To identify the most accurate method of predicting peak height velocity in adolescents.Systematic review.A comprehensive literature search of six electronic databases and reference lists was conducted. Studies that met selection criteria of (1) observational longitudinal cohort study (2) reproducible method/s of predicting peak height velocity (3) minimum six-month follow-up (4) healthy male and/or female adolescent subjects, with the exception of participants with adolescent idiopathic scoliosis, were considered for review. Studies were screened using a modified quality assessment checklist, with only those scoring >50% included. The type of surrogate measure of peak height velocity, its reliability and ability to predict peak height velocity were extracted from the year or stage immediately preceding peak height velocity. We defined "predict" as when both the estimates of effect and 95% confidence intervals of the surrogate occurred prior to the actual age of PHV.The nine included studies examined three anthropometric, three equation and four radiographic-based surrogates for PHV. Of these, the radiographic measures were reported to exhibit moderate to high intra- and inter-rater reliability. Three of the four radiographic surrogates predicted PHV. Two anthropometric measures also predicted PHV but reliability of the measures is unknown. All equation-based methods predicted the timing of PHV to occur later than it actually happened when applied in the year prior to expected PHV.In the year/stage immediately preceding peak height velocity, radiograph-based methods appear to be accurate and reliable surrogates.