The ovarian hormones as all steroid hormones are synthesized from cholesterol, and due to its lipophilicity, have easy access to all cells and organs, including the central nervous system (CNS) [1]. At the cerebral level, these hormones influence the function of many nervous cells, playing an important role in coordinating a number of physical and behavioral changes related to reproductive cycle [2]. However, studies are highlighting the critical role that ovarian hormones may have on the organization of non-reproductive behavior, especially in response to stress and anxiety [3-6].

Progesterone, in the corpus luteum, is converted into, among other metabolites, allopregnanolone (3a-hydroxy-5a-pregnan-20- one), which is secreted under stimulation of the luteinizing hormone (LH) and readily crosses the blood brain barrier [3,7,8]. It has been demonstrated that the 3a-hydroxysteroid is a potent anxiolytic, anticonvulsant, sedative/hypnotic and anesthetic, which exerts its effects through allosteric modulation of the gamma-aminobutyric acid (GABA) receptor complex. By binding to GABA_A receptors, allopregnanolone (alloP) increases the binding of benzodiazepines and GABA to neuronal membranes resulting in increases in the influx of Cl-. Thus, it has been suggested that alloP enhances GABA-mediated inhibition during states of hyperexcitability of the CNS, such as stress or anxiety [9-13].

Estrogens stimulate the hypothalamic-pituitary-adrenal axis (HPA). This can be observed by: 1) the presence of high levels of free cortisol, both in the morning and in the evening, by pregnant women or women receiving high doses of estrogens [14]; 2) the greater responsiveness of the HPA axis in women compared to men [14]; and 3) the acute administration of estradiol to healthy men in a psychosocial stress situation, which results in hyper responsiveness of the HPA axis and norepinephrine [15]. Furthermore, administration of estradiol to ovariectomized mice increases anxiety in situations of potential threat [16].

In addition to all these effects, estrogens and progesterone can change the density of GABA_A receptors in certain brain regions [17- 19] and, as a consequence, they may also alter the effects of certain drugs, such as benzodiazepines. Considering that women are naturally exposed to monthly fluctuation of these hormones, it is possible that their response to benzodiazepines also change over the ovarian cycle.

Studies in rodents have demonstrated changes in the sensitivity to the anxiolytic effect of benzodiazepines during the estrous cycle [20- 22]. Moreover, short-term exposure (48-72 h) of female rats to high concentrations of alloP (10 mg/kg in sesame oil) results in increased expression of the a4 subunit of the GABA_A receptor, with subsequent behavioral and pharmacological changes of GABAergic function, represented by increased anxiety and insensitivity to benzodiazepine [23,24] whereas long-term exposure of female rats to pregnancyinduced high concentrations of alloP results in anxiolysis [25].

In women, the response to benzodiazepines also seems to vary along the menstrual cycle. According to a preliminary study by Sundströn et al. [26], the sensitivity of healthy women to the sedative effect of diazepam, measured by saccadic eye movement velocity, is greater in the luteal phase compared to the follicular phase.

These observations acquire clinical importance when one considers that women suffer more of anxiety disorders [27-29], receiving, as a result, more benzodiazepine prescriptions [30,31].

Taking all these points into consideration, the aim of this study was to investigate the possible influence of the ovarian cycle in the acute effect of diazepam, in healthy women submitted to an experimental model of anxiety.

Participants

Fifty-eight undergraduate and graduate student volunteers, aged between 18 and 35, were selected for inclusion in the study using the following: a clinical structured questionnaire, a translated and adapted version of a Premenstrual Assessment Form [32]; a translated and adapted version of the State-Trait Anxiety inventory - Part II (STAItrait) [33], a translated and adapted version of the Social Phobia Inventory (SPIN) [34]. Individuals possibly presenting pathologies that could interfere with the results or scoring above 19 on SPIN, women presenting premenstrual complaints and men scoring above 50 on STAI-trait (for comparison with a previous study) were excluded.

From the 58 volunteers, 40 women (20 in follicular phase and 20 in luteal phase) were selected for Experiment I, and 18 men for Experiment II.

This study was conducted with approval from the Ethics Committee in Research with Humans of the Universidade Federal de Sergipe, Brazil. Written informed consent was obtained from all participants.

