Attention deficit hyperactivity and oppositional defiant disorder symptoms in adolescence and risk of substance use disorders—A general population‐based birth cohort study

Abstract Background Externalizing symptoms are associated with risk of future substance use disorder (SUD). Few longitudinal studies exist using general population‐based samples which assess the spectrum of attention deficit hyperactivity disorder (ADHD) and oppositional defiant disorder (ODD) symptoms. Aims/Objectives We aimed to study the associations between adolescent ADHD symptoms and subsequent SUD and additionally examine whether the risk of SUD is influenced by comorbid oppositional defiant disorder (ODD) symptoms. Methods The Northern Finland Birth Cohort 1986 was linked to nationwide health care register data for incident SUD diagnoses until age 33 years (n = 6278, 49.5% male). ADHD/ODD‐case status at age 16 years was defined using parent‐rated ADHD indicated by Strengths and Weaknesses of ADHD symptoms and Normal Behaviors (SWAN) questionnaire with 95% percentile cut‐off. To assess the impact of ODD comorbidity on SUD risk, participants were categorized into four groups based on their ADHD/ODD case status. Cox‐regression analysis with hazard ratios (HRs) and 95% confidence intervals (CIs) were used to study associations between adolescent ADHD/ODD case statuses and subsequent SUD. Results In all, 552 participants (8.8%) presented with ADHD case status at the age of 16 years, and 154/6278 (2.5%) were diagnosed with SUD during the follow‐up. ADHD case status was associated with SUD during the follow‐up (HR = 3.84, 95% CI 2.69–5.50). After adjustments for sex, family structure, and parental psychiatric disorder and early substance use the association with ADHD case status and SUD remained statistically significant (HR = 2.60, 95% CI 1.70–3.98). The risk of SUD remained elevated in individuals with ADHD case status irrespective of ODD symptoms. Conclusions ADHD in adolescence was associated with incident SUD in those with and without symptoms of ODD. The association of ADHD and SUD persisted even after adjustment for a wide range of potential confounds. This emphasizes the need to identify preventative strategies for adolescents with ADHD so as to improve health outcomes.


| INTRODUCTION
Attention Deficit Hyperactivity Disorder (ADHD) is a neurodevelopmental disorder featuring symptoms of inattention, hyperactivity, and impulsivity with onset prior to age of 12. 1 The community prevalence of ADHD is 6%-7% in youths 2 and some individuals with ADHD continue to display fluctuating levels of ADHD symptoms and impairments in social functioning in young adulthood. 3 Past research in longitudinal samples consistently report associations between childhood/adolescent ADHD and early use of licit or illicit substances [4][5][6] and substance use disorders (SUDs) 4,7-14 substantiated by meta-analyses. [15][16][17][18] However, these data are typically based on clinically diagnosed samples or are relatively small (except for 12,13 ). There are very few community-based studies that have assessed the full spectrum of ADHD symptoms and include participants who have not been clinically diagnosed but have nonetheless elevated symptoms. This is important as ADHD symptoms are associated with significant impairment even at an undiagnosed or subthreshold level. 19,20 Thus, there remains a gap in knowledge concerning persons with elevated ADHD symptoms but who may not have been included in clinical registers.
Whether comorbidity with disruptive behavior disorders (DBD) including conduct disorder (CD) / oppositional defiant disorder (ODD) may account for or have an impact

Significant Outcomes
• ADHD case status in adolescence was associated with risk of incident SUD. • This association was independent of multiple confounders such as adolescent substance use and sex. • The association of ADHD case status and SUD persisted in individuals with and without symptoms of ODD.

