n engl j med 391;24 nejm.org December 19/26, 2024 2315 T h e n e w e ngl a nd j o u r na l o f m e dic i n e The authors’ full names, academic de‑ grees, and affiliations are listed in the Appendix. Dr. Hesseling can be contact‑ ed at annekeh@ sun . ac . za or at Desmond Tutu TB Centre, Department of Pediat‑ rics and Child Health, Clinical Bldg., Rm. 0070, Faculty of Medicine and Health Sciences, Stellenbosch University, Parow, Cape Town 8001, South Africa. N Engl J Med 2024;391:2315-26. DOI: 10.1056/NEJMoa2314318 Copyright © 2024 Massachusetts Medical Society. BACKGROUND Worldwide, approximately 2 million children younger than 15 years of age are in- fected with multidrug-resistant (MDR) Mycobacterium tuberculosis, with MDR tubercu- losis developing in approximately 30,000 annually. Evidence from randomized, controlled trials on tuberculosis preventive treatment in persons exposed to MDR tuberculosis is lacking. METHODS In this community-based, multisite, double-blind, cluster-randomized, placebo- controlled trial in South Africa, we assessed the efficacy and safety of levofloxacin as preventive treatment in children with household exposure to an adult with bacterio- logically confirmed MDR pulmonary tuberculosis. Children younger than 5 years of age were eligible for inclusion regardless of interferon-γ release assay result or human immunodeficiency virus (HIV) status, and children 5 to 17 years of age were eligible if they had a positive interferon-γ release assay or HIV infection. Households were randomly assigned to a trial regimen, and children in the household received levo- floxacin or placebo once daily for 24 weeks. The primary efficacy end point was inci- dent tuberculosis, which included death from tuberculosis, by week 48 after random- ization. The primary safety end point was any adverse event of grade 3 or higher during the treatment period that was at least possibly related to the trial regimen. RESULTS Of 922 participants from 497 households, 453 were assigned to receive levofloxacin and 469 to placebo; 91.0% of the participants were younger than 5 years of age. At least 80% of the assigned doses of levofloxacin or placebo were received by 86% of the participants in each trial group. By week 48, tuberculosis had developed in 5 partici- pants (1.1%) in the levofloxacin group and in 12 participants (2.6%) in the placebo group (hazard ratio, 0.44; 95% confidence interval [CI], 0.15 to 1.25). The results of sensitivity analyses were consistent with those of the primary analysis. Grade 3 or higher adverse events during the treatment period that were considered to be at least possibly related to the trial regimen occurred in 4 participants in the levofloxacin group and in 8 participants in the placebo group (hazard ratio, 0.52; 95% CI, 0.16 to 1.71). Grade 2 tendonitis occurred in 1 child in the levofloxacin group. CONCLUSIONS Although preventive treatment with levofloxacin led to a lower incidence of tubercu- losis than placebo among children with household exposure to MDR tuberculosis, the difference was not significant. (Supported by Unitaid and others; TB-CHAMP ISRCTN Registry number, ISRCTN92634082.) A BS TR AC T Levofloxacin Preventive Treatment in Children Exposed to MDR Tuberculosis A.C. Hesseling, S.E. Purchase, N.A. Martinson, L. Fairlie, H.S. Schaaf, J. Brigden, S. Staples, D.M. Gibb, A. Garcia‑Prats, F. Conradie, C. McGowan, C. Layton, E. Batist, A.‑M. Demers, S. Nyamathe, L. Frigati, R. Turner, T. Duong, and J.A. Seddon Original Article The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on January 27, 2025. For personal use only. No other uses without permission. Copyright © 2024 Massachusetts Medical Society. All rights reserved. https://doi.org/10.1186/ISRCTN92634082 n engl j med 391;24 nejm.org December 19/26, 20242316 T h e n e w e ngl a nd j o u r na l o f m e dic i n e Each year, multidrug-resistant (MDR) tuberculosis develops in approximately half a million people worldwide.1 MDR tuber- culosis is caused by Mycobacterium tuberculosis strains that are resistant to isoniazid and ri- fampin. Treatment of MDR tuberculosis is more challenging than treatment of drug-susceptible tuberculosis and is expensive and complex for health services and families.2 An estimated 2 million children worldwide are currently infected with MDR M. tuberculosis, with MDR tuberculosis developing in approximately 30,000 annually.3 Current approaches to controlling the MDR tuberculosis epidemic largely focus on diagnosing the disease in symptomatic persons and initiat- ing effective treatment. However, modeling stud- ies suggest that contact management (investiga- tion of close contacts of persons exposed to a patient with infectious tuberculosis) and delivery of tuberculosis preventive treatment to healthy persons at high risk for disease progression4 are needed to reach the ambitious targets of the End TB Strategy.5 Good evidence is available on the efficacy and safety of tuberculosis preventive treatment for persons exposed to drug-susceptible M. tuberculosis, but data from randomized, controlled trials for MDR tuberculosis are lacking.6 Observational data suggest that regimens containing fluoro- quinolones are associated with a low risk of tuberculosis.7 The World Health Organization currently conditionally recommends the consid- eration of tuberculosis preventive treatment for selected high-risk household contacts of patients with MDR tuberculosis on the basis of an indi- vidualized risk assessment8 and calls for data from clinical trials owing to the low quality of evidence for this recommendation. Young children, particularly those under 5 years of age, are an important population for tuberculo- sis preventive treatment owing to the higher risks of faster disease progression after exposure and severe forms of tuberculosis than among older children and adults.9 Older children and adolescents with evidence of M. tuberculosis infec- tion and those with human immunodeficiency virus (HIV) infection are also at increased risk.10,11 Among children younger than 5 years of age with exposure to M. tuberculosis followed by the onset of disease, tuberculosis developed within 1 year after exposure in 80%.10 One of the most efficient strategies for iden- tifying children who have recently been exposed to tuberculosis involves tracing household con- tacts of patients with newly