A four week randomised control trial of adjunctive medroxyprogesterone and tamoxifen in women with mania

Psychoneuroendocrinology (2014) 43, 52—61

Available online at www.sciencedirect.com

ScienceDirect j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / p s y n e u e n

A four week randomised control trial of adjunctive medroxyprogesterone and tamoxifen in women with mania Jayashri Kulkarni a,*, Michael Berk b, Wei Wang a, Ling Mu a, Elizabeth Scarr c, Tamsyn E. Van Rheenen a, Roisin Worsley a,d, Caroline Gurvich a, Emorfia Gavrilidis a, Anthony de Castella a, Paul Fitzgerald a, Susan R. Davis d a

Monash Alfred Psychiatry Research Centre, Central Clinical School, Monash University and the Alfred Hospital, Melbourne, Victoria 3004, Australia b IMPACT Strategic Research Centre, School of Medicine, Deakin University, Barwon Health, Ryrie Street, Geelong, Victoria 3220, Australia c The Florey Institute of Neuroscience and Mental Health Victoria, Parkville, Victoria 3052, Australia d Women’s Health Research Program, School of Public Health and Preventive Medicine, Monash University and Alfred Hospital, Victoria 3004, Australia Received 10 December 2013; received in revised form 5 February 2014; accepted 6 February 2014

KEYWORDS Tamoxifen; Mania; Treatment; Bipolar disorder; Schizoaffective disorder; Medroxyprogesterone acetate

Summary Emerging research has suggested that hormone treatments such as selective oestrogen receptor modulators (SERMs) or progestins may be useful in the treatment of mania. The current pilot study compared the use of the SERM tamoxifen and the progestin medroxyprogesterone acetate (MPA), as an adjunct to mood stabiliser medications, for the treatment of mania symptoms in 51 women in a 28-day double blind, placebo controlled study. The primary outcome was the change between baseline and day 28 mania scores as measured by the Clinician Administered Rating Scale for Mania (CARS-M). Adjunctive MPA treatment provided greater and more rapid improvement in mania symptoms compared with adjunctive placebo and tamoxifen treatment. Adjunctive therapy with MPA may be a potentially useful new treatment for persistent mania, leading to a greater and more rapid resolution of symptoms compared with mood stabiliser treatment alone. # 2014 Elsevier Ltd. All rights reserved.

* Corresponding author at: Monash Alfred Psychiatry Research Centre, Level 4, 607 St Kilda Road, Melbourne 3004, Australia. Tel.: +61 3 9076 6924; fax: +61 3 9076 8545. E-mail addresses: [email protected], [email protected] (J. Kulkarni). http://dx.doi.org/10.1016/j.psyneuen.2014.02.004 0306-4530/# 2014 Elsevier Ltd. All rights reserved.

Tamoxifen and medroxyprogesterone treatment in mania

1. Introduction Mania is characterised by a prolonged period of excessively elevated or irritable mood (American Psychiatric Association, 2000). It is the hallmark of bipolar affective disorder (BPAD) and is also seen in schizoaffective disorder. Currently, the biological processes involved in mania pathophysiology are unclear, and its treatment remains complex (Geddes and Miklowitz, 2013). Women with BPAD experience fluctuating mood across the menstrual cycle, with an exacerbation of symptoms reported during the premenstrual period (Rasgon et al., 2003, 2005; Payne et al., 2007; Dias et al., 2011). Further, evidence suggests that women with BPAD have an increased risk of mood disorder during menopause, a time of decreasing oestrogen levels (Freeman et al., 2002; Marsh et al., 2012). Thus, in women, there is suggestive evidence that endogenous oestrogen levels may mediate manic symptoms. In previous work, we found that a small subgroup of schizoaffective women with acute mania who received 50 mcg of adjunctive transdermal estradiol for 28 days experienced a worsening of manic symptoms (Kulkarni et al., 2001). We hypothesised this as the effect of an increase in Central Nervous System (CNS) oestrogen exposure. This hypothesis is supported by case reports of improvement in mania with treatments that reduce estradiol levels by inducing anovulation, such as danazol (Goldstein, 1986; Nelson, 1988) and medroxyprogesterone acetate (MPA: Chouinard et al., 1987). In early animal studies, MPA has been shown to cause down regulation of oestrogen receptors in the hypothalamus and pituitary (Blaustein and Brown, 1984) and has an ‘anti-estrogenic’ effect in the brain generally (Irwin et al., 2011). Tamoxifen is a selective oestrogen receptor modulator (SERM), with tissue specific effects on oestrogen receptors as well as second messenger pathways (Lam, 1984). Tamoxifen has mainly ‘anti-estrogenic’ effects in many parts of the CNS, although studies are largely inconclusive (Paganini-Hill and Clark, 2000; Buwalda and Schagen, 2013). Several small studies have also demonstrated a clinical benefit of tamoxifen in the treatment of mania (Manji and Lenox, 1999; Bebchuk et al., 2000; Zarate et al., 2007; DiazGranados and Zarate, 2008; Yildiz et al., 2008; Amrollahi et al., 2011; Yildiz et al., 2011). A small study found tamoxifen to be superior to MPA and clomiphene in preventing amphetamine induced hyper locomotion in an animal model of mania (Pereira et al., 2011). Our group has previously reported results of a three arm pilot study comparing adjunctive tamoxifen, MPA and placebo in the treatment of 13 women with manic symptoms (Kulkarni et al., 2006). The women treated with tamoxifen had a greater improvement than the MPA group, and both groups showed a greater improvement than placebo. While both tamoxifen and MPA have oestrogen antagonism effects, tamoxifen is also known to be a potent protein kinase-C (PKC) inhibitor. Elevated PKC levels are seen in mania, and the resolution of mania is associated with reductions in PKC levels (Friedman et al., 1993; Kulkarni et al., 2006; Einat et al., 2007; DiazGranados and Zarate, 2008). The study reported here expands our pilot study. The aim was to compare the use of two adjunctive agents (tamoxifen and MPA) in a 28-day three-arm double blind, placebo-controlled study in the treatment of acute mania or hypomania

