Journal of Affective Disorders 184 (2015) 249–255

Contents lists available at ScienceDirect

Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad

Research report

Childhood abuse, family history and stressors in older patients with bipolar disorder in relation to age at onset C.S. Thesing a, M.L. Stek a, D.S. van Grootheest a, P.M. van de Ven c, A.T. Beekman b, R.W. Kupka b, H.C. Comijs a, A. Dols a,n a

Department of Old Age Psychiatry GGZ inGeest/VUmc, Amsterdam, The Netherlands Department of Psychiatry GGZ inGeest/VUmc, Amsterdam, The Netherlands c Department of Epidemiology and Biostatistics, VUmc, Amsterdam, The Netherlands b

art ic l e i nf o

a b s t r a c t

Article history: Received 29 September 2014 Received in revised form 29 May 2015 Accepted 31 May 2015 Available online 11 June 2015

Objectives: The aim of this study is to explore the family history of psychiatric disorders, childhood abuse, and stressors in older patients with Bipolar Disorder (BD) and the association of these variables with the age at onset of BD. Methods: The Questionnaire for Bipolar Disorder (QBP) and the Mini International Neuropsychiatric Interview (MINI-Plus) were obtained from 78 patients aged 60 and over to determine diagnosis, age at onset of the first affective episode, childhood abuse, family history of psychiatric disorders and past and recent stressful life events. Results: Increased family history of psychiatric disorders was the only factor associated with an earlier age at onset of BD. Less family history of psychiatric disorders and more negative stressors were significantly associated with a later age at onset of the first (hypo)manic episode. Limitations: Age at onset, history of childhood abuse, and past stressful life events were assessed retrospectively. Family members of BD patients were not interviewed. Conclusions: Our findings suggest that age at onset can define distinct BD phenotypes. More specifically there was a stronger heredity of BD and other psychiatric disorders in patients with an early age of onset of BD. Negative stressors may play a specific role in patients with a late age at onset of a first (hypo)manic episode. & 2015 Elsevier B.V. All rights reserved.

Keywords: Bipolar disorder Age at onset Childhood abuse Stressors Family history

1. Introduction Studies have suggested a multifactorial etiology for Bipolar Disorder (BD), including both biological and psychosocial factors, with genetic and environmental factors interacting (Brietzke et al., 2009; Caspi and Moffitt, 2006; Gama et al., 2013). Identification of the pathophysiological determinants of BD is a major challenge with implications for the early detection, prevention, and treatment strategies of this disorder (Etain et al., 2008). Subtyping bipolar patients by age at onset may help identify different phenotypes that require specific treatment strategies, since an early onset (EO) has a less favorable course (Geoffroy et al., 2013). For example, a study showed that early onset of BD in general was associated with a more severe historical disease course, as reflected in greater comorbidity for most Axis I disorders, greater chronicity represented by more mood episodes and a greater n

Corresponding author. Tel.: þ 31 207885565; fax þ 31 20 7885577. E-mail address: [email protected] (A. Dols).

http://dx.doi.org/10.1016/j.jad.2015.05.066 0165-0327/& 2015 Elsevier B.V. All rights reserved.

proportion of days depressed, and greater lifetime risk of suicide attempts (Perlis et al., 2004). Also functioning and quality of life at study entry was significantly poorer among early- and very-earlyonset subjects (Perlis et al., 2004). The cut off for defining early onset bipolar disorder (EOBD) and late onset bipolar disorder (LOBD) ranges across studies from 30 to 65 years, but is most often set at age 50 (Depp and Jeste, 2004). However, this cut off at age 50 is debated upon as it is in the middle of the later peak of life time incidence rates of BD (Kroon et al., 2013). Some of the pathophysiological determinants for BD may be more associated with EOBD than LOBD. Studies found that a family history of affective disorder, psychotic disorder or substance abuse is more common in EOBD than in LOBD (Leboyer et al., 2005; Post et al., 2013b). Moreover, a family history of affective disorder is associated with a more severe course of BD, including an earlier onset, more episodes, more suicide attempts and lower quality of life (Antypa and Serretti, 2014; Berutti et al., 2014). In addition to hereditary factors, severe childhood abuse is also associated with an earlier onset of BD (Etain et al., 2008; Garno et al., 2005; Suppes et al., 2001). In younger adults with BD it was found that the

