Effect of SR49059, an orally active V1a vasopressin receptor antagonist, in the prevention of dysmenorrhoea


Correspondence: Professor M. Åkerlund, Department of Obstetrics and Gynaecology, University Hospital, S-22185 Lund, Sweden.


Objective To investigate the clinical effect of SR49059 when given shortly before the onset of menstruation as a preventative treatment of dysmenorrhoea.

Design A double-blind, randomised, placebo-controlled, cross-over trial in complete block design (three periods, three treatments).

Setting A clinical research organisation in Paris, France.

Participants Women aged 18–35 years suffering from primary dysmenorrhoea.

Interventions In each of three menstrual cycles, women reported to the study centre and were given a daily dose of either placebo, 100 mg or 300 mg SR49059 from a minimum of 4 hours up to a maximum of three days before the onset of bleeding and/or menstrual pain. If this did not control the pain, women were allowed once a day to take a second dose of study treatment providing that at least 4 hours had passed since the first drug intake.

Main outcome measures Intensity of menstrual pain recorded by means of a visual analogue scale. Rating of symptoms of dysmenorrhoea (mainly back and pelvic pain) in relation to functional capacity (Sultan score). Self-assessment of menstrual blood loss in a menstrual diary record.

Results Analysis of intensity of menstrual pain, as recorded by visual analogue scale and Sultan pain score (back and pelvic pain) during the first 24 hours of dysmenorrhoea, showed a dose-related effect of SR49059. The 300 mg dose of SR49059 was significantly more effective than placebo. Similarly, a dose-related effect of SR49059 was shown on total Sultan score. SR49059 was well tolerated and no significant effect on the bleeding pattern was noted.

Conclusions This study showed for the first time a therapeutic effect of an orally active vasopressin V1a receptor antagonist in the prevention of dysmenorrhoea. Further studies are required to examine effect mechanisms and determine effective doses.


Women with primary dysmenorrhoea have myometrial hyperactivity and reduced uterine blood flow during pain, the endometrial blood flow often fluctuating in a pattern which could coincide with a pulsatile vasopressin secretion1. The average circulating level of this hormone is increased during painful menses2–4. Furthermore, in women who begin to experience pain before the onset of bleeding, plasma vasopressin levels are already raised at this time3. The details of the contribution of vasopressin in mechanisms of menstruation and dysmenorrhoea are still unknown. However, the peptide has pronounced stimulatory effects on smooth muscle activity of both myometrium and uterine arteries via the vasopressin V1a receptor5,6, which is distinctly different from the V2 type regulating kidney function. In women with dysmenorrhoea, a slight further increase in the circulating level of the hormone, which can be induced by infusion of hypertonic saline solution, causes increased uterine contractions and pain4.

SR49059 is an orally active, nonpeptide compound which specifically and selectively antagonises the effect of vasopressin on the V1a receptor in animals7 and in humans8. The drug has been shown to have an excellent safety profile in single and repeated dose toxicological studies in animals and, up to now, no toxicological target organ and no maximal tolerated dose have been identified in humans. In the human uterus in vitro, SR49059 caused a dose-dependent inhibition of vasopressin V1a receptor-mediated activity of myometrial strips and isolated uterine arteries9,10. In vivo in nonpregnant women, an inhibition of vasopressin-induced uterine activity has been observed11. SR49059 has no influence on the effect of vasopressin on the V2 receptors regulating osmolality function8,11.

In the present study, treatment with SR49059 was tested in women suffering regularly from moderate to severe dysmenorrhoea, the drug being given between 4 and 72 hours before the onset of symptoms. The hours immediately preceding the onset of menstruation and pain may be of particular interest in the evaluation of pathophysiological mechanisms of primary dysmenorrhoea. Only relatively young women participated, to ensure as far as possible that only dysmenorrhoea of the primary type was under study.


A total of 73 women were included in this study, which was performed in a clinical research organisation in Paris, France. Their age range was 18 to 35 years (mean 23.8 years). They were of normal weight and had a history of at least six months of regularly occurring moderate to severe dysmenorrhoea warranting pharmacological treatment. Their menstrual cycles were in the range of 26 to 35 days (mean 29.6 days) and menses lasted three to seven days (mean 4.6 days). The pain started before the onset of bleeding in about two-thirds of the women and at the beginning of menstruation in the remainder. All but nine women were nulliparous, and another nine had undergone legal abortion in the first trimester of pregnancy. The women in this study had not used any oral contraceptives or intrauterine contraceptive devices during at least the two menstrual cycles preceding the study and did not use any during the course of the study.