53 to further explore dosing effects in larger sample. Extrapolating from our pilot results, we hypothesised that tamoxifen therapy would be associated with a significant reduction in mania symptoms compared to the MPA and placebo groups. We also attempted to understand more about the mechanisms underpinning any reduction in mania symptoms. In particular, we were interested in exploring whether tamoxifen reduced mania symptoms through PKC inhibition or by oestrogen antagonism. To do this we included the MPA arm as a direct comparator for an ‘‘anti -estrogen’’ effect, and examined for changes in PKC levels across the trial.

2. Method This three arm double blind randomised controlled trial was approved by The Alfred Hospital Ethics Committee and Barwon Health Research and Ethics Advisory Committee (Alfred Project Number 77/02). All participants provided written informed consent. The registration number of this trial (at clinicaltrials.gov) is NCT00206544.

2.1. Participants Women were assessed for eligibility across two Australian sites (the Alfred Hospital, a tertiary hospital and Barwon Health, a regional hospital) between 2004 and 2007. Eligible women were aged between 18 and 65; able to give informed consent; taking a mood stabiliser (lithium, sodium valproate or carbamazepine) and/or a mood — stabilising antipsychotic and had a diagnosis of schizoaffective disorder or BPAD made by their treating psychiatrist which was confirmed by the Structured Clinical Interview for DSM-IV Disorders (SCID-IV). Participants were also required to be currently manic as defined by a score of 15 or above on the Clinician Administered Rating Scale for Mania (CARS-M). Women were excluded if they were receiving oestrogen or progestin therapy including the oral contraceptive pill; were pregnant or lactating; had a history of hyperthyroidism or any unstable neurological or other serious medical condition; were taking drugs known to interact with tamoxifen or MPA including warfarin, aminoglutethimide, diuretics, methyldopa, theophylline, fluoxetine, calcium channel blockers and non-steroidal anti-inflammatory drugs; or had a history of substance abuse or dependence in the previous six months. Six of the final sample of 51 women were postmenopausal, as classified by a baseline FSH of more than 25 (Harlow et al., 2012) and amenorrhoea.

2.2. Treatment All participants were randomised by the Alfred Clinical trials pharmacy to receive either tamoxifen 40 mg daily (n = 15); MPA 20 mg daily (n = 18); or placebo capsules (n = 18) according to a computer generated randomisation list.1 Bebchuck et al. (2000) reported that a dose of up to 80 mg of oral tamoxifen was effective in reducing manic symptoms in both

1 The same doses of tamoxifen and MPA were used in our pilot study (see Kulkarni et al., 2006).

54 males and females. However, given our sample comprised only women, and as tamoxifen increases the risk of hot flushes in women, a more tolerable, lower dose of 40 mg/ day oral tamoxifen was used in this study. The placebo was non-active milk powder. The Alfred Clinical Trials Pharmacy identically encapsulated all medications. All participants and study personnel remained blind to treatment allocation for the duration of each participant’s enrolment in the study.2 To ensure that hot flushes/menstrual irregularities did not interfere with treatment blinding, all women were told that menstrual irregularities were possible with any of the treatments (although none of the women experienced disturbing hot flushes). Study medication was given as an adjunct to each participant’s standard treatment. To ensure medication adherence, patients were questioned by the researchers about their medication at each visit using a medication adherence rating scale. Study medications were given to patients at each visit and compliance was assessed by the return of unused capsules and verbally by interview during weekly visits. This was documented in the participants case report form.

2.3. Primary outcome measures: mania Mania was assessed at baseline and then weekly for four weeks using the CARS-M (Altman et al., 1994). The CARS-M is a 15-item scale that assesses manic and psychotic symptoms over the last 7 days, with scores ranging from 0 to 74. The CARS-M mania subscale (CARSMANIA) includes the ten items specific to mania and is scored from 0 to 50, with a score of 26 or more indicating severe pathology.

2.4. Secondary outcome measures: PKC analysis Blood was drawn at days 7, 14, 21 and 28 to enable platelet PKC analysis, hormone analysis and weekly analysis of mood stabiliser serum levels. Details about platelet preparation and PKC assay techniques are included as Appendix 1.