Ovarian cycle phase determination

The women were randomly allocated to two different groups: 1) Follicular Phase; and 2) Luteal Phase. Subsequently, at least two of their ovarian cycles were monitored to check for length and regularity. Then, the experimental sessions were scheduled to coincide with one of the two ovarian phases as follows:

1. Follicular Phase: 6 to 10 days from the first day of the cycle (after menses);

2. Luteal Phase: 5 to 10 days after detection of urinary LH peak by an ovulation predictor kit (Baby Hopes^®), which was used by the volunteers at home, after instructions.

Experimental model of anxiety

The Video-Monitored Stroop Color-Word Test, as standardized by Teixeira-Silva et al. [35], was used to elicit anxiety in the volunteers. In short, this test consists of presenting a board to the participant with one hundred of the color naming words blue, yellow, red, green and violet organized randomly in a 10 X 10 matrix. Each word is printed in a color different to its meaning, for example the word “red” printed in yellow ink. This board corresponds to the “Color-Word” card of the Stroop test [36]. To perform the task, the subject has to say, as quickly as possible and in the sequence presented, the names of the colors being seen (i.e. the color of the ink), but not the colors designated by the words. The task has to be performed in 2 minutes (maximum) and any errors are signaled with a bell. Skipping a color’s sequence, hesitation in saying the color, and saying the color’s “word” instead of its “ink” are all considered to be errors. The whole test is videoed and presented to the subject on a monitor during the test.

Instructions were given to the subject using a CD recording which led them to believe that a group of professionals, located in another room, were observing them and would evaluate their performance.

Psychological measurements

State-Trait Anxiety Inventory (STAI) [33]: A validated Portuguese version was used [37].

Self-evaluation of tension level: This was performed using an analogical scale from 0 (totally relaxed) to 10 (extremely tense).

Visual Analogue Mood Scale (VAMS) [38]: This scale is composed of 16 pairs of opposite adjectives. Each pair is separated by a 100-mm horizontal line on which the subjects are requested to indicate with a vertical mark the point that best represents their feelings at that moment. The 16 items are distributed into four categories: (1) mental sedation (alert/drowsy, muzzy/clear headed, mentally slow/ quick witted, attentive/dreamy); (2) physical sedation (strong/feeble, well-coordinated/clumsy, lethargic/energetic, incompetent/proficient); (3) tranquilization (calm/excited, contented/discontented, troubled/ tranquil, tense/relaxed), and (4) other feelings and attitudes (happy/ sad, antagonistic/amicable, interested/bored, withdrawn/gregarious). The range of values, for each category, is 0 – 400 mm. A Portuguese version, translated and validated by Zuardi and Karniol [39] and adapted by Del Porto et al. [40], was used.

Physiological measurements

Heart rate: Derived from two active Ag/AgCl electrodes placed on each side of the thorax, just above the nipple.

Electromyogram activity: Derived from two Ag/AgCl electrodes placed on the gastrocnemius muscle (part of the fight/flight response) at the nondominant leg.

1. The skin was cleaned with a mixture of alcohol/ether (90:10, V/V) before placement of the electrodes.

2. The recordings were made using a computerized system for monitoring physiological responses (I-330-C2+ Physiological Monitoring System, J&J Engineering, USA). All tests were performed in a quiet room maintained at a temperature between 22 and 25°C.

Drugs

Placebo: Corn Starch - 10 mg (Souza Farmácia de Manipulação).

Diazepam: Diempax - 10 mg (Sanofi Winthrop).

Procedure

Two experiments were performed. The first one aimed to investigate the possible influence of the ovarian cycle in the anxiolytic and hypnotic/sedative action of diazepam in volunteers submitted to the VMSCWT. The second one aimed simply to verify the validity of the experimental conditions in our laboratory.

In both experiments, the selected volunteers attended the laboratory on two consecutive days. The first day was used for adaptation, and the second for the actual test.

Adaptation day: The participants were taken to the experiment room, which was already organized and equipped with the necessary apparatus for the execution of the test. After filling in a consent form and then resting for about 5 min, the participants were submitted to the psychological evaluations and then to 5 min of physiological recordings.

Test day: The participants were again taken to the experiment room, where they rested for about 5 min and were then submitted to the actual test, which underwent the same experimental design in both experiments.