Limitations
• Only 3.1% of the sample was diagnosed with SUD which is likely to be underestimate. • We used parent-rated ADHD symptom data with cut-offs to indicate ADHD case status which does not constitute a clinical diagnosis of ADHD. • Possible role of childhood or familial adversity could not be accounted for.
on substance use trajectories with ADHD, remains a subject for debate. To our knowledge, two meta-analyses have addressed this issue with inconsistent findings. The meta-analysis by Serra-Pinheiro et al. 21 concluded that ADHD alone was not sufficient to increase the risk of illicit substance use/SUD beyond the effect of CD/ODD. 21 However, even with the combined outcome of both illicit substance use and SUD, the authors acknowledged there were power issues that limited interpretation of their findings. In contrast, the more recent meta-analysis by Groenman et al. 17 reported that comorbid ODD/CD with ADHD did not influence the association with SUD. 17 However, the studies included in the meta-analysis typically focused on the influence of comorbid CD with ADHD and the evidence base for the effect of ODD is weak. Past research, which is limited to clinical samples, suggests that comorbid ODD with ADHD is not associated with an increased risk for developing SUD. 9,22 However, the association remains unexamined in community samples, which are less affected by selection bias than clinical samples (e.g., access to care, severity of symptoms, social-economic factors). Thus, there is a need for large studies to elucidate whether comorbid ODD with ADHD influences the risk of SUD further. Furthermore, there are only few studies 8,23 that have been able examine the natural course of ADHD and SUD in an unselected way, for example, by using birth cohort sample. Using this method, all consenting participants from the community are assessed for ADHD and ODD symptoms. Northern Finland Birth Cohort 1986 (NFBC 1986) is a large prospective general population birth cohort, 24 where individual data for the cohort members are linked to several nationwide registers for all lifetime psychiatric diagnoses. Data used in this study gives opportunity to consider a number of potential confounders such as sex, family structure, parental psychiatric disorders and adolescent substance use when studying the complex associations between adolescent ADHD/ ODD and subsequent SUD. Furthermore, NFBC 1986 (over 6000 at follow-up) is large compared to the previous Dunedin, 23 and Christchurch Health and Development Studies 8 that include up to 1200 individuals at enrolment. The substantially larger sample size of this dataset provides more certainty and power to the analysis.
In this sample ADHD / ODD are defined as probable case statuses based on the parent-rated The Strengths and Weaknesses of ADHD symptoms and Normal Behaviors (SWAN) questionnaire. Using this birth cohort data our aim was to investigate: (1) whether ADHD case status in adolescence is associated with clinical SUD diagnosis up until age 33 years and (2) whether ODD influences the association between ADHD case status and SUD diagnosis.

| MATERIAL AND METHODS
NFBC 1986 is an ongoing multidisciplinary birth cohort study comprising 99% of all live-born children (n = 9432) with an expected date of birth between July 1, 1985 and June 30,1986, from the two northernmost provinces in Finland. 25 Parents and offspring have been followed-up in regular intervals including clinical studies at child ages 7-8 years and 15-16 years old. This study concerns the data collected in 2001-2002 when study members were aged 15-16 years. The data collection for the adolescent follow-up entailed participants and their parents with known addresses. They were sent self-report questionnaires in separate envelopes (n = 9215). Adolescents answered questions concerning their physical health and psychosocial wellbeing (n = 7344) and substance use for those participating in the clinical study (n = 6798) that took place after the initial postal questionnaire. Parents reported family factors and adolescent ADHD symptoms (n = 6985). Participants and parents who signed the informed consent form and with data available on ADHD symptoms were included in the analyses. Figure 1 shows the data flow and the final sample available for analysis (N = 6278) after exclusion of those who had been diagnosed with any prior psychiatric disorder (ICD-10: F00-F99) before the age of 16 to account for baseline psychiatric comorbidity.
NFBC 1986 study is approved by the Ethics committee of the Northern Ostrobothnia Hospital District in Finland with latest version dated on January 15, 2018 (EETTMK 108/2017). The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.

| Outcome variable: substance use disorder
Participants' data were linked to register-based substance use disorder (SUD) diagnoses (F1x.x) from age of 16 years until the end of 2018 when participants were aged 33 years. The onset age of each disorder was based on the first record of the diagnosis in registers. SUD diagnoses were obtained from linkage to four nationwide registers providing extensive coverage on diagnosed psychiatric disorders and only minimal attrition.