Medical examination performed within four weeks of treatment administration showed all women to be healthy. This medical screening included a general physical, neurological and gynaecological examination, as well as pelvic echography. Blood pressure (measured by an automatic sphygmomanometer), a 12 lead electrocardiogram, and routine blood and urine safety parameters of biochemistry, haematology, qualitative drug screen and serology were also found to be normal.

Treatment regimen

Each woman received 100 mg SR49059, 300 mg SR49059 and placebo over three, usually consecutive, menstrual cycles according to a double-blind, crossover design12. The administration of study drugs was blinded and the appearance of placebo and SR49059 tablets was indistinguishable. The lower dose of SR49059 was chosen in order to achieve a plasma concentration (about 15 ng/mL) similar to that which inhibited the effect of lysine vasopressin challenge on uterine contractility in a previous study11, taking into account the increased bioavailability of the presently used formulation. The higher dose of 300 mg SR49059 was selected in order to enable verification of a dose-dependent effect of the drug. The women were requested to come to the study centre every day for a minimum of one day and a maximum of three days before menstruation and during menstruation, for drug administration and evaluation of pain. The total duration of treatment with SR49059 during each cycle was at least two days and could not exceed five days, in view of the limited toxicological results available at the time of the study. A negative pregnancy test was required before any administration of study drug.

Study treatment began between 4 and 72 hours before the expected onset of menstruation and was continued for a maximum of three days, with the same oral daily dose, until menstruation started. After the start of dysmenorrhoea, defined as the onset of vaginal bleeding or the onset of pain, whichever occurred first, treatment was prolonged for up to three days. If menstruation did not occur within three days of expected bleeding, this time being estimated on the basis of six previous menstrual cycle records, the treatment was discontinued for that period and a pregnancy test repeatedly taken until the onset of menstruation.

Primary outcome measure

The primary outcome measure was intensity of menstrual pain, as assessed by the women by means of a visual analogue scale and self-rating of symptoms associated with dysmenorrhoea (mainly back and pelvic pain) using a French-validated Sultan score13 adequate for the population of this trial. For the visual analogue scale recording, a 100 mm ungraded scale was used ranging from 0 (= no pain at all) to 100 (= worst ever experienced). The Sultan score ranged from zero (absence) to three (maximum discomfort). This form was modified compared with the original one in as much as absenteeism from work was not included, since the women were asked to come to the centre every day. Visual analogue scale and Sultan score ratings were performed by the women just before and every two hours during the six hours following each drug intake and then every six hours during day time and eight hours at night. Due to the expected variable duration of pain, the area under the curve for visual analogue scale and Sultan scores was derived over the period between onset of dysmenorrhoea (Hdysm0) and 24 hours thereafter (Hdysm24). A validated menstrual diary with information on size and extent of soiling of the sanitary protection used (towels and tampons)14 was also completed by the women. In this record they also noted all intake of analgesic drugs if received. During each treatment period women were allowed once a day to take a second dose of test medication, provided that at least four hours had passed since the first intake. If pain persisted for more than two hours following this second administration, women were allowed to take additional 500 mg paracetamol (Doliprane, Rhône Poulenc, Lyon, France) every four hours in addition to the randomised treatment, to a maximum of 3 g per day.

Plasma concentrations of vasopressin were measured by radioimmunoassay (RIA Kit Bühlmann lab. S RKARI) and SR49059 by a validated high performance liquid chromatography—mass spectometry/mass spectometry method (limit of quantification 0.2 ng/mL; Sanofi Recherche). Blood samples were taken and handled as previously described11. Blood pressure and pulse rate in supine and standing positions were recorded before and 30 minutes after each study drug administration. Electrocardiograms were recorded at Hdysm0, as well as at the pre- and post-study visits.

Statistical methods

The sample size was based on the results of a previous Sultan score study13 and calculated using a two-sided significance level of 0.05. The within-subject standard deviation was estimated from the results of the Sultan total score (the item ‘absenteeism from work’ was not taken into account in the calculation). A sample size of 60 was used as it was estimated to give at least a 90% chance of a true difference of 1.55 points or more between two treatments being deemed significant if the within-subject standard deviation was not > 2.6.

Analysis of efficacy parameters took into account measurements associated with the first and second intake of SR49059 or placebo. Parameters were analysed by means of cross-over analyses of variance, investigating for subject, period and treatment effects. When a significant treatment effect was observed, pairwise comparison of doses was carried out using the contrast method. All statistical tests were two-tailed and used a 5% significance level. The statistical package used for the analysis was SAS version 6.09 (SAS Institute Inc, Cary, North Carolina, USA). Regarding plasma concentrations of arginine vasopressin, cross-over analysis of variance was performed on the difference between H0 (pre-dose) and H1 (one hour post-dose). Only descriptive statistics were performed for pharmacokinetic data.