250

C.S. Thesing et al. / Journal of Affective Disorders 184 (2015) 249–255

prevalence of any type of abuse ranges from 36% (Leverich et al., 2001) to 49.9% (Hyun et al., 2000). In studies reporting different types of abuse, verbal abuse is more frequent (50.3–52.0%), compared to physical (23.0–33.8%), or sexual abuse (27.0–34.1%) either during childhood or adolescence (Leverich et al., 2001; Post et al., 2013a; Suppes et al., 2001). Causal relationships between BD and psychosocial stressors are complex to assess, as it is often difficult to establish whether a stressor was the cause or consequence of the illness (Alloy et al., 2005; Hosang et al., 2012; Johnson, 2005). Still stressors contribute to the onset of bipolar disorder as well as to subsequent episodes (Tsuchiya et al., 2003; Urosevic et al., 2008). A recent study found that the numbers of stressful life events in childhood (including abuse) and in the year prior to the last episode of BD were significantly higher in more severe BD patients with an earlier onset (Post et al., 2013a). The observation of an earlier age at onset in patients with a positive family history of severe affective disorders can be explained by a higher genetic risk, but alternatively, the emotional distress of having an affected parent or other close family member could also trigger an earlier onset of symptoms. The burden of a positive family history may be enhanced by exposure to more stressors and possible childhood abuse early in life. Post et al. (2013a) found an association between an early age at onset and physical or sexual abuse in childhood in BD patients with a parental load for affective disorder. Alternatively, severe affective symptoms during childhood may predispose to disrupted interactions with parents, increasing the risk for emotional or even physical abuse. The present study aims to explore associations between age at onset of BD and a family history of psychiatric disorders, childhood abuse, and past stressors in bipolar patients aged 60 and over. We hypothesize that patients with an EOBD have significantly more psychiatric disorders among first and second degree family members and have experienced more childhood abuse than patients with a later onset. Further, it is hypothesized that patients with LOBD have experienced significantly more negative or positive stressors prior to their first episode of BD, given the association of late-life mania with somatic or environmental causes (Shulman, 1997; Van Gerpen et al., 1999). To the best of our knowledge this is the first study exploring both hereditary and environmental aspects of late life BD with respect to the onset of BD.

2. Methods 2.1. Subjects In this cross-sectional study all patients (both in- and outpatients) aged 60 and over with a diagnosis of BD who were treated in our institute during the year 2012 were recruited by a search of the computerized record keeping system of the Mental Health Organization (GGZ inGeest, Amsterdam, The Netherlands) and were considered eligible for inclusion (for details see (Dols et al., 2014)). Patients were included in the study only if the diagnosis of Bipolar I Disorder (DSM-IV-TR: 296.00-.06, 296.40.46, 296.50-.56, 296.60-.66, 296.7), Bipolar II Disorder (DSM-IV-TR: 296.89), or Bipolar Disorder NOS (Not Otherwise Specified, DSMIV-TR: 296.80) was confirmed with the Mini International Neuropsychiatric Interview Plus (MINI-Plus) (van Vliet and de Beurs, 2007). The MINI-Plus interview is common clinical practice during the diagnostic phase of patient care and the results of this interview are documented in a computerized record keeping system. Patients were excluded if they were unable to provide written informed consent (requested by their psychiatrist or community psychiatric nurse) due to inability to communicate in Dutch or

English, mental retardation (IQ below 70, estimated on the basis of the patients’ medical record), dementia (Mini Mental State Examination [MMSE] below 18 obtained during the interview) or because they were in a very unstable psychiatric condition (e.g. compulsory admission). Of all 139 eligible patients, 25 patients were excluded because of a language barrier (n ¼1), mental retardation (n ¼4), dementia (n¼ 8), very unstable psychiatric condition (n ¼1), death (n ¼ 1), a different diagnosis (n ¼10: mood disorder NOS [n¼ 3], personality disorder [n¼ 2], schizoaffective disorder [n ¼3], cyclothymic disorder [n ¼1], recurrent depression [n ¼1]). Thirty-six patients were not willing to provide written informed consent for interviews, but of these patients 23 provided consent for medical record review (including a MINI-Plus interview). Seventy-eight patients were interviewed by the principal investigator or one of three trained psychology students. The study was approved by the Medical Ethics Committee (METC) of the VU University Medical Center, Amsterdam, the Netherlands. 2.2. Assessments Demographic data (including age, sex, marital status, nation of birth, level of education, and living situation) were derived from the patients' medical records and confirmed during interview. The Global Assessment of Functioning score (GAF-score) is a measure by which the psychological, social and occupational functioning of a person is indicated by a score between 0 and 100 (American Psychiatric Association, 2003). GAF scores were separated for symptoms (GAF-s) and functioning (GAF-p). The ‘Center for Epidemiologic Studies Depression Scale’ (CES-D), a depression selfreport questionnaire, was used to examine if participants experienced a depressive episode at the moment of the interviews (Radloff, 1977). Age at onset of BD, defined as the onset of the first affective episode, either (hypo)manic or depressive, and the age at first (hypo)manic episode were obtained from the MINI-Plus. The Dutch version of the Questionnaire for Bipolar Disorder (QBP) is an adaptation of the Enrollment Questionnaire as previously used in the Stanley Foundation Bipolar Network (Leverich et al., 2001; Suppes et al., 2001). The QBP-NL was used by the interviewer to investigate a family history of six mental disorders (depression, bipolar disorder, alcoholism, suicide or serious suicide-attempt, drug abuse, and psychosis) in first degree relatives (parents, siblings and offspring) and second degree relatives (grandparents) of the participants. The response options were: 1 (no), 2 (possible), 3 (probably), 4 (certainly: diagnosed or treated), or 5 (inapplicable). A specific psychiatric disorder was scored as ‘disorder present in the family’ if ‘possible’, ‘probably’ or ‘certainly’ was answered and as ‘disorder not present in the family’ if ‘no’ or ‘unknown’ was answered. Family history was quantified as the number of different types of mental disorders present in the whole family (not per family member), ranging between the minimum of 0 and maximum of 6. The following aspects of childhood abuse were investigated: verbal or emotional abuse, physical abuse or ill treatment and assault or sexual abuse. In the QBP the following 3 questions were formulated: “Have you ever been verbally or emotionally abused (such as intimidation, threats, humiliation or serious insult which caused you some serious emotional damage) as a child?”, “Have you ever been physically attacked or abused (i.e., are there experiences with physical injury or damage caused by another by beating, punches, kicking, biting, burning out, strangling or an attack with a weapon) as a child?” and “Have you ever been raped or sexually abused (i.e., are there experiences with sexual violence, sexual assault or forced sexual activity) as a child?”. The data obtained by the QBP about abuse occurring in adolescence or adulthood were not used in this study. The response options were 0 (never), 1 (rarely), 2 (occasionally) or 3 (frequently). A childhood

