SummaryObjective Growth hormone (GH) and insulin-like growth factor I (IGF-I) are the principal biomarkers used to assess disease activity in acromegaly, and any discrepancy between them renders interpretation of results inconclusive. Purpose of this study was to assess the frequency of this discrepancy and identify parameters that might affect its occurrence. Design A systematic review of MEDLINE and Scopus was performed followed by a meta-analysis to address the frequency of discrepant results between GH and IGF-I levels. Meta-regression and subgroup analyses were performed assessing the effects of the year of publication, the different types of GH testing and GH assays used, as well as the impact of treatment with somatostatin analogues (SSAs) on the occurrence of this discrepancy. Results The analysis retrieved 39 eligible studies totalling 7071 patients. The pooled discordance rate between GH and IGF-I was 25Á7% (95% CI: 22Á3-29Á4), and the predominant format was that of elevated IGF-I with normal GH levels (15Á3%, 95% CI: 12Á5-18Á7). No significant correlation between the discordance rate and the year of publication was shown; whereas, the use of ultrasensitive GH assays resulted in higher discordance rates (30Á7%, 95% CI: 25Á9-35Á9 vs 19Á8%, 95% CI: 14Á1-27Á2, P = 0Á04) as did treatment with SSAs (32Á5%, 95% CI: 27Á8-37Á4) vs (21Á6%, 95% CI: 17Á8-25Á6, P = 0Á001).Conclusions Discrepancy between GH and IGF-I results is encountered in a quarter of treated patients with acromegaly, especially when using ultrasensitive GH assays or in patients receiving SSAs, a fact that the clinician should take into consideration when making clinical decisions.