Introduction

Neuroticism seems to reflect an enduring vulnerability to major depressive disorder (MD) partly reflecting shared genetic risk factors [1]. Most of the genetic risk for MD expressed via personality is captured by neuroticism, with a modest influence of conscientiousness, and with small influence of openness, extroversion, and agreeableness [2], [3]. When neuroticism decreases in patients with depression treated with antidepressants, it has been difficult to clearly distinguish the treatment effect on neuroticism from the treatment effect on the depressive disorder, as remission of depressive symptoms is associated with partial normalization of neuroticism [4]. Decrease in neuroticism scores during paroxetine treatment of patients with MD, even after controlling for depression improvement, has been observed in a large group of depressed patients [4]. Thus, it is possible that response to selective serotonin re-uptake inhibitors (SSRIs) may be mediated at least partly via a decrease in neuroticism [4], [5]. Higher neuroticism is associated with higher thalamic serotonin binding [6]. Furthermore, a recent study from our group has suggested that familial risk of depression and neuroticism interact in their relation to the degree of specific serotonin transporter binding [7]

Two randomized trials have investigated the effect of SSRI on behavior and aspects of personality with some relation to neuroticism in healthy participants without a family history of MD. Knutson et al. found that four weeks intervention with paroxetine 20 mg/day (n = 26) versus placebo (n = 25) significantly increased social affiliation and decreased negative affect [8]. Tse et al. found that two weeks intervention with citalopram 20 mg/day (n = 11) compared with placebo (n = 9) resulted in a statistically significant increase in self-directedness [9]. Results from these trials suggest that SSRIs may affect aspects of personality even in the absence of clinical depression. SSRIs do not seem to influence mood in healthy individuals [10]–[13]. Results from a number of studies, although not all [14], have suggested increased levels of neuroticism in healthy first-degree relatives of patients with MD compared to healthy individuals without a family history of MD [15]. However, no trial has investigated the effect of SSRIs on neuroticism and other personality dimensions in healthy individuals with a family history of MD [16].

Thus, to examine the effect of escitalopram on neuroticism and other aspects of personality, and excluding an effect on depression, we recruited healthy first-degree relatives of patients with MD for the AGENDA (Associations between Gene-polymorphisms, Endophenotypes for Depression and Antidepressive Intervention) trial. The trial is the first to investigate the effect of a four week self-administered daily escitalopram versus placebo on personality traits, as measured with the Revised Neuroticism-Extroversion-Openness-Personality Inventory (NEO-PI-R) and the Eysenck Personality Inventory (EPQ), in healthy individuals . We tested the hypothesis that intervention with escitalopram for a month compared with placebo decreases symptoms of neuroticsm in healthy first-degree relatives of patients with MD.

Materials and Methods

Probands and participants

The selection of diseased probands and healthy participants has previously been described in details [17]. Probands were patients with a diagnosis of MD from psychiatric hospital in- or out-patient contact in Denmark who participated in ongoing studies at the Department of Psychiatry, Rigshospitalet, Denmark. The proband's diagnoses were validated by face-to-face interviews including the semi-structured interview Schedules for Clinical Assessment in Neuropsychiatry (SCAN) by trained clinicians [20]. Probands were asked to permit a contact to their adult children and/or siblings who were the eligible participants for the AGENDA trial. The probands (n = 466) gave written permission to contact 359 first-degree relatives, who were the potential participants in the trial. The participant flow is shown in Figure 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Flowchart consort for Agenda Trial Neuroticism.

https://doi.org/10.1371/journal.pone.0031980.g001

Individuals of either sex, aged 18–60 years with Danish ethnicity (i.e., born in Denmark with Danish parents and European grandparents) were eligible for the trial. Ethnicity was used to get a genetically homogeneous sample. We excluded individuals with somatic illnesses or a handicap that made participation in the trial impossible while six individuals with stable, treated milder medical conditions were included: hypertensio arterialis (three), pancreatitis antea (one), hypothyroidism (one), and acne vulgaris (one). Furthermore, we excluded individuals with a daily intake of drugs interfering with corticosteroids or escitalopram (cipralex), including birth control pills or any kind of corticosteroids, and individuals who were allergic to the study drug or placebo. Additionally, former medical or psychological treatment for diseases in the affective or schizophrenic spectrum and current abuse of alcohol or psychotropic medication led to exclusion. Women who were trying to conceive, or who were pregnant or breastfeeding were excluded. Women were postmenopausal or in the lutheal phase of the menstrual cycle at the time of randomization. Women taking birth control pills were instructed to discontinue these six weeks prior to entering the trial. Furthermore, all women were carefully instructed to use double barrier birth control methods and pregnancy tests were performed both before and after the intervention [17].