| Exposure variable: ADHD and ODD symptoms in adolescence reported by parent
The Strengths and Weaknesses of ADHD symptoms and Normal Behaviors (SWAN) questionnaire 29,30 is a revised version of SNAP-IV developed by Swanson and his colleagues. SWAN measures problems in attention, hyperactivity/impulsivity, and disruptive behavior. 29,30 This study used a SWAN version where the items are based on the 18 ADHD symptoms and 8 ODD symptoms described in the DSM-IV-TR for example "Gives close attention to detail and avoids careless mistakes." Parents rate these items on a 7-point scale anchored to average behavior (i.e., Far Below Average = 3, Below Average = 2, Somewhat Below Average = 1, Average = 0, Somewhat Above Average = À1, Above Average = À2, and Far Above Average = À3) resulting in normally distributed behavioral ratings. Individual SWAN items were used to compute means for a Combined ADHD scale score (C), using all 18 items, an Inattentive score (I) using 9 Inattentive items, and a Hyperactive-Impulsive score (HI) using 9 Hyperactive-Impulsive items and Oppositional defiant disorder scale (ODD) using 8 Oppositional defiant disorder items. The SWAN scores were used to identify probable ADHD "cases" by requiring SWAN scores to exceed the >95th percentile on either C or I or HI scales. 31 Participants who scored below the 95th percentile on all three ADHD scales were classified as "controls." We excluded participant data if there were two or more items missing in one of the SWAN scales. If there was only one item missing in any SWAN scale, the missing value was replaced by the mean value of the items of that particular scale for that person. To study the effect of comorbid ODD, we defined ODD case status as exceeding 95th percentile of the symptom distribution in SWAN ODD. The data on ADHD / ODD case statuses were then categorized into four categories (1) Positive ADHD and ODD case statuses, (2) Positive ADHD case status only, (3) Positive ODD case status only, and (4) controls (participants below 95% percentile in each of these categories). For the questionnaire used in this study, see NFBC website: https://www.oulu.fi/nfbc/node/18149.

| Covariates
Early risk factors for substance use disorders have been previously described in the literature [32][33][34] ; these variables have also been found to associate with ADHD. 35,36 Therefore, to clarify the association between ADHD during To assess lifetime cannabis use until follow-up at age 16 years, adolescents were asked: "Have you used marihuana or hashish?" with options "never," "once," "two to four times," "five or more times," or "I use regularly." These options were pooled as "lifetime cannabis use" yes (=1) or no (=0) to provide sufficient sample size.
Adolescents were also asked "Have you tried or used any of the following substances?-Ecstasy, heroin, cocaine, amphetamine, LSD or other similar intoxicating T A B L E 1 Association of covariates and ADHD case status and incident substance use disorder (SUD) in Northern Finland Birth Cohort 1986. drugs?" and "Have you ever tried or used any of the following substances?-Sniffing thinner, glue, etc. for intoxication" with options "never," "once," "two to four times," "five or more times," or "I use regularly." The data on other illicit substances other than cannabis and use of inhalants were pooled into a binary variable "other lifetime substance use" yes (=1) or no (=0) to provide sufficient sample size.

| Family structure
Information on family structure was collected by combining data from parents at birth and from the clinical study in [2001][2002]. These data were categorized as "family with two biological parents (=0)," where both biological parents lived together with the participant, and "other (=1)," which consisted of all other family types.

| Statistical methods
The statistical analyses were performed using SPSS statistical software (IBM SPSS Statistics, version 28; IBM Co., Armonk, New York, USA) and R (R Foundation for Statistical Computing, version 4.2.0; R Core Team., Armonk, Vienna, Austria) packages Epi and survival. Association of between covariates, ADHD case status and incident SUD diagnosis was assessed with chi-square test. The association ADHD case status and risk of SUD was examined using Cox regression analysis with hazard ratios (HR) and 95% confidence intervals (95% CI). Times at emigration outside the country (n = 247) and death (n = 60) were used as censoring points. Statistical significance was defined as p ≤ 0.05. We used the following models to adjust for potential confounders: Model 1 crude; Model 2: sex, family structure and any parental psychiatric disorder; Model 3: additionally, lifetime cannabis use, other lifetime substance use, frequent alcohol intoxication. Furthermore, we studied risk of SUD in different ADHD/ODD case status categories by using Cox-regression analysis. The Aalen-Johansen cumulative incidence curves were computed for the Cox-regression models. Linear regression and multicollinearity diagnostics with variance inflation factor (VIF) scores were used to examine possible correlation between multiple covariates (variables in Model 3). We considered known attrition due to nonparticipation in this sample. 37 Fewer males (64% vs. 71%; p < 0.001), individuals living in urban areas (66% vs. 71%, p < 0.001) and individuals with parental psychiatric disorder (58% vs. 69%, p < 0.001) were less likely to participate in the 15-16-year follow-up study. 37 We addressed this attrition by weighing our main analyses by sex, parental psychiatric disorder, and urbanicity by using inverse probability weighting 38 and analyzed these data with logistic regression analysis and odds ratios (OR) ( Figure S1).
To assess the stability of our results, we conducted a set of sensitivity analyses. First, we used in Cox-regression analysis with hazard ratios (HR) and 95% CI in Model 3 without restricting the sample for participants psychiatric disorder prior age 16 years ( Figure S1). Second, we studied the association of ADHD case status and SUD in Models 1-3 using SWAN 90th percentile as cut-off ( Figure S1), which reflects symptoms at a sub-threshold level.

