Introduction
Attention-deficit/hyperactivity disorder (ADHD) is a common childhood psychiatric disorder that is defined by developmentally inappropriate levels of inattention, and/or hyperactivity and impulsivity [
2]. Previously identified risk factors for ADHD include both genetic and environmental factors. However, the genetic risk factors identified thus far explain only a small percentage of the heritability [for extensive reviews see:
27,
50]. Environmental risk factors for ADHD have shown to have more substantial effect sizes than genetic factors [
27], and may have more immediate relevance for clinical treatment. While some environmental factors can be used to provide early identification of at-risk individuals, others can possibly be counteracted by education, training, or interventions. Environmental factors can be subdivided into pre- and perinatal factors, transgenerational influences and postnatal factors. Well-documented pre- and perinatal factors for ADHD include premature birth, low birth weight and maternal smoking during pregnancy [
32]. Low birth weight is one of the most investigated and consistently reported risk factors for ADHD, and might even (partly) explain the association between maternal smoking during pregnancy and ADHD [
39,
41]. Well-documented transgenerational influences and postnatal risk factors include a family history of ADHD and higher levels of family conflict [
36,
49], although a family history of ADHD is likely to comprise both environmental and genetic influences.
A condition which is frequently comorbid with ADHD is oppositional defiant disorder (ODD), occurring in up to 60% of individuals with ADHD [
14,
21]. ODD is defined by a frequent and persistent pattern of irritable and angry mood, vindictiveness, and developmentally inappropriate, negativistic, defiant, and disobedient behaviour toward authority figures [
2]. Individuals with both ADHD and ODD have a considerably worse prognosis than individuals with either one of the disorders in terms of an increased risk to develop anxiety and depressive disorders as well as conduct disorder and even antisocial personality disorder later in life [
4,
35]. Furthermore, the comorbid group shows an earlier onset with more functional impairments and exhibits more physical aggression and delinquency than individuals with ADHD or ODD alone [
4,
34,
35]. This emphasises the need to not only identify risk factors for ADHD, but especially for ODD comorbid with ADHD. The identification of these factors can contribute to the development of early preventive interventions.
Compared with ADHD, relatively few studies have investigated risk factors for comorbid ODD in ADHD. Reported risk factors for ODD, which are arguably also implicated in the development of comorbid ODD, include both risk factors overlapping with those reported for ADHD and risk factors specific for ODD. Overlapping risk factors for ODD and ADHD encompass maternal smoking during pregnancy, a family history of ADHD or ODD, and higher levels of family conflict [
6,
36,
39]. Specific risk factors for ODD, compared with ADHD, include deviant peer affiliation, harsh or inconsistent parenting, low levels of parental affection, and exposure to family violence [
6,
36,
43]. Studies into specific risk factors for comorbid ODD in ADHD have mainly focused on transgenerational influences, such as parental psychopathology and parenting styles, and reported significant associations of those factors with ODD, rather than with ADHD [for reviews see:
23,
37]. The relative paucity of studies that investigated other environmental risk factors for the development of comorbid ODD in individuals with ADHD is remarkable, given the high prevalence of this comorbid condition.
The aim of the current study was to investigate potential risk factors for the development of comorbid ODD in individuals with ADHD (ADHD + ODD), and to identify whether risk factors differed for individuals with ADHD-only and individuals with ADHD + ODD. To this end, we assessed pre- and perinatal risk factors (pregnancy duration, birth weight, maternal smoking during pregnancy), transgenerational influences (parental ADHD, for ADHD-only and ADHD + ODD parental warmth and parental criticism as well), and postnatal risk factors (socioeconomic status [SES], adverse life events, deviant peer affiliations) in three groups: ADHD + ODD, ADHD-only, and typically developing controls. All groups were matched for age and gender, and the diagnostic groups were additionally matched for IQ and ADHD-subtype. We hypothesised that for both ADHD + ODD and ADHD-only (compared with controls), pre- and perinatal adversities and negative transgenerational influences would be risk factors [
6,
33]. In differentiating between ADHD + ODD and ADHD-only, we hypothesised that postnatal risk factors would be more strongly related to ADHD + ODD [
1,
23]. Finally, we expected less parental warmth and more parental criticism to be predictive for ADHD + ODD group membership, compared with ADHD-only [
23,
43].