Results

A total of 80 participants were randomized. The characteristics of the participants can be seen in Table 1. The mean age of non-participants was 37 (SD11) years and 58% were women. The reasons for non-participation are presented in Figure 1. A statistically significant correlation was found between EPQ neuroticism and NEO-PI-R neuroticism reported at entry (Pearsson = 0.8; p<0.0005). There was no statistical significant difference in the change of HAM-D (four weeks - entry) between the escitalopram group and the placebo group (p = 0.6). No severe adverse reactions or serious adverse events were obseved. The side effect measure UKU total for participants of the escitalopram and placebo groups did not differ significantly at four weeks (Table 1). However, sexual adverse effects showed a statistically significant increase and insomnia showed a statistically significant decrease in the escitalopram group compared with the placebo group [18].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Clinical and demographic characteristics of participants in the Agenda Trial.

https://doi.org/10.1371/journal.pone.0031980.t001

Effects on neuroticism

The dataset was complete with the exception of one man and one woman in the escitalopram group, who left the trial prior to the intervention, and two men in the placebo group in whom data collection failed for both EPQ and NEO-PI-R in one and for EPQ in another, see Figure 1. There was no statistical significant difference in the change in reported neuroticism scores according to the NEO-PI-R (p = 0.09) and the EPQ (p = 0.7) for participants who got escitalopram compared to those who got placebo (Table 2). No stastistically significant correlations were found between change in neuroticism measured using EPQ or NEO-PI-R, and age, sex, years of education, or plasma escitalopram. The exclusion of two participants with immeasurable concentrations did not change the results substantially.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Changes in personality scores in the escitalopram and the placebo group following four weeks of treatment.

https://doi.org/10.1371/journal.pone.0031980.t002

Further analyses showed no statistically significant correlations between: change in EPQ neuroticism and BDI-21 at entry (rho = −0.26; p = 0.06), change in EPQ neuroticism and HAM-D at entry (rho = 0.12; p = 0.3), change in NEO-PI-R neuroticism and BDI-21 at entry (rho = −0.10; p = 0.4), and change in NEO-PI-R neuroticism and HAM-D at entry (rho = −0.05; p = 0.7).

Furthermore, no statistically significant differences were shown in changes in EPQ extraversion (p = 0.2), EPQ psychoticism (p = 1.0), NEO-PI-R extraversion (p = 0.9), NEO-PI-R openness (p = 0.3), and NEO-PI-R conscientiousness (p = 0.07) between escitalopram and placebo participants. However, a statistically significant difference was found in the change in NEO-PI-R agreeableness between escitalopram (mean: 2.38, SD; 8.09) and placebo (mean −1.32, SD: 7.94; p = 0.046 (Table 2).

Discussion

The results of the AGENDA trial do not support our hypothesis that a four-week long intervention with escitalopram as compared with placebo would decrease EPQ or NEO-PI-R neuroticism total scores in healthy first-degree relatives of patients with MD. However, we found statistically significant changes on the agreeableness dimension of personality on the NEO-PI-R, (p = 0.046) but none of the other dimensions of personality measured by EPQ or NEO-PI-R were significantly affected by the intervention. We found a trend for change in NEO-PI-R conscientiousness (p = 0.07) between escitalopram and placebo participants. We had no hypothesis regarding conscientiousness, and greater statistical power might have been needed to evaluate this item. We found no significant correlation between change in neuroticism during intervention, and age, sex, education or plasma escitalopram concentration, respectively.

Our present trial is the first to suggest an effect of escitalopram on agreeableness in healthy individuals without depressive symptoms. The trial has several advantages. The participants were studied in a randomized clinical trial blinded in all phases including the statistical analyses and conclusion phase. The trial and the analyses were carried out as planned in advance and the completion and compliance in the trial was very high. The registered diagnosis of depression for the probands was verified by a face-to-face psychiatric research interview by trained medical doctors. The participants were assessed and diagnosed by validated and frequently used multi-dimensional methods. Further, the participants were subjected to four weeks of intervention thus including the interval in which clinical improvement has been reported in trials with patients with MD [28]. Personality variables were assessed with the NEO-PI-R and EPQ, both widely used self-report measures based on the Five-Factor Model of Personality [29].

Despite these advantages our trial also has a number of potential limitations. Firstly, a large number of women were excluded from our trial due to oral contraceptives and pregnancy, thus the trial population is characterized by an overrepresentation of men. Secondly, we cannot exclude that the dosage of 10 mg escitalopram was too low although this has been suggested as the optimum dose for treatment of moderate depression [30]. Even though the participants received weekly phone calls to optimise adherence, several of the participants in the escitalopram group were found to have low plasma escitalopram concentrations. We considered using a higher dosage, but escitalopram 20 mg daily might have given more adverse effects, eventually jeopardizing blinding and adherence, thus we decided to use 10 mg daily. This dose of escitalopram resulted in well-known adverse effects, such as sexual adverse effects, as described in a prior paper from the study [18]. Further, the exclusion of the two participants with immeasurable concentrations did not change the results substantially. Thirdly, it may be argued that four weeks of treatment is too short a period to affect aspects of personality. However, two prior studies on the effect of SSRI on personality revealed positive effects within a four week [8] and a two week [9] treatment period, respectively. Thus, these trials reported changes in relatively complex tasks or in specific measures (hostility, submissiveness) rather than the broad personality self-report descriptors as the NEO-PI or EPQ. Instead neuroticism was found to be affected after 8 weeks of administration by Tang et al. 2009. Moreover, we have not compared healthy individuals with a family history of MDD to healthy individuals without a family history of MD. Finally considering the multitude of statistical significance tests performed a p value of 0.046 may very well occur by chance and correction for multiple testing would render it insignificant. It should be noted that when adjusting for multiple analyses the p value for agreeableness was p = 0.28 using Holm's test. Thus the result needs to be confirmed in additional studies.

