Age of onset of bipolar disorder: Combined effect of childhood adversity and familial loading of psychiatric disorders

Journal of Psychiatric Research 81 (2016) 63e70

Contents lists available at ScienceDirect

Journal of Psychiatric Research journal homepage: www.elsevier.com/locate/psychires

Age of onset of bipolar disorder: Combined effect of childhood adversity and familial loading of psychiatric disorders Robert M. Post a, b, *, Lori L. Altshuler c, d, Ralph Kupka e, Susan L. McElroy f, g, Mark A. Frye h, Michael Rowe a, Heinz Grunze i, Trisha Suppes j, k, Paul E. Keck Jr. e, f, Gabriele S. Leverich a, Willem A. Nolen l a

Bipolar Collaborative Network, Bethesda, MD, USA Department of Psychiatry and Behavioral Sciences, George Washington University, Washington D.C., USA Department of Psychiatry and Biobehavioral Sciences, David Geffen School of Medicine, University of California, Los Angeles, CA, USA d Department of Psychiatry, VA Greater Los Angeles Healthcare System, West Los Angeles Healthcare Center, Los Angeles, CA, USA e Department of Psychiatry & Behavioral Neuroscience, University of Cincinnati College of Medicine, Cincinnati, OH, USA f Lindner Center of HOPE, Mason, OH, USA g Biological Psychiatry Program, University of Cincinnati Medical College, Cincinnati, OH, USA h Department of Psychiatry, Mayo Clinic, Rochester, MI, USA i Paracelsus Medical University, Salzburg, Austria j Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Palo Alto, CA, USA k V.A. Palo Alto Health Care System, Palo Alto, CA, USA l University Medical Center, University of Groningen, Groningen, The Netherlands b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 5 August 2015 Received in revised form 18 April 2016 Accepted 10 June 2016

Background: Family history and adversity in childhood are two replicated risk factors for early onset bipolar disorder. However, their combined impact has not been adequately studied. Methods: Based on questionnaire data from 968 outpatients with bipolar disorder who gave informed consent, the relationship and interaction of: 1) parental and grandparental total burden of psychiatric illness; and 2) the degree of adversity the patient experienced in childhood on their age of onset of bipolar disorder was examined with multiple regression and illustrated with a heat map. Results: The familial loading and child adversity vulnerability factors were significantly related to age of onset of bipolar and their combined effect was even larger. A heat map showed that at the extremes (none of each factor vs high amounts of both) the average age of onset differed by almost 20 years (mean ¼ 25.8 vs 5.9 years of age). Limitations: The data were not based on interviews of family members and came from unverified answers on a patient questionnaire. Conclusions: Family loading for psychiatric illness and adversity in childhood combine to have a very large influence on age of onset of bipolar disorder. These variables should be considered in assessment of risk for illness onset in different populations, the need for early intervention, and in the design of studies of primary and secondary prevention. © 2016 Elsevier Ltd. All rights reserved.

Keywords: Family history Depression Substance abuse Childhood abuse Age of onset of bipolar disorder Heat map

1. Introduction Previous studies have reported an earlier age of onset of bipolar disorder in those with a positive family history of bipolar disorder

* Corresponding author. Bipolar Collaborative Network, 5415 West Cedar Ln. Suite 201B, Bethesda, MD 20814, USA. E-mail address: [email protected] (R.M. Post). http://dx.doi.org/10.1016/j.jpsychires.2016.06.008 0022-3956/© 2016 Elsevier Ltd. All rights reserved.

(Leverich et al., 2002, Pavuluri et al., 2005, Post et al., 2014a) and in those with a history of adversity in childhood (Brown et al., 2005, Garno et al., 2005, Larsson et al., 2013, Leverich et al., 2002). More recently it has also been revealed that the burden of other psychiatric illnesses, beyond that of just bipolar disorder, in the direct progenitors (parents and grandparents) of patients also contributed to the vulnerability early onset bipolar disorder (Post et al., 2015c,d). However, the combined effects of these two vulnerability factors – genetic/familial loading of psychiatric

64

R.M. Post et al. / Journal of Psychiatric Research 81 (2016) 63e70

illnesses and psychosocial adversity in childhood – as they affect age of onset of bipolar disorder has not been examined in detail. Here we present statistical and graphical evidence in the form of a heat map of how these two risk factors independently and in combination affect the age of onset of bipolar disorder. The heat map allows one to directly visualize how each factor alone and their combined presence (when both factors converge and are present in the same individuals), influence the mean age of onset of bipolar disorder. We hypothesized that the two vulnerability factors for early onset bipolar disorder –genetic/familial burden of psychiatric illness and psychosocial adversity in childhood – would combine in a clinically robust fashion, further helping to explain some of the very large variations in the age of onset of bipolar disorder. We further discuss some of the mechanistic and clinical implications of these inter-relationships. 2. Methods 2.1. Patients 968 outpatients (average age 41) with bipolar disorder (about 75% BP I) diagnosed by SCID interview were recruited from advertisements and local clinics in four cities in the United States (Los Angeles, Dallas, Cincinnati, Bethesda) and three in Europe (Utrecht, the Netherlands and Freiburg and Munich, Germany) from 1995 to 2002. Patients gave informed consent for participation in the network and completed a detailed self-rated questionnaires on family history, psychosocial adversity in childhood, and their retrospective course of illness (Leverich et al., 2003, Leverich et al., 2002, Post et al., 2014a, Post et al., 2013a, Post et al., 2010a, Post et al., 2014b, Post et al., 2010b). 2.2. Family history The family history diagnoses rated included: 1) unipolar depression, 2) bipolar disorder, 3) history of a suicide attempt or completed suicide, 4) alcohol abuse, 5) drug abuse, and 6) “other illness” with the specific examples given as: “(i.e., anxiety, panic attacks, eating disorders, attention-deficit disorder, behavioral problems, obsessive compulsive disorder, autism, etc.)”. Each parental and grandparental diagnosis was rated by the patient as definite, likely, unlikely, or not present, and a definite or likely rating was taken as a positive diagnosis for that relative (Post et al., 2014a,b). The same history ratings were also inquired about for the patients’ siblings, spouse, and offspring as described elsewhere. Since a history of a suicide attempt is not a formal diagnosis, we also refer to these 6 categories as illnesses or difficulties. The total score of the presence or absence of these difficulties in either parent or any of the four grandparents was then used as the measure of total illness burden in two generations of the direct progenitors of the patients. We had previously seen that the grandparental loading of multiple illnesses also conveyed vulnerability to the development of an adverse course of bipolar disorder in the patient (Post et al, 2015c,d). The maximum score for family loading of illness could be 36, but the range was typically 0 to 16 with one patient having a score of 28. 2.3. Age of onset of bipolar disorder The questionnaire also elicited answers pertaining to demographics, stressors in childhood, and course of illness characteristics, including the age of onset of bipolar disorder. This was described as the age of onset of the first major depression associated with dysfunction or the first manic or hypomanic episode. Age of onset was inquired about for the first onset of depression, mania

