Adverse mood effects of combined oral contraceptives in relation to personality traits

1. Introduction 

Combined oral contraceptive (COC) pill is the most common type of contraception among young women, with user frequencies in Sweden of almost 56% within the ages of 18–24 and 29% in ages 25–34 [1]. Increased irritability, mood swings and depressive symptoms have always, in spite of decreasing hormone concentrations over time, been major reasons for discontinuing treatment [2], [3], [4]. When studied prospectively, 7% of women on COC report increased anxiety and 10% report increased depressive mood [5], whereas retrospective studies display somewhat higher rates of adverse mood effects [6]. Discontinuation rates due to adverse mood symptoms have been reported to be 14–21% among COC users, with the highest mood-related discontinuation rate in the oldest age group [7].

As many women refrain from using safe contraception because of adverse mood effects when on COCs, it is important to elucidate the underlying causes for these changes in affect. It can be assumed that not only the active ingredients, ethinylestradiol and progestins, in the oral contraceptive pill will affect the psychological experiences and adverse effects of treatment, but also psychiatric history, personality traits, interpersonal relationships and socioeconomic factors. Thus far, most efforts to elucidate these issues have relied on retrospective collection of data. A previous depressive episode is significantly associated with worsening of mood during COC, whereas prior dysmenorrhea and early onset of premenstrual syndrome are associated with improved mood [6].

The possibility that certain personality traits render subjects vulnerable to experience adverse mood symptoms, or may increase the likelihood of reporting adverse mood while on COC has thus far not been explored. High levels of personality traits of neuroticism, introversion, interpersonal dependency and lack of self-confidence have all been proposed to be risk factors for the development of depression [8], [9], [10], [11]. Because prior depression is associated with adverse mood symptoms on COC, it is plausible that these personality traits are also associated with development or reporting of negative mood changes during COC use.

The aim of this study was to compare personality traits in women with ongoing or previous use of COCs and with different experiences from these compounds with respect to adverse mood symptoms.

2. Methods 

2.2. Personality inventory 

Assessments in the COC groups were made during the COC cycle on an arbitrary day of active treatment. The pCOC groups were evaluated during the luteal phase, the last 7 days before onset of menstruation. Luteal phase was confirmed by records of next onset of menstrual bleeding on the CD scale and from progesterone levels sampled on the testing day.

The Swedish universities Scales of Personality (SSP) is a self-rating questionnaire, based on the Karolinska Personality Scales (KSP) [13]. Compared to KSP, SSP has a reduced number of items and has an improved psychometric quality as concerns face validity, internal consistency and response differentiation [14]. The SSP contains 91 items divided into 13 scales, each of seven items: Somatic Trait Anxiety, Psychic Trait Anxiety, Stress Susceptibility, Lack of Assertiveness, Embitterment, Trait Irritability, Mistrust, Detachment, Impulsiveness, Adventure Seeking, Social Desirability, Verbal Trait Aggression and Physical Trait Aggression (Table 1) [14]. The participants rated the items on a scale from 1 to 4, where 1 equals “does not apply at all” and 4 equals “applies completely”. The SSP results were normalized to the population, and gender- and age-adjusted t-scores were calculated (mean 50, S.D. 10) [14].

Table 1. The Swedish universities Scales of Personality (SSP) items

	SSP subscales	Description of subjects with high scores adapted from [14]
	Somatic trait anxiety (STA)	Autonomic disturbances, restless, tense
	Psychic trait anxiety (PsTA)	Worries, anticipates, lacks self-confidence
	Stress susceptibility (SS)	Easily fatigued, feels uneasy when urged to speed up
	Lack of assertiveness (LA)	Lacks assertiveness in social situations
	Impulsiveness (I)	Acts on the spur of the moment, non-planning, impulsive
	Adventure seeking (ASA)	Avoids routine, has a need for change and action
	Detachment (D)	Avoids involvement with others, withdrawn, ‘schizoid’
	Social desirability (SD)	Socially conforming, friendly, helpful
	Embitterment (E)	Dissatisfied, blames and envies others
	Trait irritability (TI)	Irritable, lacks patience
	Mistrust (M)	Suspicious, distrust people's motives
	Verbal trait aggression (VTA)	Gets into arguments, berates people when annoyed
	Physical trait aggression (PhTA)	Gets into fights, starts fights, hits back

In addition, all subjects filled out the self-rated version of the Montgomery–Åsberg Depression Rating Scale (MADRS-S). The MADRS-S scores depressive symptoms during the preceding 3 days. Mild depression was defined as a score ≥11 on the MADRS-S scale and possible major depression as a score >20 [15].