As regards efficacy, a woman was considered evaluable for one treatment period if the onset of dysmenorrhoea (i.e. the first vaginal bleeding or the first feeling of pain) appeared between four hours and four days after the first drug intake. A total of 65 women were available for at least one study period (per protocol population), whereas only 45 women completed the three study periods (complete case population). Of the 65 women, 15 were not evaluable for one or two periods due to delayed menses (i.e. with no occurrence of menstruation or pain from the day of expected menstruation and four days onwards), three were not evaluable due to deviation from the protocol (pain present before the first drug intake in two women, concomitant intake of paracetamol in one woman), and four were not available due to permanent treatment discontinuation (for three women, period three was not performed in order to comply with the planned end of the study, for one woman the reason for withdrawal was personal). As a consequence, a total of 20 women were not available for one or two periods (note: a subject could belong to two different categories).

In the analysis of visual analogue scale rating according to the per protocol analysis (Fig. 1), overall comparison of mean area under the curve of visual analogue scale scores obtained over the first 24 hours of dysmenorrhoea (i.e. between Hdysm0 and Hdysm24) revealed a dose-related effect of SR49059 on pain intensity (cross-over analysis of variance, treatment effect significant with P= 0.02). Mean area under the curve was significantly lower at the 300 mg dose compared with placebo (P= 0.005), but not at the 100 mg dose. No statistically significant difference was observed between the two SR49059 doses. Analysis of area under the curve of visual analogue scale rating according to the complete case analysis (45 women) revealed similar results.

Figure 1.

Effect of placebo, 100 and 300 mg SR49059 on pain as measured by visual analogue scale in women with dysmenorrhoea (n= 55, n= 54 and n= 54, respectively, per protocol population). The area under the curve during the first 24 hours of dysmenorrhoea was used for the calculation. Mean and standard error of the mean for each treatment are shown.

Administration of SR49059 resulted in a dose-related effect on back and pelvic pain intensity, as assessed by Sultan score in the per protocol population (Fig. 2; treatment effect significant with P= 0.002). Mean area under the curve of back and pelvic pain score was significantly lower in the 300 mg group than in the 100 mg (P= 0.05) and placebo (P= 0.0003) groups, but no significant difference was observed between 100 mg and placebo (Fig. 2). Analysis of the complete case population gave similar results.

Figure 2.

Effect of placebo, 100 mg and 300 mg SR49059 on back and pelvic pain as measured by Sultan score in women with dysmenorrhoea (n= 55, n= 54 and n= 54, respectively, per protocol population). The area under the curve during the first 24 hours of dysmenorrhoea was used for the calculation. Mean and standard error of the mean for each treatment are shown.

Regarding the total Sultan score ratings over 24 hours in the protocol population significantly lower scores were obtained in the 300 mg SR49059 group compared with 100 mg of this drug and with placebo (P= 0.003 and P= 0.05, respectively). For the item ‘feeling of malaise’, only a trend was observed (overall treatment effect; P= 0.07). A dose-related effect was shown on the total Sultan score, with the difference between 100 mg and 300 mg of SR49059 versus placebo being significantly different (P= 0.05 and P= 0.003, respectively). There was a strong positive linear relationship between the Sultan pain score and the visual analogue scale score (Pearson's correlation coefficient: r= 0.87; P < 0.001).

A second dose of study drug after treatment with placebo, 100 and 300 mg SR49059 was taken by 46%, 45% and 35% of women, respectively, the difference between groups not being statistically significant. Additional paracetamol treatment was taken by 32%, 20% and 13% of women in the three groups, respectively, the difference between placebo and 300 mg SR49059 being statistically significant (P= 0.006). Furthermore, the amount of paracetamol taken by the women treated with 300 mg SR49059 was significantly lower than that taken by placebo treated women (mean intakes 182 and 438 mg, respectively; P= 0.005).

No significant differences, expressed as mean (standard deviation) between groups were observed as regards bleeding duration, which was on average 59 (3), 60 (3) and 57 (3) hours after placebo, 100 and 300 mg SR49059, respectively. Nor were there any differences between the three treatment groups as regards blood clot size or extent of soiling of the sanitary protection. Time to onset of menstrual bleeding in the placebo, 100 and 300 mg SR49059 groups was 46.7, 49.6 and 53.6 hours, respectively, with no significant difference between the three treatment groups.

The average concentration of vasopressin in plasma before the first administration of placebo, 100 and 300 mg SR49059 was 2.0 (0.3), 1.7 (0.2) and 1.7 (0.1) pg/mL, respectively, and one hour after the first drug administration 1.8 (0.2) pg/mL in each treatment group (i.e. no significant treatment effect was observed). The plasma concentrations of SR49059 one hour after the first intake of 100 and 300 mg SR49059 were 17.2 (0.4) and 30.4 (4.3) ng/mL, respectively.