or both in the same year. The age of onset used here was the age at which ever phase came first. 2.4. Childhood adversity Stressors in childhood included a total score for the report of verbal, physical, and sex abuse, each rated as never ¼ 0, rarely ¼ 1, occasionally ¼ 2, and frequently ¼ 3 (Leverich et al., 2002; Post et al., 2015b). A maximum score of 9 would thus reflect the most frequent experience of all three types of abuse. We examined all three together, based on previous analyses that verbal abuse in addition to what is usually considered more severe forms of abuse e physical and sexual e contributed independently to the impact on age of onset and other poor prognosis factors (Post et al., 2015b). The same questions were repeated for abuse experienced in adolescence and again in adulthood, but only those experienced in childhood were utilized here, since this provided the greatest likelihood that these stressors would have occurred prior to the onset of the bipolar disorder. The relationship between the burden of family illness (sum of parental and grandparental illnesses) and childhood adversity score to age of onset independently and as they combined was examined by multiple regressions controlling for country and sex. 2.5. Heat map A heat map illustrating these relationships to age of onset of bipolar disorder was constructed using the full range of total family history positivity for psychiatric illness (0e28) and total amount of adversity the patient reported experiencing in childhood (0e9). A heat map is a graphical representation of the data where values are color coded in a matrix. The heat map supplements the statistical analysis by providing a way of visualizing the impact of each vulnerability factor and their combined effect on age of onset of bipolar disorder in a single graphic which also allows for specific examination of the actual mean ages of onset involved. For ease of visualization, mean ages of onset were color coded to reflect the usual separation of ages of onset in the literature for childhood, adolescent, early adulthood and late adulthood (Perlis et al., 2004, Post et al., 2010a). Mean ages of onset in childhood (prior to age 13) were coded in red; in adolescence (prior to age 19) were coded in pink; in young adulthood (prior to age 30) were coded in blue; and in older adulthood (age 30 or greater) were coded in grey. As the numbers of patients involved in the mean ages of onset of bipolar disorder varied considerably, a separate illustration of the Ns for each category is presented and color-coded similarly to the first figure for ease of visualization. 2.6. Statistics The statistical significance of the hypothesized relationships of family burden of illness and childhood adversity to age of onset of bipolar disorder was tested with a multiple regression with the effects of sex (gender) and country (US versus European) accounted for. The US versus European “country” variable was entered as previous data had shown a very substantially earlier age of onset of bipolar in the US than in Germany and the Netherlands (Post et al., 2014a). Sex was examined as sexual abuse is more likely to occur in women than in men. We studied the interaction of these two variables, family burden of illness and childhood adversity, on age onset in a separate regression. This was necessary because we were unable to include both the individual terms and the interaction in the same model, as including both resulted in a failure to meet the underlying assumptions required to run a regression; these diagnostics could include linearity, normality, and homoscedasticity.

R.M. Post et al. / Journal of Psychiatric Research 81 (2016) 63e70

Instead the interaction and main effects are presented separately and any interpretation of them should be interpreted cautiously with the knowledge they could not be combined into a valid model.”

65

Table 1b Multiple regression showing the interaction of the burden of family illness and degree of abuse in childhood on the age of onset of bipolar disorder. Mean Age of Onset Coef.

Std. Err.

t

P>|t|

[95% Conf. Interval]

1.58 7.16 0.08

0.61 0.65 0.01

2.60 11.00 12.64

0.01 0.00 0.00

0.39 5.88 0.06

3. Results The percentage of patients reporting a history of any type of abuse in childhood was 56.5%. 21.2% experienced verbal abuse only, while those that any verbal (ie also in combination with the other types) was 50.3%. Physical abuse alone occurred in 1.3%, while any physical abuse occurred in 24.5%. Sexual abuse alone occurred in 5.0%, while any sexual abuse occurred in 21.9%. 10.6% of the patients had experienced all three types of abuse (Post et al., 2013a). As illustrated in Table 1a, both the reported total familial loading for psychiatric difficulties (in parents and grandparents) and total adversity in childhood (frequency/severity of verbal, physical, and sexual abuse) were independently related to an earlier age of onset of bipolar disorder (F (3 905) ¼ 54.72, P < 0.001, adjusted r ¼ 0.192). The separate regression run with the interaction of family psychiatric difficulties and childhood abuse instead of the individual terms produced a model which explained age of onset to a slightly greater extent (F (2, 905) ¼ 142.08, p < 0.001, adjusted r ¼ 0.232). While early onsets were more common in the cohort from the US than those from the Netherlands and Germany (Post et al., 2014a), the relationship of family loading and childhood adversity to early age of onset of bipolar disorder remained significant in both populations. When we examined annual income and degree of educational attainment as proxy measures for socio-economic status (SES) in the US (because SES has often been inversely linked to a higher incidence of psychiatric illness), these were not found to be statistically independently related to an earlier age of onset (results not shown). Annual income showed a negligible correlation with age of onset (r ¼ 0.05, p ¼ n.s.) or the degree of abuse in childhood (r ¼ 0.08, p ¼ n.s.) (Table 1b). The heat maps illustrated the average age of onset (Fig. 1) and number of subjects in each category (Fig. 2). The age of onset averaged 25.8 years in the 123 patients with no reported family history or abuse in childhood (bottom left, blue) and decreased to a range of 3e9 years of age (mean ¼ 5.9 years) in the 5 patients with many positives in the parental and grandparental family history and much abuse in childhood (top right, red). Thus, at the extremes, average age of onset of bipolar disorder differed by about 20 years in those with the least compared to the most of these two vulnerability factors. Overall, the red colors reflecting an age of onset in childhood (ie < 13) are largely grouped in the top right of the graph associated with the presence of substantial amounts of both vulnerability factors. Conversely the blue colors in the bottom left reflecting an average age of onset in young adulthood (ages 19 to 29) are almost exclusively in those with both a very low family illness burden and minimal adversity in childhood. If one grouped the child adversity scores into four categories, in those with no family history of psychiatric illnesses, the age of

