Participants in this study were recruited from the Calgary community in Alberta, Canada as part of a larger longitudinal cohort exploring predictors of psychosocial outcomes related to symptoms of ADHD. Eligibility criteria for this larger study included sufficient fluency in English to complete all questionnaires, age 18 or older, and normal or corrected-to-normal hearing and vision. Stroke or dementia were exclusionary; MCI was not. Participants were not required to have ADHD to participate, but all were asked upon enrollment whether they had a confirmed or suspected ADHD diagnosis (yes or no). To receive a formal diagnosis of ADHD in Canada, thorough assessment by a healthcare practitioner is required (only family doctors, pediatricians, psychiatrists, nurse practitioners, and psychologists can provide this service). Practitioners normally apply the Canadian ADHD Practice Guidelines (45), which indicate that the assessing clinician must be licensed and adequately trained in the application of Diagnostic and Statistical Manual Fifth Edition [DSM-5; (46)] diagnostic criteria for ADHD. The patient must meet DSM−5 criteria for ADHD, which require at least five symptoms of inattention or impulsivity/hyperactivity that are longstanding since childhood, cause impairment in multiple domains of functioning, and are not better explained by an alternate or circumstantial condition. The Canadian guidelines further indicate that the clinician must assess the frequency and severity of symptoms and impairment using a comprehensive clinical interview (which must include a complete childhood developmental history and review of past medical records) in combination with valid, reliable and sensitive rating scales and incorporating corroborating reports from knowledgeable informants. The assessment must also reflect an understanding of multi-systemic issues that may confound or complicate the ADHD diagnosis. ADHD diagnosis was not used as a variable of interest in this study's statistical analyses, but this information was collected simply to characterize the sample.

Regardless of their diagnostic status, all participants in this completed ADHD symptom measures, described below, to estimate symptom presence and severity. All participants gave written informed consent to participate, and study procedures were carried out in accordance with the Declaration of Helsinki and were approved by the University of Calgary Conjoint Faculties Research Ethics Board (CFREB#20-1103).

Participants received an email link to complete an online questionnaire, programmed in the Qualtrics survey environment, in which they were asked to provide basic sociodemographic data (age, gender, education level, and ethnic background) and to complete various cognitive and behavioral measures, described below.

All statistical analyses were conducted in SPSS v.26 for Windows. Participants’ sociodemographic characteristics were summarized using descriptive statistics. Participant age, ADHD symptoms and cognitive and emotional complaints were compared between men and women using Student's t tests. Confirmed ADHD diagnosis by a healthcare professional (% yes), Education (levels) and ethnicity (% White) were compared across men and women using chi-square. Two multivariate general linear models were built with ADHD symptoms (CAARS ADHD Index) as continuous covariates (i.e., independent variables) and cognitive (CFQ, BDEFS) and emotional difficulties (DERS) as the dependent variables; both models included an age-by-ADHD symptom interaction term to test the modifying effect of age on ADHD-related outcomes. The model can be formulated as:Y(n×m)=X(n×k+1)B(k+1×m)+E(n×m)where Y is a matrix of n observations on m outcomes (in this case, CFQ and BDEFS scores in one model, and DERS scores in a second model); X is a model matrix for k predictors (in this case, age, CAARS ADHD Index, and the age-by-ADHD Index interaction in both models) plus a regression constant; B is a matrix of the regression coefficients associated with the predictors and the constant; and E is a matrix of errors (67). The models were gender-stratified to compare potential age effects across men and women. Adjusted models controlled for symptoms of depression and anxiety. All model residuals were normally distributed. Normality of residuals was ascertained by visually inspecting the standardized residuals of the predicted values. Homogeneity of variances across men and women could not be calculated directly because the models were gender-stratified (i.e., gender was a not a between-subjects factor) but the spread of dependent variables in both groups was found to be generally comparable by visual inspection of the boxplots.

Table 2 shows a correlation matrix of age and all variables of interest. Correlation coefficients r < .3, reflect weak correlations, those.3 < r < .7 reflect moderate correlations, and those r > .7 are considered strong. In the overall sample, the ADHD Index was significantly correlated with all cognitive (.39 < r < .68) and emotional outcomes (.21 < r < .64). In general, correlations were stronger in men than in women. Age, in particular, was moderately negatively correlated with multiple measures of cognitive and emotional complaints in men, but only weak or non-significant correlations for women.

Table 3 summarizes ADHD symptoms and cognitive and emotional complaints in men and women. Effect sizes are reported as η_p² where values of.01,.06 and.14 indicate weak, moderate and strong effects, respectively. Generally, men and women reported similar levels of ADHD symptom severity, with the exception of inattentive ADHD symptoms which were greater in women (M = 69.57, SD = 11.30) than in men (M = 63.77, SD = 10.80, p = .021, d = 0.52), and overall cognitive failures which were also greater in women (M = 60.35, SD = 15.76) than in men (M = 51.65, SD = 17.24, p = .017, d = 0.54).

