Objectives:The aim of this study was to assess the level and correlates of biomarker-confirmed adherence to isoniazid (INH) preventive therapy (IPT) among children with HIV (CLHIV).Design:This prospective cohort study assessed adherence among CLHIV on IPT in public sector HIV clinics from 2019 through 2020.Methods:Adherence was assessed by pill counts or caregiver or self-reports, and urine biomarkers (in-house dipstick and Isoscreen). Both urine biomarker tests detect INH metabolites within 48 h of ingestion. Consistent adherence was defined as having positive results on either biomarker at all visits. Correlates of biomarker-confirmed nonadherence at each visit were evaluated using generalized estimating equations. The in-house dipstick was validated using Isoscreen as the reference.Results:Among 97 CLHIV on IPT with adherence assessments, median age was 10 years (IQR 7–13). All were on ART at IPT initiation (median duration 46 months [IQR 4–89]); 81% were virally suppressed (<1000 copies/ml). At all visits, 59% (n = 57) of CLHIV reported taking at least 80% of their doses, while 39% (n = 38) had biomarker-confirmed adherence. Viral nonsuppression (adjusted risk ratio [aRR] = 1.65; 95% confidence interval [95% CI] 1.09–2.49) and the sixth month of IPT use (aRR = 2.49; 95% CI 1.34–4.65) were independent correlates of biomarker-confirmed nonadherence at each visit. Sensitivity and specificity of the in-house dipstick were 98.1% (94.7–99.6%)and 94.7% (88.1–98.3%), respectively, versus Isoscreen.Conclusion:Biomarker-confirmed adherence to IPT was sub-optimal and was associated with viral nonsuppression and duration of IPT. Urine dipstick testing may be useful in assessing adherence to IPT in clinical care.