Sex Country Total family illness  Abuse

2.77 8.44 0.09

Legend: The interaction of family illness and abuse in childhood on age of onset of bipolar disorder remained significant when accounting for both sex and country (US vs European patients).

onset bipolar disorder would decrease from an average of 25.9 years with minimal adversity scores of 0e2; to an average of 24.8 years with some adversity (3e5); to 17.3 years much adversity (6e8); and to 13.0 years in those with the most severe childhood adversity F (17, 912) ¼ 7.61, p < 0.001. Similarly, in the absence of any child adversity, an earlier age of onset of bipolar disorder occurred with increasing family burden The age of onset decreased from 25.8 years in the 123 patients with very low family history of illness burden (0e2) to 22.8 years in the 131 patients with a low family burden (3e5), to 16.6 years in the 33 patients with a moderate illness burden (6e8), to 10.2 years in 5 patients with a high family burden (although 2 single individuals with the highest number of illnesses in the family (11 and 15) and no childhood adversity had an age of onset of 20 and 15 years; ANOVA: total family illness F (9, 931) ¼ 12.1, p < 0.0001). 4. Discussion There were dramatic influences on the age of onset of bipolar disorder of both the direct lineage of parental and grandparental loading for psychiatric difficulties and the amount of psychosocial adversity experienced in childhood. Not only were these each independently statistically significant, but their combined presence produced a heat map with the oldest (25.8 years) average age of onset in the very large group of patients with neither of these vulnerability factors, and a very early average age of onset (5.9 years) in the small number of individuals with the highest amounts of both factors. These data are of substantial clinical importance from multiple clinical, research, and public health perspectives. 4.1. Limitations Several limitations to the data and caveats to their interpretation need to be considered prior to further discussion. Since the main effects of family history and childhood adversity on age of onset could not be tested in the same model with their interaction, the results should be interpreted with caution and we have been conservative in referring to their combined effects rather than their interaction. Importantly, all of the data in this study depended on the patient’s answers to a detailed patient questionnaire, such that

Table 1a Multiple regression showing effects of country*, burden of family illness, and childhood abuse on age of onset of bipolar disorder. Mean age of onset

Sex Country* Total burden of family illnesses Total adversity in childhood

Coef.

Std. Err.

t

P>|t|

[95% conf. interval]

0.69 4.47 0.69 0.86

0.64 0.73 0.12 0.13

1.07 6.14 5.83 6.49

0.29 0.00 0.00 0.00

0.57 3.04 0.92 1.13

Country* refers to the comparison of the 4 sites in the United States (N ¼ 678) compared to the 3 sites in the Netherlands and Germany (N ¼ 292).

1.94 5.89 0.46 0.60

66

R.M. Post et al. / Journal of Psychiatric Research 81 (2016) 63e70

Fig. 1. Heat Map of the Interaction of the Amount of Family History of Psychiatric Illness and Abuse in Childhood on Mean Age on Onset of Bipolar Disorder. Legend: Mean ages of onset of bipolar illness are graphed as a function of the burden of family illness in the patients’ parents and grandparents and the patients’ experience of abuse in childhood. Red colors refer to means ages of onset in childhood (<13 years of age); pink in adolescence (<19 yrs); blue in early adulthood (<29 yrs); and grey in late adulthood (30 yrs and older).

Fig. 2. The Number of Patients Contributing to the Mean Age of Onset At the Intersections of Family Illness and Childhood Abuse Legend: The number of patients in each intersecting bin contributing to the mean ages of onset of bipolar disorder illustrated in Fig. 1 are illustrated and color coded in an identical fashion as in Fig. 1 for ease of visualization. Large numbers of patients had relatively low levels of both family illness burden and childhood adversity (bottom, left, blue), while relatively few patients had high levels of both vulnerability factors (top, right, red).

R.M. Post et al. / Journal of Psychiatric Research 81 (2016) 63e70

age of onset, degree of adversity in childhood, and parental and grandparental history of psychiatric illness were self-reported by the patient and not subject to independent verification. However, our US data on age of onset (Post et al., 2014a; Post et al., 2014b) are virtually identical to that obtained by the STEP-BD network in other US academic sites and cities (Perlis et al., 2004), and the earlier age of onset of bipolar disorder in US compared to multiple European countries were replicated by Bellivier et al. (2014) and Etain et al. (2012) suggesting their likely reliability. Patient reported family history has been used as a surrogate for interview-based diagnoses with some success, although others have criticized the practice especially for more specific gene association studies. However, multiple studies have converged with the finding that even based on interviews, children with early onset bipolar disorder are at increased genetic risk as revealed by more bipolar disorder in their relatives than those with adult onset bipolar disorder (Baldessarini et al., 2012, Pavuluri et al., 2005). In addition, the incidence of various psychiatric illnesses in the parents of the patients from the United States were highly similar to those reported by Birmaher et al. (2010, 2014, 2013) which were obtained with direct interviews. Similarly, multiple studies have shown a relationship of adversity in childhood to an earlier age of onset or a more adverse course of bipolar disorder (Brown et al., 2005; Garno et al., 2005; Leverich et al., 2003; Leverich et al., 2002). Moreover, some studies suggest that self-report may underestimate, rather than overestimate, the degree of abuse in childhood (Hardt and Rutter, 2004; Kessler et al., 2010, Shonkoff and Garner, 2012). Given the likelihood, based on the considerations above, that our findings will to a large extent be replicated by others using other methods and with data independently verifiable, we comment on some of the potential implications of these striking findings. Multiple risk factors for early onset bipolar disorder have been identified, but two of the most robust and well replicated are those we have examined here. Emerging data reveal that in the US, bipolar illness is one of childhood and adolescent onset in two thirds of adults with an unequivocal diagnosis bipolar disorder (Perlis et al., 2004; Post et al., 2015b; Post et al., 2013a). 4.2. Mechanistic implications The lower age of onset of bipolar disorder in the US is very much in line with previous findings of greater degrees of family loading (Post et al., 2015a) and greater amounts of psychosocial adversity (Post et al., 2013a) in the US compared to Germany and the Netherlands. The relationship of family loading for multiple psychiatric disorders, and not just bipolar disorder, to an early age of onset is convergent with the accepted view that bipolar disorder is polygenic (Craddock et al., 2009, Nurnberger et al., 2014, Song et al., 2015) and factors from other illnesses may contribute to vulnerability to bipolar disorder. Antypa and Serretti (2014) also reported that patients with a positive family history of any mood disorder had an earlier age of onset and more severe form of bipolar disorder. Geller et al. (2006) found an earlier age of onset in offspring of bipolar parents who also had a diagnosis of ADHD, as opposed to those without ADHD. These data on other illnesses also conveying risk for bipolar disorder are also consistent with the recent findings that many of the same genes appear involved in multiple different psychiatric illnesses, including schizophrenia, bipolar disorder, depression, and autism (Lee et al., 2013). Early onset bipolar disorder and substance abuse have been closely linked with the suggestion that they are genetically related (Lin et al., 2006). Early onset bipolar disorder may be a separate genetic variant as Preisig et al. (2016) reported that only those patients with bipolar disorder who have an early onset to their