2.5. Additional measures Side effects were monitored weekly using the following scales: the 21-item Adverse Symptoms Checklist (ASC) where each item rated 0—3 [0 = absent; 1 = slight; 2 = moderate; 3 = severe]). The 10 item Simpson-Angus Rating Scale for Extrapyramidal Side Effects (SAS: Simpson and Angus, 1970) where total scores range from 0 to 40. The 12 item Abnormal Involuntary Movement Scale (AIMS: Guy, 1976) where total scores range from 0 to 28. Adverse effects of hormone therapy such as hot flushes or erratic menstrual bleeding were monitored at each visit and recorded in the case report form.

2

We chose to run the study over a 28 day period and no longer, to ensure maximal recruitment and retention of this difficult group. Twenty-eight days, which is a standard menstrual cycle, was also chosen to also minimise menstrual irregularities over several cycles and hence maintain the blinding as much as possible.

J. Kulkarni et al.

2.6. Statistical analysis The sample size was determined according to power analyses based on our pilot study. This power analysis was performed using a difference of 8 points on baseline CARS-M as a meaningful difference. At a = 0.01, minimum n = 50 gives power of 0.992.3 Demographic differences between groups were assessed by one way analysis of variance (ANOVA) or Chi square tests. Multilevel modelling was used to assess changes in the outcome measures (Hox, 2010). There are a number of advantages of using multilevel modelling over repeated measure MANOVA (Van Der Leeden, 1998): (1) it makes no requirements for measurements assessed at the same time points or for the time points to be equally spaced, (2) less problematic with missing data and small sample size, (3) it takes into account the individual growth, (4) the within-subject covariance structure can be modelled and (5) time-varying covariates can be included. Multilevel modelling implemented via Mplus version 7 (Muthe ´n and Muthe ´n, 1998—2013) was used to model the changes on the three outcome variables, namely the CARS-M Total score, the CARSMANIA Sub-scale and PKC. These repeated measures were viewed as hierarchical structure, with observations over time at level 1 nested within patients at level 2 (Hox, 2010). In a two-level model, the level 1 model specifies polynomial growth curves for each patient. At the level 2, it models the growth parameters as a mean growth of all patients plus their growth variations, which may be explained by the covariates. Fig. 1 presents the hierarchical structure of the analysis. In the multilevel model for representing change on the three outcome variables (CARS-M, CARSMANIA, and PKC), the changes are predicted by time and serum levels of mood stabiliser at level 1. The regression of each outcome variable on time and serum levels of mood stabiliser (time-varying covariate) is assumed to have variation across individuals around mean slope (rate of change), and variation across individuals around mean intercept (baseline level). The random intercept and slopes are predicted by the time-invariant covariates including hospital site, tamoxifen, MPA, age, diagnosis, and mood stabiliser at level 2.

3. Results 3.1. Descriptive analyses There were no significant differences in mean age, diagnosis, age at illness onset or number of hospitalisations between the treatment groups (Table 1), nor were there significant differences in the distribution of women taking mood stabilisers (x2(8) = 13.18, p = 0.11) or antipsychotics (x2(2) = 1.81, p = 0.40) at baseline.

3

To determine if menopausal status had an effect on the different treatments, we ran a Fishers Exact Test to estimate the relationship between these variables. The results did not reach significance ( p =0.345) and thus subsequent analyses were run including all six postmenopausal women.

Tamoxifen and medroxyprogesterone treatment in mania

55

Figure 1 Mplus notation for the hierarchical model. MPA = medroxyprogesterone; s1 = the effect of time on the outcome measures (e.g. CARS-M, CARSMANIA, PKC); s2 = the effect of level of stabiliser on the outcome measures (e.g. CARS-M, CARSMANIA, PKC).

Table 1

Demographic baseline data for the 51 participants.

Demographics

Tamoxifen (n = 15) M(SD)

MPA (n = 18) M(SD)

Placebo (n = 18) M(SD)

Group comparisons

Mean age [years] Mean age of illness onset [years] Mean number of hospital admissions since illness onset Postmenopausal # Diagnosis # Bipolar disorder Schizoaffective disorder Current mood stabiliser [% of sample per group] Lithium Sodium valproate Carbamazepine None Current mood-stabilising antipsychotic treatment Yes No

36.80 (8.29) 24.46 (6.49) 8.45 (14.87) 1

40.72 (11.10) 27.5 (9.8) 7.3 (11.9) 4

39.22 (9.27) 25.80 (6.20) 7.72 (8.57) 1

F(2,48) = 0.67, p = 0.52 F(2,42) = 0.55, p = 0.58 F(2,38) = 0.03, p = 0.97 x2(2) = 2.94, p = 0.23

13 2

17 1

12 6

x2(2) = 5.05, p = 0.08 x2(8) = 13.18, p = 0.11

53.3 26.7 0 13.3

38.9 61.1 0 0

55.6 22.2 11.1 11.1

14.3 85.7

22.2 72.2

5.6 94.4

x2(2) = 1.81 p = 0.40

3.2. Primary outcome measures: CARS-M total and subscale analyses There were no significant differences in CARS-M total score at baseline. The CARS-M Total score and CARSMANIA Subscale means scores decreased across the treatment period in the treatment groups (Fig. 2). This occurred more quickly in

the MPA group relative to the tamoxifen and placebo groups. Using multilevel modelling, there was a significant interaction effect between MPA treatment and the effect of the mood stabiliser on the CARS-M total score (b = 0.16, p < 0.05) (Table 2), such that there was a statistically significant decrease in CARS-M in the MPA group compared with the tamoxifen and placebo groups.