No effect of SR49059 was shown on electrocardiogram, blood pressure or pulse rate, and no serious adverse event was noted. Eleven of the women discontinued their participation in the study (seven of these women were replaced) because of unwanted pregnancy (in one woman) or personal reasons. No significant effect on blood and urine safety parameters was observed.


This study demonstrates for the first time a therapeutic benefit of an orally active vasopressin V1a receptor antagonist in primary dysmenorrhoea. However, it should be kept in mind that the study focused on prevention of dysmenorrhoea and that the women were monitored extensively during the day preceding menstruation and up to two to three days after the onset of menstruation. This study design was important for studying the vasopressin involvement in dysmenorrhoea, a very early event in the onset of menstruation and dysmenorrhoic pain.

In a recent pilot study performed by our group (unpublished) with SR49059 given after the onset of menstrual pain, the pain reduction after treatment by SR49059 was similar to that induced by placebo. The assessment of pain in the present study was probably more reliable since recording of pain started within 24 hours of the expected start of menstruation, thus reducing the placebo effect.

A significant effect of SR49059 on pain intensity during the first 24 hours of dysmenorrhoea was observed in the visual analogue scale recording, and this effect was dose-related. Furthermore, in the modified Sultan score ratings of symptoms associated with dysmenorrhoea, back and pelvic pain decreased significantly with 300 mg SR49059, compared with placebo. The validity of a dose-related effect of SR49059 in primary dysmenorrhoea is further strengthened by the very good correlation between pain recordings in the Sultan and visual analogue scale scorings. This observation also emphasises the validity of using Sultan scoring in estimating pain of dysmenorrhoea.

The significant differences in the percentage of subjects taking additional paracetamol treatment support the potential therapeutic effect of SR49059 in the prevention of dysmenorrhoea. Nonsteroid anti-inflammatory drugs were not used as analgesics in the present study in order to avoid a potential bias due to interference with prostaglandin synthesis. Endometrial synthesis of prostaglandin F2a appears to be involved as one of the final steps in the induction of uterine hyperactivity of primary dysmenorrhoea3.

The women with dysmenorrhoea in this study had circulating levels of vasopressin which were on average four times higher than those observed in healthy women at the same time of the menstrual cycle3. The effect of the study drug was presumably due to interference with the uterine effects of this hormone at the receptor level. Thus, the results presented here support the aetiological importance of vasopressin in primary dysmenorrhoea, as has been suggested by a long series of investigations with recordings of uterine activity and blood flow, estimations of plasma levels of the hormone, and receptor studies1–6. The vasopressin levels one hour after intake of the study drug were not significantly decreased from before, an observation which confirms that the action of the study drug was blockade of V1a receptors and not inhibition of vasopressin release. The therapeutic effect of SR49059 in dysmenorrhoea via blocking the receptor is also in agreement with that observed with the peptide vasopressin V1a antagonist 1-deamino-2-D-Tyr (OEt)-4-Thr-8-Orn-oxytocin when given intravenously to women with this condition15.

The effect of vasopressin on the nonpregnant uterus in vivo is about five times more pronounced than that of the other posterior pituitary hormone, oxytocin, and furthermore it increases premenstrually6. The uterine concentrations of vasopressin V1a and oxytocin receptors differ to the same extent6. However, an additional contribution of oxytocin in the pathophysiology of dysmenorrhoea can not be completely excluded and is not contradictory to the therapeutic effects seen with 1-deamino-2-D-Tyr (OEt)-4-Thr-8-Orn-oxytocin15 and with SR49059 in the present study. It is well established that the former oxytocin analogue acts as an oxytocin receptor inhibitor as well16,17. This may also be the case for SR49059, since it also binds, although with less affinity, to the oxytocin receptor17.

We have previously shown that SR49059 selectively inhibits the response to vasopressin of human uterine arteries of different diameters10. However, this does not seem to influence mechanisms regulating vasoconstriction at menstruation, since no significant effect on the bleeding pattern was observed in the present study in women with dysmenorrhoea. The present study showed no effect of SR49059 on blood pressure and pulse rate, and no other significant adverse effects were observed. These observations are in agreement with our previous clinical experience of this drug8,11,18.

In a previous study11 it was observed that SR49059 blocks the effects of vasopressin on uterine contractions and the present study has demonstrated a therapeutic effect of the drug in dysmenorrhoea. Studies are now required to clarify the mechanisms by which vasopressin causes primary dysmenorrhoea and how SR49059 interferes with this mechanism.