In men, age significantly moderated the relationship between ADHD symptoms and BDEFS self-restraint difficulties, whereby the association was stronger in younger than in older men (B = −0.009, p = .021, η_p² = .23). Age was not a significant moderator of this association in women (B = 0.001, p = .415, η_p² < .01) (Figure 1). In women, age only moderated the relationship between ADHD symptom severity and total cognitive failures, which was stronger in younger than in older women (B = −0.017, p = .030, η_p² = .05). Although this interaction was not statistically significant in men, the moderating effect was of similar magnitude (B = −0.025, p = .197, η_p² = .08) (Figure 2) and confidence intervals overlapped with those of women (Table 4). All associations survived adjustments for depression and anxiety symptoms and the magnitude of the effects remained similar. None of the other cognitive outcomes evidenced a moderating effect of age. Adjusted model parameters are summarized in Table 4.

Age significantly moderated the relationship between ADHD symptoms and DERS impulse control difficulties in men, whereby the association was stronger in younger than in older men (B = −0.017, p = .012, η_p² = .28) (Figure 3). This association survived adjustments for depression and anxiety symptoms and the magnitude of the effect remained similar. There was no evidence of age moderation on any of the other emotional outcomes in women or men. Adjusted model parameters are summarized in Table 5.

In men, age significantly moderated the impact of ADHD symptoms on two measures, both of which index impulse control. The BDEFS Self-Restraint subscale refers to one's inability to inhibit reactions or responses, the tendency to make impulsive comments to others, and to act without thinking. The DERS Impulse Control Difficulties subscale describes feeling out of control when upset and being unable to control overwhelming emotional experiences. Thus, both these subscales refer to self-regulatory control of behavior or emotion, processes which are subserved by the prefrontal cortex (68). Although men and women showed comparable mean scores for both measures, the moderating effect of age was only evidence in men: the association between ADHD symptoms and impulse control was weaker in older men than in younger men. Effect sizes were large for both BDEFS and DERS indices of impulse control. In other words, for a given level of ADHD symptoms, older men reported much fewer impulse control difficulties relative to younger men, but this was not true for women.

It is possible to tentatively infer from these results that impulsivity may improve with age in men with ADHD, a finding which would be consistent with other reports of age-related decreases in some forms of impulse control in neurotypical men [e.g., (69)]. Although this can only be confirmed in longitudinal investigations, the possibility of age-related impulse control improvements in ADHD may have optimistic implications for aging in men with ADHD. Impulsivity is associated with many adverse outcomes in ADHD, including increased substance abuse problems and risky sensation seeking (70) as well as occupational, criminal, driving, and financial difficulties (71), and improvements in self-regulatory processes may indicate better outcomes in these domains in later life. Long-term follow-up of men with ADHD would be valuable to ascertain the extent and impact of possible changes in impulsivity.

The picture emerging from the data in the present study is one of stability in women's subjective cognitive and emotional complaints related to ADHD symptoms across age groups. Women of different ages in this sample reported generally similar levels of difficulties, regardless of their ADHD symptom severity. The exception to this pattern was a small- to medium-sized age-related decrease in the association between ADHD symptoms and total CFQ score in women. The CFQ asks about everyday cognitive mistakes related to distractibility, forgetfulness, inattention, and word slips. These cognitive concerns were reported frequently—of a total possible score of 100, women's average score was 60—and were moderately correlated with ADHD symptom severity. Rather surprisingly, CFQ scores were inversely correlated with age for both genders [despite known cognitive declines in normal aging: (72)], and age further moderated the association between ADHD symptoms and CFQ, whereby older women reported slightly but significantly fewer cognitive complaints than younger women with an equivalent level of ADHD symptoms. In men, this trend was also apparent and of comparable moderate magnitude but did not reach statistical significance, likely due to sample size. Previously, there have been other published reports of older adults reporting fewer cognitive difficulties on the CFQ relative to younger adults (73–75). As explanations for these counterintuitive findings, authors have suggested the possibility that older adults may actually experience fewer everyday cognitive slips as a result of reduced demands following retirement and lifestyle changes, or alternatively that they do experience cognitive difficulties but are inaccurate in monitoring and reporting them (73). In the present sample, another possibility could be that older participants have already been experiencing ADHD-related cognitive slips for many decades, and do not identify them as ‘complaints’ but rather as how their brain has always functioned. It is impossible for the present study to adjudicate this question without any objective cognitive measures or informant reports to use as ancillary evidence. However, this question is one that deserves further investigation because it has implications for how age-related cognitive processes are conceptualized in ADHD and the extent to which we can rely on subjective reports of difficulty in aging.

