Supplementation with high-dose EPA/DHA concentrate resulted in significant modifications of fatty acids, particularly a significant improvement in the AA:EPA ratio in the isolated plasma phospholipids and improvements in behavior assessed by a psychiatrist (blinded to protocol adherence and supplement dosage adjustments) in this small pilot sample of children with ADHD.
At baseline fatty acid analysis of the isolated plasma phospholipids from the children in this study were similar to that of previous studies of children with ADHD and thirst/skin symptoms of EFA deficiency [3, 5, 11]; however, we did not assess EFA deficiency symptoms. Children with ADHD and thirst/skin symptoms of EFA deficiency had lower AA and DHA levels in the plasma phospholipids compared to control groups. Both the AA and DHA mean levels from previous studies [3, 5, 11] were within the 95% CI (8.98–10.05; 1.63–2.97, respectively) of this study's mean AA and DHA levels.
Supplementation of high-dose EPA/DHA concentrates resulted in marked changes in fatty acid levels of the isolated plasma phospholipids. EPA and DHA levels in the isolated plasma phospholipids were used to monitor compliance. We chose the AA:EPA ratio as an important marker because of it's relationship with depression , as depression is often associated with ADHD . In this study, there was indeed a significant positive correlation between the AA:EPA ratio and severity of illness.
Although the EPA and DHA supplementation dosages used in this study were high compared to previous studies with children, there was no serious adverse effect except one case of loose stools that was corrected with a lower dose. Young et al.  supplemented adults with ADHD with high-dose EPA/DHA concentrates (approximately 36 g EPA and DHA per day) with no reported serious adverse effects other than loose stools and fishy burps. The average AA:EPA ratio after 12 weeks of the high-dose EPA/DHA supplementation in adults with ADHD was 1.4 ± 0.6 ; however, behavior was not assessed in this study.
Stevens et al.  supplemented children with ADHD and thirst/skin symptoms with 480 mg DHA, 80 mg EPA, 40 mg AA and 60 mg GLA per day. At these levels, the AA:EPA ratio was reduced to15.19 after four months , which remains 2.5 times greater than the mean AA:EPA ratio obtained in this study. Stevens et al.  did monitor behavior and found improvements in conduct assessed by the parents and attention assessed by the teachers in the PUFA group compared to olive oil placebo. When assessed clinically, the parental rating scales were also evaluated based on diagnostic criteria, and a significant PUFA treatment effect was reported for oppositional/defiant disorder. The findings by Stevens et al.  supports our data in that we also found improvements in oppositional behaviors rated by the parents and improvements in both oppositional/defiant behaviors and conduct assessed clinically by the psychiatrist, however, a psychiatrist did not assess behavior in Stevens et al.'s study.
This study found a statistically significant improvement in the psychiatrist's report of inattention, hyperactivity, oppositional/defiant behavior and conduct disorder based on the ADHD SC-4 questionnaire. Scores for inattention, hyperactivity and oppositional/defiant behavior continued to improve from week four to eight, even with the EPA/DHA concentrate dosage adjustment. The dosage adjustment, however, did not bring the AA:EPA ratio above 3 suggesting the importance of monitoring fatty acids and the AA:EPA ratio in particular rather than EPA/DHA dosage alone. The severity of illness scale demonstrated a positive improvement from an average of moderately symptomatic to mildly symptomatic. This improvement was similar regardless of medication use or lack there of. The percent change in severity of illness also correlated with percent decrease in the AA:EPA ratio, suggesting a connection between the clinical improvement observed by the psychiatrist and the improvements in the AA:EPA ratio.
Data from Stevens et al.  in children and Young et al.  in adults with ADHD suggest that greater amounts of both EPA and DHA may be required to decrease the AA:EPA ratio to between 1.5 and 3. The mean AA:EPA ratio at the end of this study was 5.95 ± 7.35 for all participants. When the two participants who were non-compliant were removed, the AA:EPA ratio was 2.52 ± 2.91, suggesting a daily dose between 8.1 g and 16.2 g of EPA/DHA concentrate may be appropriate to decrease the AA:EPA ratio to between 1.5 and 3 and to observe improvements in behavior in children with ADHD.
There are a number of limitations to this pilot study and therefore interpretation of results requires caution. The study is limited in that there was no placebo group for reference comparisons as this was a pilot study to determine appropriate dosage for protocol adherence and to maintain AA:EPA levels between 1.5 and 3. Dietary intake was not recorded at baseline or monitored throughout the study; therefore, we are unable to decipher intake of fatty acids from the diet. Also related to diet, we advised the children to eat more fruits and vegetables and consume meals and snacks that are balanced with protein, carbohydrates (preferably fruit and vegetables) and "healthy" monounsaturated fats. Advice for following both a "healthy diet" and high-dose fish oil supplementation may have been confounding factors. However, the dose-response relationship between percent change in AA:EPA ratio and the reduction in the severity of ADHD suggest the behavioral changes were due to, at least in part, the intake of high-dose EPA/DHA concentrates. The lack of behavioral change or regression to pre-study status in those subjects who were least compliant to supplementation also suggest that behavioral changes were associated with intake of the LC omega-3 fatty acids.
EPA/DHA concentrate dosage adjustments themselves can be viewed as a limitation since some, but not all participants' daily intake dosage was modified at week four. The supplement intervention adjustment was based on the AA:EPA ratio, therefore those whose AA:EPA ratio dropped below the goal range was adjusted upward and by using this ratio as our goal, we also avoided most adverse events. The lack of a proper means to monitor supplement intake, such as weight of returned bottles, was also a limitation of this study. However, this was compensated for by use of isolated plasma phospholipids levels as a means to monitor protocol adherence.