This study examined the persistence of attention deficit hyperactivity disorder (ADHD) into adulthood. We analyzed data from published follow-up studies of ADHD. To be included in the analysis, these additional studies had to meet the following criteria: the study included a control group and it was clear from the methods if the diagnosis of ADHD included subjects who did not meet full criteria but showed residual and impairing signs of the disorder. We used a meta-analysis regression model to separately assess the syndromatic and symptomatic persistence of ADHD. When we define only those meeting full criteria for ADHD as having 'persistent ADHD', the rate of persistence is low, approximately 15% at age 25 years. But when we include cases consistent with DSM-IV's definition of ADHD in partial remission, the rate of persistence is much higher, approximately 65%. Our results show that estimates of ADHD's persistence rely heavily on how one defines persistence. Yet, regardless of definition, our analyses show that evidence for ADHD lessens with age. More work is needed to determine if this reflects true remission of ADHD symptoms or is due to the developmental insensitivity of diagnostic criteria for the disorder.