3. Results 

The socio-demographic and clinical variables of the study groups are shown in Table 2. There were no differences between the groups in terms of socio-demographic or clinical variables. None of the pCOC groups differed with respect to the timing of assessment for luteal phase (cycle day −5.3±0.6 vs. −4.4±0.5). All pCOC subjects were considered to have had ovulatory cycles when evaluated, which was also confirmed by progesterone levels >15.0nmol/l. The pCOC groups did not differ with respect to the progesterone levels in the late luteal phase (pCOC-fine 19.3±3.6nmol/l vs. pCOC-mood 23.0±2.6nmol/l).

Table 2. Demographic data for the study groups

		COC-fine (n=30)	COC-mood (n=28)	pCOC-fine (n=27)	pCOC-mood (n=33)
	Age, years (mean±S.D.)	24.6±2.1	24.8±2.7	26.4±4.3	25.2±2.9
	BMI, kg/m2 (mean±S.D.)	21.6±2.1	21.1±2.2	22.9±3.2	21.7±3.8
	Steady relationship, n	20 (66.7%)	21 (75.0%)	14 (51.9%)	24 (71.7%)
	Education, years (mean±S.D.)	15.2±1.5	14.6±2.6	15.1±2.1	14.8±2.3
	Students, n	27 (90.0%)	21 (75.0%)	18 (66.7%)	30 (90.9%)
	Subjects with children, n	0.0	1 (3.6%)	2 (7.4%)	3 (9.1%)
	Smokers, n	3 (10.0%)	1 (3.6%)	2 (7.4%)	3 (9.1%)

18 women (15.3%) had MADRS-S scores indicating mild or major depression (only one subject had scores indicating major depression), but they were equally distributed between groups (COC-fine 2 (6.7%), COC-mood 5 (17.9%), pCOC-fine 2 (7.4 %) and pCOC-mood 9 (27.3%)).

Among the COC-fine patients, 6 (20.0%) were currently on COCs with estrogen profile and 23 (76.7%) on COCs with progestagen profile. In the COC-mood group the distribution was 12 (42.8%) on COCs with estrogen profile and 14 (50.0%) on COCs with progestagen profile. In three subjects, the profile of their preparation was unknown. The COC-mood group had on average tried 2.4 (range 1–4) different COC brands and pCOC-mood subjects had tried 2.0 (range 1–3) brands before discontinuing treatment.

3.1. Personality 

The mean personality scores in the groups varied between 41.7 and 53.0 and had standard deviations ranged between 6.1 and 12.4.

Personality traits assessed by the SSP revealed several statistically significant differences between subject currently on COC with and without ongoing adverse mood symptoms (Table 3). The women who were experiencing mood-related side effects on their current COC exhibited higher mean scores on the somatic anxiety trait (p<0.026) and on the stress susceptibility trait (p<0.026) compared to the women who did not experience any mood-related side effects from their current COCs.

Table 3. Mean±S.D. of personality trait t-scores for the groups on current COC treatment

		COC-fine (n=30)	COC-mood (n=28)	p-Value
	Somatic anxiety	47.2±6.1	52.0±9.7	ns
	Psychic anxiety	44.5±6.7	50.0±9.3	<0.026
	Stress susceptibility	45.6±8.0	52.1±12.4	<0.026
	Lack of assertiveness	46.5±8.5	50.5±9.0	ns
	Impulsiveness	46.2±9.3	42.3±9.8	ns
	Adventure seeking	49.4±9.3	46.0±9.0	ns
	Detachment	45.3±6.2	46.1±8.5	ns
	Social desirability	48.9±7.6	48.3±10.6	ns
	Embitterment	44.5±7.7	44.4±7.2	ns
	Trait irritability	49.3±9.5	51.8±8.0	ns
	Mistrust	45.7±8.5	44.0±8.0	ns
	Verbal trait aggression	50.0±9.3	47.6±10.2	ns
	Physical trait aggression	46.7±8.4	43.2±8.7	ns

Women who had discontinued COC treatment because of adverse mood effects had higher scores of detachment (p<0.026) and mistrust (p<0.026) compared to women who had discontinued COC for reasons unrelated to mood effects (Table 4).

