Does fludrocortisone influence autobiographical memory retrieval? A study in patients with major depression, patients with borderline personality disorder and healthy controls

Abstract

There is evidence that stimulation of mineralocorticoid receptors (MR) enhances memory in healthy subjects and in patients with major depression (MDD). In contrast, in patients with borderline personality disorder (BPD), this effect seems to be task dependent. The aim of this study was to investigate the effect of MR stimulation on autobiographical memory retrieval in healthy individuals, patients with MDD, and patients with BPD. We conducted a placebo-controlled study in an intra-individual cross-over design. Twenty-four patients with MDD, 37 patients with BPD, and 67 healthy participants completed an autobiographical memory test after receiving 0.4 mg fludrocortisone, a mineralocorticoid receptor preferring agonist, or placebo in a randomized order. Healthy subjects, patients with MDD, and patients with BPD did not differ in their autobiographical memory retrieval. Furthermore, the administration of fludrocortisone had no effect on autobiographical memory. In conclusion, the stimulation of MR does not influence autobiographical memory retrieval in healthy subjects, patients with MDD, and patients with BPD. Our results do not support a role of MR in autobiographical memory.

• Major depression
• borderline personality disorder
• memory
• mineralocorticoid receptor

Introduction

There is evidence that stress or the acute administration of hydrocortisone impairs memory retrieval (de Quervain et al., 2009). Cortisol binds to two subtypes of receptors: the mineralocorticoid receptor (MR), which is mainly found in limbic structures, including the hippocampus and prefrontal cortical regions, and the glucocorticoid receptor (GR), which is expressed throughout the brain (de Kloet et al., 2005). While most of these effects of cortisol and stress on cognition have been attributed to GR function, recent studies emphasize the importance of MR (de Kloet, 2013).

Autobiographical memory retrieval is strongly associated with activation of the prefrontal cortex and the hippocampus (Svoboda et al., 2006) and might, therefore, be sensitive to cortisol effects. Indeed, two studies found that administration of 10 mg hydrocortisone impaired autobiographical memory retrieval in healthy subjects (Buss et al., 2004; Schlosser et al., 2010). Consistent with the hypothesis of reduced GR sensitivity in major depressive disorder (MDD), hydrocortisone did not lead to a decrease in autobiographical memory performance in these patients (Schlosser et al., 2010). In contrast, in patients with borderline personality disorder (BPD), autobiographical memory retrieval was enhanced after hydrocortisone compared with placebo (Wingenfeld et al., 2013).

As hydrocortisone stimulates both receptor types, it remains unclear whether the effect on memory is due to GR or MR stimulation. We recently showed that fludrocortisone, a MR preferring agonist, leads to improved verbal memory and executive function compared with placebo in healthy individuals and in patients with MDD (Otte et al., 2015). In contrast, patients with BPD performed worse after fludrocortisone compared with placebo in verbal memory and visuospatial memory while working memory was improved (Wingenfeld et al., 2015).

The effect of MR stimulation on autobiographical memory, which is strongly related to hippocampal function, has not yet been investigated. Therefore, the aim of this study was to investigate whether MR stimulation influences autobiographical memory performance in MDD patients, patients with BPD, and healthy subjects. Based on our recent results (Otte et al., 2015; Wingenfeld et al., 2015), we hypothesized that fludrocortisone would improve autobiographical memory retrieval in patients with MDD and healthy controls but would impair autobiographical memory retrieval in patients with BPD.

Methods

Participants

We recruited 24 patients with major depression, 37 patients with borderline personality disorder, and 70 healthy participants at the age of 18–65. Only women were included because of higher prevalence rates of BPD in women (Lieb et al., 2004). Inclusion criteria for patients were diagnoses of major depression or borderline personality disorder based on the MINI-International Neuropsychiatric Interview (Sheehan et al., 1998) or the SCID-II (Structured Clinical Interview for DSM-IV, Wittchen et al., 1997). Exclusion criteria for healthy participants were a former or a present DSM-IV axis I disorder according to the MINI-interview. All participants and patients were excluded if they met one of the following criteria: intake of psychotropic medication, serious medical conditions associated with adrenal dysfunction or well-known impact on HPA activity or cognitive function, steroid use, pregnancy, and nursing.

Patients were recruited at Charité University Berlin, Campus Benjamin Franklin, whereas healthy participants were recruited by local advertising. The study was approved by the local ethics committee. All procedures were carried out with full understanding of the participants. Written informed consent was obtained.

Procedures

We used a double-blind cross-over design with a randomized test order. Each participant was tested twice with a test re-test interval of 3 d. Participants received either a dosage of fludrocortisone (0.4 mg) or placebo. The order of fludrocortisone and placebo administration was counterbalanced. Drugs were administered 90 min prior to testing which took place between 14:00 h and 16:00 h. Two parallel versions of an Autobiographical Memory Test were presented to each participant with both versions being counterbalanced across the two treatment conditions.