C.S. Thesing et al. / Journal of Affective Disorders 184 (2015) 249–255

abuse index was constructed by recoding the answers from the 3 childhood abuse questions from 0 (never), 1 (rarely), 2 (occasionally) and 3 (frequently) to the frequency scores: 0 (never), 1 (once or sometimes) and 2 (often). These frequency scores of the three types of abuse were summed up, resulting in a childhood abuse index that ranges from 0 to 6, with higher scores indicating more severe childhood abuse (Comijs et al., 2012). The occurrence of negative or positive stressors is part of the QBP by noting the severity of 16 negative and 8 positive stressors that could have been present during the period of 12 months prior to patient’s current episode of BD (or most recent episode of BD if there is no current episode) and during the 12 months prior to the patient’s first episode of BD. There were sixteen negative stressors assessed which were divided into three groups: social, financial or medical. The social stressors were: problems with partner, no support of family, loss of an important person by death, loss of social support, lack of counselor, cannot meet the many demands placed within social and occupational functioning, training problems and other psychosocial problems. The financial stressors were: unemployment problems, problems with housing, problems at work (e.g. threat of unemployment, a stressful schedule, difficult working conditions, dissatisfied with the job, change of job, disagreement with boss or colleagues), financial problems and judicial problems. The medical stressors were: problems associated with other medical illnesses, insufficiently insured for medical care and problems with access to health care. There were eight positive life events assessed. Five of them were social ones: start of school/study, final exam/graduation, engagement, marriage and birth of child, and three concerned financial positive stressors: promotion, bought a house, and change of job. There were no positive medical stressors included. The response options were: 0 (stressor not present), 1 (no impact), 2 (slight impact), 3 (moderate impact), and 4 (severe impact). We dichotomized the variables into 0 (no life event) and 1 (life event present; scores 1–4). The variable ‘number of negative stressors prior to the 12 months before first episode of BD’ is a variable ranging from 0 to 16, and the variable ‘number of positive stressors prior to the 12 months before first episode of BD’ is a variable ranging from 0 to 8. All data has been obtained by oral interviews, including the QBP, which is commonly a self-report questionnaire. 2.3. Statistical analyses First, we investigated whether participants who gave their written informed consent to view their medical records (n ¼ 23) and participants willing to provide consent for interviews (n ¼ 78) differed significantly in socio-demographic and clinical characteristics. Statistical tests were performed based on chi-square statistics for categorical variables and t-tests for continuous variables, at a significance level of 5%. Chi-square tests and Fisher's exact tests were conducted to see whether having experienced a type of abuse (never, once or sometimes or often) was associated with a family history of psychiatric disorders and to see whether the occurrence of different types of abuse were associated. Hierarchical multiple regression analyses were used to assess the relation between family history, childhood abuse, negative stressors and positive stressors and age at onset of BD (first episode of BD, whether this was a depressive or (hypo)manic episode) and age at onset of first mania, after controlling for gender and education. As age at onset was the dependent variable, age at time of study could not be a confounding variable and hence there was no need to control for it in the analyses. Preliminary analyses were performed to ensure no assumptions (e.g. level of measurement, independence of measurements, normality, linearity and homoscedasticity) were violated. We tested for multicollinearity before performing regression

251

analysis by examining the strength of association between abuse, family history of psychiatric disorders and positive and negative life events by Pearson Correlation coefficients and in regression analysis using Variance Inflation Factors and Tolerance scores. Post-hoc analysis (independent samples t-tests) are performed to investigate whether childhood abuse index scores and the number of positive stressors experienced in the 12 months prior to the first episode of BD differ between patients with a CES-D score beneath the cut-off score for depression and patients with a CES-D score above the cut-off score for depression. Data were analyzed using the Statistical Package of the Social Sciences (SPSS, version 20.0, SPSS Inc., Chicago, IL). A significance level of 0.05 was used.