67

illness conveyed increased risk of bipolar disorder to their offspring. It is likely that this reflects a subgroup of patients with bipolar disorder conveying relatively greater genetic/familial vulnerability, and that with larger number of subjects and longer periods of follow up one would see vulnerability conveyed by patients with bipolar disorder also beginning later in life. While evidence is accumulating about the polygenetic nature of bipolar disorder, it is also clear that parental illness can convey increased risk to the next generation via an environmental/psychosocial route based on behavioral interactions with their offspring and the associated stressors experienced. Wickramaratne et al. (2011) showed that mothers with depression who were treated to remission had offspring with fewer psychiatric difficulties than those who were treated but remained symptomatic and did not achieve remission. The data linking childhood adversity to an earlier onset of bipolar disorder provide further evidence of an epigenetic basis for some of the illness transmission. Data in laboratory animals and in humans clearly show epigenetic changes associated with the occurrence of stressors in childhood (Labonte et al., 2013, McGowan et al., 2009, 2011, Mehta et al., 2013, Roth et al., 2009, Suderman et al., 2012, Yehuda et al., 2015). Such epigenetic changes in methylation of DNA; in histone acetylation, methylation and other chemical markers; and in microRNA can be based on direct environmental exposure to altered behavior or stressors (Meaney et al., 2013, Roth et al., 2009; Weaver et al., 2004), but new data also indicate that some of these epigenetic marks can escape erasure and be maintained in sperm and ova of animals and passed to the next generation in absence of direct behavioral contact. This has been most robustly demonstrated with exposure of male rats to stressors, toxins, or drugs of abuse such that altered reactivity to these experiences is passed on to the offspring in the absence of any contact with them (Bale, 2014, Dias and Ressler, 2014, Dietz and Nestler, 2012, Gapp et al., 2014, Szutorisz et al., 2014, Vassoler et al., 2013). Thus epigenetic changes based on either direct effects of the environment or indirect transgenerational effects in the absence of paternal contact can pave the way for potential geneenvironmental interactions. Environmental stressor effects interacting with specific gene variants have been shown to have important effects on manifestation of illness, such as with depression vulnerability in those with the short allele of the serotonin transporter who had been exposed to childhood adversity and then experienced subsequent stressors in adulthood (Caspi et al., 2003, Karg et al., 2011, Risch et al., 2009). Multiple other examples of gene-environmental interactions for vulnerability to unipolar depression have been reported on such targets as brain derived neurotrophic factor, glucocorticoid receptors (GR), and the GR co-chaperone FKBP5 (Mandelli and Serretti, 2013), and likely similar interactions will ultimately emerge in bipolar disorder. Conversely, resilience to stressors can be mediated by other types of epigenetic effects on the GR chaperone dynamics (Jochems et al., 2015). In our patients we do not have access to their gene sequences (or their epigenetic signatures), so specific types of geneenvironmental interactions cannot be studied. However, in the absence of knowing the genetic makeup of our patients, we have observed here a robust combined effect of reported childhood stressors and familial loading for psychiatric illness on the age of onset in a general, unselected population of bipolar patients. One could presume that within this overall combined effect, there might ultimately emerge more specific and greater magnitude effects in those with known or to-be-revealed gene variant vulnerabilities. Additionally, many other contributors to vulnerability to early

68

R.M. Post et al. / Journal of Psychiatric Research 81 (2016) 63e70

onset bipolar disorder have been postulated or identified and their role in moderating the gene-environmental effect observed here remains to be assessed. These could include: solar insolation; in utero infections and other difficulties; shorter gestation times; parental age at conception; low socioeconomic status; poor diet, and the like. Another potential mechanism for childhood stressors interacting with familial vulnerability is on telomere length. Childhood adversity is known to increase the percentage of short telomeres, which increases the risk for a multitude of medical and psychiatric illnesses (Epel et al., 2004, Shonkoff and Garner, 2012; Wolkowitz et al., 2011). Based on our data we cannot determine whether the combined effect of familial loading for psychiatric illness and environmental adversity in childhood provide more than additive effects on early age of onset of bipolar disorder, but they do clearly indicate that the two combined provide dramatic differences on age of onset. The difference in average age of onset of bipolar disorder is about 10 years in those with lowest to highest amounts of each vulnerability factor alone, but the magnitude of the difference is about 20 years when both familial loading and childhood adversity co-occur. In addition, since the combined effect was significant in both the US and European patients, this suggests that the effect occurs both in populations with high levels of both vulnerability factors (such as the US) and in those with relatively low levels of both factors (as in Germany and the Netherlands). Whatever the precise neurobiological mechanisms turn out to be, the magnitude of the effect appears to be substantial and of considerable clinical relevance. 4.3. Clinical implications for recognition, early intervention, and prevention Within the general populations of adult outpatients with bipolar disorder studied in multiple settings, it now appears that both reported family loading of psychiatric illness and degree of psychosocial stress in childhood account for large differences in the range of average age of onset of bipolar disorder. Those without either of these reported factors have an average age of onset of approximately 26 years, while those with either high familial loading alone or childhood adversity alone have an average age of onset of approximately 15 years. Those with extensive amounts of both vulnerability factors, as shown in the heat map, have very early average age onsets of their bipolar disorder (usually in childhood before age 13, and in 10 instances at or below 6 years of age). This perspective may help account for the very different and somewhat controversial ages of onset of bipolar disorder in different samples, communities, and countries as reviewed elsewhere (Bellivier et al., 2014; Etain et al., 2012; Post et al., 2015b). The magnitude of the differences in age of onset suggest that grandparental and parental loading for psychiatric illness and a history of childhood adversity could be used in the assessment of risk of early onset bipolar disorder, especially in those already showing prodromal symptoms (Post et al., 2013b). In this fashion one could begin to assess the effectiveness of early interventions and attempts at prevention in those at ultra-high risk because of the presence of three vulnerability factors: high degree of familial loading; psychosocial stress in childhood; and premonitory syndromes or prodromal symptoms (Post, 2015; Post et al., 2013b). Not only does early onset illness predict a poorer long term outcome (Azorin et al., 2013, Birmaher et al., 2009, Birmaher et al., 2014, Carlson et al., 2002, Carter et al., 2003, DelBello et al., 2007, Ernst and Goldberg, 2004, Geoffroy et al., 2013, Perlis et al., 2009, Perlis et al., 2004, Peters et al., 2014, Post et al., 2014b, Post et al., 2010b, Suominen et al., 2007), but it is also associated with an increased duration of the delay to first treatment (Wang et al., 2005), and this delay is longer in the US than in Europe (Post