56

J. Kulkarni et al. Tamoxifen

MPA

Placebo

45 CARS-M Total score

40 35 30 25 20 15

*

10

^ *

5 0

Baseline

Day 7

Day 14

Day 21

Tamoxifen

MPA

Placebo

Day 28

CARSMANIA Sub-scale

35 30 25 20 15 *

10 5 0

^ Baseline

Day 7

Day 14

Day 21

Day 28

Figure 2 Means at the measurement occasions for CARS-M total score and CARSMANIA subscale score. Note MPA = medroxyprogesterone; *p < 0.05; ^p = 0.07. MPA < Placebo; error bars represent standard deviations.

3.3. PKC activity, hormone analysis and serum mood stabiliser analysis Table 3 shows descriptive statistics for the secondary outcome variable, the platelet PKC enzyme levels, at four time points for each treatment group. The mean PKC levels decreased in the MPA and tamoxifen groups, but increased in the placebo group over the 28 day period, although this was not significant. The parameter estimates for PKC in Table 4 suggests that PKC levels remained more stable in the MPA group as the effect of time on PKC was negatively related to treatment using MPA (b = 88.41, p < 0.05). Table 5 presents the mean hormone and serum mood stabiliser level changes from baseline to day 28 of the trial across the three groups. There was no change in serum mood stabiliser levels, indicating that all women remained compliant with their mood stabiliser medication for the duration of the trial. The only significant group difference in hormone levels was found in the oestrogen change from baseline, with an increase in oestrogen levels over time for the tamoxifen group compared to the placebo and MPA groups.

3.4. Additional measures: adverse events There were no significant changes, over time or by group, on the AIMS and SAS, nor was there a significant main effect of group (F(2,33) = 1.96, p = 0.16) or an interaction between group and time on the ASC (F(1,33) = 0.01, p = 0.99), suggesting all groups improved in a similar manner.

4. Discussion This study suggests that adjunctive MPA, but not tamoxifen, was beneficial in treating acute mania in women. Women in the MPA group had a greater reduction in symptoms and responded more rapidly. No statistically significant effect was seen in women treated with adjunctive tamoxifen. These findings are different to our original, smaller pilot study, which indicated adjunctive tamoxifen to be superior to MPA and placebo (Kulkarni et al., 2006). We suspect that the differences between our earlier findings and this second, larger, pilot study may be due to the larger sample size in the current study. Indeed, the current size of each group was over three times as large as the pilot groups, and the current results are therefore likely to better reflect the true efficacy of tamoxifen versus MPA treatment for mania. In the present data, the lack of effect in the tamoxifen group may be the result of the dosage used. As tamoxifen is known to cause the side effect of hot flushes in women, we used the lower dose of 40 mg/day in our study as we considered it to be safer in women. However, Bebchuk et al. (2000) described significant improvements in mania symptoms in men and women using a predominant dose of 60— 80 mg/day of tamoxifen, which is substantially larger than the dose of tamoxifen used here. Other studies reporting the efficacy of tamoxifen for the treatment of mania tended to increase/vary the initial 40 mg baseline dose over the course of the trials (Yildiz et al., 2008; Zarate et al., 2007). Although the 40 mg dose appeared to be sufficient in our initial pilot

Tamoxifen and medroxyprogesterone treatment in mania Table 2

Parameter estimates of the multilevel model for the primary outcome measures. Fix effect

Estimate

Tamoxifen ! CARS-M

CARS-M

CARSMANIA

se

z

Random effect Residual variance level 1 CARS-M Residual variance level 2 CARS-M Residual variance level 2 s1 Residual variance level 2 s2

8.92

21.76

0.41

Tamoxifen ! s1

1.90

9.47

0.20

Tamoxifen ! s2

0.03

0.03

0.90

MPA ! CARS-M

66.70

38.92

1.71

MPA ! s1 MPA ! s2 Intercept of CARS-M Intercept of s1 Intercept of s2

5.92 0.16 51.78 53.17 0.26

15.81 0.08 59.67 37.84 0.15

0.37 2.06 * 0.87 1.41 1.73

19.58

33.61

0.58

3.20

9.06

0.35

Site ! s1

23.55

11.07

2.12 *

Tamoxifen ! s2

0.07

1.29

0.06

32.09 7.53 0.12 18.66 57.88 0.37

40.35 8.90 2.92 74.14 32.18 1.24

0.80 0.85 0.04 0.25 1.80 0.30

14.57

156.91

0.09

Tamoxifen ! s1

26.31

26.34

1.00

MPA ! PKC

41.39

185.19

0.22

88.41 508.55 260.27 0.77 1165.04 14.08

43.53 154.36 120.10 0.22 364.33 84.31

2.03 * 3.29 ** 2.17 3.58 ** 3.20 0.17

Tamoxifen ! CARSMANIA Tamoxifen ! s1

MPA ! CARSMANIA MPA ! s1 MPA ! s2 Intercept of CARSMANIA Intercept of s1 Intercept of s2 PKC

57

Tamoxifen ! PKC

MPA ! s1 Site ! PKC Diagnosis ! PKC Mood stabiliser ! PKC Intercept of PKC Intercept of s1