Experimental design

In both experiments I and II, a capsule containing placebo (PBO), or 10 mg of diazepam (DZP) was administered to the volunteers. Before treatment, the adaptation procedure from the first day was repeated and the psychological and physiological data collected were labeled as the “Time 1” experimental situation. Following treatment, the volunteers were left in a comfortable waiting room for a period of 60 min, during which they could watch television or read magazines. Immediately after this time, before being given the test instructions, the participant had his/her psychological and physiological parameters evaluated. The psychological tests used were the STAI-state, self-evaluation of tension level, and VAMS while the physiological measurements made were heart rate and gastrocnemius electromyogram activity. These data were labeled as the ‘‘Time 2’’ experimental situation. After listening to the recorded instructions, the participant then performed the task, during which his/her physiological measurements were recorded. After 50 words, a pause was made for a third set of psychological evaluations. These new data were labeled as the ‘‘Time 3’’ experimental situation. Immediately following the evaluations, the test was restarted and continued up to the last color or until the end of the scheduled time. The participant then rested for 5 min, after which all the physiological and psychological parameters were again evaluated. This final set of data was labeled as the ‘‘Time 4’’ experimental situation.

Statistical analyses

The data collected during the adaptation phase were not analyzed as this situation was only intended to habituate the participants to the environment and apparatus that would be used on the following day.

For both experiments, the results collected on the test day were first analyzed using Kolmogorov–Smirnov’s test for normal distribution and Bartlett’s test for the homogeneity of variances. No impediments to the use of parametric tests were found for any of the evaluated parameters.

All the data obtained were analyzed using a two-factor analysis of variance (ANOVA) for repeated measures followed by Tukey’s test for post hoc comparisons. For the cases in which the interaction between factors presented a significant difference, the analyses were followed by an ANOVA of simple main effect conducted on the situation variable for each group of volunteers.

Mauchly’s test for sphericity (an assumption in within-participants ANOVA) was also performed on the repeated-measures factor of the ANOVA. Data which failed this test were subjected to the Huynh– Feldt correction which adjusts the degrees of freedom. An effect of this correction is that he degrees of freedom are subsequently expressed as decimals.

The STAI-trait and SPIN scores used in the selection process were analyzed by Student’s t test. All significance tests were two tailed and were performed at the 5% significance level.

To aid visualization of the results, the graphs shown here represent the profile of the responses presented by the volunteers, with the Time 2 being considered as the ‘‘zero’’ point for the changes induced by the test. As the analyses involved repeated measures of the same individuals, representation of the dispersion of the data provides little information and therefore error bars have been omitted from the graphs. Nevertheless, the means and standard deviations of the absolute values of the results for each of the psychological and physiological parameters are presented in Tables 1 and 2.

Table 1: Summary of the absolute values obtained in Experiment I.

Table 2: Summary of the absolute values obtained in Experiment II.

Experiment I

In this experiment, women in different ovarian cycle phases (follicular and luteal) were analyzed separately by two-way ANOVA, with treatment (PBO, DZP) and situation (Time 1, Time 2, Time 3, Time 4) as factors. There were 20 women for each cycle phase, 10 randomly allocated to the DZP group and 10 to the PBO group.

Psychological measurements

STAI-state: (Figure 1).

Figure 1: State anxiety levels of women, treated with diazepam (DZP) or placebo (PBO), in response to the anxiogenic test. TEST SITUATION

Follicular phase: The interaction effect between situation and treatment was not significant and so the two main effects were analyzed. The situation effect was significant [F(2.31, 41.67) = 12.42; p < 0.000001]. State-anxiety level was higher during Time 1 (p = 0.001), Time 3 (p = 0.0001), and Time 4 (p = 0.02), in relation to Time 2. The treatment effect was not significant.

Luteal phase: The interaction effect between situation and treatment was not significant and so the two main effects were analyzed. The situation effect was significant [F(2.14, 38.51) = 21.33; p = 0.00005]. State-anxiety level was higher during Time 3 (p = 0.0001), but not during Time 4, in relation to Time 2. Pre- and post-drug (Times 1 and 2) values did not differ significantly. The treatment effect was not significant.

Self-evaluation of tension level: (Figure 2).

Figure 2: Subjective tension levels of women, treated with diazepam (DZP) or placebo (PBO), in response to the anxiogenic test.