Women's cognitive and emotional challenges linked to ADHD symptoms were otherwise stable across different age groups in the present study. At the outset, it was expected that self-reported difficulties would be worse in older women relative to younger women and to men. This hypothesis was based on known mid-life changes in cognitive and emotional health that arise concurrent to menopause (33, 34), and it was anticipated that existing ADHD-related difficulties would be exacerbated by menopausal reductions in estrogen and dopamine levels (40). The fact that cognitive or emotional challenges were not markedly worse in older women in this sample is relatively more consistent with descriptions of potential menopausal changes as being isolated to the perimenopausal period, rather than persistent and lasting (32). Yet, this finding is in direct contradiction to results from a reader survey of more than 1,500 women conducted by ADDitude Magazine, in which 94% of respondents reported experiencing the most severe ADHD-related impairments of their life during perimenopause and menopause (76). Clearly, much remains to be understood about women's age-related changes in ADHD symptoms and associated impairments linked to menopause, and there is a critical need for further general research about links between ADHD and sex hormones in girls and women.

ADHD symptoms were broadly associated with subjective cognitive and emotional complaints. The lack of any observed age-related increases in cognitive complaints is reassuring from a dementia risk perspective, but the robustness of this finding will need to be replicated in larger samples. It would also be valuable for future work to use objective measures of cognitive processes alongside subjective estimates, as evidence from younger samples suggests that both types of data provide complementary information about cognitive performance (77). It will be useful to establish the validity of this phenomenon in older cohorts.

ADHD symptoms were also significantly associated with participants’ self-reports of nonacceptance of emotional responses, difficulties engaging in goal-directed behavior when upset, difficulties controlling emotional impulses, lack of emotional awareness, difficulties using emotion-regulation strategies, and lack of emotional clarity. None of these challenges—with the exception of emotional impulse control for men—showed any meaningful age-related change in this sample. We may cautiously interpret from this finding the possibility that emotional processes in adults with ADHD do not follow the age-related improvements predicted by socioemotional selectivity theory (27). Indeed, qualitative interviews of adults aged 50+ with ADHD highlight peer rejection and family conflict as the most frequently mentioned challenges by participants (78, 79). Although these challenges do not necessarily imply emotional dysregulation per se, poor friendship and romantic relationship quality in adults with ADHD are significantly predicted by emotion regulation problems (80). Future studies seeking to confirm reduced socioemotional selectivity among older adults with ADHD should employ a range of tasks previously employed in the broader literature, including those that measure the construct implicitly.

This evidence presented here should be considered preliminary in light of several methodological limitations that may be addressed in future work.

First, the cross-sectional design precludes any conclusions about causality and introduces the possibility that results may be driven by cohort bias wherein older men and women may self-report (or under-report) certain kinds of complaints because of generational influences unrelated to aging. While cross-sectional studies are valuable to uncover associations between variables of interest, they only capture a snapshot of the data at one time point and do not provide information about the directionality of these associations. The time point at which data were collected is not necessarily an accurate reflection of the groups’ more general behaviors. Thus, the gender differences observed in this study will need to be tested in cohorts followed over time to robustly ascertain how ADHD-related cognitive and emotional complaints evolve longitudinally as individuals age.

Second, all outcome measures in the present study were self-reported, an approach deliberately taken to explore participants’ subjective difficulties. However, women may be slightly more likely to endorse ADHD-related impairment (81), and it is possible that age-related “decreases” in cognitive and emotional impulsivity in men actually reflect lower self-disclosure in older men relative to women. It is also possible that people of different ages may have different biases in how they perceive their own symptoms. Older adults, for example, may downplay their symptoms as a result of an age-related positivity bias (26) or because of memory changes that compromise their retrospective estimates of symptoms or behaviors (6). Data collected from informants (e.g., family members) may have provided different results from those observed in the present study.

Third, various clinical presentations unrelated to ADHD can include features of inattention, hyperactivity and impulsivity, as well as dysfunction in memory, executive abilities and emotional regulation. These features are nonspecific to ADHD and it is therefore possible that pathological processes unrelated to ADHD may have driven some of the results reported here. Even so, we estimate that this is not likely to be a significant threat to the findings given that most participants reported having received a formal diagnosis of ADHD from their healthcare provider (71%) or scored above the ASRS clinical threshold (78%) that is 94% specific to ADHD. Nonetheless, the ASRS alone cannot be used to confirm an ADHD diagnosis, and we cannot definitively confirm that Canadian ADHD Practice Guidelines were rigorously applied in all cases where a formal diagnosis was self-reported. Future work should seek to confirm the observed associations in a sample with confirmed ADHD diagnosis.

Finally, results are interpreted cautiously in the context of relatively small sample size, particularly with regard to male participants. This sample was underpowered to explore age effects across specific developmental stages, but future work in this regard would provide a richer, more nuanced perspective on ADHD outcomes across the lifespan.

Some cognitive and emotional difficulties associated with ADHD symptoms showed age-related declines in men but not women, tentatively suggesting that women with ADHD may experience greater challenges as they age relative to men with ADHD. Altogether, results support the need for a gender-specific lens when considering the lifespan impacts of ADHD symptoms and potentially point to women with ADHD as an especially vulnerable segment of the population regarding cognitive and emotional health.