diagnosed tubercu- losis and offering preventive treatment to the contacts after tuberculosis has been ruled out. The TB-CHAMP (Tuberculosis Child Multidrug- Resistant Preventive Therapy) trial was designed to assess the efficacy and safety of levofloxacin as compared with placebo as tuberculosis pre- ventive treatment in children and adolescents with household exposure to an adult with infectious MDR tuberculosis. Me thods Trial Oversight The funders had no role in the design or conduct of the trial or the dissemination of the data. Levofloxacin and placebo were purchased from the manufacturer (Macleods Pharmaceuticals), which did not provide funding for the trial. The trial was approved by institutional review boards in each center and by national ethics com- mittees in each country. All the authors vouch for the accuracy and completeness of the data and for the fidelity of the trial to the protocol, which is available with the full text of this article at NEJM.org. Additional details about trial over- sight are provided in the Supplementary Appendix (available at NEJM.org). Trial Design and Participants The TB-CHAMP trial was a double-blind, cluster- randomized, placebo-controlled trial conducted at 5 sites in South Africa between September 26, 2017, and January 21, 2023.12 Adult index pa- tients with a routine diagnosis of bacteriologi- cally confirmed pulmonary tuberculosis and mul- tidrug resistance that was detected on the basis of genotypic or phenotypic approaches were identi- fied and assessed for eligibility. Index patients were eligible if they provided written informed consent and had lived in the same household with at least one child younger than 5 years of age during the previous 6 months. Children younger than 5 years of age were eligible regardless of interferon-γ release assay result or HIV status if their parents or caregivers provided written in- formed consent and they had been exposed to an index patient during the preceding 6 months. A Quick Take is available at NEJM.org The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on January 27, 2025. For personal use only. No other uses without permission. Copyright © 2024 Massachusetts Medical Society. All rights reserved. n engl j med 391;24 nejm.org December 19/26, 2024 2317 Levofloxacin Preventive Ther apy for MDR Tuberculosis Under version 3.0 of the protocol (September 2021), children 5 to 17 years of age were eligible if they had a positive interferon-γ release assay (QuantiFERON-Gold Plus, Qiagen) or HIV infec- tion and their parents or caregivers provided writ- ten informed consent; children 7 years of age or older also needed to provide assent. Before enrollment, children were evaluated for prevalent tuberculosis with the use of history taking, physical examination, and plain-film chest radiography. Respiratory samples for mi- crobiologic testing for M. tuberculosis were ob- tained from children with symptoms or signs consistent with tuberculosis, abnormal findings on chest radiography, or both. Children were given broad-spectrum antibiotic agents on the basis of the clinical judgment of the treating physician and reevaluated. Only children with no evidence of tuberculosis were recruited. Households were randomly assigned in a 1:1 ratio to the levofloxacin group or the placebo group, with stratification according to trial site. Randomization was conducted according to a computer-generated randomization list that was prepared by the trial statistician with the use of block randomization with varying block sizes and stored in a centralized Web-based database. Additional details about the trial design and participants are provided in the Supplementary Appendix. Trial Procedures Children were seen at baseline; at weeks 4, 8, 12, 16, 24, 48, and 72 after randomization; and at interim visits as clinically indicated. Disease as- sessments were conducted during the trial visits if the child had symptoms that were potentially consistent with tuberculosis, including cough or fever for more than 2 weeks and poor weight gain or weight loss, as well as any symptoms without an alternative explanation. Children were also evaluated for adverse events related to levofloxacin or placebo with the use of National Institutes of Health Division of AIDS toxicity tables. Additional information about the trial procedures is provided in the Supplementary Appendix. Trial Regimens Scored 250-mg tablets of levofloxacin (given at a dose of 15 to 20 mg per kilogram of body weight; maximum dose, 750 mg) and matched placebo (primarily microcrystalline cellulose) were admin- istered once daily for 24 weeks. Children were weighed during the trial visits, and doses were adjusted as appropriate. Levofloxacin and placebo were given by parents or caregivers, and adher- ence was assessed with the use of questionnaires administered during the trial visits, treatment cards, and pill counts. Treatment interruptions of up to 4 weeks overall were allowed. End Points The primary end point was incident tuberculosis (microbiologically confirmed or clinically diag- nosed), which included death from tuberculosis, by week 48 after randomization. An independent end-point review committee whose members were unaware of the trial group assignments adjudi- cated the primary end-point events with the use of international consensus case definitions.13 Discor- dant assessments were resolved by discussion. The prespecified primary safety end point was any grade 3 or higher adverse event during the treatment period that was at least possibly related to the trial regimen. Secondary end points included incident tuberculosis by week 72, death from any cause, any adverse event of grade 3 or higher between the start of the trial regi- men and up to 30 days after the last dose, seri- ous adverse events up to 30 days after the last dose, discontinuation of the trial regimen be- cause of adverse events, selected predefined ad- verse events (see the Supplementary Appendix for details), and adherence to treatment. Statistical Analysis We calculated that 1009 participants would pro- vide the trial with 80% power to detect a 60% reduction in the incidence of tuberculosis in the levofloxacin group as compared with the placebo group by week 48, at a two-sided significance level of 5%. We based this calculation on the as- sumption of a cumulative tuberculosis incidence of 7% in