Follicular phase: The interaction effect between situation and treatment was not significant and so the two main effects were analyzed. The situation effect was significant [F(3, 54) = 18.96; p <0.000001]. Tension level was higher during Time 1 (p = 0.001) and Time 3 (p = 0.0001), but not during Time 4, in relation to Time 2. The treatment effect was not significant.

Luteal phase: The interaction effect between situation and treatment was not significant and so the two main effects were analyzed. The situation effect was significant [F(2.36, 42.48) = 29.10; p <0.000001]. Tension level was higher during Time 1 (p = 0.04) and Time 3 (p = 0.0001), but not during Time 4, in relation to Time 2. The treatment effect was not significant.

VAMS—mental sedation: (Figure 3).

Figure 3: Interaction between test situation and treatment for mental sedation levels of women, treated with diazepam (DZP) or placebo (PBO), in response to the anxiogenic test.

Follicular phase: The interaction effect between situation and treatment was significant [F(3, 54) = 3.11; p = 0.03]. Analysis of the situation as a single factor for the PBO group did not show any significant differences, although this analysis did reveal differences for the DZP group [F(3, 27) = 4.60; p = 0.009]. Self-reported mental sedation scores were greater during Time 2 (p = 0.02) in relation to Time 1.

Luteal phase: The interaction effect between situation and treatment was not significant and so the two main effects were analyzed. Neither the situation nor the treatment effects were significant.

VAMS-physical sedation: (Figure 4).

Figure 4: Interaction between test situation and treatment for physical sedation levels of women, treated with diazepam (DZP) or placebo (PBO), in response to the anxiogenic test.

Follicular phase: The interaction effect between situation and treatment was significant [F(2.17, 39.09) = 3.18; p = 0.05]. Analysis of the situation as a single factor for the PBO group did not show any significant differences, although this analysis did reveal differences for the DZP group [F(2.20, 38) = 4.40; p = 0.02], as the mental sedation scores were greater during Time 2 (p = 0.02) in relation to Time 1.

Luteal phase: The interaction effect between situation and treatment was not significant and so the two main effects were analyzed. Neither the situation nor the treatment effects were significant.

VAMS—tranquilization

Follicular phase: The interaction effect between situation and treatment was not significant and so the two main effects were analyzed. The situation effect was significant [F(2.57, 46.33) =12.53; p = 0.00001]. Tranquilization scores were smaller during Time 3 (p = 0.0001), but not during Time 4, in relation to Time 2. Pre- and post-drug (Times 1 and 2) values did not differ significantly. The treatment effect was not significant.

Luteal phase: The interaction effect between situation and treatment was not significant and so the two main effects were analyzed. The situation effect was significant [F(2.20, 39.50) = 5.07; p = 0.01]. Tranquilization scores were smaller during Time 3 (p = 0.002), but not during Time 4, in relation to Time 2. Pre- and post-drug (Times 1 and 2) values did not differ significantly. The treatment effect was not significant.

VAMS—other feelings and attitudes

Follicular phase: The interaction effect between situation and treatment was not significant and so the two main effects were analyzed. Neither the situation nor the treatment effects were significant.

Luteal phase: The interaction effect between situation and treatment was not significant and so the two main effects were analyzed. Neither the situation nor the treatment effects were significant.

Physiological measurements

Heart rate

Follicular phase: The interaction effect between situation and treatment was not significant and so the two main effects were analyzed. The situation effect was significant [F(1.60, 25.67) = 33.52; p < 0.000001]. Heart rate was higher during Time 1 (p = 0.0007) and Time 3 (p = 0.0001), but not during Time 4, in relation to Time 2. The treatment effect was not significant.

Luteal phase: The interaction effect between situation and treatment was not significant and so the two main effects were analyzed. The situation effect was significant [F(1.37, 24.64) = 26.29; p = 0.000006]. Heart rate was higher during Time 1 (p = 0.05) and Time 3 (p = 0.0001), but not during Time 4, in relation to Time 2. The treatment effect was not significant.

Gastrocnemius electromyogram

Follicular phase: The interaction effect between situation and treatment was not significant and so the two main effects were analyzed. The situation effect was significant [F(1.31, 23.66) = 10.17; p = 0.002]. Muscular tension was higher during Time 3 (p = 0.0002), but not during Time 4, in relation to Time 2. Pre- and post-drug (Times 1 and 2) values did not differ significantly. The treatment effect was not significant.