Table 4. Mean±S.D. of personality trait t-scores for the groups with previous experience of COCs

		pCOC-fine (n=27)	pCOC-mood (n=33)	p-Value
	Somatic anxiety	46.1±6.9	48.1±6.8	ns
	Psychic anxiety	47.5±7.2	49.6±8.3	ns
	Stress susceptibility	49.4±9.6	51.9±10.5	ns
	Lack of assertiveness	52.2±10.1	48.8±9.1	ns
	Impulsiveness	46.2±7.9	49.3±7.5	ns
	Adventure seeking	50.1±8.1	47.3±7.2	ns
	Detachment	44.8±8.0	50.3±8.7	<0.026
	Social desirability	53.0±6.8	50.2±8.7	ns
	Embitterment	46.0±7.5	45.1±7.4	ns
	Trait irritability	46.8±9.6	49.6±10.6	ns
	Mistrust	41.7±7.3	46.8±10.2	<0.026
	Verbal trait aggression	46.6±9.6	48.2±8.7	ns
	Physical trait aggression	40.6±7.3	44.2±10.7	ns

Finally, although not included in the primary aims of the study, subjects who continued COC use in spite of adverse mood symptoms were compared to women who had discontinued because of these side effects. Women who had discontinued COC treatment due to adverse mood symptoms had higher scores of impulsiveness (p<0.001) than women who continued COC despite their experience of adverse mood.

4. Comment 

The main finding of the present study is that higher scores on specific personality traits such as somatic anxiety and stress susceptibility are found in women with ongoing experience of adverse mood symptoms from COC. Likewise, higher scores of mistrust and detachment are more common among women who have discontinued COC treatment due to adverse mood effects.

These traits, together with psychic anxiety, embitterment, trait irritability and lack of assertiveness, are all included in the SSP domain neuroticism [16]. High levels of neuroticism, high levels of interpersonal dependency, lack of self-confidence, and introversion are known to be associated with major depression [8], [9], [10], [11], [17]. Also, patients with premenstrual syndrome have been reported to have increased scores on somatic anxiety, although they also displayed increased scores of indirect aggression, verbal aggression and lower scores on socialization than healthy controls [18]. Although the majority of subjects in the current study was not depressed, the personality traits displayed by women suffering from adverse mood symptoms during COC treatment are to some extent similar to the ones encountered among depressed subjects and among patients with premenstrual syndrome.

Hence, it is possible that treatment with COCs may evoke symptoms in a group of women who, because of high neuroticism, are at risk of developing a depressive episode later in life or that women with high neuroticism are more likely to report adverse mood to health care professionals.

Personality traits have previously been shown to be fairly stable [19]. However, bearing in mind that many women start treatment with COCs at a very young age, the possibility that a continued use of oral contraceptives could affect personality traits in adult women cannot be ruled out. Relatively short-term use of SSRI has previously been shown to modulate dimensions of normal personality, even though subjects were not clinically depressed when treatment started [20]. However, the present study was not designed to detect changes in personality traits following the use of COC, and longitudinal studies would be required to verify any influence of that kind.

Finally, our study also indicates that there are differences in personality traits between those women who remain with their COC use although they experience mood-related side effects compared to those who discontinue due to the side effects. Possibly those staying on the pill are more anxious of the outcome of a discontinued use, i.e. an unwanted pregnancy, while those discontinuing give into their more impulsive personality. The results illuminate the importance of adequate counseling of women with adverse mood effects from their COC. Unless advice is based on evidence and sound expectations, the woman might be left with no other choices than to discontinue treatment.

There are a number of limitations to the present study that may interfere with the interpretations of our findings. First, most of the women recruited for the study were university students so our findings might not be generalized to other populations of women. Second, patients were grouped according to their own, retrospective reports of adverse mood symptoms during ongoing or prior COC use. Although the design of the present study is relevant for the clinical practice where patients seek care for adverse mood effects from COCs, prospective confirmation of mood deterioration would have been preferable. With a longitudinal design, where subjects are assessed prior to start of COC use, it would be possible to distinguish subjects who deteriorate due to the COCs and exclude subjects who have other pre-existing causes for their mood symptoms. A longitudinal design was, however, considered unsuitable as most women report unchanged or improved mood during COC treatment, thus requiring the screening of a substantially larger study population.

In conclusion, the present study has indicated that certain personality traits such as somatic anxiety, stress susceptibility, mistrust and detachment are more common among women who report adverse mood effects from COCs. These personality traits are also associated with increased risk of developing depression later in life. Based on our findings, it can be assumed that emotional side effects from oral contraceptive pills also have a psychological etiology.