the placebo group by week 48, the en- rollment of a mean of two participants per house- hold, a household intraclass correlation coeffi- cient of 0.10, and a loss to follow-up of 10%. The primary efficacy analysis was performed in the modified intention-to-treat population, which comprised all the participants who under- went randomization except those with late screen- ing failure (defined as the identification of a The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on January 27, 2025. For personal use only. No other uses without permission. Copyright © 2024 Massachusetts Medical Society. All rights reserved. n engl j med 391;24 nejm.org December 19/26, 20242318 T h e n e w e ngl a nd j o u r na l o f m e dic i n e participant’s ineligibility for the trial after the participant had been assigned to a trial regimen) due to undiagnosed tuberculosis at baseline. Kaplan–Meier methods were used to assess the time from randomization to incident tuberculo- sis. A prespecified window (±6 weeks) was al- lowed for the week 48 visit. Follow-up data for participants without incident tuberculosis by the week 48 visit were censored at week 54 after ran- domization, on the date of last follow-up visit, or on the date of death, whichever occurred first. We performed Cox regression analysis to es- timate the hazard ratio for incident tuberculosis in the levofloxacin group as compared with the placebo group. Intrahousehold correlation was accounted for with the use of a cluster-robust variance estimator. We adjusted for trial site and age group (<5 vs. 5 to 17 years) using the inverse probability of treatment weighting method, ow- ing to the small number of primary end-point events.14 The analysis was conducted under the assumption of noninformative censoring, but oth- erwise there were no missing data in the model. We assessed for evidence of nonproportional hazards using the Grambsch–Therneau test based on scaled Schoenfeld residuals.15 Safety analyses included all the participants who received at least one dose of levofloxacin or placebo. Follow-up time was censored at 30 days after the last dose of levofloxacin or placebo for the analysis of relevant end points. Treatment adherence was classified according to the per- centage of the assigned doses that were received (<80%, 80 to <90%, or ≥90%). Prespecified sen- sitivity analyses of the primary end point in- cluded an analysis with adjustment for prespeci- fied baseline covariates (interferon-γ release assay result, HIV status, and weight-for-age z score), an analysis of incident tuberculosis events that were considered to be adjudicated at the trial site ow- ing to the attending clinician’s decision to start treatment, an analysis that included only chil- dren younger than 5 years of age, and a per- protocol analysis restricted to participants who adhered to the assigned trial regimen and did not have late screening failure.16 Confidence in- tervals for the safety, secondary, and prespeci- fied sensitivity analyses were not adjusted for multiplicity and should not be used to infer de- finitive treatment effects. The independent data monitoring committee reviewed unblinded interim data four times during the trial, including findings from formal interim efficacy analysis performed after 500 children had been followed for at least 24 weeks (October 2021). The Haybittle–Peto criterion for efficacy (P<0.001) was used to guide recommendations to stop or modify the trial. Analyses were per- formed with Stata, version 16.0 or later (Stata- Corp). R esult s Participants We identified 5063 index patients with MDR tuberculosis, of whom 631 were screened for eligibility (Fig. 1). A total of 1120 children who were household contacts of eligible index pa- tients were screened for eligibility, and 922 chil- dren (household contacts of 497 index patients from 497 households) were enrolled between September 27, 2017, and July 29, 2022. Baseline characteristics of the 497 index patients who were household contacts of the participants are provided in Table S5.1 in the Supplementary Appendix. A total of 453 participants were from house- holds assigned to the levofloxacin group, and 469 participants were from households assigned to the placebo group. Late screening failure oc- curred in 29 participants, including 6 partici- pants with undiagnosed tuberculosis at baseline (Table S4.4). Thus, 916 participants were included in the modified intention-to-treat population and assessed in the primary efficacy analysis. Baseline characteristics of the participants from randomized households were reasonably similar in the trial groups (Table 1). The median age of the participants was 2.8 years (interquartile range, 1.3 to 4.2 years), 91.0% were younger than 5 years of age, and 93.8% had received bacille Calmette– Guérin vaccine. Overall, 49.2% of the participants were boys, 2.1% had HIV infection, and 33.9% had been exposed to HIV but were not infected. Of the 815 children younger than 5 years of age with available test results, 163 (20.0%) had a positive interferon-γ release assay. Follow-up Retention was similar in the trial groups, with 81% of the participants followed for at least 48 weeks and 73% followed to week 72 (Table S6.1); 11% of the participants left the trial before their last scheduled visit (Table S6.3). One participant The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on January 27, 2025. For personal use only. No other uses without permission. Copyright © 2024 Massachusetts Medical Society. All rights reserved. n engl j med 391;24 nejm.org December 19/26, 2024 2319 Levofloxacin Preventive Ther apy for MDR Tuberculosis did not start the trial regimen because of early withdrawal. Primary End Point Incident tuberculosis as adjudicated by the end- point review committee occurred in 21 of 916 participants (2.3%) during the follow-up period; 10 of the participants had microbiologically con- firmed tuberculosis. In 2 participants, incident tuberculosis occurred after the week 72 visit. Incident tuberculosis by week 48 (primary end point) occurred in 5 of 451 participants (1.1%) in the levofloxacin group and in 12 of 465 partici- pants (2.6%) in the placebo group (Table S8.1), corresponding to 1.2 cases per 100 person-years (95% confidence interval [CI], 0.5 to 2.9) and 2.9 cases per 100 person-years (95% CI, 1.6 to 5.2), respectively (hazard ratio, 