Contraception
46:253-268,
1992
ORAL CONTRACEPTIVES IN PREMENSTRUAL SYNDROME: A RANDOMIZED COMPARISON OF TRIPHASIC AND MONOPHASIC PREPARATIONS TorbjBrn
BBckstrCim,
Cavalli-Bjbrkman, Department
Yvonne
Stig
of Obstetrics
Hansson-Malmstr6m,
Bengt-Ake
Lindhe.
Bengt
Nordenstrdm and Gynaecology,
Stockholm private Gynaecological
University
H&pita1
of Ume&
Sweden;
group, Stockholm, Sweden and Organon Sweden AB,
Gliteborg
ABSTRACT Thirty-seven women with cyclical mood changes, either only in the premenstrual phase (‘pure PMS’) or during the entire cycle with premenstrual aggravation (‘PMA’), were recruited to participate in a randomized study investigating the effect of three different oral contraceptives (OCs) on mood symptoms; 32 out of the 37 women completed the study. The monophasic ethinylestradiol (EE)/desogestrel (DSG) OC was compared, in a single (doctor)-blind cross-over design, with a monophasic and a triphasic levonorgestrel (LNG)-containing combined OC. The women kept a record of their symptoms and complaints by noting daily ratings using a validated visual analogue scale. One pretreatment cycle was followed by four treatment cycles, two cycles on each OC. All OCs had a beneficial effect on the PMS symptoms compared to the pre-treatment period. There were no changes between consecutive cycles. Cyclical symptom changes were noted during all OC treatment. The monophasic desogestrel pill provoked less changes in mood parameters than the monophasic and triphasic levonorgestrel OCs. However, physical complaints were less frequently reported during the use of the triphasic preparation as compared to the monophasic desogestrel preparation. Women with ‘pure’ PMS (premenstrual syndrome) were more consistent in their reactions on OCs compared to women with PMA. Address correspondence to: T. Backstrom. MD, Ph.D. Dept Obstetrics and Gynaecology Academic Hospital S-757 85 Uppsala Fax no. Sweden Submitted for publication Accepted for publication
Copyright
+46-l 8559775
October 16, July 2, 1992
0 1992 Butterworth-Heinemann
1991
254
Contraception
Introduction It is well known that certain women suffer from severe mood changes when using oral contraception, and that this has been an important factor in the arguments for the discontinuation of the Pill. Earlier studies have shown that mood changes as side effects of OC mainly occur in women suffering from premenstrual syndrome (PMS) (1). Cullberg noted that only women suffering from PMS reacted negatively to the Pill. This suggests that women with PMS are more susceptible to hormonal provocation than women without PMS. A high percentage of women with PMS report negative mood changes while using an OC (2,3,4). Others have noted that premenstrual syndromelike mood changes are less frequently reported in women on the Pill (4,5). This can be interpreted to mean that women with PMS would be less likely to continue the use of an OC. There are few prospective studies into the effect of OCs on mood in women with PMS. One study showed a greater beneficial effect from a monophasic OC containing norgestrel than a triphasic preparation with the same progestogen (6). Another placebo-controlled prospective study showed no beneficial effect on premenstrual symptoms when taking a triphasic OC (7). One of the main causes for women to discontinue the use of an OC is the occurrence of undesirable changes in mood. More knowledge is necessary on whether different types of progestogens have different effects in provoking ‘negative’ mood changes, and also whether the quantity of the progestogen is of importance. That the progestogen is of interest has been shown in postmenopausal women using a sequential estrogen/progestogen treatment. They develop cyclical mood changes similar to that seen during the menstrual cycle in women with PMS (8). The effect of OCs on mood is best studied in women who have hormone-related cyclical mood changes, as differences between the preparations are probably more easily detected in this group. Women with PMS are probably more susceptible to developing undesirable changes in mood during OC use, than women in general (1,2,3). Therefore, we have chosen to study, in a randomized design, the effects of three different OC preparations in a group
Contraception
of women suffering from PMS. These women noted daily ratings in order to record cyclical mood changes during treatment cycle and the four following treatment cycles.