3. Results 3.1. Socio-demographic and clinical characteristics The mean age of the interviewed participants was 68.5 years (range 61–98, SD 7.8), and 48.7% was female (Table 1). Bipolar I disorder was diagnosed in 53.8% of all patients, and 80.8% had an (early) onset of first affective symptoms before the age 50. No patient was diagnosed with Bipolar Disorder NOS. More detailed characteristics are described elsewhere (Dols et al., 2014). Age, gender, age at onset and clinical diagnosis of Bipolar I or II were not significantly different in participants that only provided written informed consent to review their medical records (n¼23) versus interviewed participants (n¼78). Mean GAF-s and GAF-p scores were significantly higher in interviewed participants (mean GAF-s 64.6, SD 12.0, mean GAF-p 65.0, SD 11.2) compared to the participants that only provided written informed consent to review their medical records (mean GAF-s 57.9, SD 9.6, mean GAF-p 58.1, SD 9.9) (Z¼  2.722 p¼.006, Z¼  2.671 p¼0.008 respectively). 3.2. Childhood abuse Table 2 shows the proportion of patients that experienced different types of abuse in childhood. Verbal abuse was associated with sexual abuse (Fishers Exact Test ¼14.79, p o.01) but not with physical abuse (Fishers Exact Test ¼7.59, p ¼.086). Physical abuse was not associated with sexual abuse (Fishers Exact Test¼ 5.94, p¼ .392). None of the three types of childhood abuse was significantly associated with the presence or absence of a family history of psychiatric disorders ( depression, bipolar disorder, alcoholism, suicide or serious suicide-attempt, drug abuse and/or psychosis) (Table 2.) 3.3. Family history Of all patients 76.9% had one or more first degree relatives (parents, siblings or offspring) with depression, bipolar disorder, alcoholism, psychosis, drug abuse and/or a first degree relative that committed suicide or did a serious suicide attempt. One or more grandparents with one or more of these disorders were present in the families of 25.6% of all patients. Of all disorders depression was the most common disorder in first degree relatives (followed by bipolar disorder) and bipolar disorder was the most common disorder in second degree relatives (evenly followed by depression and alcoholism). Psychosis, completed or attempted suicide and drug abuse were the psychiatric disorders that were least present in the family histories of the participants (Table 3). Of all participants, 43.6% (n ¼34) did not have any parent with a disorder, whereas 46.1% (n ¼36) had one parent (father or mother) with at least one disorder and only 10.3% (n ¼8) had both parents with one or more disorders.

252

C.S. Thesing et al. / Journal of Affective Disorders 184 (2015) 249–255

Table 1 Socio-demographic and clinical characteristics. All n ¼78

Gender Female (%) Age mean ( 7 SD) 60–69 years (%) 70–79 years (%) 80 years and over (%) GAF-s mean ( 7SD) range

48.7 68.5 ( 7 7.8) 66.7 23.1 10.3 64.6 ( 7 12.0) 35–85 65.0 ( 7 11.2) 40–91

GAF-p mean ( 7 SD) range DSM IV-TR diagnosis Bipolar I (%) Bipolar II (%) Age at first mania mean ( 7 SD) range %o 50 years

53.8 46.2 39.9 ( 7 16.4) 10–84 67.9 b Age at first depression mean ( 7 SD) range 35.8 ( 7 15.1) %o 50 years 8–63 79.1 Age at onset BD mean ( 7SD) range %o 50 years 33.4 (15.2) 8–72 80.8

Patients

Statistics

FH without psychiatric disorders (n¼ 16)

FH with psychiatric disorders (n ¼62)

X2/t/Fisher df

37.5 69.38 ( 7 7.6) 56.3 31.2 12.5 68.38 (7 14.4) 35–85 67.94 ( 7 13.0) 45–85

51.6 68.23 ( 7 7.9) 69.4 21.0 9.7 63.68 ( 7 11.3) 35-85 64.29 ( 7 10.6) 40–91

0.528 0.521 1.327

1 76

.468 .604

1.400

76

.562a .166

1.169

76

.246

50.0 50.0 44.44 (17.4) 18–84 62.5 42.0 (11.6) 24–60 72.7 39.06 (12.6) 18–61 81.3

45.2 54.8 38.66 (16.0) 10–77 69.4 31.2 (14.5) 8–63 80.4 31.87 (31.87) 8–72 80.6

0.004

1

.948

1.263

76

.211

2.327

65

.765a o .05 .686a

1.710

76

P (2-sided)

.091 1.000a

Notes: Statistical tests based on X2 statistics for categorical variables and t-tests for continuous variables, at a significance level of 5%. GAF: global assessment of functioning. a b

Fisher's exact test. n¼ 67.

Table 2 Incidence type of abuse. Patients

Statistics

All (n¼ 78) FH without psychiatric disorders (n ¼16) No abuse n (%) 58 (74.4) Verbal abuse Never, n (%) 63 (80.8) Once or sometimes, n (%) 7 (9.0) Often, n (%) 8 (10.2) Physical abuse Never, n (%) 72 (92.3) Once or sometimes, n (%) 2 (2.6) Often, n (%) 4 (5.1) Sexual abuse Never, n (%) 72 (92.3) Once or sometimes, n (%) 4 (5.1) Often, n (%) 2 (2.6)

13 14 1 1 15 1 0 15 1 0

(81.3) (87.5) (6.3) (6.3) (93.8) (6.3) (0.0) (93.8) (6.3) (0.0)

3.4. Stressors prior to illness onset and prior to the last episode The percentages of participants experiencing negative and positive stressors prior to their first and last mood episode of BD are presented in Table 4. The most common stressors before both the first and last affective episode are social stressors, followed by financial ones before the first affective episode and medical stressors before the last affective episode For the positive stressors, also the most common positive stressors before both the first and last affective episode are of a social nature. 3.5. Age at onset in relation to family history, childhood abuse, and stressors All Pearson Correlation coefficients between independent variables that we considered were below.20, except for the correlation between positive and negative life events which was 0.78. Multicollinearity of the independent variables positive and