et al., 2014a). The delay to first treatment itself is a further independent predictor of a poor outcome, particularly a longer duration and greater severity of depression assessed prospectively in adulthood (Post et al., 2010b). This delay to first treatment is a modifiable risk factor, which could readily be shortened with better awareness and recognition of the illness and public education. Miklowitz et al. (2013) reported significant improvement over treatment as usual with family focused intervention (FFT) in those who were both at high risk because of a positive family history of bipolar disorder and the presence of a depression, cyclothymia, or BP-NOS diagnosis. Since FFT and other related family-based therapies focus on increased intrafamilial communication and reduction in excessive expressed negative emotion, they could have an added benefit of decreasing family chaos and verbal abuse which by itself is a risk factor for early onset bipolar disorder even in the absence of a history of physical or sexual abuse (Post et al., 2015b). One would thus postulate that FFT would not only improve depression and anxiety symptoms in those at high risk (Miklowitz et al., 2013), but it might even delay the onset of bipolar disorder because of likely reduction in childhood adversity and related stressors in association with increased levels of parental support. Greater amounts of psychiatric difficulties in the US compared to the Netherlands and Germany appear to span four generations. These include that reported in: 1) grandparents and 2) parents of patients, 3) the bipolar patients themselves and their siblings, and 4) their offspring (Post et al., 2014a; Post et al., 2016). This suggests that greater psychiatric difficulty and more complex bipolar disorder in the US has been present for multiple generations and is not likely to be self-correcting. The transgenerational vulnerability to early onset and more severe illness are likely mediated by both genetic and epigenetic mechanisms (Bale, 2014; Meaney et al., 2013). Thus, greater attention to the risk factors for early onset bipolar disorder and earlier and more effective attempts at intervention (Correll et al., 2007, McNamara et al., 2010, Post et al., 2013b, Viana and Andrade, 2012) would appear to be critical clinical and public health strategies in attempting to achieve better treatment of early onset bipolar disorder and a more benign outcome. A better long-term outcome with exposure to early expert specialty care has been demonstrated by Kessing et al. (2013). They randomized youngsters with a first hospitalization for mania to two years in a specialty clinic versus two years of treatment as usual (TAU). Not only did those randomized to the specialty clinic have fewer hospitalizations over this two year period, but the differences were maintained and amplified over the next four years even though both groups received TAU over this latter time span. Thus, initial proper pharmacotherapy, psychotherapy, psycho-education, and illness monitoring had enduring positive effects on the course of bipolar illness. A new treatment research agenda for children with early onset mood and behavioral disorders in high risk families will also be required to ensure good long-term outcome, as the best approaches to clinical therapeutics are only just beginning to be elucidated and most consensus recommendations are based on an inadequate treatment literature (Geller et al., 2012, Kowatch et al., 2005). Role of funding agency No funding agency played any roll in the analysis or presentation of this work. Contributors Each of the coauthors participated in the design, conduct, data

