Subgroup considerations Children and adolescents: Levofloxacin can be used in children and adolescents, in whom completion and tolerability in the TB CHAMP trial (which included only individuals aged < 18 years) were much better than in the V-QUIN trial (in which 97% of participants were aged ≥ 15 years). There is no requirement to test for TBI before starting levofloxacin in children who are contacts of people with MDR/RR-TB. Although there has been concern about use of fluoroquinolones in children because of retardation of cartilage development shown in juvenile animals exposed to these agents (119), similar effects have not been found in humans (120,121). Pregnancy and breastfeeding: TPT with levofloxacin in pregnancy requires a risk–benefit assessment and an informed choice by pregnant woman on whether to take TPT or to defer TPT to the end of pregnancy. The advice should depend on the circumstances (e.g. first trimester versus later). Pregnancy increases the risk of progression from infection to disease and the risk of poor maternal and fetal outcomes should TB disease occur. MDR/RR-TB in pregnancy is a serious condition, and some of the drugs used to treat MDR-TB may be toxic to the fetus. Observations from studies in animals exposed to levofloxacin have limited its use in pregnancy; however, one meta-analysis of observational studies with 2800 pregnant women given fluoroquinolones for any indication (e.g. urinary tract infection) found no difference in the incidence of birth defects, spontaneous abortion or prematurity from that in unexposed pregnant women (122). The concentrations of levofloxacin in breastmilk appeared to be far lower than the dose for infants and would not be expected to cause adverse effects in breastfed infants (123). Its use should therefore not be suspended during breastfeeding. While effects of fluoroquinolones on bone and cartilage observed in animals have not been seen in humans, the data and follow-up times of infants are limited. Recent alerts have, however, highlighted safety concerns associated with prolonged use of fluoroquinolones in humans (124–126). HIV infection: Levofloxacin can be used in people with HIV. No specific drug–drug interaction with ART has been observed in people with HIV exposed to MDR/RR-TB, and there is no need to test for infection before starting levofloxacin. Contraindication: Levofloxacin should not be given to people who are allergic to fluoroquinolone, who have another contraindication to the same class of drugs or when there is potential drug–drug interaction. Levofloxacin should be discontinued if the person develops a serious or severe adverse drug reaction. (See below for other TPT regimen options in such a case.) Implementation considerations The strong recommendation reflects the GDG opinion that the benefits of levofloxacin outweigh the potential harm in most people who are eligible to receive it. Health programmes and clinicians should strictly ensure eligibility for its use, maximize the likelihood of treatment completion as expected and ensure that contacts are followed up regardless of whether TPT was completed. Contacts of people with RR-TB are usually treated as for MDR-TB, unless susceptibility to isoniazid is reliably confirmed in the index person, in which case isoniazid may be considered an effective TPT option. The GDG considered that levofloxacin could be used in any setting, regardless of TB burden, provided that the health infrastructure can ensure that treatment is given correctly without creating inequity, and that TB disease can be excluded reliably before initiation of treatment. Levofloxacin is widely available as a generic drug, in both adult and paediatric formulations. As for other TPT, the GDG noted that treatment can be self-administered and that a requirement for direct observation could be a significant barrier to implementation. Digital adherence technologies (e.g. electronic medication monitors) may be used, but few studies have been conducted on their use for TPT. The GDG noted that the 6-month duration of levofloxacin treatment may appear long to patients and caregivers when compared with the shorter, 4- or 12-week TPT regimens that are now available for prevention of drug-susceptible TB. People receiving TPT should also be provided with advice on treatment and management of adverse events. Levofloxacin is the preferred fluoroquinolone for use in TPT, and it was used in both trials. Instructions on dosage are provided in the WHO operational handbook on TPT (12). While there are no comparable data on alternatives, moxifloxacin can be used if levofloxacin is not available. Drug-susceptibility testing of the source case strain would provide important additional information, especially in situations where fluoroquinolone resistance is known to be high. If the strain in the source patient is resistant to these medicines, other TB drugs (e.g. ethionamide, ethambutol) can be used as TPT according to the best available information on the drug susceptibility profile of the presumed strain. In this case, the certainty of the effectiveness of TPT is much lower than with levofloxacin (see also below). A positive test for TBI before starting TPT for MDR/RR-TB is not required for child contacts or people with immunocompromising conditions. In other populations, this would be desirable but not mandatory. Lack of availability of testing should not be a barrier to providing TPT to individuals who are at risk. Screening of all household and other close contacts for co-prevalent TB disease will be important. The approach to screening and ruling out TB in contacts is otherwise no different from that described earlier (see section 1.2). Provision of TPT with levofloxacin should include consideration of factors such as age, risk of toxicity or interaction, co-morbidity, the susceptibility to drugs of the strain of the most likely source case, background resistance to fluoroquinolones in MDR/RR-TB strains, availability and the individual’s preferences. The capacity of a programme to provide TPT for MDR/RR-TB should be carefully planned to ensure that all the necessary resources are in place, including programme capacity to rule out TB disease, perform quality-assured testing for drug susceptibility in the presumed source case, deliver the necessary medications and closely monitor adverse events and emergence of TB disease. Engagement of stakeholders in the community is important, as for other means to address constraints to implementation. A paediatric formulation of levofloxacin can be used. Instructions on dosage are provided in the WHO operational handbook on TPT (12). If fluoroquinolones cannot be used because of intolerance or resistance in the strain from the presumed source case, treatment with the other TB drugs used in some studies may be considered (e.g. ethambutol, ethionamide), although the evidence for their efficacy is much less certain (127,128). While ethambutol is considered safe in pregnancy, ethionamide has been associated with teratogenic potential at high doses in experimental animals, although there are minimal data on human pregnancy. There is limited evidence for the optimal duration of MDR-TB preventive treatment, which should be based on clinical judgement. In the studies conducted so far, levofloxacin was given for 6, 9 or 12 months. None of studies included studies of pharmacokinetics or safety in pregnancy or a comparison of risks for adverse events, although one reported no serious adverse events attributable to fluoroquinolone-based preventive treatment (104).