FH with psychiatric disorders (n¼62) 45 49 6 7 57 1 5 57 3 2

(72.6) (79.0) (9.7) (11.3) (91.9) (1.6) (6.5) (91.9) (4.8) (3.2)

Fishers Exact Test

df P (2-sided)

.352

.749 1.000

1.967

.353

.580

1.000

negative life events was checked when performing the regression analyses, where Variances Inflation Factors and Tolerances did not indicate severe multicollinearity (although it may have been present to some extent as indicated by the increased standard deviation for negative life events when positive life events were added to the model in the final step). The results of the regression analysis with age at onset of BD as the dependent variable (first episode of BD, whether this was a depressive or (hypo)manic episode) are presented in Table 5a. Family history of psychiatric disorders turned out to be the only variable that was significantly associated to age at onset and this remained significant when the model was expanded by adding the other covariates. In all the models with age at first mania as the dependent variable, a family history of psychiatric disorders turned out to be significantly associated with age at first mania (Table 5b) and this remained significant when the model was expanded by adding the other covariates. In the fully adjusted model negative stressors also became significantly associated with age at onset (model 5).

C.S. Thesing et al. / Journal of Affective Disorders 184 (2015) 249–255

Table 3 Incidence of family history.

Table 5a Hierarchical regression analysis predicting the age at onset of BD from family history of psychiatric disorders, childhood abuse, negative life events and positive life events prior to first episode.

Relatives First degreea Grandparents None, n(%) Depression

Bipolar disorder

Alcoholism

Suicide or serious suicideattempt

Psychosis

18 (23.1) 41 (52.6) 12 (15.4)

58 (74.4) 70 (89.7) 5 (6.4)

25 (32.1) 46 (59.0) 13 (16.7)

3 (3.8) 68 (87.2) 6 (7.7)

19 (24.4) 51 (65.4) 16 (20.5)

4 (5.1) 70 (89.7) 5 (6.4)

11 (14.1) 65 (83.3)

3 (3.8) 75 (96.2)

2 (2.6)

0 (0.0)

11 (14.4) 68 (87.2) 4 (5.1)

3 (3.8) 78 (100.0) 0 (0.0)

6 (7.7) 61 (78.2) 6 (7.7)

0 (0.0) 77 (98.7) 0 (0.0)

11 (14.1)

1 (1.3)

Parents, siblings and/or children.

Table 4 Percentages of participants reporting negative and positive stressors and mean of negative and positive stressors prior to illness onset and the latest episode of BD. Stressors

Negative

Positive

First episode

Total Social Financial Medical Total Social Financial

Statistics Model Predictors

B (SE)

1 2 3

 3.686  3.271  3.183  1.415  3.394  1.801 0.649  3.422  1.820 0.769  0.289

FH of psychiatric disorders FH of psychiatric disordersa FH of psychiatric disorders Childhood abuse FH of psychiatric disorders Childhood abuse Negative stressors FH of psychiatric disorders Childhood abuse Negative stressors Positive stressors

4

5

(1.047) (1.103) (1.109) (1.620) (1.101) (1.613) (0.372) (1.113) (1.625) (0.593) (1.105)

t

p-Value R2

 3.521  2.966  2.870  0.874  3.084  1.116 1.746  3.075  1.120 1.296  0.261

o .01 o .01 o .01 .385 o .01 .268 .085 o .01 .267 .199 .795

.140 .162 .171 .204

.205

FH¼ family history. Possible-probably, n(%) Certainly, n(%) No, n(%) Possible-probably, n(%) Certainly, n(%) No, n(%) Possible-probably, n(%) Certainly, n(%)

Drug abuse

a

No, n(%) Possible-probably, n(%) Certainly, n(%) No, n(%) Possible-probably, n(%) Certainly, n(%) No, n(%) Possible-probably, n(%) Certainly, n(%) No, n(%)

253

Latest episode M (SD)

n (%)

M (SD)

48 47 15 5 21 17 6

1.78 1.47 0.24 0.06 0.35 0.26 0.09

47 40 16 17 9 7 5

1.56 1.06 0.26 0.24 0.17 0.09 0.08

(2.16) (1.74) (0.56) (0.25) (0.64) (0.52) (0.33)

(60.3) (51.3) (20.5) (21.8) (11.5) (9.0) (6.4)

(1.68) (1.33) (0.55) (0.49) (0.49) (0.29) (0.31)

3.6. Depressed mood as possible cause for over- or underreporting In a post-hoc analysis independent samples t-tests show that there was no significant difference in childhood abuse score between patients with a CES-D score beneath the cut-off score for depression (M¼ .49, SD ¼.98) and patient with a CES-D score above the cut-off score for depression (M¼ .62, SD¼ 1.12), t (76) ¼  .491, p ¼ .625. There was also no significant difference in the number of positive events prior to the first episode of BD experiences by patients with a CES-D score beneath the cut-off score for depression (M¼1.07, SD ¼2.05) and patient with a CES-D score above the cut-off score for depression (M ¼2.19, SD¼ 3.03), t (27.067) ¼  1.569, p¼ .128

4. Discussion The results of our study suggest that an earlier age at first affective episode of BD or first (hypo)mania is associated with a family history of psychiatric disorders and that later age at first mania is associated with the number of negative stressors in the

Models 2–5 are adjusted for gender and education.