Discussion
The aim of the current study was to investigate risk factors for the development of comorbid ODD along with ADHD. Therefore, we assessed pre- and perinatal factors, transgenerational influences and postnatal factors. We hypothesised that pre- and perinatal adversities and negative transgenerational influences would act as risk factors for both ADHD and ADHD + ODD, and postnatal adversities to act primarily as risk factors for ADHD + ODD compared with ADHD-only [
1,
6,
23,
33]. Additionally, in differentiating between ADHD-only and ADHD + ODD, we hypothesised that postnatal adversities and negative transgenerational influences would be more strongly related to ADHD + ODD than to ADHD-only [
43]. Our models identified several risk factors for ADHD + ODD and for ADHD-only, compared with controls, with high levels of explained deviance of 55.2 and 62.5%, respectively. Our model for risk factors differentiating between ADHD + ODD and ADHD-only showed an explained deviance of almost 15.3%. All three models showed good sensitivity (90–98%), and the models for the control group versus both the ADHD + ODD and the ADHD-only groups also showed good specificity (80–87%). We found no interaction between age and any of the risk factors, indicating that predictors are equally important during all stages of development in our sample, and independent of age of the participants (age 7–24 years).
Our first hypothesis that negative transgenerational influences and pre- and perinatal adversities would act as risk factors for the diagnostic groups was supported by our findings, since we found that parental ADHD acted as a relatively major risk factor within our models, showing the highest explained deviance for both diagnostic groups relative to the control group. This is in line with the many studies showing significant heritability rates for ADHD [
10,
22], and large effects of environmental influences associated with parental ADHD on the development of ADHD in the child [
8]. In terms of pre- and perinatal adversities, maternal smoking during pregnancy acted as a relatively minor risk factor within our model for ADHD-only, while
higher (not lower) birth weight acted as a relatively minor risk factor for ADHD + ODD, relative to controls. This supports the notion that there may be an optimum birth weight in terms of the development of behavioural problems such as ADHD, as previously suggested in other studies [
17,
25,
54]. We were not able to replicate previously reported findings of lower birth weight or pregnancy duration as risk factors for ADHD [
50], which may be due to the small number of individuals with a low birth weight or premature birth in our sample (8 and 12, respectively). To conclude, our findings show parental ADHD as a significant risk factor and suggest that the relationship between birth weight and pregnancy duration and the development of ADHD might only hold true for values below a certain threshold.
We also found support for our second hypothesis of postnatal adversities acting as risk factors for ADHD + ODD rather than for ADHD-only. For both ADHD-only and ADHD + ODD, adverse life events, which included parental divorce and family conflicts, acted as a risk factor. However, adverse life events acted as a stronger risk factor for ADHD + ODD than for ADHD-only, as stressed by its differentiating ability between the ADHD + ODD and ADHD-only groups. The mechanism by which adverse life events may affect ODD is still unclear, and may vary between types of event; potential explanations include (a) negative effects on maturation of cerebral brain structures in the child due to stress, (b) teaching individuals to use antisocial strategies to cope with stressful situations, and (c) causing an overactive sympathetic nervous system [
7,
29]. All these factors have been implicated in the development of ODD and receive extensive support, suggesting a combination of these risk factors to operate in ODD [
7,
29]. While no other risk factors were observed for the development of ADHD-only, more deviant peer affiliations and lower SES did act as additional risk factors for ADHD + ODD, compared with controls. This is consistent with previous studies showing that more deviant peer affiliations reinforce an individual’s own antisocial behaviours [
18,
30,
48]. SES acted as a relatively minor risk factor within our model (1.5–5.6%), presumably exerting its effect through poor parenting and deviant socialisation processes that are associated with lower parental SES [
42]. The relatively weak effect of SES may be due to its relationship to parental ADHD, which was also included in the model (e.g. lower parental mental health has been associated with lower SES) [
45]. Since both deviant peer affiliations and SES differentiated between ADHD + ODD and ADHD-only, these risk factors seem especially important for the development of comorbid ODD.