R.M. Post et al. / Journal of Psychiatric Research 81 (2016) 63e70

collection and interpretation of this work except for Michael Rowe who was the statistician who analyzed the data and assisted in the presentation and write-up. Conflict of interest I have no conflicts of interest that are directly pertinent to the topic of this article, although I have been a speaker for Astra Zeneca, Sunovion, and Validus in the past year. Acknowledgements The Stanley Medical Research Institute supported the conduct of the clinical work and data collection from 1995 to 2002. References Antypa, N., Serretti, A., 2014. Family history of a mood disorder indicates a more severe bipolar disorder. J. Affect. Disord. 156, 178e186. Azorin, J.M., Bellivier, F., Kaladjian, A., Adida, M., Belzeaux, R., Fakra, E., et al., 2013. Characteristics and profiles of bipolar I patients according to age-at-onset: findings from an admixture analysis. J. Affect. Disord. 150, 993e1000. Baldessarini, R.J., Tondo, L., Vazquez, G.H., Undurraga, J., Bolzani, L., Yildiz, A., et al., 2012. Age at onset versus family history and clinical outcomes in 1,665 international bipolar-I disorder patients. World Psychiatry 11, 40e46. Bale, T.L., 2014. Lifetime stress experience: transgenerational epigenetics and germ cell programming. Dialogues Clin. Neurosci. 16, 297e305. Bellivier, F., Etain, B., Malafosse, A., Henry, C., Kahn, J.P., Elgrabli-Wajsbrot, O., et al., 2014. Age at onset in bipolar I affective disorder in the USA and Europe. World J. Biol. Psychiatry Off. J. World Fed. Soc. Biol. Psychiatry 15, 369e376. Birmaher, B., Axelson, D., Goldstein, B., Monk, K., Kalas, C., Obreja, M., et al., 2010. Psychiatric disorders in preschool offspring of parents with bipolar disorder: the Pittsburgh Bipolar Offspring Study (BIOS). Am. J. Psychiatry 167, 321e330. Birmaher, B., Axelson, D., Monk, K., Kalas, C., Goldstein, B., Hickey, M.B., et al., 2009. Lifetime psychiatric disorders in school-aged offspring of parents with bipolar disorder: the Pittsburgh Bipolar Offspring study. Archives General Psychiatry 66, 287e296. Birmaher, B., Gill, M.K., Axelson, D.A., Goldstein, B.I., Goldstein, T.R., Yu, H., et al., 2014. Longitudinal trajectories and associated baseline predictors in youths with bipolar spectrum disorders. Am. J. Psychiatry 171, 990e999. Birmaher, B., Goldstein, B.I., Axelson, D.A., Monk, K., Hickey, M.B., Fan, J., et al., 2013. Mood lability among offspring of parents with bipolar disorder and community controls. Bipolar Disord. 15, 253e263. Brown, G.R., McBride, L., Bauer, M.S., Williford, W.O., 2005. Impact of childhood abuse on the course of bipolar disorder: a replication study in U.S. veterans. J. Affect. Disord. 89, 57e67. Carlson, G.A., Bromet, E.J., Driessens, C., Mojtabai, R., Schwartz, J.E., 2002. Age at onset, childhood psychopathology, and 2-year outcome in psychotic bipolar disorder. Am. J. Psychiatry 159, 307e309. Carter, T.D., Mundo, E., Parikh, S.V., Kennedy, J.L., 2003. Early age at onset as a risk factor for poor outcome of bipolar disorder. J. Psychiatric Res. 37, 297e303. Caspi, A., Sugden, K., Moffitt, T.E., Taylor, A., Craig, I.W., Harrington, H., et al., 2003. Influence of life stress on depression: moderation by a polymorphism in the 5HTT gene. Science 301, 386e389. Correll, C.U., Penzner, J.B., Lencz, T., Auther, A., Smith, C.W., Malhotra, A.K., et al., 2007. Early identification and high-risk strategies for bipolar disorder. Bipolar Disord. 9, 324e338. Craddock, N., O’Donovan, M.C., Owen, M.J., 2009. Psychosis genetics: modeling the relationship between schizophrenia, bipolar disorder, and mixed (or “schizoaffective”) psychoses. Schizophr. Bull. 35, 482e490. DelBello, M.P., Hanseman, D., Adler, C.M., Fleck, D.E., Strakowski, S.M., 2007. Twelvemonth outcome of adolescents with bipolar disorder following first hospitalization for a manic or mixed episode. Am. J. Psychiatry 164, 582e590. Dias, B.G., Ressler, K.J., 2014. Parental olfactory experience influences behavior and neural structure in subsequent generations. Nat. Neurosci. 17, 89e96. Dietz, D.M., Nestler, E.J., 2012. From father to offspring: paternal transmission of depressive-like behaviors. Neuropsychopharmacology. official Publ. Am. Coll. Neuropsychopharmacol. 37, 311e312. Epel, E.S., Blackburn, E.H., Lin, J., Dhabhar, F.S., Adler, N.E., Morrow, J.D., et al., 2004. Accelerated telomere shortening in response to life stress. Proc. Natl. Acad. Sci. U. S. A. 101, 17312e17315. Ernst, C.L., Goldberg, J.F., 2004. Clinical features related to age at onset in bipolar disorder. J. Affect. Disord. 82, 21e27. Etain, B., Lajnef, M., Bellivier, F., Mathieu, F., Raust, A., Cochet, B., et al., 2012. Clinical expression of bipolar disorder type I as a function of age and polarity at onset: convergent findings in samples from France and the United States. J. Clin. Psychiatry 73, e561ee566. Gapp, K., von Ziegler, L., Tweedie-Cullen, R.Y., Mansuy, I.M., 2014. Early life epigenetic programming and transmission of stress-induced traits in mammals: how and when can environmental factors influence traits and their