| RESULTS
The final sample for analysis included 6278 (49.5% male) individuals as shown in Figure 1. Of these individuals F I G U R E 2 The Aalen-Johansen cumulative incidence curve for the association between ADHD case status and substance use disorder. 552 (8.8%) presented with ADHD case status at the age of 16 years. During the follow-up from 2001 to 2018, that is, from age 16 to 33 years, 154/6278 (2.5%) individuals were diagnosed with SUD.
Those reaching case status for ADHD at 16 years were more likely to be male, come from families where both biological parents did not live with the participant, and were also more likely to live in families with parental psychiatric disorder. Furthermore, they were also more likely to report lifetime substance use and frequent alcohol intoxication past year and to be diagnosed with incident SUD compared to those who did not (Table 1).
Cumulative incidence curve for the association between ADHD case status and SUD is presented in Figure 2. In the crude analyses (Model 1) ADHD case status was associated with SUD during the follow-up at a statistically significant level (HR = 3.84, 95% CI 2.69-5.50). After adjustments for sex, family structure, and parental psychiatric disorder (Model 2), the association of ADHD case status and SUD attenuated but remained statistically significant (HR = 2.97, 95% CI 2.04-4.32). The association attenuated further but remained still statistically significant (HR = 2.60, 95% CI 1.70-3.98) after adjustment for frequent alcohol intoxication past year and lifetime cannabis and other substance use until adolescence (Model 3, see Table 2). In these final models, frequent alcohol intoxication, lifetime cannabis use and other substance use, family structure, parental psychiatric disorder, and male sex were also associated with SUD (Table S1). The results of our primary analyses aligned with our sensitivity analysis that were not restricted for baseline psychiatric diagnoses or use of the SWAN 90% cutoff when defining ADHD case status ( Figure S1). Furthermore, all the statistically significant ORs in unweighted analyses were also statistically significant in the weighted analyses, and the strength of the associations were of similar magnitude (see Figure S1). Multicollinearity was not seen (Model 3, all VIFs <1.2).
Complete SWAN ADHD and ODD data were available for 6165 individuals of which 152 (2.5%) were diagnosed with SUD during the follow-up. There were 188 individuals (3.0% of the sample) individuals with ADHD +/ODD+ case status, 355 individuals (5.7% of the sample) with ADHD+/ODD-case status, 130 individuals T A B L E 2 Association of ADHD case status and risk of substance use disorder (SUD). F I G U R E 3 The Aalen-Johansen cumulative incidence curve for the association between ADHD/ODD case status categories and substance use disorder.

