Background
Although it is now well established that individuals with neurodevelopmental disorders often meet diagnostic criteria for more than one disorder, this fact has generally not been reflected in longitudinal outcome studies of attention-deficit/hyperactivity disorder (ADHD) in adulthood [
1]. Outcome studies have commonly focused on ADHD severity and measures of intelligence, but overlooked other important co-occurring symptoms and neurodevelopmental diagnoses that may contribute to prognosis. Thorough neurodevelopmental assessments will almost always detect impairing symptoms from multiple disorders, more so with repeated assessments across the life span [
2]. Such clinical observations have been corroborated by family- twin- and genetic studies wherein a wide array of neurodevelopmental symptoms co-aggregate [
3‐
5].
With a view to emphasizing this coexistence of disorders, the concept of ESSENCE was launched by Gillberg [
2]. ESSENCE, the acronym for Early Symptomatic Syndromes Eliciting Neurodevelopmental Clinical Examinations, embraces all kinds and severities of neurodevelopmental disorders (NDD), e.g., intellectual disability, autism spectrum disorder, ADHD, developmental coordination disorder (DCD) language disorders, specific learning disorders, Tourette syndrome and early onset epilepsies. The ESSENCE concept highlights the importance of considering all functional areas and describing the individual’s total amount of functional deficits and disabilities. The concept also emphasizes the need for etiological concerns. This holistic view would create the best possible conditions for follow-up and intervention.
Most long-term outcome studies of ADHD report an association with lower academic and occupational achievement, psychiatric comorbidity, substance use/abuse, risk of accidents and relationship problems [
1,
6]. Putative predictors for long-term outcome include IQ, conduct problems, symptom severity and parental factors such as psychopathology and parental practices [
6]. These factors (or predictors) were also highlighted in a meta-analysis of ADHD-persistence into adulthood by Faraone et al. [
7], wherein eight studies followed participants into their twenties, and only one study into their thirties. Besides the relatively short follow-up time beyond early adulthood, included studies have been further limited by clinic-based sampling bias and by failure to account for effects of comorbid problems (e.g. autism spectrum disorder was considered an exclusion criterion for ADHD according to DSM-III and IV).
DCD is characterized by a delay in the development of gross and fine motor skills, poor motor planning and coordination, resulting in difficulties or inability to acquire common, everyday skills dependent on integration of executive, cognitive and emotional processes [
8].
Although recognition of DCD varies with diagnostic practices across countries, when ascertained it occurs in a severe form in 5% of children [
9,
10], and can be considered a marker of attention deficits [
9,
11,
12], autistic traits [
9,
12] and poorer reading skills [
10,
13], as well as of depressive symptoms [
11] persisting into teenage years.
Among very few studies reporting on the trajectory of DCD into adulthood, a recent prospective cohort study of women with autism and/or ADHD revealed that one in four had a history of DCD in addition to ADHD and/or ASD. In this group chronic pain was reported by 77% [
14]. Aligned with studies engaging patients with a chief complaint of pain, both retrospective studies of adults [
15], and cross-sectional studies of children [
16] report increased rates of neurodevelopmental disorder symptoms of which ADHD has been most commonly assessed.
Aims of the study
In addition to describing the outcome in adulthood of ADHD with coexisting DCD, the aim was also to prospectively test whether an association to pain of unknown etiology and prescription of analgesic medications can be replicated.
Discussion
In this prospective cohort study of long-term outcomes, the diagnosis of ADHD+DCD made after a general population screening at school entry was associated with increased rates of sick pension, welfare dependence, psychiatric disorder, and prescription of psychotropic medications in adulthood. At the time of the initial study a diagnosis in the study did not confer referral to care, because specific interventions for ADHD+DCD (e.g. parental programs, stimulant treatment) were almost non-existent in the area. Hence, because only 25% of those with ADHD+DCD received a formal clinically registered diagnosis of NDD (e.g. ADHD or ASD) in primary pediatric or psychiatric care services, and a minority (10%) was prescribed stimulants after 2005, we consider the study to mainly portray the natural course of the “syndrome”.
In comparisons, results generally exhibited a uniform direction of increase or decrease, with the PM group at the better end, RM in between, and ADHD+DCD group at the poorer end. When the uniform direction was present but differences were statistically insignificant, this may likely have been due to insufficient power rather than being a true null-result.
In terms of education and occupational proficiency, the ADHD+DCD group exhibited a shift toward lower attainment. This is consistent with a recent Danish report of completed educational grades in final high-school examinations, wherein ADHD was associated with both fewer completed grades and lower grades (Cohens
d − 0.62) [
25]. Out of 629,622 children in the Danish cohort, 38,001 (6%) had a mental disorder, and 4% had ADHD. In contrast, DCD, with an expected prevalence of 5%, was diagnosed in 0.02% of children. Although participants had co-existing disorders diagnosed (number of diagnoses is almost twice as many as the number of participants), the role of co-occurring impairments was not accounted for, nor discussed. While the large registry data set provides power and statistical precision, the study by Dalsgaard et al. also highlights two important limitations of registry-based studies. First, relying on regular care diagnoses of neurodevelopmental disorders confer systematic bias due to poor recognition of many conditions, as illustrated by the low rate of DCD in their cohort. Second, this limitation makes inferences of the relative contribution from neurodevelopmental disorders on outcome unreliable. Thus, there is still a need for in-depth clinical studies to better gauge the role of co-existing factors.
The proportion reported to have no occupational skill (ADHD+DCD 16%, PM 6%, RM 4%, Fig.
