BACKGROUND: For human immunodeficiency virus (HIV)-infected adolescents facing lifelong antiretroviral therapy (ART), short-cycle therapy (SCT) with long-acting agents offers the potential for drug-free weekends, less toxicity, better adherence and cost savings. OBJECTIVES: To determine whether or not efavirenz (EFV)-based ART in short cycles of 5 days on and 2 days off is as efficacious (in maintaining virological suppression) as continuous EFV-based ART (continuous therapy; CT). Secondary objectives included the occurrence of new clinical HIV events or death, changes in immunological status, emergence of HIV drug resistance, drug toxicity and changes in therapy. DESIGN: Open, randomised, non-inferiority trial. SETTING: Europe, Thailand, Uganda, Argentina and the USA. PARTICIPANTS: Young people (aged 8-24 years) on EFV plus two nucleoside reverse transcriptase inhibitors and with a HIV-1 ribonucleic acid level [viral load (VL)] of < 50 copies/ml for > 12 months. INTERVENTIONS: Young people were randomised to continue daily ART (CT) or change to SCT (5 days on, 2 days off ART). MAIN OUTCOME MEASURES: Follow-up was for a minimum of 48 weeks (0, 4 and 12 weeks and then 12-weekly visits). The primary outcome was the difference between arms in the proportion with VL > 50 copies/ml (confirmed) by 48 weeks, estimated using the Kaplan-Meier method (12% non-inferiority margin) adjusted for region and age. RESULTS: In total, 199 young people (11 countries) were randomised (n = 99 SCT group, n = 100 CT group) and followed for a median of 86 weeks. Overall, 53% were male; the median age was 14 years (21% ≥ 18 years); 13% were from the UK, 56% were black, 19% were Asian and 21% were Caucasian; and the median CD4% and CD4 count were 34% and 735 cells/mm(3), respectively. By week 48, only one participant (CT) was lost to follow-up. The SCT arm had a 27% decreased drug exposure as measured by the adherence questionnaire and a MEMSCap(™) Medication Event Monitoring System (MEMSCap Inc., Durham, NC, USA) substudy (median cap openings per week: SCT group, n = 5; CT group, n = 7). By 48 weeks, six participants in the SCT group and seven in the CT group had a confirmed VL > 50 copies/ml [difference -1.2%, 90% confidence interval (CI) -7.3% to 4.9%] and two in the SCT group and four in the CT group had a confirmed VL > 400 copies/ml (difference -2.1%, 90% CI -6.2% to 1.9%). All six participants in the SCT group with a VL > 50 copies/ml resumed daily ART, of whom five were resuppressed, three were on the same regimen and two with a switch; two others on SCT resumed daily ART for other reasons. Overall, three participants in the SCT group and nine in the CT group (p = 0.1) changed ART regimen, five because of toxicity, four for simplification reasons, two because of compliance issues and one because of VL failure. Seven young people (SCT group, n = 2; CT group, n = 5) had major non-nucleoside reverse transcriptase inhibitor mutations at VL failure, of whom two (n = 1 SCT group, n = 1 CT group) had the M184V mutation. Two young people had new Centers for Disease Control B events (SCT group, n = 1; CT group, n = 1). There were no significant differences between SCT and CT in grade 3/4 adverse events (13 vs. 14) or in serious adverse events (7 vs. 6); there were fewer ART-related adverse events in the SCT arm (2 vs. 14; p = 0.02). At week 48 there was no evidence that SCT led to increased inflammation using an extensive panel of markers. Young people expressed a strong preference for SCT in a qualitative substudy and in pre- and post-trial questionnaires. In total, 98% of the young people are taking part in a 2-year follow-up extension of the trial. CONCLUSIONS: Non-inferiority of VL suppression in young people on EFV-based first-line ART with a VL of < 50 copies/ml was demonstrated for SCT compared with CT, with similar resistance, safety and inflammatory marker profiles. The SCT group had fewer ART-related adverse events. Further evaluation of the immunological and virological impact of SCT is ongoing. A limitation of the trial is that the results cannot be generalised to settings where VL monitoring is either not available or infrequent, nor to use of low-dose EFV. Two-year extended follow-up of the trial is ongoing to confirm the durability of the SCT strategy. Further trials of SCT in settings with infrequent VL monitoring and with other antiretroviral drugs such as tenofovir alafenamide, which has a long intracellular half-life, and/or dolutegravir, which has a higher barrier to resistance, are planned. TRIAL REGISTRATION: Current Controlled Trials ISRCTN97755073; EUDRACT 2009-012947-40; and CTA 27505/0005/001-0001. FUNDING: This project was funded by the National Institute for Health Research (NIHR) Health Technology Assessment programme (projects 08/53/25 and 11/136/108), the European Commission through EuroCoord (FP7/2007/2015), the Economic and Social Research Council, the PENTA Foundation, the Medical Research Council and INSERM SC10-US19, France, and will be published in full in Health Technology Assessment; Vol. 20, No. 49. See the NIHR Journals Library website for further project information.