0.44; 95% CI, 0.15 to 1.25; P = 0.12). Incident tuberculosis occurred during the 24-week treatment period in 1 partici- pant in the levofloxacin group and in 10 partici- pants in the placebo group (P = 0.11 by the test for nonproportional hazards) (Fig. 2). Secondary End Points By the week 72 visit, incident tuberculosis had occurred in 6 participants (1.3%) in the levo- floxacin group and in 13 participants (2.8%) in the placebo group (hazard ratio, 0.49; 95% CI, 0.18 to 1.30) (Table S8.6). Safety analyses included 921 participants who started the trial regimen. Four of 452 participants (0.9%) in the levofloxacin group and 8 of 469 participants (1.7%) in the placebo group had a grade 3 or higher adverse event during the treatment period that was as- sessed by the site investigator to be at least pos- sibly related to the trial regimen (hazard ratio, 0.52; 95% CI, 0.16 to 1.71; P = 0.29) (Table 2); 14 participants (3.1%) and 24 participants (5.1%), respectively, had a grade 3 or higher adverse event that occurred up to 30 days after the last dose of levofloxacin or placebo (hazard ratio, 0.64; 95% CI, 0.32 to 1.28; P = 0.21). Two deaths occurred during follow-up. One infant in the levofloxacin group who was 12 months of age died from cardiac arrest caused by a congenital heart defect after week 39, and 1 infant in the placebo group who was 11 months of age died from viral pneumonia after week 11. Neither death was considered by the site investi- gator to be related to the trial regimen or tuber- culosis. One participant in the levof loxacin group had grade 2 tendonitis, which resolved 21 days after stopping treatment. Seventy-five partici- pants (16.6%) in the levofloxacin group and 80 participants (17.1%) in the placebo group perma- nently discontinued the trial regimen early for any reason (Table S7.2). Six participants (1.3%) in the levofloxacin group and 1 participant (0.2%) in the placebo group discontinued the trial regimen be- cause of adverse events (hazard ratio, 5.00; 95% CI, 0.61 to 41.32; P = 0.14). In each trial group, 86% of the participants received at least 80% of the as- signed doses of levofloxacin or placebo (Table S7.3). Poor adherence to the levofloxacin regimen was not associated with an increased incidence of tuberculosis during follow-up. Sensitivity Analyses Results of prespecified sensitivity analyses were generally consistent with those of the primary analysis, although the analysis based on the adju- dication of incident tuberculosis at the trial site showed a slightly stronger treatment effect than in the primary analysis (Fig. 3). We found no evi- dence of heterogeneity of treatment effect in pre- specified exploratory subgroup analyses, includ- ing analyses according to sex, interferon-γ release assay result, and HIV status (Table S8.5), but the numbers of events were small. Discussion In this pediatric trial of preventive treatment for MDR tuberculosis, we recruited nearly 1000 chil- dren with household exposure to MDR tuber- culosis. Incident tuberculosis as adjudicated by the independent end-point review committee oc- curred in 17 children by week 48. Although tu- berculosis occurred in a smaller percentage of children in the levofloxacin group than in the placebo group, the difference was not signifi- cant. We have shown that levofloxacin treatment taken once daily for 6 months did not lead to any safety concerns in the children. Adherence to the trial regimen was 86% in the two groups, despite the use of an adult formulation. We anticipated that the effect size of levo- floxacin would be similar to that in the trial, but we had assumed that the risk of incident tuber- culosis in the placebo group would be twice that in the trial, owing to results of observational studies of drug-susceptible and drug-resistant The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on January 27, 2025. For personal use only. No other uses without permission. Copyright © 2024 Massachusetts Medical Society. All rights reserved. n engl j med 391;24 nejm.org December 19/26, 20242320 T h e n e w e ngl a nd j o u r na l o f m e dic i n e 922 Child participants from 497 households were enrolled and assigned to a trial regimen 5063 Adult index patients with multidrug- resistant tuberculosis were identified 198 Were not eligible 32 Had presumptive tuberculosis, had initiation of treatment for tuberculosis, or both 103 Had abnormal findings on chest radiography with or without symptoms suggestive of tubercu- losis and were referred for further investigation 8 Were ≥5 years of age (before protocol amendment) 6 Had insufficient exposure to index patient 4 Had received treatment for tuberculosis in the past 12 mo 14 Were IGRA negative and ≥5 yr of age 3 Had been receiving preventive treatment for tuberculosis for ≥14 days 7 Had concurrent exposure to INH-susceptible tuberculosis 8 Had informed consent declined 1 Weighed <3 kg 2 Had other medical conditions 4 Did not return for enrollment 6 Had no additional information 453 (from 248 households) Were assigned to receive levofloxacin 469 (from 249 households) Were assigned to receive placebo 378 Were included in the per-protocol population 407 Were included in the per-protocol population 631 Were screened 1120 Children were screened 2 Were excluded owing to late screening failure with tuberculosis at baseline 4 Were excluded owing to late screening failure with tuberculosis at baseline 451 Were included in the modified intention-to-treat population for the primary efficacy analysis 465 Were included in the modified intention-to-treat population for the primary efficacy analysis 73 Were excluded from the per-protocol population 1 Incorrectly received placebo 41 Were nonadherent 16 Had missing informa- tion on treatment adherence 15 Had other late screening failure 58 Were excluded from the per-protocol population 32 Were nonadherent 18 Had missing informa- tion on treatment adherence 8 Had other late screening failure The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on January 27, 2025. For personal use only. No other uses without permission. Copyright © 2024 Massachusetts Medical Society. All rights reserved. n engl j med 391;24 nejm.org December 19/26, 2024 2321 Levofloxacin Preventive Ther apy for MDR Tuberculosis tuberculosis.16-19 The lower-than-expected inci- dence of tuberculosis in the placebo group has several