2) could be due to only working on short-term contracts (for which subjects are not reported as permanently hired and visible in the registry), but is in light of low rates in the comparison groups most parsimoniously explained by
absence of any significant work experience. Further, participants differ strikingly in terms of average recipiency of welfare benefits in adulthood, where median sums are almost 10 times higher in the ADHD+DCD group compared to the PM group. In contrast, the occurrence of welfare recipiency in the last year of follow-up showed no meaningful difference (Table
1). Therefore, rather than the occurrence per se, it is the weight of impairment and need for welfare support
over time that singles out the ADHD+DCD group. This can be conceived as adult endpoints of the childhood impairments apparent in academic, sports and everyday tasks associated with DCD [
10,
13].
Groups did not differ meaningfully in rates of accidents or hospitalizations, (Table
1, Kaplan-Meier curve in
Additional file), contrary to previous studies of ADHD showing elevated risks for accidents [
35]. We think that the phenotype with predominately attention deficit and coexisting DCD more likely subsume to a sedentary lifestyle, in which impulsivity is not a hallmark. This is in line with childhood studies of DCD, wherein participants are at risk of obesity and non-participation in vigorous physical activity [
36].
Rates of medication prescriptions in the comparison groups are on par with a recently reported national Swedish male conscript cohort of similar age span (
N = 414,595) [
24]. On the contrary, the ADHD+DCD group had significantly more prescriptions of psychotropic medications, and nominally higher rates of prescriptions for virtually all subclasses (Table
1, Kaplan-Meier curves in
Additional file). Besides stimulants, the increase was most pronounced in the class of antidepressants and anxiolytics, drugs with a primary indication of anxiety and depression. Subtracting the somewhat lower rates of
registered diagnoses (register did not cover primary care) of affective disorders from these prescriptions (Table
1), it can be inferred that participants with ADHD+DCD in adulthood were diagnosed with anxiety and depression in primary care. This is in line with previous reports on an association between DCD and affective symptoms in early adolescence and suggests persistence into adulthood [
11]. The rates of pain diagnoses and analgesic prescriptions did not differ significantly between groups, although the ADHD+DCD group had the highest rates of both. A longer follow-up period may give a more comprehensive picture. For instance, fibromyalgia was diagnosed at a mean age of 37 years in the study by van Rensburg et al. [
15]. Additionally, there may be sex differences in the magnitude of this association. Asztély et al. reported chronic pain in 77% of adult women with ADHD and/or ASD [
14], and both of these studies predominately included women, whereas only 19% were women in our study.
When comparing results to the only previous follow-up study of ADHD+DCD conducted at 22 years of age [
22], differences across groups are similar, but less pronounced in this study. For example, rates of criminal offences and indices of substance use did not differ meaningfully across groups in this study, whereas it was evident in their cohort, together with even more pronounced differences in educational attainment [
22]. Because drug use in Sweden is less common in rural settings, we hypothesize that county differences (i.e. previous study conducted in an urban area, and the current in rural municipalities) may partly explain the lower rates of substance use in our cohort [
37]. Rates of sick pension were similar, as were the proportion with no poor outcome in the composite measure (around 40%). This is consistent with the notion that a substantial minority of those with ADHD as children will have few symptoms and little impairment as adults [
1]. Barbaresi et al. reported on adult outcome of a population-based cohort with and without ADHD, recruited by retrospective case ascertainment through medical records [
38]. Of 367 participants, 63% (
n = 232) participated in a clinical follow-up at a mean age of 27 years. Similarly, no disorder was diagnosed at adult follow-up in 38% of participants with ADHD in childhood. ADHD persisted in 38%, of which 57% had a co-occurring psychiatric disorder (in descending frequency, alcohol abuse, antisocial personality disorder, substance abuse, hypomanic episodes, anxiety and depression), compared with 35% of controls. However, differences in measurement methods both at study inclusion (screening and clinical assessment vs retrospective case ascertainment) and follow-up (registry data vs clinical assessment) precludes definite conclusions, especially regarding psychiatric characteristics. Positive predictive value of childhood NDD:s in the NPR is generally good (> 80%) in recent years [
39,
40]. But the discrepancy between registered diagnoses of ADHD in the RM group (0.3%) and the expected rates reported in research studies of ADHD in adulthood (~ 3%) indicates that NDD:s for the period at study largely were undetected in clinical practice, and a low negative predictive value for adulthood NDD:s in the NPR for this time period [
41]. In light of aforementioned limitations with reliance on registry data only, we think our results provide a minimum level description of adverse outcome for the group with ADHD+DCD.
Strengths of the study includes the population-based sampling, providing the full panorama of the condition at study, the long follow-up time, valid comparison groups and low attrition rate. However, the study is somewhat limited by the modest sample size of the index cohort with ADHD+DCD. This did not allow for meaningful analyses subgrouping ADHD with or without DCD, and made inferences on rare, but important events (such as severe criminality, suicide or death) impossible. Absence of face-to-face assessments precludes information from subtle clinical findings or repeated neuropsychiatric assessments, physical exam and the experiences reported from participants themselves. Apart from prescription medications, data sources are lacking on the study populations utilization of primary care. Ideally, longitudinal studies combine registry data and clinical assessments for comprehensive profiling of participants.
In conclusion, ADHD+DCD diagnosed at school-entry was associated with a less favorable outcome in early adulthood compared both to matched peers without NDD and a registry-drawn group from the same county. The objective nature of neuromotor function tests opens up the possibility that children at risk of adverse outcome in adulthood might be detected prior to school-entry and receive proper support early in education.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.