A modified version of the autobiographical memory test (AMT, for a more detailed description see Buss et al., 2004; Schlosser et al., 2010) was used. In this test, emotional adjectives that represent an emotional state are presented to the participants consecutively on cards. The participants were instructed to write down a specific event from their past in response to the adjective and were asked to describe the location, time, persons involved and activities carried out at this event. After completing the AMT, two independent and trained raters evaluated the answers regarding specificity of the response. A memory was considered as specific if at least three out of four criteria (time, location, persons, and activity) were met. If the participant described two or less criteria, a memory was considered as categorical and, therefore, non-specific. We also recorded whether participants could not produce a memory in response to an adjective which is considered as non-memory. Before completing the test, participants practiced on an additional cue word with a neutral valence to check if they fully understood the instruction.

We used two parallel versions of the AMT. Both versions are identical in instruction and procedure and consisted of six emotional adjectives (two of each neutral, positive, and negative) whereby the adjectives differed between test versions. Participants completed one of the versions at each test session.

Statistical analyses

Demographic characteristics between all groups were compared using univariate analysis of variance (ANOVA) for continuous variables and Chi-square tests for dichotomous variables. To analyze the number of specific memories, categorical memories, and non-memory answers, a repeated measures (rm) – ANOVA with treatment as within-subject factor and group as between factor was conducted for each outcome variable separately. To examine the influence of valence, an rm-ANOVA was conducted with treatment and valence as within-subject factors and group as between-subject factors. All analyses were repeated with age, years of education, smoking, and cigarettes a day as covariates and the use of oral contraceptives as additional between-subject factor because all groups differed significantly on those variables.

Results

Demographic data

Demographic characteristics are presented in Table 1. Patients and healthy subjects differed significantly regarding age and years of education. Patients with major depression were significantly older than healthy controls and patients with borderline personality disorder. The borderline personality disorder group also showed significantly fewer years of education compared with patients with major depression and healthy controls. All groups differed significantly in smoking habits and in use of oral contraceptives. Both patient groups included significantly more smokers than the control group, whereas the control subjects use oral contraceptive more often compared with patients with MDD as well as patients with BPD. All patients and participants were medication free.

Table 1. Sample characteristics (means and standard deviations).

	Major depression (N=24)	Borderline personality disorder (N=37)	Healthy subjects (N=67)	Statistical analysis
Age	35.8 (14.8)	24.8 (5.8)	30.2 (13.1)	F(2,127) = 6.3, p=0.002
Years of education	11.9 (1.5)	10.2 (1.5)	11.4 (1.5)	F(2,124) = 9.3, p<0.001
Body mass index	22.8 (4.1)	26.6 (6.9)	23.31 (4.4)	F(2,127) = 5.5, p=0.005
Smoking (yes, %)	12 (50)	25 (67.6)	23 (34.3)	χ^2(2) = 13.9, p=0.001
Cigarettes a day	6.2 (10.3)	11.5 (12.3)	3.4 (5.5)	F(2, 125) = 9.8, p<0.001
Oral contraceptives (%)	3 (12.5)	7 (18.9)	25 (37.3)	χ^2(2) = 12.5, p=0.002

Autobiographical memory test

With regard to the number of specific memories, ANOVA revealed neither a significant main effect for group (F(2,125) = 0.74, p=0.18) nor for treatment (F(1,125) = 0.26, p=0.61) nor a significant treatment by group interaction (F(2,125) = 1.37, p=0.26). When controlling for covariates such as age, years of education, BMI and intake of oral contraceptives, results remained the same. Results are presented in Figure 1.

Figure 1. Number of specific memories in patients with major depression (MDD), patients with borderline personality disorder (BPD) and healthy controls (HC) after placebo and fludrocortisone administration in means.

When taking word valence into account, ANOVA revealed a significant main effect for valence (F(2,250) = 7.63, p=0.001), but no significant group effect (F(2,125) = 1.74, p=0.18). Post hoc tests indicated that across groups, participants recalled more specific memories for positive words than for neutral (p < 0.001) or negative words (p < 0.05). There was no significant difference between neutral and negative words (p > 0.05). When controlled for the above-mentioned covariates, the main effect for valence failed to reach significance (F(2,202) = 2.14, p=0.12).

Regarding the number of categorical memories, ANOVA revealed a significant main effect of group (F(2,158) = 5.09, p=0.195) but no main effect of treatment (F(1,158) = 2.39, p=0.12) nor a significant group × treatment interaction (F(2,158) = 1.65, p=0.19). Post hoc t-tests revealed that healthy controls recalled more categorical memories when compared with patients with BPD (p=0.007) but after controlling for all covariates, the main effect of group failed to reach significance (p=0.156).

With regard to the number of non-memory answers, ANOVA revealed neither a significant main effect of treatment (F(1,158) = 0.56, p=0.45) or the group F(2,158) = 1.71, p=0.184) nor a significant interaction of the treatment and the group F(2,158) = 0.68, p=0.51).

Discussion

This study investigated the influence of fludrocortisone, a MR preferring agonist, on autobiographical memory retrieval in patients with MDD, patients with BPD and healthy subjects. Contrary to our hypothesis, there was no significant effect of fludrocortisone on autobiographical memory retrieval in any of the three groups. Furthermore, groups did not differ in autobiographical memory performance.

