are traditionally considered to have paucibacillary tuberculosis, and as they generally have a poor tussive force, they are considered to pose a low infection risk. However, our data indicate that children with culture-conrmed MDRtuberculosis frequently have severe disease, are somewhat older than those with drug-susceptible tuberculosis [22, 39], and, coupled with high rates of smear positivity, may prove

References
1. World Health Organization (WHO). Multidrug and extensively drugresistant TB (M/XDR-TB) 2010 Global report on surveillance and response. 2010: (WHO/HTM/TB/2010.3). 2. Orenstein EW, Basu S, Shah NS, et al. Treatment outcomes among patients with multidrug-resistant tuberculosis: systematic review and meta-analysis. Lancet Infect Dis 2009; 9:15361. 3. van Rie A, Beyers N, Gie RP, Kunneke M, Zietsman L, Donald PR. Childhood tuberculosis in an urban population in South Africa: burden and risk factor. Arch Dis Child 1999; 80:4337. 4. Schaaf HS, Shean K, Donald PR. Culture conrmed multidrug resistant tuberculosis: diagnostic delay, clinical features, and outcome. Arch Dis Child 2003; 88:110611. 5. Drobac PC, Mukherjee JS, Joseph JK, et al. Community-based therapy for children with multidrug-resistant tuberculosis. Pediatrics 2006; 117:20229. 6. Feja K, McNelley E, Tran CS, Burzynski J, Saiman L. Management of pediatric multidrug-resistant tuberculosis and latent tuberculosis infections in New York City from 1995 to 2003. Pediatr Infect Dis J 2008; 27:90712.

MDR-Tuberculosis in Children

CID 2012:54 (15 January)

165

7. Padayatchi N, Bamber S, Dawood H, Bobat R. Multidrug-resistant tuberculous meningitis in children in Durban, South Africa. Pediatr Infect Dis J 2006; 25:14750. 8. Mendez Echevarria A, Baquero Artigao F, Garcia Miguel MJ, et al. Multidrug-resistant tuberculosis in the pediatric age group. Anales de Pediatria 2007; 67:20611. 9. Schluger NW, Lawrence RM, McGuiness G, Park M, Rom WN. Multidrug-resistant tuberculosis in children: two cases and a review of the literature. Pediatr Pulmonol 1996; 21:13842. 10. Suessmuth S, Bange FC, Gappa M. Multidrug resistant tuberculosis in a 6 year old child. Paediatr Respir Rev 2007; 8:2658. 11. Pinon M, Scolfaro C, Bignamini E, et al. Two pediatric cases of multidrugresistant tuberculosis treated with linezolid and moxioxacin. Pediatrics 2010; 126:e12536. 12. Fairlie L, Beylis NC, Reubenson G, Moore DP, Madhi SA. High prevalence of childhood multi-drug resistant tuberculosis in Johannesburg, South Africa: a cross sectional study. BMC Infect Dis 2011; 11:28. 13. Thomas TA, Shenoi SV, Heysell SK, et al. Extensively drug-resistant tuberculosis in children with human immunodeciency virus in rural South Africa. Int J Tuberc Lung Dis 2010; 14:124451. 14. Kjollerstrom P, Brito MJ, Gouveia C, Ferreira G, Varandas L. Linezolid in the treatment of multidrug-resistant/extensively drug-resistant tuberculosis in paediatric patients: experience of a paediatric infectious diseases unit. Scand J Infect Dis 2011; 43:5569. 15. Tochon M, Bosdure E, Salles M, et al. Management of young children in contact with an adult with drug-resistant tuberculosis, France, 20042008. Int J Tuberc Lung Dis 2011; 15:32630. 16. Mukherjee JS, Joseph JK, Rich ML, et al. Clinical and programmatic considerations in the treatment of MDR-TB in children: a series of 16 patients from Lima, Peru. Int J Tuberc Lung Dis 2003; 7:63744. 17. Beyers N, Gie RP, Schaaf HS, et al. Delay in the diagnosis, notication and initiation of treatment and compliance in children with tuberculosis. Tuber Lung Dis 1994; 75:2605. 18. Marais BJ, Gie RP, Hesseling AC, et al. A rened symptom-based approach to diagnose pulmonary tuberculosis in children. Pediatrics 2006; 118:e13509. 19. Marais BJ, Hesseling AC, Gie RP, Schaaf HS, Enarson DA, Beyers N. The bacteriologic yield in children with intrathoracic tuberculosis. Clin Infect Dis 2006; 42:e6971. 20. Loebstein R, Koren G. Clinical pharmacology and therapeutic drug monitoring in neonates and children. Pediatr Rev 1998; 19:4238. 21. Health Systems Trust. Incidence of TB (all types) per 100000. 2009. Available at: http://www.hst.org.za/healthstats/16/data. Accessed 24 February 2011. 22. Schaaf HS, Marais BJ, Hesseling AC, Brittle W, Donald PR. Surveillance of antituberculosis drug resistance among children from the Western Cape Province of South Africaan upward trend. Am J Public Health 2009; 99:148690. 23. Warren RM, Gey van Pittius NC, Barnard M, et al. Differentiation of Mycobacterium tuberculosis complex by PCR amplication of

