Objectives: This systematic review examines the agreement between assessed skeletal age by the Greulich and Pyle atlas (GP skeletal age) and chronological age.
Methods: We searched electronic databases until January 2017 for studies reporting GP skeletal age and confirmed chronological age in healthy individuals aged 10-25 years. Results are presented as forest plots and meta-analyses (random-effects models).
Results: In separate meta-analyses for each age group and sex (14-18 years for girls, 14-19 years for boys), the pooled mean differences between GP skeletal age and chronological age varied from -0.52 years to 0.47 years. In individual studies, age group and sex-specific mean differences between GP skeletal age and chronological age rarely exceeded 1 year, but between-study heterogeneities were large in most age groups. Few studies examined mean chronological age and distribution for each GP skeletal age. One study of good methodological quality indicates that 95% prediction intervals for chronological age from given GP skeletal ages are typically around 4 years.
Conclusions: There is still good correlation between GP skeletal age and mean chronological age in modern populations. However, the individual variation of development within a population and heterogeneities between studies are substantial.
Key points: • The GP atlas still corresponds well with mean chronological age in modern populations. • The substantial variation within a population must be considered. • The heterogeneity between studies is relatively large and of unknown origin.
Keywords: Age determination by skeleton; Forensic sciences; Hand bones; Meta-analysis.