Estimate

se

z

at

4584.70

2596.09

at

9899.26

97.29

101.75 **

at

674.36

552.36

1.22

at

0.03

2.23

0.01

Residual variance at level 1 CARSMANIA Residual variance at level 2 CARSMANIA Residual variance at level 2 s1 Residual variance at level 2 s2

4820.23

2810.49

1.72

14,093.49

21,679.23

0.65

635.09

291.95

2.18 *

0.03

198.99

0.00

30,824.56

12,309.34

2.50 *

37,295.19

5.14 **

3192.14

2.05 *

Residual variance at level 1 PKC Residual variance at level 2 PKC Residual variance at level 2 s1

— 6547.70

1.77

Note: There was a significant interaction effect between MPA treatment and the effect of the mood stabiliser on the CARS-M total score. * p < 0.05. ** p < 0.01. se = standard error; z = z score; s1 = the effect of time on the outcome measures; s2 = the effect of level of mood stabiliser on the outcome measure.

Table 3

Descriptive statistics for the secondary outcome measure over time. Placebo Mean

PKC enzyme in platelets 111.95 Day 7 Day 14 65.20 111.42 Day 21 153.60 Day 28

Tamoxifen 40 mg/day

MPA 20 mg/day

SD

Mean

SD

Mean

SD

98.20 133.16 171.73 398.53

125.75 111.62 145.32 88.12

84.72 114.44 235.12 97.86

152.64 150.65 61.08 96.16

295.56 240.95 37.80 148.33

Note: There were no significant differences in PKC levels between groups at the four time points (all p’s < 0.05).

58 Table 4

J. Kulkarni et al. Parameter estimates of the multilevel model for the secondary outcome measure.

Fix effect

Estimate

se

z

Random effect

Estimate

se

z

PKC Tamoxifen ! PKC Tamoxifen ! s1 MPA ! PKC MPA ! s1 Site ! PKC Diagnosis ! PKC Mood stabiliser ! PKC Intercept of PKC Intercept of s1

14.57 26.31 41.39 88.41 508.55 260.27 0.77 1165.04 14.08

156.91 26.34 185.19 43.53 154.36 120.10 0.22 364.33 84.31

0.09 1.00 0.22 2.03 * 3.29 ** 2.17 3.58 ** 3.20 0.17

Residual variance at level 1 PKC Residual variance at level 2 PKC Residual variance at level 2 s1

30,824.56 — 6547.70

12,309.34 37,295.19 3192.14

2.50 * 5.14 ** 2.05 *

Note: There was a significant interaction between time and MPA on PKC * p < 0.05. ** p < 0.01. se = standard error; z = z score; s1 = the effect of time on the outcome measures; s2 = the effect of level of mood stabiliser on the outcome measures.

study of only 13 women, when taken in combination with our current findings which are arguably more reliable in view of the larger sample, it appears that there is no good evidence of improvement using 40 mg tamoxifen. Thus it appears that larger doses may be necessary to see a treatment effect.

Table 5 Group

It is also possible that the specific treatment effects evident here may be related to these baseline medication differences, given that there was a greater proportion of women in the tamoxifen and placebo groups taking lithium and a greater proportion of women in the MPA group taking

Mean hormone and mood stabiliser levels at baseline and day 28 for the three groups. Hormone

Baseline

Day 28

Change from baseline *

M

SD

M

SD

M

SD

Tamox

Oestrogen Progesterone LH FSH DHEA Prolactin Testosterone Serum level of mood stabiliser

200.47 2.21 8.65 11.32 6.84 1290.79 1.65 241.98

204.55 1.41 12.83 13.17 4.30 1375.32 0.92 324.14

1071.85 15.24 8.04 9.61 7.28 1472.83 1.83 230.20

1002.65 27.04 10.87 11.36 3.08 1401.23 1.03 278.64

859.38 12.77 0.05 0.32 0.49 152.73 0.15 11.79

823.61 26.66 4.51 4.17 2.52 449.35 1.31 94.27

MPA

Oestrogen Progesterone LH FSH DHEA Prolactin Testosterone Serum level of mood stabiliser

221.53 6.36 9.76 16.79 4.54 1144.67 1.38 331.01

228.86 12.29 8.80 20.32 3.22 1219.68 1.09 258.05

111.86 1.71 2.79 9.40 3.88 1154.21 1.14 389.88

84.08 1.49 1.35 8.26 2.43 1121.98 0.64 358.98

113.92 1.12 5.09 5.34 0.27 129.29 0.34 28.84

183.90 3.90 7.38 9.58 1.24 987.69 0.79 256.56

Placebo

Oestrogen Progesterone LH FSH DHEA Prolactin Testosterone Serum level of mood stabiliser

326.17 3.80 8.79 17.34 5.00 919.41 1.33 211.01

406.38 4.03 11.09 17.37 5.39 831.87 1.49 261.31

195.25 4.65 10.54 28.57 4.31 1197.89 1.26 196.67

228.15 4.65 10.54 28.57 4.31 1197.89 1.26 296.23

145.35 0.13 1.99 1.05 0.09 74.07 0.02 0.60

342.60 7.82 10.21 12.77 2.72 572.56 0.90 154.45

Note: There were no significant differences in hormone or serum mood stabiliser levels between groups at baseline. * Indicates change from baseline across groups significant at p = 0.001 for oestrogen only (Tamoxifen > MPA, p = 0.001; Tamoxifen > Placebo, p = 0.001; MPA = Placebo, p = 0.99).