Luteal phase: The interaction effect between situation and treatment was not significant and so the two main effects were analyzed. Neither the situation nor the treatment effects were significant.

STAI-trait

Follicular phase: The level of trait anxiety was not significantly different between the groups PBO (mean ± S.D. = 40.0 ± 9.1) and DZP (mean ± S.D. = 40.2 ± 9.4).

Luteal phase: The level of trait anxiety was not significantly different between the groups PBO (mean ± S.D. = 44.7 ± 10.7) and DZP (mean ± S.D. = 41.9 ± 8.6).

SPIN

Follicular phase: The scores of social phobia were not significantly different between the groups PBO (mean ± S.D. = 13.6 ± 4.7) and DZP (mean ± S.D. = 10.8 ± 5.9).

Luteal phase: The scores of social phobia were not significantly different between the groups PBO (mean ± S.D. = 12.9 ± 5.2) and DZP (mean ± S.D. = 11.4 ± 5.7).

Experiment II

In this experiment, men were analyzed using a two-way ANOVA, with treatment (PBO and DZP) and situation (Time 1, Time 2, Time 3 and Time 4) as factors. There were nine volunteers per group.

Psychological measurements

STAI-state: (Figure 5).

Figure 5: Interaction between test situation and treatment for state-anxiety levels of men, treated with diazepam (DZP) or placebo (PBO), in response to the anxiogenic test.

The interaction effect between situation and treatment was significant [F(3, 48) = 3.16; p = 0.03]. Analysis of the situation as a single factor for the PBO group revealed significant differences [F(3, 24) = 9.29; p = 0.0002], as the anxiety level was higher during Time 3 (p = 0.0003), but not during Time 4, in relation to Time 2. Pre- and post-drug (Times 1 and 2) values did not differ significantly. The same analyses for the DZP group did not reveal any significant differences.

Self-evaluation of tension level: (Figure 6).

Figure 6: Interaction between test situation and treatment for tension levels of men, treated with diazepam (DZP) or placebo (PBO), in response to the anxiogenic test.

The interaction effect between situation and treatment was significant [F(3, 48) = 2.97; p = 0.04]. Analysis of the situation as a single factor for the PBO group revealed significant differences [F(3, 24) = 13.51; p = 0.00002], as the anxiety level was higher during Time 3 (p = 0.0001), but not during Time 4, in relation to Time 2. Pre- and post-drug (Times 1 and 2) values did not differ significantly. The DZP group also presented significant differences [F (1.88, 15.10) = 5.96, p = 0.01], as the anxiety level was higher during Time 1 (p = 0.03) and Time 3 (p = 0.01), in relation to Time 2, but these differences were less pronounced than for the PBO group.

VAMS—mental sedation: (Figure 7).

Figure 7: Mental sedation levels of men, treated with diazepam (DZP) or placebo (PBO), in response to the anxiogenic test.

The interaction effect between situation and treatment was not significant and so the two main effects were analyzed. Neither the situation nor the treatment effects were significant.

VAMS—physical sedation: (Figure 8).

Figure 8: Physical sedation levels of men, treated with diazepam (DZP) or placebo (PBO), in response to the anxiogenic test.

The interaction effect between situation and treatment was not significant and so the two main effects were analyzed. Neither the situation nor the treatment effects were significant.

VAMS—tranquilization: The interaction effect between situation and treatment was not significant and so the two main effects were analyzed. The situation effect was significant [F(3, 48) = 11.87; p = 0.000006]. Tranquilization scores were smaller during Time 3 (p = 0.0001), but not during Time 4, in relation to Time 2. Pre- and postdrug (Times 1 and 2) values did not differ significantly. The treatment effect was not significant.

VAMS—other feelings and attitudes: The interaction effect between situation and treatment was not significant and so the two main effects were analyzed. Neither the situation nor the treatment effects were significant.

Physiological measurements

Heart rate: The interaction effect between situation and treatment was not significant and so the two main effects were analyzed. The situation effect was significant [F(2.66, 42.60) =27.41; p <0.000001]. Heart rate was higher during Time 3 (p = 0.0001), but not during Time 4, in relation to Time 2. Pre- and post-drug (Times 1 and 2) values did not differ significantly. The treatment effect was not significant.