Table 5b Hierarchical regression analysis predicting the age at onset of first mania from family history of psychiatric disorders, childhood abuse, negative life events and positive life events prior to first episode. Statistics Model Predictors

B (SE)

1 2 3

 3.855  3.616  3.532  1.355  3.794  1.832 0.804  3.923  1.920 1.363  1.340

FH of psychiatric disorders FH of psychiatric disordersa FH of psychiatric disorders Childhood abuse FH of psychiatric disorders Childhood abuse Negative stressors FH of psychiatric disorders Childhood abuse Negative stressors Positive stressors

4

5

n (%) (61.5) (60.3) (19.2) (6.4) (26.9) (21.8) (7.7)

a

(1.135) (1.208) (1.216) (1.776) (1.199) (1.758) (0.405) (1.203) (1.756) (0.641) (1.194)

t

p-Value R2

 3.398  2.995  2.905  .763  3.163  1.043 1.986  3.261  1.094 2.126  1.122

o .01 o .01 o .01 .448 o .01 .301 .051 o .01 .278 o .05 .265

.132 .137 .144 .188

.202

FH¼ family history. a

Models 2–5 are adjusted for gender and education.

year before onset. Very few studies have focused on family history in late life BD, and the literature is far from conclusive about rates of familial affective disorder by age of onset in older adults (Depp and Jeste, 2004), most likely due to differences in definitions on age at onset and family history of psychiatric disorders. However, our finding that a family history of psychiatric disorders is associated with an earlier onset of BD is in line with findings in younger adults (Leboyer et al., 2005). The association of negative stressors in the year before onset and a later age at first mania is in line with our hypothesis that LOBD is linked to somatic or environmental causes, however our findings need replication in larger cohorts. Besides the fact that there may be no association between childhood abuse and age and onset or family history of psychiatric disorders in older European samples, in the present study the percentage of participants that experienced childhood abuse may be too low to investigate these associations. There may be some logical explanations for these relatively low incidence rates of childhood abuse. In the present study 6 of 78 (7.7%) participants experienced sexual abuse during childhood. This is lower than found in depressed elderly (22.6%) and slightly higher than in nondepressed older controls (5.3%) (Comijs et al., 2012). A similar trend is observed for verbal and physical abuse, suggesting that childhood abuse is a less prominent factor in late life BD compared to late life unipolar depression. Studies in younger adults with BD

254

C.S. Thesing et al. / Journal of Affective Disorders 184 (2015) 249–255

with an earlier age at onset found higher percentages of any type of childhood abuse (Hyun et al., 2000; Leverich et al., 2001; Post et al., 2013a; Suppes et al., 2001), thereby possibly underscoring an association between age at onset and abuse. The numbers of stressful life events in childhood (including abuse) and in the year prior to the last episode of BD appear to be significantly higher in more severe patients from the U.S. compared to patients from Europe (Post et al., 2013a), which indicated that the relatively low rates of abuse found in the present study are more in line with the findings in younger bipolar patients from Europe than those from the U.S. Moreover, the majority of our participants, namely 90.7%, had at least one healthy parent (without a psychiatric disorder), which may be a protective factor that makes them less vulnerable for childhood abuse. The percentage of patients reporting having a first-degree family member with a mood disorder in the USA (75.5%) is much higher than in our sample (Antypa and Serretti, 2014). Another possible explanation for the low rates of abuse and stressful live events in the present study could be that patients with BD who experienced childhood abuse or stressful live events are more likely to stay at a mental institution as their course of disease is more severe (Alvarez et al., 2011; Brown et al., 2005), and thereby they are not in our reach to contact for participation in this study. Furthermore childhood abuse is linked to increased risk of suicide (Perich et al., 2014), eliminating these vulnerable patients from a cohort of older bipolar patients. The present study has some important strengths. To the best of our knowledge this is the first study exploring both hereditary and environmental aspects of BD in older patients. By including elderly we were able to explore the association with early as well as later age at onset. The diagnoses, age at onset of BD, data on childhood abuse, family history and stressors were systematically and carefully assessed in structured interviews and confirmed by medical records. Although our study is limited by its small sample size, it must be noted that the majority of older bipolar patients attending our mental health institution (68% of 114 eligible subject) participated in the interviews and thereby our findings probably can be generalized to a broad range of BP patients. However, in the present study the interviewed participants had higher GAF-scores than the participants that only gave informed consent to view their medical records, possibly limiting generalizability of our results to those who are more impaired. By studying a sample of older patients we had, in contrast to other studies, a broad range in age at onset so we were better able to explore the association with age at onset. Also, older patients have more first-degree relatives than younger patients, and their offspring has reached adulthood which implies that they are likely to have developed frank disorders. Our study has also limitations. First of all, while testing for multicollinearity the association between the variable positive life events and the variable negative life events turned out to be.78 which may have affected the results of the hierarchical multiple regression analyses to the extent that the regression coefficients for negative and positive life events in the full model may be difficult to interpret individually. Further, most of the information was assessed retrospectively. There may be underreporting of childhood abuse caused by memory problems, unwillingness to report embarrassing events or to disclose painful memories. On the other hand, there may also be the possibility of overreporting childhood abuse and underreporting of positive stressors by patients who were experiencing a depressive episode at the time of interviewing (Gorman, 1993). However, in the present study no significant differences were found in the childhood abuse index scores and in number of positive stressors experienced in the 12 months prior to the first episode of BD between patients with a