24.

25.

26.

27.

28.

29. 30.

31. 32. 33.

34.

35.

36.

37.

38.

39.

genomic regions of difference. Int J Tuberc Lung Dis 2006; 10:81822. Schaaf HS, Marais BJ, Whitelaw A, et al. Culture-conrmed childhood tuberculosis in Cape Town, South Africa: a review of 596 cases. BMC Infect Dis 2007; 7:140. Schaaf HS, Krook S, Hollemans DW, Warren RM, Donald PR, Hesseling AC. Recurrent culture-conrmed tuberculosis in human immunodeciency virus-infected children. Pediatr Infect Dis J 2005; 24:68591. World Health Organization (WHO). Guidelines for the programmatic management of drug-resistant tuberculosisemergency update. 2008: (WHO/HTM/TB/2008.402). World Health Organization (WHO). Guidance for national tuberculosis programmes on the management of tuberculosis in children. 2006: (WHO/HTM/TB/2006371, WHO/FCH/CAH/20067). Caminero JA, Sotgiu G, Zumla A, Migliori GB. Best drug treatment for multidrug-resistant and extensively drug-resistant tuberculosis. Lancet Infect Dis 2010; 10:6219. Chiang CY, Schaaf HS. Management of drug-resistant tuberculosis. Int J Tuberc Lung Dis 2010; 14:67282. Schaaf HS, Marais BJ. Management of multidrug-resistant tuberculosis in children: a survival guide for paediatricians. Paediatr Respir Rev 2011; 12:318. Al-Dabbagh M, Lapphra K, McGloin R, et al. Drug-resistant tuberculosis: pediatric guidelines. Pediatr Infect Dis J 2011; 30:5015. Grant A, Gothard P, Thwaites G. Managing drug resistant tuberculosis. BMJ 2008; 337:a1110. Schaaf HS, Donald PR. The epidemiology and management of drugresistant tuberculosis in childhood. In: Spiegelburg DD, ed. New topics in tuberculosis research. New York, NY: Nova Science Publishers, 2007: 71102. Schaaf HS, Victor TC, Engelke E, et al. Minimal inhibitory concentration of isoniazid in isoniazid-resistant Mycobacterium tuberculosis isolates from children. Eur J Clin Microbiol Infect Dis 2007; 26:2035. World Health Organization (WHO). WHO case denitions of HIV for surveillance and revised clinical staging and immunological classication of HIV-related disease in adults and children. Geneva, Switzerland: WHO, 2007. Marais BJ, Gie RP, Schaaf HS, et al. A proposed radiological classication of childhood intra-thoracic tuberculosis. Pediatr Radiol 2004; 34:88694. World Health Organization (Headquarters and Regional Ofces), the International Union Against Tuberculosis and Lung Disease (IUATLD), and the Royal Netherlands Tuberculosis Association (KNCV). Revised international denitions in tuberculosis control. Int J Tuberc Lung Dis 2001; 5:2135. Hesseling AC, Walzl G, Enarson DA, et al. Baseline sputum time to detection predicts month two culture conversion and relapse in nonHIV-infected patients. Int J Tuberc Lung Dis 2010; 14:56070. Schaaf HS, Marais BJ, Hesseling AC, Gie RP, Beyers N, Donald PR. Childhood drug-resistant tuberculosis in the Western Cape Province of South Africa. Acta Paediatr 2006; 95:5238.

Downloaded from http://cid.oxfordjournals.org/ at HINARI Bolivia Administrative Account on May 3, 2012

166

CID 2012:54 (15 January)

Seddon et al