This study did not find more severe ADHD characteristics to be associated with more severe ED symptoms in women diagnosed with ED. Although about one fourth of the sample presented at least subsyndromal ADHD, neither the severity of inattention nor the severity of hyperactivity was associated with any of the ED symptoms’ severity. The only statistically significant correlation (when controlling for inflated type I error due to multiple testing in the sample) was found between the degree of impulsivity and avoidance of fattening food. In addition to the preponderance of these small and non-significant correlations between the severity of ADHD and ED features, the observed associations were often counterintuitive; positive correlations were found where negative correlations were expected and vice versa. For instance, higher impulsivity was not only statistically significant related to the avoidance of fattening food, but it also tended to be positively correlated with limiting caloric intake, dieting, and excessive fasting. These behaviors all require strict discipline and thus seem not to be very compatible with impulsivity. These preliminary findings, though unexpected, could elucidate an important peculiarity of ED: the extreme control, which is a central symptom of AN patients (e.g., when encountering fattening food) could be interpreted as an answer to a primary impulsivity, and explain the frequent diagnostic cross over from AN to BN [16–20]. Although we usually associate impulsivity with BN, a high prevalence of ADHD was also found in a sample with mainly AN subjects .
Whereas our findings suggest that impulsivity might be increased in ED patients compared to healthy controls [2, 3], it does not appear to be a predictor of the severity of eating-related impulsive behavior such as binge eating or purging, though one would expect impulsivity to abet sudden decisions to binge and purge. Interestingly, such lack of an association between impulsivity and ED symptom severity has already been reported in previous studies among BN patients [27, 28].
Given the examination of few (N = 32) patients in our study, one could suppose that the lack of significant correlations is related to the low statistical power in such a small sample. However, the correlation coefficients, which (unlike significance tests) constitute a measure that is independent of the sample size, in general indicated only small associations between the severity of ADHD and ED characteristics (most r’s < 0.30). For instance, the highest correlation between an indicator of impulsivity (the BIS-11 sum score) and a measure of binge eating (the severity of binge eating episodes) equaled r = 0.218. That is, knowing BIS-11 impulsivity scores enables explanation of less than 5% (r
= 0.2182 = 0.048) of the variance of the severity of binge eating episodes. It is therefore unlikely that impulsivity scores simply reflect the presence or absence of impulsive eating patterns. In line with this, Davies et al.  did not find higher BIS total scores in BN patients when compared to AN patients. To provide yet another example, there was also no substantial association between hyperactivity and excessive exercise (r = 0.137), which was also contrary to our expectations.
Considering these unexpected findings, one could argue that adopting a “transdiagnostic” view on ED–and hence aggregating results across different specific ED diagnoses–may have masked associations between ADHD features and ED symptoms in specific ED types. Correlational analyses within the specific ED subtypes in our sample (AN-restrictive type, AN-binge/purging type, BN-binge/purging type, and EDNOS), however, did not provide any indication of ADHD-ED symptom severity relations in specific ED subtypes. Furthermore, though longitudinal diagnostic flux between specific ED diagnoses is frequent and recurrent, the overreaching diagnosis of ED and the characteristics of ADHD are both relatively stable over time [17–19]. If there were striking associations between ADHD features and the severity of ED symptoms, one might expect them to be revealed even in a diagnostically mixed sample.
In summary, the findings of this study indicate that the associations between the severity of ADHD features and the severity of specific ED symptoms are low in women diagnosed with ED. This is in line with some previous studies that failed to find increased rates of ADHD in ED [14, 15]. And even if there should be an increased prevalence of ADHD in ED individuals [5–7, 13], the relation between these two disorders might be less straightforward and more complex than expected based on clinical features shared by both disorders. Therefore, regarding clinical implications, if ADHD key features are salient in ED, both symptomatologies seem to require equal clinical attention. Focusing primarily on ADHD symptoms and expecting that ED symptoms will be automatically alleviated by such intervention might result in disappointing outcomes. Several case reports suggested that methylphenidate and other psychostimulants may be effective in reducing BN symptoms in patients with comorbid ADHD and BN [43–46]. These case reports should be interpreted with caution since adequate rating scales to evaluate improvement in bulimic symptoms were not used in the majority of them. Some of the reports included patients who were diagnosed not only with BN and ADHD, but with cluster B personality disorders as well. In our sample only two participants had such personality disorder. Therefore, the lack of an association between ADHD features and the severity of ED symptoms observed in our study calls into question the promise of interventions targeting on ADHD features as a treatment for ED symptomatology.
Among the limitations of this study are the small sample size and the aggregation across specific ED, both of which were discussed earlier. The problem of the relatively small number of patients in this study is accentuated when performing analyses within specific ED subtypes. Thus, our null-findings resulting from subgroup analyses should be treated with caution. They maybe just reflect the need for larger samples. Further research is needed to examine associations between the severity of specific ADHD and ED symptoms in ED subtypes.
As a further potential limitation of this study, we used self-report questionnaires (ADHD-SR and BIS-11) to assess ADHD characteristics. The Structured Clinical Interview for DSM-IV Disorders, which is the most common instrument to asses mental disorders, does not encompass ADHD. To date there is no standardized method for identifying subjects with ADHD , and existing interviews to assess neuropsychiatric disorders with onset during childhood are extremely time-consuming . Nevertheless, clinicians’ ratings of the severity of ADHD core features might have resulted in different findings than the self-report questionnaires used in this study. Based on information from the ADHD-SD, none of our participants was diagnoses with an ADHD when applying the strict research criteria (an ADHD symptom was considered to be present if its intensity was rated at least “moderate” by the participant). However, when using the less strict clinical criteria (a symptom is already assumed to be present if it is rated at least “mild (occasionally present)” ), the proportion of ADHD diagnoses in our ED sample increased to 29%. Thus, at least subsyndromal ADHD was present in many patients of our sample . Finally, normative data of the ADHD-SR questionnaire is not yet available for healthy controls. So it remains unclear whether our ED patients reported increased scores on the ADHD-SR scales compared to healthy controls. Future research on associations between ADHD and ED would benefit from efficient and thoroughly validated instruments to assess ADHD.