69

transgenerational inheritance? BioEssays: News Rev. Mol. Cell. Dev. Biol. 36, 491e502. Garno, J.L., Goldberg, J.F., Ramirez, P.M., Ritzler, B.A., 2005. Impact of childhood abuse on the clinical course of bipolar disorder. Br. J. Psychiatry J. Ment. Sci. 186, 121e125. Geller, B., Luby, J.L., Joshi, P., Wagner, K.D., Emslie, G., Walkup, J.T., et al., 2012. A randomized controlled trial of risperidone, lithium, or divalproex sodium for initial treatment of bipolar I disorder, manic or mixed phase, in children and adolescents. Archives General Pychiatry 69, 515e528. Geller, B., Tillman, R., Bolhofner, K., Zimerman, B., Strauss, N.A., Kaufmann, P., 2006. Controlled, blindly rated, direct-interview family study of a prepubertal and early-adolescent bipolar I disorder phenotype: morbid risk, age at onset, and comorbidity. Archives General Psychiatry 63, 1130e1138. Geoffroy, P.A., Etain, B., Scott, J., Henry, C., Jamain, S., Leboyer, M., et al., 2013. Reconsideration of bipolar disorder as a developmental disorder: importance of the time of onset. J. Physiology, Paris 107, 278e285. Hardt, J., Rutter, M., 2004. Validity of adult retrospective reports of adverse childhood experiences: review of the evidence. J. Child Psychol. Psychiatry, Allied Discip. 45, 260e273. Jochems, J., Teegarden, S.L., Chen, Y., Boulden, J., Challis, C., Ben-Dor, G.A., et al., 2015. Enhancement of stress resilience through histone deacetylase 6-mediated regulation of glucocorticoid receptor chaperone dynamics. Biol. Psychiatry 77, 345e355. Karg, K., Burmeister, M., Shedden, K., Sen, S., 2011. The serotonin transporter promoter variant (5-HTTLPR), stress, and depression meta-analysis revisited: evidence of genetic moderation. Archives General Psychiatry 68, 444e454. Kessing, L.V., Hansen, H.V., Hvenegaard, A., Christensen, E.M., Dam, H., Gluud, C., et al., 2013. Treatment in a specialised out-patient mood disorder clinic v. standard out-patient treatment in the early course of bipolar disorder: randomised clinical trial. The British journal of psychiatry. J. Ment. Sci. 202, 212e219. Kessler, R.C., McLaughlin, K.A., Green, J.G., Gruber, M.J., Sampson, N.A., Zaslavsky, A.M., et al., 2010. Childhood adversities and adult psychopathology in the WHO world mental health surveys. Br. J. psychiatry J. Ment. Sci. 197, 378e385. Kowatch, R.A., Fristad, M., Birmaher, B., Wagner, K.D., Findling, R.L., Hellander, M., 2005. Treatment guidelines for children and adolescents with bipolar disorder. J. Am. Acad. Child Adolesc. Psychiatry 44, 213e235. Labonte, B., Suderman, M., Maussion, G., Lopez, J.P., Navarro-Sanchez, L., Yerko, V., et al., 2013. Genome-wide methylation changes in the brains of suicide completers. Am. J. Psychiatry 170, 511e520. Larsson, S., Aas, M., Klungsoyr, O., Agartz, I., Mork, E., Steen, N.E., et al., 2013. Patterns of childhood adverse events are associated with clinical characteristics of bipolar disorder. BMC Psychiatry 13, 97. Lee, S.H., Ripke, S., Neale, B.M., Faraone, S.V., Purcell, S.M., Perlis, R.H., et al., 2013. Genetic relationship between five psychiatric disorders estimated from genome-wide SNPs. Nat. Genet. 45, 984e994. Leverich, G.S., Altshuler, L.L., Frye, M.A., Suppes, T., Keck Jr., P.E., McElroy, S.L., et al., 2003. Factors associated with suicide attempts in 648 patients with bipolar disorder in the Stanley Foundation Bipolar Network. J. Clin. Psychiatry 64, 506e515. Leverich, G.S., McElroy, S.L., Suppes, T., Keck Jr., P.E., Denicoff, K.D., Nolen, W.A., et al., 2002. Early physical and sexual abuse associated with an adverse course of bipolar illness. Biol. Psychiatry 51, 288e297. Lin, P.I., McInnis, M.G., Potash, J.B., Willour, V., MacKinnon, D.F., DePaulo, J.R., et al., 2006. Clinical correlates and familial aggregation of age at onset in bipolar disorder. Am. J. Psychiatry 163, 240e246. Mandelli, L., Serretti, A., 2013. Gene environment interaction studies in depression and suicidal behavior: an update. Neurosci. Biobehav. Rev. 37, 2375e2397. McGowan, P.O., Sasaki, A., D’Alessio, A.C., Dymov, S., Labonte, B., Szyf, M., et al., 2009. Epigenetic regulation of the glucocorticoid receptor in human brain associates with childhood abuse. Nat. Neurosci. 12, 342e348. McGowan, P.O., Suderman, M., Sasaki, A., Huang, T.C., Hallett, M., Meaney, M.J., et al., 2011. Broad epigenetic signature of maternal care in the brain of adult rats. PloS one 6, e14739. McNamara, R.K., Nandagopal, J.J., Strakowski, S.M., DelBello, M.P., 2010. Preventative strategies for early-onset bipolar disorder: towards a clinical staging model. CNS Drugs 24, 983e996. Meaney, M.J., Aitken, D.H., Bodnoff, S.R., Iny, L.J., Tatarewicz, J.E., Sapolsky, R.M., 2013. Early postnatal handling alters glucocorticoid receptor concentrations in selected brain regions. Behav. Neurosci. 127, 637e641. Mehta, D., Klengel, T., Conneely, K.N., Smith, A.K., Altmann, A., Pace, T.W., et al., 2013. Childhood maltreatment is associated with distinct genomic and epigenetic profiles in posttraumatic stress disorder. Proc. Natl. Acad. Sci. U. S. A. 110, 8302e8307. Miklowitz, D.J., Schneck, C.D., Singh, M.K., Taylor, D.O., George, E.L., Cosgrove, V.E., et al., 2013. Early intervention for symptomatic youth at risk for bipolar disorder: a randomized trial of family-focused therapy. J. Am. Acad. Child Adolesc. Psychiatry 52, 121e131. Nurnberger Jr., J.I., Koller, D.L., Jung, J., Edenberg, H.J., Foroud, T., Guella, I., et al., 2014. Identification of pathways for bipolar disorder: a meta-analysis. JAMA Psychiatry 71, 657e664. Pavuluri, M.N., Birmaher, B., Naylor, M.W., 2005. Pediatric bipolar disorder: a review of the past 10 years. J. Am. Acad. Child Adolesc. Psychiatry 44, 846e871. Perlis, R.H., Dennehy, E.B., Miklowitz, D.J., Delbello, M.P., Ostacher, M., Calabrese, J.R., et al., 2009. Retrospective age at onset of bipolar disorder and outcome during