CES-D score beneath and above the cut-off score for depression. A further limitation is that family members were not interviewed, and thus this aspect of the study depended on the patients knowledge of the psychiatric disorders in first and second degree family members which may have also led to under- or overreporting of family history. Finally, experiencing life events may be age dependent and the association between age at onset and life events found in the present study may be explained by the increasing risk of experiencing negative life events while aging or the fact that stressors preceding an earlier onset are more subject to recall bias causing underreporting of stressors. The explained variances of the models in this study that included the variable ‘family history of psychiatric disorders’ were modest. The QBP-NL included a limited number of mental disorders, whereas other psychiatric disorders that have been associated with endophenotypes in bipolarity in family members, namely anxiety disorders, ADHD and some personality disorders, may also be important mental disorders when investigating family history. In addition, there may be other, unidentified, factors that influence the age at onset of BD which are not included in this study. Further research should be aimed at exploring both the biological and environmental aspects of BD together, like family history of psychiatric disorders, childhood abuse and stressors, but also neurobiological determinants (e.g. neurodegeneration, vascular factors, inflammation, metabolic factors, and physical illness, disability and pain) in relation to age at onset of BD. From a clinical perspective a history of childhood abuse and an early age at onset could identify a different phenotype that requires specific treatment strategies. However, separating EOBD from LOBD has been proven to be difficult as they are possibly a continuum where LOBD-patients have their first mania as a result of a stressor or a somatic cause but with a possible vulnerability for developing an affective disorder going back to childhood (abuse, stressor or family history). Later age at onset was associated with negative stressors prior to the first episode, possibly requiring a clinical emphasis on preventing stressors and/or strengthening patients in their abilities to deal with stressors to prevent new affective episodes. Role of funding source No funding was received for this study.

Conflict of interest CS: none, MS: none, DG: none, PV: none, AB has received honoraries as a speaker from Eli Lilly and Lundbeck, RK: has receives unrestricted research grants from AstraZeneca, honoraria as a speaker from AstraZeneca, Lundbeck, BristolMyersSquibb, and royalties from Wolters Kluwer Health, HC: none, AD: none.

Acknowledgments We thank Piet Eikelenboom for his helpful comments on the earlier versions of the manuscript.

References Alloy, L.B., Abramson, L.Y., Urosevic, S., Walshaw, P.D., Nusslock, R., Neeren, A.M., 2005. The psychosocial context of bipolar disorder: environmental, cognitive, and developmental risk factors. Clin. Psychol. Rev. 25, 1043–1075. Alvarez, M.J., Roura, P., Oses, A., Foguet, Q., Sola, J., Arrufat, F.X., 2011. Prevalence and clinical impact of childhood trauma in patients with severe mental disorders. J. Nerv. Ment. Dis. 199, 156–161. American Psychiatric Association, 2003. Diagnostic and statistical manual of mental disorders: DSM-IV-TR. Amercian Psychiatric Association, Washington D.C.. Antypa, N., Serretti, A., 2014. Family history of a mood disorder indicates a more severe bipolar disorder. J. Affect. Disord. 156, 178–186. Berutti, M., Nery, F.G., Sato, R., Scippa, A., Kapczinski, F., Lafer, B., 2014. Association between family history of mood disorders and clinical characteristics of bipolar