Gastrocnemius electromyogram: The interaction effect between situation and treatment was not significant and so the two main effects were analyzed. The situation effect was significant [F(1.43, 22.84) = 5.85; p = 0.01]. Muscular tension was higher during Time 3 (p = 0.005), but not during Time 4, in relation to before. Pre- and post-drug (Times 1 and 2) values did not differ significantly. The treatment effect was not significant.

STAI-trait

The level of trait anxiety was not significantly different between the groups PBO (mean ± S.D. = 38.6 ± 3.9) and DZP (mean ± S.D. = 37.4 ± 9.5).

SPIN

The scores of social phobia were not significantly different between the groups PBO (mean ± S.D. = 10.7 ± 6.9) and DZP (mean ± S.D. = 9.5 ± 5.4).

The aim of this work was to investigate the possible influence of the ovarian cycle in the classical effects of diazepam in healthy women submitted to an experimental model of anxiety.

The results of Experiment I showed that women in the follicular ovarian phase presented no response to the anxiolytic action of diazepam, since the increase in anxiety promoted by the anxiogenic VMSCWT was neither prevented nor reduced by the drug, as demonstrated by the STAI-state scores, the self-evaluated tension levels and the values of the tranquilization factor of the VAMS. In contrast, these women did show an increase in their levels of mental and physical sedation in response to diazepam. These results replicate the findings of Teixeira-Silva et al. [35].

Surprisingly, diazepam also did not present an anxiolytic effect in women in the luteal ovarian phase neither presented a hypnotic/ sedative action. These intriguing results show a complete lack of response to the effects of the acute administration of diazepam by women in the luteal phase. In order to verify if this finding was not due to flaws in the execution of the methods or in the drug used, Experiment II was conducted, only with men. Now, the diazepam action profile was exactly what expected, taking into consideration a previous study [35]. The drug prevented the increase in anxiety during the test, as demonstrated by the STAI-state and by the self-evaluation of tension level, without causing sedation. Therefore, the results obtained with women cannot be attributed to failure in the experimental procedure.

So what could explain the different response profiles to acute diazepam administration between men and women and between women in different phases of the ovarian cycle?

Possible pharmacokinetic differences could account for all different response patterns observed here. However, it has been demonstrated that there is no difference in diazepam bioavailability between men and women, when orally administered [41], while data about benzodiazepine pharmacokinetic changes across the ovarian cycle are scarce. To the best of our knowledge, there are no studies concerning the influence of the menstrual cycle on diazepam pharmacokinetics. Anyhow, it has been demonstrated that the pharmacokinetics of orally administered alprazolam does not change through the ovarian cycle [42]. The same has been observed for the intravenous administration of midazolam [43].

Having said all that, it is not possible to rely on pharmacokinetics to explain our results. Thus, pharmacodynamic differences may be responsible for these findings.

What would make women in the follicular phase, unlike men, irresponsive to the anxiolytic action of diazepam?

This would suggest a decreased sensitivity to diazepam of women in the follicular phase in comparison to men. In fact, it has been shown that diazepam has an anxiolytic effect in men in a lower dose (5 mg), which did not have any effect at all in women in the follicular phase [35]. A possible explanation of this could be the low levels of progesterone and it metabolites in this phase, especially alloP, which potentiates the action of benzodiazepines [44]. However, with 10 mg of diazepam in the present study, as in a previous one [35], these women already felt the hypnotic/sedative effect of the drug, which is not seen in men. Therefore, there was a difference in type of response and not just a dose-response curve dislocation.

Fluctuations of the sex steroid levels during the ovarian cycle are strongly associated with changes in the of the GABA_A receptor subunits composition in the brain. Evidence in rodents, monkeys and cell models show that estrogens, progesterone and alloP influence the expression pattern of GABA_A receptor subunits [45-51]. Variable subunits expression patterns define the GABA_A/BZD receptor subtype. Those containing a1ß?₂ subunits correspond to BZ1 receptor, which mediates the sedative effect of diazepam, while those containing a₂, a₃ or a₅ in combination with the ß and ?₂ subunits correspond to BZ2 receptor, which is responsible for the anxiolytic effect [52-54]. Considering this, although there are no studies to support this idea, it is tempting to speculate that women in the follicular phase present more BZ1and less BZ2 receptors. Future research could investigate this further. On the other hand, neither the anxiolytic nor the hypnotic/ sedative effects were present in women in the luteal phase.