potential explanations. First, 20% of the children who were younger than 5 years of age had a positive interferon-γ release assay, a per- centage considerably lower than in previous studies in South Africa, in which 40 to 50% of children with household exposure had a positive result.17 Second, given the wide rollout of the Xpert MTB/RIF assay in South Africa, it is pos- sible that persons with MDR tuberculosis now begin appropriate therapy sooner after disease onset, leading to a decrease in the duration of household exposure and the risk of M. tuberculosis transmission. Third, with the widespread rollout of more effective treatments for MDR tuberculo- sis (particularly bedaquiline and linezolid) since 2018, it is possible that patients with MDR tu- berculosis who receive these agents become noninfectious more rapidly. Fourth, to exclude children with prevalent tuberculosis from the trial, we performed rigorous assessments at base- line, including thorough review of medical his- tory, physical examination, and chest radiogra- phy, with microbiologic testing performed in those with any symptoms or signs of tuberculo- sis or abnormal findings on chest radiography. Children with late screening failure owing to the presence of tuberculosis at baseline may not have been detected with less comprehensive screening and would otherwise have been con- sidered to have incident disease when symptoms and signs became more pronounced later during the trial. The timing of incident tuberculosis differed in the trial groups. Most cases in the placebo group occurred during the first 12 weeks after randomization. This result suggests that although many of the children with incident tuberculosis did not have clinically overt disease at baseline, they had some form of subclinical disease that rapidly progressed to symptomatic disease. In the levofloxacin group, incident tuberculosis oc- curred in 1 child during preventive treatment, whereas most cases occurred around 1 year after randomization. Levofloxacin may therefore have treated subclinical disease and prevented disease progression, but it appeared to have had mini- mal effect after the treatment period. This find- ing has implications for household-contact man- agement and provision of tuberculosis preventive treatment early after the diagnosis of tuberculo- sis in the index patient. The similar incidence of tuberculosis in both groups after week 48 could have been due to subsequent reinfection in this high-burden setting. A systematic review from 2017 showed that the incidence of MDR tuberculosis among ex- posed persons who received preventive treatment was 90% lower than that among those who did not receive preventive treatment.7 Several studies and trials assessing preventive treatment in house- hold contacts of persons with tuberculosis have since been initiated. An observational study in Pakistan showed that the use of a fluoroquinolone together with either ethambutol or ethionamide as preventive therapy resulted in a decrease of 65% in the risk of drug-resistant tuberculosis among 172 study participants as compared with historical controls.18 The VQUIN MDR trial in Vietnam com- pared the efficacy of levofloxacin with that of placebo for the prevention of tuberculosis in 2041 household contacts, mainly adults; results of the trial are reported in this issue of the Journal.20 The Protecting Households on Exposure to Newly Di- agnosed Index Multidrug-Resistant Tuberculosis Patients trial is an open-label trial comparing delamanid with isoniazid for the prevention of tuberculosis in adults and children with house- hold exposure to MDR tuberculosis, with recruit- ment ongoing at multiple sites worldwide.19 Figure 1 (facing page). Screening and Trial-Group Assignment. Reasons for screening ineligibility among the index patients included inability of the trial team to establish contact, death or relocation, age of less than 18 years, presence of rifampicin‑monoresistant tuberculosis, re‑ ceipt of tuberculosis treatment for more than 6 months, presence of nonpulmonary tuberculosis, and no chil‑ dren younger than 5 years of age reported to be living in the household during the past 6 months. In Sep‑ tember 2021, the protocol was amended to extend eligibility to children 5 to 17 years of age if they had a positive interferon‑γ release assay (IGRA) or human immunodeficiency virus infection and their parents or caregivers provided written informed consent. Cluster randomization was performed at the household level. Late screening failure was defined as the identification of a participant’s ineligibility for the trial after the par‑ ticipant had been assigned to a trial regimen. INH de‑ notes isoniazid. The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on January 27, 2025. For personal use only. No other uses without permission. Copyright © 2024 Massachusetts Medical Society. All rights reserved. n engl j med 391;24 nejm.org December 19/26, 20242322 T h e n e w e ngl a nd j o u r na l o f m e dic i n e Table 1. Characteristics of the Participants at Baseline.* Characteristic Levofloxacin (N = 453) Placebo (N = 469) Total (N = 922) Male sex — no. (%) 213 (47.0) 241 (51.4) 454 (49.2) Age Median (IQR) — yr 3.0 (1.4 to 4.3) 2.6 (1.3 to 4.1) 2.8 (1.3 to 4.2) Range — yr 0.1 to 17.9 0.1 to 17.4 0.1 to 17.9 Distribution — no. (%) <1 yr 85 (18.8) 83 (17.7) 168 (18.2) 1 to <3 yr 140 (30.9) 175 (37.3) 315 (34.2) 3 to <5 yr 180 (39.7) 176 (37.5) 356 (38.6) 5 to <10 yr 18 (4.0) 17 (3.6) 35 (3.8) 10 to <15 yr 20 (4.4) 13 (2.8) 33 (3.6) 15 to <18 yr 10 (2.2) 5 (1.1) 15 (1.6) Race or ethnic group — no. (%)† Black 362 (79.9) 381 (81.2) 743 (80.6) South African mixed 81 (17.9) 78 (16.6) 159 (17.2) Mixed race 8 (1.8) 8 (1.7) 16 (1.7) Other 2 (0.4) 2 (0.4) 4 (0.4) IGRA result — no. (%)‡ Negative 297 (65.6) 335 (71.4) 632 (68.5) Positive 134 (29.6) 108 (23.0) 242 (26.2) Indeterminate 12 (2.6) 12 (2.6) 24 (2.6) Missing data 10 (2.2) 14 (3.0) 24 (2.6) HIV status — no. (%) Positive 10 (2.2) 9 (1.9) 19 (2.1) HIV‑exposed, uninfected 153 (33.8) 160 (34.1) 313 (33.9) HIV‑unexposed 288 (63.6) 298 (63.5) 586 (63.6) Missing data 2 (0.4) 2 (0.4) 4 (0.4) Previous BCG vaccination — no. (%)§ Yes 423 (93.4) 442 (94.2) 865 (93.8) No 28 (6.2) 25 (5.3) 53 (5.7) Missing data 2 (0.4) 2 (0.4) 4 (0.4) Previous tuberculosis treatment — no. (%) Yes 10 (2.2) 8 (1.7) 18 (2.0) No 443 (97.8) 461 (98.3) 904 (98.0) Current tuberculosis preventive treatment — no. (%) Yes 9 (2.0) 6 (1.3) 15 (1.6) No 444 (98.0) 463 (98.7) 907 (98.4) Median weight‑for‑age z score (IQR)¶ −0.4 (−1.2 to 0.3) −0.4 (−1.2 to 0.4) −0.4 (−1.2 to 0.3) Median height‑for‑age z score (IQR)‖ −0.9 (−1.6 to −0.2) −0.9 (−1.8 to −0.2) −0.9 (−1.7 to −0.2) Trial site — no. (%) Desmond Tutu TB Centre 222 (49.0) 230 (49.0) 452 (49.0) The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on January 27, 2025. For personal use only. No other uses without permission. Copyright © 2024 Massachusetts Medical Society. All rights reserved. n engl j med 391;24 nejm.org December 19/26, 2024 2323 Levofloxacin Preventive Ther apy for MDR Tuberculosis An important consideration when assessing the efficacy of tuberculosis preventive treatment is the contribution of the preventive treatment it- self as compared with that of the household- contact management strategy. In our trial, tuber- culosis preventive treatment was compared with no treatment, but in both trial groups, the index patients with MDR tuberculosis were identified, the household contacts were rigorously screened for prevalent disease, and the participating child contacts were followed regularly. In most areas with a high burden of tuberculosis, evaluation and treatment of close contacts of patients with MDR tuberculosis in routine care is probably less complete. In a modeling study of the effect of household-contact management and tuberculo- sis preventive treatment in children exposed to MDR tuberculosis, screening for prevalent dis- ease was responsible for 65% of the total num- ber of deaths prevented.21 Moreover, modeling work indicates that the long-term population- level effects of preventive treatment for MDR tuberculosis, including the reduction in onward transmission owing to the prevention of cases in adults, could be greater than suggested by trial end points alone.22 In addition, fluoroquinolones may have substantial negative effects on gram- positive and gram-negative bacteria, potentially leading to the development of antimicrobial re- sistance and alterations in the gut microbiome23; these possible effects are being evaluated. Safety results of our trial were reassuring, with no con- cerning patterns of grade 3 or higher adverse events identified. One episode of tendonitis (grade 2 in severity) was reported. Our trial has limitations. First, the incidence of tuberculosis in the placebo group was lower than anticipated. Second, findings of the trial Characteristic Levofloxacin (N = 453) Placebo (N = 469) Total (N = 922) Shandukani 92 (20.3) 80 (17.1) 172 (18.7) Matlosana 117 (25.8) 142 (30.3) 259 (28.1) Isanga Lethembo 3 (0.7) 4 (0.9) 7 (0.8) TB and HIV Investigative Network 19 (4.2) 13 (2.8) 32 (3.5) * Percentages may not sum to 100 because of rounding. IQR denotes interquartile range. † Race or ethnic group was reported by the parents or caregivers. ‡ Mycobacterium tuberculosis infection status at baseline was assessed with the use of the QuantiFERON‑Gold Plus (Qiagen) interferon‑γ release assay (IGRA). § Previous bacille Calmette–Guérin (BCG) vaccination was verified on the basis of a review of medical records, the pres‑ ence of a vaccine scar, or both. ¶ Weight‑for‑age z scores for participants 10 years of age or younger were calculated with the use of the World Health Organization (WHO) reference standard. Scores were calculated with the use of the U.K. reference standard for partici‑ pants older than 10 years of age, because WHO weight‑for‑age growth charts are only available for children 10 years of age and younger. ‖ Height‑for‑age z scores were calculated with the use of the WHO reference standard. Table 1. (Continued.) Figure 2. Incident Tuberculosis by Week 48 According to Trial Group. Shown are Kaplan–Meier estimates of the cumulative incidence of tubercu‑ losis. A prespecified window (±6 weeks) was allowed for scheduled visits. Therefore, follow‑up time in the analysis of incident tuberculosis, which included death from tuberculosis, by week 48 after randomization (primary end point) was censored at 54 weeks. Shaded regions indicate confidence bands, which were adjusted for household clustering. C um ul at iv e In ci de nc e (% ) 100 80 90 70 60 40 30 10 50 20 0 0 12 24 36 48 54 6 5 4 3 1 2 0 0 2412 36 48 54 Weeks since Randomization Hazard ratio, 0.44 (95% CI, 0.15–1.25) P=0.12 Placebo Levofloxacin 465 451 379 368 357 339 334 323 Levofloxacin Placebo No. at Risk Placebo Levofloxacin 3 1 7 0 0 0 0 1 441 425 414 412 2 3 No. of Primary End-Point Events The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on January 27, 2025. For personal use only. No other uses without permission. Copyright © 2024 Massachusetts Medical Society. All rights reserved. n engl j med 391;24 nejm.org December 19/26, 20242324 T h e n e w e ngl a nd j o u r na l o f m e dic i n e Table 2. Safety End Points.* End Point Levofloxacin (N = 452) Placebo (N = 469) Hazard Ratio (95% CI) P Value no. of participants (%) Grade ≥3 adverse event during treatment period at least possibly related to trial regimen 4 (0.9) 8 (1.7) 0.52 (0.16–1.71) 0.29 Any grade ≥3 adverse event† 14 (3.1) 24 (5.1) 0.64 (0.32–1.28) 0.21 Serious adverse event 9 (2.0) 8 (1.7) 1.22 (0.45–3.34) 0.69 Adverse event that led to permanent discon‑ tinuation of trial regimen 6 (1.3) 1 (0.2) 5.00 (0.61–41.32) 0.14 Prespecified adverse events‡ Arthritis, arthralgia, or tendinopathy of any grade, alone or in combination 6 (1.3) 4 (0.9) 1.32 (0.35–4.98) 0.69 Central nervous system disorders 7 (1.5) 11 (2.3) 0.59 (0.21–1.62) 0.30 Severe rash or cutaneous reaction 1 (0.2) 0 — — * Data are for participants who received at least one dose of levofloxacin or placebo. The confidence intervals were not adjusted for multiplicity and should not be used to infer definitive treatment effects. Additional details about the reported adverse events for each safety end point are provided in Tables S9.9 through S9.11. † Data are for events that occurred up to 30 days after receipt of the last dose of levofloxacin or placebo. ‡ Central nervous system disorders included altered mood or behavior including hyperactivity, insomnia, vivid dreams, changes in vision, and headache. Drug‑related fever and peripheral neuropathy were also prespecified adverse events, but neither event occurred in the trial groups. Figure 3. Primary, Secondary, and