C.S. Thesing et al. / Journal of Affective Disorders 184 (2015) 249–255

disorder: results from the Brazilian bipolar research network. J. Affect. Disord. 161, 104–108. Brietzke, E., Stertz, L., Fernandes, B.S., Kauer-Sant’anna, M., Mascarenhas, M., Escosteguy Vargas, A., Chies, J.A., Kapczinski, F., 2009. Comparison of cytokine levels in depressed, manic and euthymic patients with bipolar disorder. J. Affect. Disord. 116, 214–217. 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, 57–67. Caspi, A., Moffitt, T.E., 2006. Gene-environment interactions in psychiatry: joining forces with neuroscience. Nat. Rev. Neurosci. 7, 583–590. Comijs, H.C., van Exel, E., van der Mast, R.C., Paauw, A., Oude Voshaar, R., Stek, M.L., 2012. Childhood abuse in late-life depression. J. Affect. Disord. 147, 241–246. Depp, C.A., Jeste, D.V., 2004. Bipolar disorder in older adults: a critical review. Bipolar Disord. 6, 343–367. Dols, A., Rhebergen, D., Beekman, A., Kupka, R., Sajatovic, M., Stek, M.L., 2014. Psychiatric and medical comorbidities: results from a bipolar elderly cohort study. Am. J. Geriatr. Psychiatry 22, 1066–1074. Etain, B., Henry, C., Bellivier, F., Mathieu, F., Leboyer, M., 2008. Beyond genetics: childhood affective trauma in bipolar disorder. Bipolar Disord. 10, 867–876. Gama, C.S., Kunz, M., Magalhaes, P.V., Kapczinski, F., 2013. Staging and neuroprogression in bipolar disorder: a systematic review of the literature. Rev. Bras. Psiquiatr. 35, 70–74. 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 186, 121–125. Geoffroy, P.A., Etain, B., Scott, J., Henry, C., Jamain, S., Leboyer, M., Bellivier, F., 2013. Reconsideration of bipolar disorder as a developmental disorder: importance of the time of onset. J. Physiol. Paris 107, 278–285. Gorman, D.M., 1993. A review of studies comparing checklist and interview methods of data collection in life event research. Behav. Med. 19, 66–73. Hosang, G.M., Korszun, A., Jones, L., Jones, I., McGuffin, P., Farmer, A.E., 2012. Lifeevent specificity: bipolar disorder compared with unipolar depression. Br. J. Psychiatry 201, 458–465. Hyun, M., Friedman, S.D., Dunner, D.L., 2000. Relationship of childhood physical and sexual abuse to adult bipolar disorder. Bipolar Disord. 2, 131–135. Johnson, S.L., 2005. Life events in bipolar disorder: towards more specific models. Clin. Psychol. Rev. 25, 1008–1027. Kroon, J.S., Wohlfarth, T.D., Dieleman, J., Sutterland, A.L., Storosum, J.G., Denys, D., de Haan, L., Sturkenboom, M.C., 2013. Incidence rates and risk factors of bipolar disorder in the general population: a population-based cohort study. Bipolar Disord. 15, 306–313.

255

Leboyer, M., Henry, C., Paillere-Martinot, M.L., Bellivier, F., 2005. Age at onset in bipolar affective disorders: a review. Bipolar Disord. 7, 111–118. Leverich, G.S., Nolen, W.A., Rush, A.J., McElroy, S.L., Keck, P.E., Denicoff, K.D., Suppes, T., Altshuler, L.L., Kupka, R., Kramlinger, K.G., Post, R.M., 2001. The Stanley foundation bipolar treatment outcome network. I. Longitudinal methodology. J. Affect. Disord. 67, 33–44. Perich, T., Mitchell, P.B., Loo, C., Hadzi-Pavlovic, D., Roberts, G., Green, M., Frankland, A., Lau, P., Corry, J., 2014. Cognitive styles and clinical correlates of childhood abuse in bipolar disorder. Bipolar Disord. 16, 600–607. Perlis, R.H., Miyahara, S., Marangell, L.B., Wisniewski, S.R., Ostacher, M., DelBello, M. P., Bowden, C.L., Sachs, G.S., Nierenberg, A.A., 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, 875–881. Post, R.M., Altshuler, L., Leverich, G., Nolen, W., Kupka, R., Grunze, H., Frye, M., Suppes, T., McElroy, S., Keck, P., Rowe, M., 2013a. More stressors prior to and during the course of bipolar illness in patients from the United States compared with Netherlands and Germany. Psychiatry Res. 210, 880–886. Post, R.M., Leverich, G.S., Kupka, R., Keck Jr., P., McElroy, S., Altshuler, L., Frye, M.A., Luckenbaugh, D.A., Rowe, M., Grunze, H., Suppes, T., Nolen, W.A., 2013b. Increased parental history of bipolar disorder in the United States: association with early age of onset. Acta Psychiatr. Scand. 129, 375–382. Radloff, L.S., 1977. The CES-D scale: a self-report depression scale for research in the general population. Appl. Psychol. Meas. 1, 385–401. Shulman, K.I., 1997. Disinhibition syndromes, secondary mania and bipolar disorder in old age. J. Affect. Disord. 46, 175–182. Suppes, T., Leverich, G.S., Keck, P.E., Nolen, W.A., Denicoff, K.D., Altshuler, L.L., McElroy, S.L., Rush, A.J., Kupka, R., Frye, M.A., Bickel, M., Post, R.M., 2001. The Stanley foundation bipolar treatment outcome network. II. Demographics and illness characteristics of the first 261 patients. J. Affect. Disord. 67, 45–59. Tsuchiya, K.J., Byrne, M., Mortensen, P.B., 2003. Risk factors in relation to an emergence of bipolar disorder: a systematic review. Bipolar Disord. 5, 231–242. Urosevic, S., Abramson, L.Y., Harmon-Jones, E., Alloy, L.B., 2008. Dysregulation of the behavioral approach system (BAS) in bipolar spectrum disorders: review of theory and evidence. Clin. Psychol. Rev. 28, 1188–1205. Van Gerpen, M.W., Johnson, J.E., Winstead, D.K., 1999. Mania in the geriatric patient population: a review of the literature. Am. J. Geriatr. Psychiatry 7, 188–202. van Vliet, I.M., de Beurs, E., 2007. The MINI-international neuropsychiatric interview. A brief structured diagnostic psychiatric interview for DSM-IV en ICD-10 psychiatric disorders. Tijdschr. Psychiatr. 49, 393–397.