| DISCUSSION
Using general population-based sample of Northern Finland Birth Cohort 1986, we report an association of parent-rated ADHD based on probable SWAN caseness and risk of SUD diagnosed in clinical practice until 33 years of age. This association was independent of sex, family structure through early life, parental psychiatric disorders, and adolescent substance use. Furthermore, this association with SUD was seen in individuals reaching ADHD case status irrespective of ODD symptoms. This study adds new knowledge pertaining to a general population sample and points to ADHD and ODD at the probable case level in adolescence to be associated with SUD up until the age of 33 years. Furthermore, our sensitivity analyses suggested that individuals with ADHD-like traits at the low and high extreme end of the distribution are at risk of SUD. Use of register data for all cohort participants were available with the benefit of long-term follow-up without confounding with ADHD/ ODD case status. Our findings align with the previous reports of positive associations of ADHD and SUD in clinical studies, 9,10 register and twin samples 4,7,12,13 and bolster previous cohort findings. 8,11 This study was able to assess individual and family level confounding due to sex, parental psychiatric disorders, family structure and adolescent substance use. These are all prominent risk factors for SUD that also associate with ADHD, and so, could potentially explain away the association between ADHD and SUD. In our analysis, the effect sizes attenuated by increasing adjustments as expected. Yet, ADHD case status remained at more than twofold risk for incident SUD. This suggests that association of ADHD and SUD was not explained by familial or individual confounding including adolescent substance use. The latter is particularly important as it is known that ADHD increases the risk of early use of licit and illicit substances [4][5][6] and this in turn is a strong predictor for future SUD. 39 Thus, these findings suggest ADHD may confer to the risk of SUD even without the influence of adolescent substance use and other prominent risk factors.
We report that cohort members with both ADHD/ ODD case statuses had the greatest point estimates and cumulative incidences of SUD during the follow-up. Yet, ADHD and ODD case statuses contributed to the risk of future SUD also individually with confidence intervals overlapping between these three groups. Thus, our findings suggest that individuals with ADHD are at risk of SUD irrespective of comorbid ODD symptoms. In this respect our findings are consistent with the most recent meta-analysis that the association of ADHD and SUD was not dependent on CD / ODD. 17 However, the novelty of this study is that we were able to focus on comorbidity with ODD using data from more than 6000 participants in contrast to the previous individual studies that are mostly based on smaller clinical samples that have typically focused on CD.
Our results have clinical implications. Based on our findings and literature to date ADHD increases the risk of SUD. Past research suggests that this comorbidity associates with further risks for general medical conditions, mortality, and disability burden. [40][41][42][43] There is evidence that stimulant treatment for ADHD mitigates the development of SUD especially when treatment is initiated early and for longer durations. 44 Further, children who had been diagnosed with ADHD during childhood are shown to have lower risks for SUD in early adulthood than those who were diagnosed with ADHD during adolescence or early adulthood (5.2% vs. 9.4% vs. 14.3%) suggesting delayed diagnoses of ADHD may increase the risk for SUD due to delayed treatment onset. 45 Backed by this evidence, our results suggest that it is vital to identify and treat ADHD to prevent or mitigate the progression of substance use to a clinical SUD. Our results further suggest that emphasis should be also placed on individuals with externalizing comorbidity such as ODD, yet the risk of SUD remains elevated in ADHD and ODD beyond their comorbidity.
The NFBC 1986 study is among the largest birth cohort studies in the world with considerable follow-up that allows for robust examination of prospective associations into adulthood. This study was able to examine the associations of ADHD/ODD symptoms and SUD by combining questionnaire data to national registers for ICD-10 diagnosis codes with a prospective general populationbased design with 18 years of follow-up. Furthermore, the extensive dataset allowed us to evaluate the association while considering several potential confounders.
There are also limitations. Information on substance use during adolescence was collected retrospectively via self-report questionnaire at age 15-16 years, which may potentially have been underreported. Although SWAN cut-offs are used as probable ADHD diagnosis, it does not constitute a full psychiatric clinical assessment.
Nonetheless, this definition sheds light on the impact of elevated ADHD symptoms regardless of clinical status. This study identified 8.8% of the sample with ADHD case status which is somewhat higher than the community prevalence of clinical ADHD (5.9%-7.2%) in a previous meta-analysis. 2 However, studies with no requirement of impairment tend to have 2.3% higher prevalence estimates that those with requirement. 46 As the ADHDÀ/ ODD+ category was small with only 130 individuals the effect size reported in this paper may be an underestimate. We did not have data on CD for the cohort and thus we were not able examine whether comorbid CD in ADHD influences SUD risk further. Despite substantial sample size, it was not possible to stratify our analyses by sex due to power issues. Furthermore, 3.1% of the whole study population was diagnosed with SUD which is likely to be an underestimate. 47 Possible reasons are underdiagnosis and the fact that part of the substance use services are administered in social services which does not contribute data to the registers used in this study. Lastly, possible role of childhood or familial adversity could not be accounted for. To conclude, ADHD case status in adolescence associated with incident SUD independent of sex, demographic factors, parental psychiatric disorder, and early substance use in the general population-based Northern Finland Birth Cohort with 18 years of follow-up. Moreover, the risk of SUD remained elevated in individuals reaching ADHD case status whether or not they had ODD symptoms. ADHD is a serious neurodevelopmental disorder requiring optimal management to prevent future SUD and other serious health problems.