Tamoxifen and medroxyprogesterone treatment in mania sodium valproate. It is difficult to say whether treatment effects may have varied had the predominant mood stabiliser in the tamoxifen group been sodium valproate and the predominant mood stabiliser in the MPA group been lithium. Unfortunately sample size restrictions prevent us from analysing each group separately according to medication type, but future studies would do well to aim to control for these potential confounds by comparing the effects of tamoxifen versus MPA in adjunct to these different medications. Unexpectedly, the PKC levels were not different between groups and estradiol levels significantly increased in the tamoxifen group. This has been seen in other studies involving premenopausal women (Cohen et al., 1999). It is possible that higher serum estradiol levels in the tamoxifen group negated any positive effect of the PKC inhibition of tamoxifen. Alternatively, this lack of PKC effect may suggest that there are pathways other than PKC inhibition that are important in the treatment of mania. The present data certainly seem to suggest that adjunctive MPA, which is also known to have oestrogen antagonistic effects, is more effective in decreasing mania symptoms than adjunctive tamoxifen. This suggests that the MPA effect on manic symptoms may be mediated through an alternative mechanism. The results of this study should be interpreted within the confines of a number of limitations; we were unable to characterise patient menstrual phases as (i) the women were all acutely unwell and (ii) only 50% of the women had regular cycles, meaning that menstrual cycle phase could not be classified in 50% of the cohort. Thus, we were unable to investigate the relationship current phase may have had with treatment responsiveness. It is certainly possible that different phase related concentrations of progesterone and oestrogen at baseline may have impacted the effect of tamoxifen/MPA on symptomatology, and future studies would do well investigate this. Further, the sample size was relatively small; mania is a difficult disorder to manage, making recruitment of women with mania into a treatment trial a very difficult task. Large numbers of women needed to be screened, which makes widespread applicability of these results difficult. However, being readily available and well-tolerated, MPA could be considered as a useful adjunctive therapy in women with refractory mania. Although epidemiological studies have not shown depot MPA therapy used for contraception to positively or negatively affect rates of hospitalisation (Berenson et al., 2011), our results suggest that treatment with adjunctive MPA at the 20 mg/day dose we trialled in this study could potentially reduce the intensity and duration of manic symptoms. However, an issue regarding the role of progestins in BPAD is their probable unidirectional profile in terms of mood symptoms; aggravation of depressive symptoms with use of agents such as norethisterone is well documented, and this understanding is important to integrate into of the potential clinical role of progestins (Lawrie et al., 1998). Fluctuations in the hypothalamic pituitary gonadal (HPG) axis can have major impacts on the mental state and hormone treatments are an important area of emerging clinical research (Craig, 2013). Our work showing oestrogen therapy as a useful adjunct for schizophrenia treatment is an example

59 of hormone modulation as an effective treatment (Kulkarni et al., 2008). Women with BPAD also appear to experience HPG axis fluctuations, but most likely in a different manner to women with schizophrenia. Hence the hormone treatment we devised in this study of mania in women was an oestrogen antagonism approach. In conclusion, women with acute mania symptoms are difficult to treat and experience a great deal of adverse psychosocial effects. Adjunctive therapy with MPA 20 mg may be a useful new treatment for persistent mania, leading to a greater and more rapid resolution of symptoms compared with mood stabiliser treatment alone.

Role of the funding sources The Stanley Medical Research Institute, Washington USA, Grant ID: 03T-415. National Health and Medical Research Council Grant ID: 284319.

Conflicts of interest ES was supported by the Royce Abbey Postdoctoral Fellow (Australian Rotary Health). MB is supported by a NHMRC SPRF.

Acknowledgements The authors would like to acknowledge the financial support of the NHMRC (Australia) and Stanley Medical Research Institute (Washington).

Appendix 1 2.4.1. Platelet preparation 1 ml of anticoagulant (26.9 mM EDTA, containing 17.1 mM NaCl) was added to 9.0 ml venous blood, which was then gently inverted to mix and stored at 4 8C until processed. The blood samples were centrifuged at 100  g, 15 min, 4 8C and the supernatant (platelet rich plasma) collected. This, in turn was centrifuged at 9000  g for 6 min, 4 8C. The supernatant was discarded and the pellet (platelets) washed 3 times with 10 volumes 25 mM TRIS (pH 7.7) 120 mM NaCl, 2.5 mM EDTA and 100 mM PMSF before being resuspended in 1 volume of the same buffer and frozen, in 50 ml aliquots at 70 8C. 2.4.2. Preparation of samples for PKC assay 100 ml of frozen platelets were defrosted on ice and 200 ml of extraction buffer (25 mM Tris—HCl, pH7.4; 0.5 mM EDTA, 0.5 mM EGTA, 0.05% Triton1 X-100, 10 mM b-mercaptoethanol, 1 mg/ml leupeptin and 1 mg/ml aprotinin) added. Once defrosted the platelets were disrupted, on ice for a total of 20 s, using a Heidolph Diax 600. The preparation was centrifuged at 15,000  g for 5 min at 4 8C and the supernatant collected. The supernatant was semipurified using DEAE sephadex A-50 columns and the elutant used in the PKC assay.