What could be the factor responsible for abolishing both anxiolytic and hypnotic/sedative effects of diazepam in women in the luteal phase?

In studies by Gulinello et al. [23,24], it was found that short-term exposure (48-72 h) to high concentrations of alloP to female rats led to a lack of lorazepam’s anxiolytic effect, while it increased the a₄ subunit expression of the GABA_A receptor, which is insensitive to the action of benzodiazepines. However, after long-term exposure (five days) to the neurosteroid, the a₄ subunit levels and the pharmacology of lorazepam returned to control values and remain unchanged until removal of the neurosteroid. Therefore, it is possible that exposure to alloP in women during the luteal phase corresponds to the short-term exposure in rats from Gulinello’s studies.

In contradiction to this hypothesis, the preliminary study by Sundström et al. [26] found that responsiveness to diazepam in healthy women, assessed by the saccadic eye movement velocity, as a monitor of sedation, was greater in the luteal phase than in the follicular phase. However, the period investigated by Sündstron et al. was the late luteal phase (1-7 days before the onset of menses), which could correspond to long-term exposure of the studies by Gulinello or a period of sex hormones decline.

At this point, it is also valid to consider the existence of differences between male and female brains, independently of the ovarian cycle, which could explain the lack of anxiolytic effect found, here, in both menstrual cycle phases. For example, a study by Ravenelle et al. [55], performed with rats, showed sex-differences in response to the anti-anxiety effects of diazepam that correlated with GABA neuron variations along stress regions. Males present a greater number of GABAergic interneurons in the amygdala, which could explain their greater response to diazepam.

Independently of its cause, which should be investigated in the future, the fact is that diazepam seems to be devoid of acute anxiolytic effect in women at any ovarian cycle phase. These findings, while intriguing, are not exactly unexpected. In another placebo-controlled study, where the anxiety was induced by a real-life stress situation (cholecystectomy) in women, the oral acute administration of diazepam (10 mg) was also unable to decrease anxiety [56].

In relation to the physiological parameters, the results showed that diazepam did not prevent the changes in heart rate and gastrocnemius tension, elicited by VMSCWT, neither in women nor men. Nevertheless, this inability of benzodiazepines to prevent or reduce anxiety-accompanied physiological alterations was already expected, as it has been shown in a variety of studies of real-life or laboratoryinduced stress [35,57-59].

Taking all these points into consideration, the data presented here not only contributes for the understanding of the important relationship among anxiety, ovarian hormones and benzodiazepines, but also indicates that care must be exercised when prescribing benzodiazepines to women, at least for acute use, which is also common. In fact, one of the indications for the use of diazepam (10 mg) is for sedation and relief of anxiety, tension and stress prior to diagnostic or therapeutic procedures, as pointed out by the Valium^® package insert itself. The acute administration of 5 to 10 mg of diazepam by oral route, usually one hour before the start of an invasive or unpleasant medical procedure is widely used as premedication or as the main pharmacological method for sedation in outpatient dental procedures [60-64].

The fact that diazepam is being prescribed to women, although it may not have the intended effect, is the reflection of the negligence in women biological research, which has negatively affected the understanding of women’s physiology [65].

A survey of studies published in nine influential medical journals, in 2004, found that from 46 clinical trials, only 37% of participants were women and this proportion declined to 24% when the analysis was restricted to drug testing. In 87% of the studies, results were not reported by gender nor were gender included as a covariate [66].

Even though women seek treatment and receive prescription of psychotropic medications more than men, the majority of phase I clinical trials, which determine the therapeutic doses, are performed only with men [67-69].

These facts make research such as the present study crucial, and the knowledge derived from them should be transferred into clinical practice and become a key consideration in treatment choice by health professionals.

In summary, despite the fact that future investigation should be performed with long-term exposure to diazepam and with different benzodiazepines, the results presented here:

1. Show that the ovarian cycle does influence the effects of a single dose of diazepam;

2. Draw attention to the importance of considering the gender and also the ovarian cycle phase when studying the effects of drugs;

3. Indicate that, contrary to expectations, a single dose of diazepam is not a good prescription choice for reducing or preventing acute anxiety in women.

This work was supported by Fundação de Apoio a Pesquisa e à Inovação Tecnológica do Estado de Sergipe (FAPITEC/SE).

Special thanks to all the men and women who were volunteers in this research.