70

R.M. Post et al. / Journal of Psychiatric Research 81 (2016) 63e70

two-year follow-up: results from the STEP-BD study. Bipolar Disord. 11, 391e400. Perlis, R.H., Miyahara, S., Marangell, L.B., Wisniewski, S.R., Ostacher, M., DelBello, M.P., et al., 2004. Long-term implications of early onset in bipolar disorder: data from the first 1000 participants in the systematic treatment enhancement program for bipolar disorder (STEP-BD). Biol. Psychiatry 55, 875e881. Peters, A., Sylvia, L.G., Magalhaes, P.V., Miklowitz, D.J., Frank, E., Otto, M.W., et al., 2014. Age at onset, course of illness and response to psychotherapy in bipolar disorder: results from the Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD). Psychol. Med. 44, 3455e3467. Post, R., Altshuler, L., Kupka, R., McElroy, S., Frye, M., Rowe, M., et al., 2015a. Multigenerational positive family history of psychiatric disorders is associated with a poor prognosis in bipolar disorder. J. neuropsychiatry Clin. Neurosci. 27 (4), 304e310. Post, R., Altshuler, L., Leverich, G., Frye, M., Suppes, T., McElroy, S., et al., 2015b. The impact of verbal abuse on the course of bipolar disorder in patients from the United States. Bipolar Disord. 17, 323e330. Post, R.M., 2015. Heading off depressive illness evolution and progression to treatment resistance. Dialogues Clin. Neurosci. 17, 105e109. Post, R.M., Altshuler, L., Kupka, R., McElroy, S., Frye, M.A., Rowe, M., et al., 2014a. More pernicious course of bipolar disorder in the United States than in many European countries: implications for policy and treatment. J. Affect. Disord. 160, 27e33. Post, R.M., Leverich, G.S., Kupka, R., Keck, P.E., McElroy, S.L., Alshuler, L.A., et al., 2015c. Increases in multiple psychiatric disorders in parents and grandparents of patients with bipolar disorder from the United States compared to the Netherlands and Germany. Psychiatr. Genet. 25 (5), 194e200. Post, R.M., Altshuler, L., Leverich, G., Nolen, W., Kupka, R., Grunze, H., et al., 2013a. More stressors prior to and during the course of bipolar illness in patients from the United States compared with the Netherlands and Germany. Psychiatry Res. 210, 880e886. Post, R.M., Altshuler, L.L., Frye, M.A., Suppes, T., Keck Jr., P.E., McElroy, S.L., et al., 2010a. Complexity of pharmacologic treatment required for sustained improvement in outpatients with bipolar disorder. J. Clin. psychiatry 71, 117686; quiz 252e3. Post, R.M., Altshuler, L.L., Kupka, R., McElroy, S.L., Frye, M.A., Rowe, M., et al., 2016. More illness in offspring of bipolar patients from the U.S. compared to Europe. J. Affect. Disord. 191, 180e186. Post, R.M., Chang, K., Frye, M.A., 2013b. Paradigm shift: preliminary clinical categorization of ultrahigh risk for childhood bipolar disorder to facilitate studies on prevention. J. Clin. psychiatry 74, 167e169. Post, R.M., Leverich, G.S., Kupka, R., Keck Jr., P., McElroy, S., Altshuler, L., et al., 2014b. Increased parental history of bipolar disorder in the United States: association with early age of onset. Acta Psychiatr. Scand. 129, 375e382. Post, R.M., Leverich, G.S., Kupka, R., Keck Jr., P.E., McElroy, S.L., Altshuler, L.L., et al., 2015d. Increases in multiple psychiatric disorders in parents and grandparents of patients with bipolar disorder from the USA compared with The Netherlands and Germany. Psychiatr. Genet. 25, 194e200. Post, R.M., Leverich, G.S., Kupka, R.W., Keck Jr., P.E., McElroy, S.L., Altshuler, L.L., et al., 2010b. Early-onset bipolar disorder and treatment delay are risk factors for poor

outcome in adulthood. J. Clin. Psychiatry 71, 864e872. Preisig, M., Strippoli, M.P., Castelao, E., Merikangas, K.R., Gholam-Rezaee, M., Marquet, P., et al., 2016. The specificity of the familial aggregation of early-onset bipolar disorder: a controlled 10-year follow-up study of offspring of parents with mood disorders. J. Affect. Disord. 190, 26e33. Risch, N., Herrell, R., Lehner, T., Liang, K.Y., Eaves, L., Hoh, J., et al., 2009. Interaction between the serotonin transporter gene (5-HTTLPR), stressful life events, and risk of depression: a meta-analysis. Jama 301, 2462e2471. Roth, T.L., Lubin, F.D., Funk, A.J., Sweatt, J.D., 2009. Lasting epigenetic influence of early-life adversity on the BDNF gene. Biol. Psychiatry 65, 760e769. Shonkoff, J.P., Garner, A.S., 2012. The lifelong effects of early childhood adversity and toxic stress. Pediatrics 129, e232ee246. Song, J., Bergen, S.E., Kuja-Halkola, R., Larsson, H., Landen, M., Lichtenstein, P., 2015. Bipolar disorder and its relation to major psychiatric disorders: a family-based study in the Swedish population. Bipolar Disord. 17, 184e193. Suderman, M., McGowan, P.O., Sasaki, A., Huang, T.C., Hallett, M.T., Meaney, M.J., et al., 2012. Conserved epigenetic sensitivity to early life experience in the rat and human hippocampus. Proc. Natl. Acad. Sci. U. S. A. 109 (Suppl. 2), 17266e17272. Suominen, K., Mantere, O., Valtonen, H., Arvilommi, P., Leppamaki, S., Paunio, T., et al., 2007. Early age at onset of bipolar disorder is associated with more severe clinical features but delayed treatment seeking. Bipolar Disord. 9, 698e705. Szutorisz, H., DiNieri, J.A., Sweet, E., Egervari, G., Michaelides, M., Carter, J.M., et al., 2014. Parental THC exposure leads to compulsive heroin-seeking and altered striatal synaptic plasticity in the subsequent generation. Neuropsychopharmacol. Off. Publ. Am. Coll. Neuropsychopharmacol. 39, 1315e1323. Vassoler, F.M., White, S.L., Schmidt, H.D., Sadri-Vakili, G., Pierce, R.C., 2013. Epigenetic inheritance of a cocaine-resistance phenotype. Nat. Neurosci. 16, 42e47. Viana, M.C., Andrade, L.H., 2012. Lifetime Prevalence, age and gender distribution and age-of-onset of psychiatric disorders in the sao Paulo Metropolitan area, Brazil: results from the sao Paulo Megacity mental health survey. Rev. Bras. Psiquiatr. 34, 249e260. Wang, P.S., Berglund, P., Olfson, M., Pincus, H.A., Wells, K.B., Kessler, R.C., 2005. Failure and delay in initial treatment contact after first onset of mental disorders in the National Comorbidity Survey Replication. Archives general psychiatry 62, 603e613. Weaver, I.C., Cervoni, N., Champagne, F.A., D’Alessio, A.C., Sharma, S., Seckl, J.R., et al., 2004. Epigenetic programming by maternal behavior. Nat. Neurosci. 7, 847e854. Wickramaratne, P., Gameroff, M.J., Pilowsky, D.J., Hughes, C.W., Garber, J., Malloy, E., et al., 2011. Children of depressed mothers 1 year after remission of maternal depression: findings from the STAR*D-Child study. Am. J. Psychiatry 168, 593e602. Wolkowitz, O.M., Mellon, S.H., Epel, E.S., Lin, J., Dhabhar, F.S., Su, Y., et al., 2011. Leukocyte telomere length in major depression: correlations with chronicity, inflammation and oxidative stressepreliminary findings. PloS One 6, e17837. Yehuda, R., Flory, J.D., Bierer, L.M., Henn-Haase, C., Lehrner, A., Desarnaud, F., et al., 2015. Lower methylation of glucocorticoid receptor gene promoter 1F in peripheral blood of veterans with posttraumatic stress disorder. Biol. Psychiatry 77, 356e364.