60 The PKC assay was carried out, in triplicate, using the SignaTECT1 protein kinase C assay kit according the manufacturer’s instructions. Protein levels of the elutants were determined using the Bio-Rad DC protein assay adapted for the microplate. From these data, using the specific activity of [g-32P]ATP, basal PKC activity (pmol ATP/minute/ug protein), activated PKC activity, PKC activity (activated minus basal) and percentage activation were calculated. In order to assess the specificity of the assay, the activated reactions were also conducted in the presence of the selective PKC inhibitor, Go ¨ 6976, at a final concentration of 0.2 mM.

References Altman, E.G., Hedeker, D.R., Janicak, P.G., Peterson, J.L., Davis, J.M., 1994. The Clinician-Administered Rating Scale for Mania (CARS-M): development, reliability, and validity. Biol. Psychiatry 36, 124—134. American Psychiatric Association, 2000. Diagnostics and Statistical Manual of Mental Disorders, text revised, 4th ed. APA, Washington, DC. Amrollahi, Z., Rezaei, F., Salehi, B., Modabbernia, A.-H., Maroufi, A., Esfandiari, G.-R., Naderi, M., Ghebleh, F., Ahmadi-Abhari, S.-A., Sadeghi, M., Tabrizi, M., Akhondzadeh, S., 2011. Double-blind, randomized, placebo-controlled 6-week study on the efficacy and safety of the tamoxifen adjunctive to lithium in acute bipolar mania. J. Affect. Disord. 129, 327—331. Bebchuk, J.M., Arfken, C.L., Dolan-Manji, S., Murphy, J., Hasanat, K., Manji, H.K., 2000. A preliminary investigation of a protein kinace C inhibitor in the treatment of acute mania. Arch. Gen. Psychiatry 57, 95—97. Berenson, A.B., Asem, H., Tan, A., Wilkinson, G.S., 2011. Continuation rates and complications of intrauterine contraception in women diagnosed with bipolar disorder. Obstet. Gynecol. 118, 1331—1336. Blaustein, J.D., Brown, T.J., 1984. Progesterone decreases the concentration of hypothalamic and anterior pituitary estrogen receptors in ovariectomized rats. Brain Res. 304, 225—236. Buwalda, B., Schagen, S.B., 2013. Is basic research providing answers if adjuvant anti-estrogen treatment of breast cancer can induce cognitive impairment? Life Sci. 93, 581—588. Chouinard, G., Steinberg, S., Steiner, W., 1987. Estrogen—progesterone combination: another mood stabilizer? Am. J. Psychiatry 144, 826. Cohen, I., Figer, A., Tepper, R., Shapira, J., Altaras, M.M., Yigael, D., Beyth, Y., 1999. Ovarian overstimulation and cystic formation in premenopausal tamoxifen exposure: comparison between tamoxifen-treated and nontreated breast cancer patients. Gynecol. Oncol. 72, 202—207. Craig, M.C., 2013. Should psychiatrists be prescribing oestrogen therapy to their female patients? Br. J. Psychiatry 202, 9—13. Dias, R.S., Lafer, B., Russo, C., Debbio, A.D., Nierenberg, A.A., Sachs, G.S., Joffe, H., 2011. Longitudinal follow-up of bipolar disorder in women with premenstrual exacerbation: findings from STEP-BD. Am. J. Psychiatry 168, 386—394. DiazGranados, N., Zarate Jr., C., 2008. A review of the preclinical and clinical evidence for protein kinase C as a target for drug development for bipolar disorder. Curr. Psychiatry Rep. 10, 510—519. Einat, H., Peixiong, Y., Szabo, S.T., Dogra, S., Manji, H.K., 2007. Protein kinase C inhibition by tamoxifen antagonizes maniclike behavior in rats: implications for the development of novel therapeutics for bipolar disorder. Neuropsychobiology 55, 123—131.