Prespecified Sensitivity Analyses of Incident Tuberculosis by Week 48. A prespecified window (±6 weeks) was allowed for the week 48 visit. Therefore, follow‑up time was censored at 54 weeks. Death from any cause was also a secondary end point; 1 participant in each trial group died. The sensitivity analyses of the primary end point included an analysis with adjustment for prespecified baseline covariates (interferon‑γ release assay result, human immunodeficiency virus status, and weight‑for‑age z score), an analysis of incident tuberculosis events that were considered to be adjudicated at the trial site owing to the attending clinician’s decision to start treatment, an analysis that included only children younger than 5 years of age, and a per‑protocol analysis restricted to participants who adhered to the assigned trial regimen and did not have late screening failure. The confidence in‑ tervals for the secondary end‑point analyses and the prespecified sensitivity analyses were not adjusted for multiplicity and should not be used to infer definitive treatment effects. 0.5000.250 1.000 2.000 Placebo BetterLevofloxacin Better Primary end point: Incident tuberculosis by wk 48 Modified intention-to-treat population Secondary end point: Incident tuberculosis by wk 72 Modified intention-to-treat population Prespecified sensitivity analyses of the primary end point Adjusted for prespecified baseline covariates Incident tuberculosis based on site assessment <5 yr of age Per-protocol population Hazard Ratio (95% CI)Analyses 0.125 0.44 (0.15–1.25) 0.49 (0.18–1.30) 0.42 (0.15–1.21) 0.33 (0.11–1.02) 0.54 (0.18–1.60) 0.50 (0.17–1.46) P Value 0.12 Levofloxacin Placebo 5/451 6/451 5/451 4/451 5/403 5/378 12/465 13/465 12/465 13/465 10/430 11/407 no. with end point/total no. The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on January 27, 2025. For personal use only. No other uses without permission. Copyright © 2024 Massachusetts Medical Society. All rights reserved. n engl j med 391;24 nejm.org December 19/26, 2024 2325 Levofloxacin Preventive Ther apy for MDR Tuberculosis may have limited generalizability. Although we recruited index patients and children from sites in South Africa with diverse populations and myco- bacterial strains, more than 90% of the partici- pating children were younger than 5 years of age, with enrollment expanded only later during the trial to selected older children. Care should be taken when extrapolating the findings of our trial to other settings and to older children and adolescents, who may have more nonhousehold exposure to tuberculosis than younger children. Fourth, we could not collect M. tuberculosis strains from all the index patients, and we were thus un- able to assess second-line drug resistance or resis- tance mechanisms. In South Africa, 13% of MDR M. tuberculosis isolates obtained between 2012 and 2014 were resistant to f luoroquinolones, so we might expect levofloxacin preventive treatment to be ineffective in contacts of a similar percentage of index patients.24 In this trial involving children with household exposure to an adult with MDR tuberculosis, tu- berculosis occurred in a smaller percentage of children in the levofloxacin group than in the placebo group. However, this difference was not significant. These results need to be integrated with results from other trials in other populations and settings to inform global policy and practice. Supported by Unitaid through a grant (2019-36-SUN-MDR to Stellenbosch University) from the BENEFIT Kids project; the Joint Global Health Trials program through a grant (MR/ M007340/1 to Stellenbosch University) funded by several U.K. agencies (the Department of Health and Social Care; the Foreign, Commonwealth, and Development Office; the Global Challenges Research Fund; the Medical Research Council [MRC]; and Wellcome) as part of the European and Developing Countries Clinical Trials Partnership 2 program supported by the Euro- pean Union; a grant (to Stellenbosch University) from the South African Medical Research Council; a grant (to Dr. Hesseling) from the South African National Research Foundation; and grants (MC_UU_00004/04 and MC_UU_00004/09 to the MRC Clinical Trials Unit at University College London) from the MRC. Disclosure forms provided by the authors are available with the full text of this article at NEJM.org. A data sharing statement provided by the authors is available with the full text of this article at NEJM.org. We thank all the participating families and communities, es- pecially given the challenges of the coronavirus disease 2019 pandemic in South Africa; the local South African government partners, including local and national tuberculosis services; and the local community advisory boards. Appendix The authors’ full names and academic degrees are as follows: Anneke C. Hesseling, Ph.D., Susan E. Purchase, M.B., Ch.B., Neil A. Martinson, M.B., Ch.B., Lee Fairlie, F.C.P., H. Simon Schaaf, Ph.D., Joanna Brigden, M.Sc., Suzanne Staples, M.B., Ch.B., Diana M. Gibb, M.D., Anthony Garcia-Prats, Ph.D., Francesca Conradie, M.B., Ch.B., Charlotte McGowan, M.Sc., Charlotte Layton, M.Sc., Elize Batist, M.P.H., Anne-Marie Demers, M.D., Samukelisiwe Nyamathe, M.B., Ch.B., Lisa Frigati, Ph.D., Rebecca Turner, Ph.D., Trinh Duong, M.Sc., and James A. Seddon, Ph.D. The authors’ affiliations are as follows: the Department of Pediatrics and Child Health, Desmond Tutu TB Centre, Stellenbosch University, Stellenbosch (A.C.H., S.E.P., H.S.S., A.G.-P., E.B., A.-M.D., S.N., J.A.S., L. Frigati), the Perinatal HIV Research Unit (N.A.M.) and the Faculty of Health Sciences, Wits Research Health Institute (L. Fairlie), University of the Witwatersrand, and Isango Lethemba TB Research Unit, Port Elizabeth, Wits Health Consortium (F.C.), Johannesburg, and the Tuberculosis and HIV Investigative Network, Durban (S.S.) — all in South Africa; the Johns Hopkins Center for TB Research, Baltimore (N.A.M.); the Medical Research Council Clinical Trials Unit at University College London (J.B., D.M.G., C.M., C.L., R.T., T.D.) and the Department of Infectious Disease, Impe- rial College London (J.A.S.) — both in London; the Department of Pediatrics, School of Medicine and Public Health, University of Wisconsin–Madison, Madison (A.G.P.); and the Division of Microbiology, Department of Laboratory Medicine, Centre Hospitalier Universitaire Sainte-Justine, and the Department of Microbiology, Immunology, and Infectious Diseases, Faculty of Medicine, University of Montreal — both in Montreal (A.