Our results in healthy subjects are in contrast to our previous results, in which verbal memory, visuospatial memory and working memory were enhanced by MR stimulation through fludrocortisone (assessed by the Verbal Learning Memory Test, the Rey Osterrieth Complex Figure Test and Digit Span, respectively; Hinkelmann et al., 2015; Otte et al., 2015) and impaired by a MR blockade (Otte et al., 2007). Furthermore, in the current study, fludrocortisone did not influence autobiographical memory performance in patients with major depression, whereas in a previous study verbal memory and executive functions assessed by the Trail Making Test were enhanced after fludrocortisone intake (Otte et al., 2015). In patients with borderline personality disorder, we did not find a significant difference in autobiographical memory after fludrocortisone compared with placebo. These results support the hypothesis that the effect of fludrocortisone might be task dependent in patients with BPD. In previous studies in BPD, working memory was enhanced whereas verbal memory and visuospatial memory were impaired by fludrocortisone (Wingenfeld et al., 2015).

However, there are methodological differences between the current and previous studies, which have to be taken into account when comparing these results. First of all, the current study is the first one to examine the influence of fludrocortisone on memory retrieval alone, whereas the former studies did not separate between memory consolidation and memory retrieval. In fact, both memory processes were influenced by fludrocortisone. Therefore, it is not clear if the increased memory performance following fludrocortisone was due to enhanced memory consolidation or retrieval. Second, in the current study, we examined autobiographical memory retrieval and, thus remote memory. In the previous studies, memory acquisition took place at the same day or only an hour before memory retrieval. Therefore, it might be possible that we could not replicate former findings because the underlying memory processes are quite different.

In contrast to fludrocortisone, hydrocortisone, which stimulates both the MR and the GR, had been shown to influence autobiographical memory differentially among healthy subjects, patients with MDD or BPD. Hydrocortisone impaired autobiographical memory retrieval in healthy subjects (Buss et al., 2004; Schlosser et al., 2010), but did not lead to a decrease in autobiographical memory in depressed patients (Schlosser et al., 2010). In patients with BPD however, autobiographical memory retrieval was even enhanced after hydrocortisone compared with placebo (Wingenfeld et al., 2013). Taken together, these results suggest that GR play a significant role in autobiographical memory but MR do not, although the database so far is limited. However, it is important to keep in mind that we used a short version of the AMT and, therefore, a direct comparison between our earlier and the current study is not possible.

There was no group effect in the current study, while previous research found impaired general and autobiographical memory retrieval in MDD patients (Van Vreeswijk & De Wilde, 2004; Williams et al., 2007). Most of the former studies included patients with psychotropic medication. It is noteworthy that patients in the current sample were free of psychotropic medication. Possibly this difference between samples led to different results. In future studies, it is of interest to further investigate to which extent psychotropic medication influences memory performance in psychiatric patients. That patients with BPD did not differ from healthy controls is in line with previous studies (Winter et al., 2014).

There are several limitations in this study. First of all, the number of participants varied between groups. Additionally, groups differed in demographic variables such as age, education level, BMI, smoking habits and usage of oral contraceptives. Although we controlled for all these variables in our analyses, we cannot definitively rule out residual confounding to some extent. In future studies, it will be important to investigate more homogenous groups. Furthermore, we only included women in the study whereas in other studies only men or both male and female subjects participated. Future studies are necessary to investigate possible gender effects regarding MR stimulation and autobiographical memory.

Fludrocortisone has some glucocorticoid potency in addition to its mineralocorticoid activity, although its MR affinity is about 150 times higher than its GR affinity (Agarwal et al., 1977). The extent of its glucocorticoid potency ranges from negligible to rather moderate (Grossmann et al., 2004; Miller et al., 2008). Because fludrocortisone has its peak of concentration 1.7 h after drug administration, we cannot disentangle fast non-genomic and slow genomic actions of fludrocortisone. This should be investigated in future studies.

The few studies, which examined cortisol effects on autobiographical memory (Schlosser et al., 2010; Wingenfeld et al., 2015), have largely reported effect sizes of a small magnitude (Cohen’s d <0.3). Therefore, and conservatively, assuming that the true effect size of fludrocortisone on autobiographical memory is rather small with Cohen’s d in the range of 0.15–0.3, our study has a power of between 86% and 99%. Thus, we are confident that our negative results are unlikely to be false negatives.

In conclusion, the stimulation of MR does not influence autobiographical memory retrieval in healthy subjects, patients with MDD, and patients with BPD. Future studies are necessary which investigate the influence of fludrocortisone on memory consolidation and memory retrieval separately to fully understand the role of MR in memory processes.

Declaration of interest

Dr. Otte was supported by the Deutsche Forschungsgemeinschaft (OT 209/7-1 and EXC 257 NeuroCure). Furthermore, this work was supported by NARSAD (Grant ID: 18766) awarded to K. W. Dr. Otte has received honoraria fees for lectures from Lundbeck and Servier and has received compensation as a member of the scientific advisory board of Lundbeck. Dr. Hinkelmann, Dr. Kühl, Dr. Wingenfeld, Dr. Roepke and Mrs. Fleischer report that they have no conflict of interest.