We planned to include 80 participants due to resources and availability of the healthy first-degree relatives of patients with MD, as previously described [17] and the AGENDA trial is the largest trial (n = 80) on healthy regarding any outcome, as shown in a recent review from our group [16]. We found a tendency for escitalopram to reduce neuroticism when measured by the NEO-PI-R, but the opposite tendency when neuroticism was measured by the EPQ. Thus, it may not seem likely that our results are due to type II errors and that a larger sample would have changed the results. Furthermore, neuroticism reported by EPQ and NEO-PI-R was closely correlated. It was not validated that the probands and the participants of the trial were indeed related. Furthermore, while our participants are at increased risk at depression many of them will not develop depression and the possible effect we found on agreeableness and conscientiousness may not be related to depression.

Results from a recent placebo-controlled trial in patients with major depression suggest that the SSRI paroxetine has a specific effect on personality traits of neuroticism and extraversion that is distinct from its effect on depression [4]. In a randomised controlled trial of patients with depression, neurotiscism and conscientiousness were significantly associated with response to a combination of medication and psychoterapy [31]. Our results suggest that escitalopram has no major direct effect on neuroticism, i.e., independently of the effect of depressive symptoms. We have chosen to recruit a sample of individuals with positive family history for MD, based on previous reports of increased neuroticism in this population. However, based on the mean scores reported in Table 1, the study sample presents with low-medium neuroticism levels thus breaking the original study design assumption. In fact, none of the subjects seem to report levels of neuroticism that would be considered as high and thus represent a risk factor for psychopathology. As familiarity for MD and neuroticism are likely to present both shared and non-shared contributions to higher vulnerability, it remains open for speculation whether a sample with positive family history but low-medium neuroticism scores (as in this trial) could in fact present some resilience towards the development of depression and perhaps a different response to antidepressants. This could also account for the negative results reported of 4 weeks antidepressant medication not affecting neuroticism levels. It could be hypothesised that changes in neuroticism by antidepressant medication only occur in the presence of baseline high neuroticism (as in a paper by Tang et al., 2009 [4]). This would be also supported by recent data showing that 1 week of antidepressant administration is able to modify negative biases in emotional processing correlated to high neuroticism [32]. According to the cognitive neuropsychological model of antidepressant drug action [33], antidepressants may work by modifying emotion processing biases and thus in turn producing changes at a phenotypic level such as symptoms and possibly personality measures.

Agreeableness is less studied in relation to depression. In a study of depressed patients (n = 53) [34] agreeableness was not significantly affected by SSRI treatment by flouoxetine. However, a study by Eskelius et al. found small but statistically significant changes in all the Karolinska Scales of Personality except the impulsiveness scale after 24 weeks of treatment with sertraline or citalopram in depressed patients. The changes (2.3–12.4%) were in the direction of normalisation, i.e., decreases in the anxiety and aggression-related scales and increases in social desirability and socialisation [35].

Studies of healthy first-degree relatives offer an excellent opportunity to determine whether personality traits represent a premorbid risk factor for subsequent onset of mood disorders or whether personality deviances are a consequence of mood episodes [15]. Future studies may explore the suggested link between the serotonergic system and the personality trait agreeableness. If the finding of changes in agreeableness is replicated, it may lead to the hypothesis that SSRI do not directly modulate neuroticism but rather mediate a different self-perception captured by changes in the scores of the facets of the personality dimension of agreeableness, which are trust, straightforwardness, altruism, compliance, modesty, and tender mindedness. Further it could be hypothesized that antidepressants first produce early changes in personality dimensions related to interpersonal measures in the direction of more positive interactions and that longer treatment is needed to affect self-reported measures of negative emotionality.

In conclusion, the AGENDA trial is the first trial to investigate if treatment with a SSRI has an effect on personality traits in healthy first-degree relatives of patients with major depressive disorder. No significant changes were found on any measure of neuroticism. In post-hoc analyses, escitalopram seemed to have an effect on agreeableness. Further, a trend was found for increased conscientiousness. Thus it is possible that SSRIs may have an effect on aspects of personality such as agreeableness and conscientiousness independent of the treatment effect on depression. The finding should be explored in future studies.

Supporting Information

Acknowledgments

We thank the diseased probands and the participants in the trial for their collaboration. Research secretary Helen Gerdrup Nielsen is thanked for practical assistance in organizing the trial. Nader Salas, CTU, is thanked for designing the randomization system. Kristian Juul is thanked for monitoring the trial according to Good Clinical Practice guidelines. The members of the data monitoring and safety committee, Associate Professor Jørgen Hilden and Professor Per Bech, are thanked for their contributions.