-M.D.). References 1. World Health Organization. Global tuberculosis report 2022. October 2022 (https://www . who . int/ teams/ global - tuberculosis - programme/ tb - reports/ global - tuberculosis - report - 2022). 2. Laurence YV, Griffiths UK, Vassall A. Costs to health services and the patient of treating tuberculosis: a systematic litera- ture review. Pharmacoeconomics 2015; 33: 939-55. 3. Dodd PJ, Sismanidis C, Seddon JA. Global burden of drug-resistant tuberculo- sis in children: a mathematical modelling study. Lancet Infect Dis 2016; 16: 1193-201. 4. Dye C, Glaziou P, Floyd K, Raviglione M. Prospects for tuberculosis elimina- tion. Annu Rev Public Health 2013; 34: 271-86. 5. World Health Organization. The end TB strategy. August 2015 (https://www . who . int/ publications/ i/ item/ WHO - HTM - TB - 2015 . 19). 6. Huaman MA, Sterling TR. Treatment of latent tuberculosis infection-an update. Clin Chest Med 2019; 40: 839-48. 7. Marks SM, Mase SR, Morris SB. System- atic review, meta-analysis, and cost-effec- tiveness of treatment of latent tuberculosis to reduce progression to multidrug-resistant tuberculosis. Clin Infect Dis 2017; 64: 1670-7. 8. World Health Organization. WHO con- solidated guidelines on tuberculosis: tuber- culosis preventive treatment. 2020 (https:// www . ncbi . nlm . nih . gov/ books/ NBK554956/ pdf/ Bookshelf _ NBK554956 . pdf). 9. Marais BJ, Gie RP, Schaaf HS, et al. The natural history of childhood intra- thoracic tuberculosis: a critical review of literature from the pre-chemotherapy era. Int J Tuberc Lung Dis 2004; 8: 392-402. 10. Martinez L, Cords O, Horsburgh CR, Andrews JR, Pediatric TB Contact Studies The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on January 27, 2025. For personal use only. No other uses without permission. Copyright © 2024 Massachusetts Medical Society. All rights reserved. https://www.who.int/teams/global-tuberculosis-programme/tb-reports/global-tuberculosis-report-2022 https://www.who.int/teams/global-tuberculosis-programme/tb-reports/global-tuberculosis-report-2022 https://www.who.int/teams/global-tuberculosis-programme/tb-reports/global-tuberculosis-report-2022 https://www.who.int/publications/i/item/WHO-HTM-TB-2015.19 https://www.who.int/publications/i/item/WHO-HTM-TB-2015.19 https://www.who.int/publications/i/item/WHO-HTM-TB-2015.19 https://www.ncbi.nlm.nih.gov/books/NBK554956/pdf/Bookshelf_NBK554956.pdf https://www.ncbi.nlm.nih.gov/books/NBK554956/pdf/Bookshelf_NBK554956.pdf https://www.ncbi.nlm.nih.gov/books/NBK554956/pdf/Bookshelf_NBK554956.pdf n engl j med 391;24 nejm.org December 19/26, 20242326 Levofloxacin Preventive Ther apy for MDR Tuberculosis Consortium. The risk of tuberculosis in children after close exposure: a systematic review and individual-participant meta- analysis. Lancet 2020; 395: 973-84. 11. Dodd PJ, Prendergast AJ, Beecroft C, Kampmann B, Seddon JA. The impact of HIV and antiretroviral therapy on TB risk in children: a systematic review and meta- analysis. Thorax 2017; 72: 559-75. 12. Seddon JA, Garcia-Prats AJ, Purchase SE, et al. Levofloxacin versus placebo for the prevention of tuberculosis disease in child contacts of multidrug-resistant tuberculo- sis: study protocol for a phase III cluster randomised controlled trial (TB-CHAMP). Trials 2018; 19: 693. 13. Graham SM, Cuevas LE, Jean-Philippe P, et al. Clinical case definitions for clas- sification of intrathoracic tuberculosis in children: an update. Clin Infect Dis 2015; 61: Suppl 3: S179-S187. 14. Morris TP, Walker AS, Williamson EJ, White IR. Planning a method for covariate adjustment in individually randomised tri- als: a practical guide. Trials 2022; 23: 328. 15. Royston P, Parmar MKB. The use of restricted mean survival time to estimate the treatment effect in randomized clini- cal trials when the proportional hazards assumption is in doubt. Stat Med 2011; 30: 2409-21. 16. Fox GJ, Nguyen CB, Nguyen TA, et al. Levofloxacin versus placebo for the treat- ment of latent tuberculosis among con- tacts of patients with multidrug-resistant tuberculosis (the VQUIN MDR trial): a protocol for a randomised controlled tri- al. BMJ Open 2020; 10: e033945. 17. Seddon JA, Hesseling AC, Godfrey- Faussett P, Fielding K, Schaaf HS. Risk factors for infection and disease in child contacts of multidrug-resistant tubercu- losis: a cross-sectional study. BMC Infect Dis 2013; 13: 392. 18. Malik AA, Gandhi NR, Lash TL, et al. Effectiveness of preventive therapy for per- sons exposed at home to drug-resistant tuberculosis, Karachi, Pakistan. Emerg Infect Dis 2021; 27: 805-12. 19. ClinicalTrial.gov. Protecting house- holds on exposure to newly diagnosed index multidrug-resistant tuberculosis pa- tients (PHOENIx MDR-TB) (https://www . clinicaltrials . gov/ study/ NCT03568383). 20. Fox GJ, Nhung NV, Binh NC, et al. Levofloxacin for the prevention of multi- drug-resistant tuberculosis in Vietnam. N Engl J Med2024; 391:2304-14. 21. Dodd PJ, Mafirakureva N, Seddon JA, McQuaid CF. The global impact of house- hold contact management for children on multidrug-resistant and rifampicin-resis- tant tuberculosis cases, deaths, and health- system costs in 2019: a modelling study. Lancet Glob Health 2022; 10(7): e1034-e1044. 22. Kasaie P, Pennington J, Gupta A, Dowdy DW, Kendall EA. The impact of preventive treatment for multidrug- and rifampin-resistant tuberculosis exceeds trial-based estimates. Clin Infect Dis 2024; 78: 133-43. 23. Croom KF, Goa KL. Levofloxacin: a review of its use in the treatment of bacte- rial infections in the United States. Drugs 2003; 63: 2769-80. 24. Ismail NA, Mvusi L, Nanoo A, et al. Prevalence of drug-resistant tuberculosis and imputed burden in South Africa: a national and sub-national cross-sectional survey. Lancet Infect Dis 2018; 18: 779-87. Copyright © 2024 Massachusetts Medical Society. The New England Journal of Medicine is produced by NEJM Group, a division of the Massachusetts Medical Society. Downloaded from nejm.org on January 27, 2025. For personal use only. No other uses without permission. Copyright © 2024 Massachusetts Medical Society. All rights reserved. https://www.clinicaltrials.gov/study/NCT03568383 https://www.clinicaltrials.gov/study/NCT03568383