J. Kulkarni et al. Freeman, M.P., Smith, K.W., Freeman, S.A., McElroy, S.L., Kmetz, G.E., Wright, R., Keck Jr., P.E., 2002. The impact of reproductive events on the course of bipolar disorder in women. J. Clin. Psychiatry 63, 284—287. Friedman, E., Hoau Yan, W., Levinson, D., Connell, T.A., Singh, H., 1993. Altered platelet protein kinase C activity in bipolar affective disorder, manic episode. Biol. Psychiatry 33, 520—525. Geddes, J.R., Miklowitz, D.J., 2013. Treatment of bipolar disorder. Lancet 381, 1672—1682. Goldstein, J.A., 1986. Danazol and the rapid-cycling patient. J. Clin. Psychiatry 47, 153—154. Guy, W., 1976. ECDEU Assessment Manual for Psychopharmacology. US Department of Health, Education and Welfare, Washington, DC534—537. Harlow, S.D., Gass, M., Hall, J.E., Lobo, R., Maki, P., Rebar, R.W., Sherman, S., Sluss, P.M., de Villiers, T.J., 2012. Executive summary of the stages of reproductive aging workshop + 10: addressing the unfinished agenda of staging reproductive aging. Fertil. Steril. 97, 843—851. Hox, J.J., 2010. Multilevel analysis: Techniques and applications, 2nd ed. Routledge, New York. Irwin, R.W., Yao, J., Ahmed, S.S., Hamilton, R.T., Cadenas, E., Brinton, R.D., 2011. Medroxyprogesterone acetate antagonizes estrogen up-regulation of brain mitochondrial function. Endocrinology 152, 556—567. Kulkarni, J., de Castella, A., Fitzgerald, P.B., 2008. Estrogen in severe mental illness: a potential new treatment approach. Arch. Gen. Psychiatry 65, 955—960. Kulkarni, J., Garland, K.A., Scaffidi, A., Headey, B., Anderson, R., de Castella, A., Fitzgerald, P., Davis, S.R., 2006. A pilot study of hormone modulation as a new treatment for mania in women with bipolar affective disorder. Psychoneuroendocrinology 31, 543—547. Kulkarni, J., Riedel, A., de Castella, A.R., Fitzgerald, P.B., Rolfe, T.J., Taffe, J., Burger, H., 2001. Estrogen — a potential treatment for schizophrenia. Schizophr. Res. 48, 137—144. Lam, H.-Y.P., 1984. Tamoxifen is a calmodulin antagonist in the activation of cAMP phosphodiesterase. Biochem. Biophys. Res. Commun. 118, 27—32. Lawrie, T.A., Justus Hofmeyr, G., De Jager, M., Berk, M., Paiker, J., Viljoen, E., 1998. A double-blind randomised placebo controlled trial of postnatal norethisterone enanthate: the effect on postnatal depression and serum hormones. Br. J. Obstet. Gynaecol. 105, 1082—1090. Manji, H.K., Lenox, R.H., 1999. Protein kinase C signaling in the brain: molecular transduction of mood stabilization in the treatment of manic-depressive illness. Biol. Psychiatry 46, 1328—1351. Marsh, W.K., Ketter, T.A., Crawford, S.L., Johnson, J.V., Kroll-Desrosiers, A.R., Rothschild, A.J., 2012. Progression of female reproductive stages associated with bipolar illness exacerbation. Bipolar Disord. 14, 515—526. Muthe ´n, B.O., 1998—2013. Mplus user’s guide, sixth ´n, L.K., Muthe ed. Muthe ´n & Muthe ´n, Los Angeles, CA Available as a download at http://www.statmodel.com/ugexcerpts.shtml. Nelson, M.V., 1988. Interaction of danazol and carbamazepine. Am. J. Psychiatry 145, 768—769. Paganini-Hill, A., Clark, L.J., 2000. Preliminary assessment of cognitive function inbreast cancer patients treated with tamoxifen. Breast Cancer Res. Treat. 64, 165—176. Payne, J.L., Roy, P.S., Murphy-Eberenz, K., Weismann, M.M., Swartz, K.L., McInnis, M.G., Nwulia, E., Mondimore, F.M., MacKinnon, D.F., Miller, E.B., Nurnberger, J.I., Levinson, D.F., DePaulo Jr., J.R., Potash, J.B., 2007. Reproductive cycle-associated mood symptoms in women with major depression and bipolar disorder. J. Affect. Disord. 99, 221—229. Pereira, M., Martynhak, B.J., Baretta, I.P., Correia, D., Siba, I.P., Andreatini, R., 2011. Antimanic-like effect of tamoxifen is not reproduced by acute or chronic administration of medroxyprogesterone or clomiphene. Neurosci. Lett. 500, 95—98.

Tamoxifen and medroxyprogesterone treatment in mania Rasgon, N., Bauer, M., Elman, S., Whybrow, P.C., 2003. Menstrual cycle related mood changes in women with bipolar disorder. Bipolar Disord. 5, 48—52. Rasgon, N.L., Reynolds, M.F., Elman, S., Saad, M., Frye, M.A., Bauer, M., Altshuler, L.L., 2005. Longitudinal evaluation of reproductive function in women treated for bipolar disorder. J. Affect. Disord. 89, 217—225. Simpson, G.M., Angus, J.W., 1970. A rating scale for extrapyramidal side effects. Acta Psychiatr. Scand. Suppl. 212, 11—19. Van Der Leeden, R., 1998. Multilevel analysis of repeated measures data. Qual. Quant. 32, 15—29.

61 ¨ ngu Yildiz, A., Guleryuz, S., Ankerst, D., O ¨r, D., Renshaw, P.F., 2008. Protein kinase c inhibition in the treatment of mania: a doubleblind, placebo-controlled trial of tamoxifen. Arch. Gen. Psychiatry 65, 255—263. Yildiz, A., Vieta, E., Leucht, S., Baldessarini, R.J., 2011. Efficacy of antimanic treatments: meta-analysis of randomized, controlled trials. Am. Coll. Neuropsychopharmacol. 36, 375—389. Zarate Jr., C.A., Singh, J.B., Carlson, P.J., Quiroz, J., Jolkovsky, L., Luckenbaugh, D.A., Manji, H.K., 2007. Efficacy of a protein kinase C inhibitor (tamoxifen) in the treatment of acute mania: a pilot study. Bipolar Disord. 9, 561—570.