Our third hypothesis, that transgenerational influences in addition to the postnatal factors would differentiate between ADHD + ODD and ADHD-only, was largely supported by our results. Parental criticism acted as a relatively strong risk factor for ADHD + ODD compared with ADHD-only, within our model. This is in line with previous studies and is presumably due to its negative influence on the child’s socialisation process [
8,
23,
37]. In addition, it has been reported that child difficulty not only increases the likelihood of maternal negative parenting, but also that maternal negative parenting heightens the child’s behavioural maladjustment that may take the form of ODD behaviours [
11]. This is in line with the coercion theory that describes a process of mutual reinforcement between the parent and child in the development of conduct problems. According to this model the parent inadvertently reinforces the child’s difficult behaviour by reacting negatively to that behaviour and therewith escalating the situation [
47]. Hence, negative parental attitudes are risk factors not only for ADHD [
23], but especially for comorbid ODD. Furthermore, when parents express low levels of support, the negative influences of deviant peers on the development of oppositional behaviour increase [
26,
53]. Against our hypothesis and contradicting previous studies, parental ADHD acted as a minor protective factor for the development of comorbid ODD in ADHD [
23]. A possible explanation may be that children with comorbid ODD are more difficult to handle, thereby increasing the likelihood of parents to seek professional help. However, this remains speculative and requires further investigation.
Even though our study has some important strengths, there are some limitations too. First, we assessed pregnancy duration, birth weight and maternal smoking during pregnancy using a retrospective parent questionnaire. Especially for maternal smoking during pregnancy, the self-report nature of our assessment may have confounded our data, due to socially acceptable answering [
24]. However, we did only investigate whether or not the mother smoked, excluding dosage effects and thereby limiting the influence of socially acceptable answering (such as reporting lower dosages). Moreover, there was a relatively large amount of missing data for pre- and perinatal information. Second, even though we assessed parental ADHD and parental psychopathology, we did not specifically assess paternal antisocial personality disorder or maternal stress, which both have been found to be related to the development of antisocial behaviour disorders [
33]. Further, we did not assess parenting styles, which would have allowed us to investigate further the alleged link between parental ADHD and deviant parenting styles. In addition, parental criticism and warmth were only assessed in the diagnostic group, limiting our findings to the diagnostic groups comparison, and thus to predictors for comorbid ODD versus ADHD-only. Third, even though we assessed robust prediction models, our findings are based on a combination of longitudinal (parental warmth and criticism), retrospective (birth weight, pregnancy duration, maternal smoking during pregnancy, adverse life events, parental ADHD, parental SES), and cross-sectional (deviant peer affiliations) data. However, the retrospective predictors were independent of the measurement period, and only deviant peer affiliation data were assessed cross-sectional. For the latter variable, our final model may have been different if it had been measured at baseline. However, although it has been suggested that the influence of deviant peer affiliations would change over development, the level of deviant peer affiliations appears to be stable over development [
28]. Fourth, especially for the control group, a relatively large proportion of siblings was included. Nevertheless, our findings did not change in terms of the predictors involved or the direction of associations when excluding these siblings, indicating that the findings are robust. Finally, since we applied strict inclusion criteria, such as excluding individuals with a comorbid conduct disorder diagnosis, our sample may represent a subsample of, rather than all, individuals with ADHD-only and ADHD + ODD. In addition, we focused on comorbid ADHD + ODD and were not able to include an ODD-only group. Therefore, the findings may not be generalizable to all individuals with ADHD or individuals with only ODD.
Overall, our study showed that postnatal risk factors (adverse life events) and transgenerational influences (parental ADHD) are important risk factors for the development of ADHD + ODD and ADHD-only. The development of comorbid ODD in individuals with ADHD was predicted by both postnatal adversities (SES, deviant peer affiliation) and negative transgenerational influences (parental criticism). These risk factors were significant for all ages. Our findings are in line with theories stating that environmental factors play an important role in the development of comorbidities such as ODD in individuals with ADHD [
5,
23,
50]. This highlights the need to take these risk factors into account when treating children with ADHD, since these factors may prove to be essential in the prevention of comorbid ODD. The development of comorbid ODD in ADHD is of concern given the lower functional outcome of this comorbid group relative to either disorder separately [
4,
35]. For example, (comorbid) ODD is reported as an important predictor for later life conduct disorder [
15]. Our findings seem to support the use of intervention programs comprising parent- and parent–child training in the prevention of comorbid ODD [
34], although we did not assess these trainings or their effects ourselves. In addition, monitoring peer affiliations of individuals with ADHD may prove useful in averting the transition from ADHD-only to the more severe ADHD + ODD [
8,
23,
37].