255
symptom one pre-
Methods Thirty-seven women with either an established PMS or PMA (premenstrual aggravation of pre-existing symptoms) were recruited for this study, to be run at four centres in Sweden. Nineteen were randomly assigned to a treatment with a monophasic preparation containing 30 mcg ethinylestradiol (EE) combined with either 150 mcg desogestrel (DSG) or 150 mcg levonorgestrel (LNG) (monophasic DSG trial). The other seventeen women were randomized to a treatment with either the monophasic DSG OC or a triphasic preparation containing EE (30/40/30 mcg) and levonorgestrel (LNG, 50/75/125 mcg) (triphasic LNG trial). After two treatment cycles, the cross-over was made, and the women were treated another two cycles with the other preparation. The drugs were foreign and the brand names were not available in Sweden. The trial medication was blinded not only for the doctor, but also for the patient as the brand names were not used in Sweden. Trial medication was supplied and randomized by Organon Sweden AD. The inclusion criteria were: women asking for oral contraception aged between 20 and 40 years with established PMS or PMA. Willingness to note daily symptom ratings. Exclusion criteria: contraindications to the use of OCs, as specified in the OCs’ “Method of Use”; hormonal treatment during the last two months; alcohol or drug abuse; a present or recent history of psychiatric disease. The study was approved by the Ethical Committee, University of Umea, Sweden. Of the 37 women who started, 32 completed the study and 5 had dropped out by the end of the study. Two women dropped out because of an increase of PMS symptoms during treatment. One woman was on monophasic EE/DSG treatment, the other on triphasic EE/LNG. Three women dropped out because of other reasons: one became pregnant before starting on the OC, the
256
Contraception
second became menopausal at the time that she was treatment, and the third woman lost her rating cards. Procedure
for
establishing
PMS
and
to
start
PMA
Before randomization, all women had daily rated their symptoms and complaints during one or two menstrual cycles, using a previously validated visual analogue scale (2). Diagnoses were made with the help of a procedure described elsewhere in detail (9). In principle, symptom scores during premenstrual and preovulatory phases were compared and the number of symptomfree preovulatory days calculated. In order to be eligible for entry into this study, the women had to meet the following criteria, apart from the ones already stated above: 1) a history of recurrent cyclical mood changes with at with maximum severity during least one psychological symptom, the premenstrual phase; 2) a significant difference in at least 3 out of 12 rated symptoms between the 9 pre-menstrual days and the 9 pre-ovulatory days of the pre-treatment cycle. This is taken as indication of a cyclical pattern of the symptoms and complaints. estimated according to earlier The severity of PMS was published definitions (lo), taking into account the effect of symptoms on family and social life as well as on the ability to work. All women had a moderateto- severe PMS, PMA. Subsequently, the women were sub-divided into two groups depending on the pattern of symptom distribution during the menstrual cycle. Women with only symptoms and complaints during the luteal phase of the cycle were diagnosed as ‘pure PMS’. Women with symptoms and complaints both in the preovulatory and luteal phase of the menstrual cycle, but with a significant aggravation of symptoms during the latter, were diagnosed as ‘PMA’. Of the 32 women who completed the study, 14 were diagnosed as having pure PMS (8 in the monophasic DSG trial, 6 in the triphasic LNG trial), and 18 with PMA (8 in DSG trial, IO in triphasic LNG trial).
257
Contraception
Rating
scale
Every evening during the 4 consecutive treatment cycles, the women filled out the daily rating scales, the same as used in the pre-treatment cycle. In total, 4 positive and 4 negative psychological mood parameters, 3 physical subjective symptoms, and 3 sexual symptoms were rated (see Tables I and II). The severity of withdrawal bleedings was also rated daily. For each item, the woman had to mark on a 10 cm long line how she had experienced a particular symptom or complaint during that day. On the scale, 0 stood for a complete absence of symptoms, whereas 10 meant maximum severity of a symptom or complaint, as compared to the woman’s personal experience in the pretreatment cycle. They also had to mark on the scale whether the symptoms or complaints had influenced their family or social life and their ability to work. Apart from this, the women were asked to describe events during the day that might have influenced their mood, as well as to record all concomitant medication taken.
Statistical
analysis
Statistical analysis was made using one- or two-way analysis of ANOVA, followed ‘ad hoc’ with the method of least variance, significant difference (LSDS) to find the differences between individual groups participating in the ANOVA. One-way ANOVA was performed first with the diagnostic cycle in the ANOVA test, thereafter without it. The other tests were made using two-way ANOVA, with the differences between treatments as the discriminating factors. The symptom scores of the 8 premenstrual days in the pretreatment cycle, and of the last 8 days in the treatment cycle were used as test variables. In order to get an overall impression of all symptoms and complaints tested, the daily scores were summarized to four summary scores per day of the pre-treatment and treatment cycles. In the summarized sexual score, the score for sexual interest and the score of positive sexual thought were while the score for negative sexual thought was added, subtracted from the summary score. The ranges for the summary scores were -10 to +20 for the sexual score, 0 to 40 for the negative mood score, 0 to 30 for the negative physical complaints and 0 to 40 for the positve score.
258
Contraception
In order to show the mean symptom and complaint scores graphically, these have been calculated for each cycle day, related to the onset of menstrual bleeding in the pre-treatment cycle, as well as to the onset of OC use in the treatment cycles. OC use started on the first day of menstrual bleeding, while the next cycles were started after a one-week pill-free interval. Because of the shorter period of the first treatment cycle, we have made one figure from the day when OC use started and for the next 4 weeks; this was done in the treatment cycles. The pretreatment cycle was compared with treatment cycles 2, 3, and 4, and figures were calculated from the onset of withdrawal bleeding with a one-week break at the beginning of the figure. RESULTS
An
analysis
of
time-related
changes
First, an analysis of the influence of time was made. The ANOVA over all cycles, pre-treatment cycle included, showed significant heterogeneity between the cycles for all symptoms and complaints (p
46:253-268,
1992
ORAL CONTRACEPTIVES IN PREMENSTRUAL SYNDROME: A RANDOMIZED COMPARISON OF TRIPHASIC AND MONOPHASIC PREPARATIONS TorbjBrn
BBckstrCim,
Cavalli-Bjbrkman, Department
Yvonne
Stig
of Obstetrics
Hansson-Malmstr6m,
Bengt-Ake
Lindhe.
Bengt
Nordenstrdm and Gynaecology,
Stockholm private Gynaecological
University
H&pita1
of Ume&
Sweden;
group, Stockholm, Sweden and Organon Sweden AB,
Gliteborg
ABSTRACT Thirty-seven women with cyclical mood changes, either only in the premenstrual phase (‘pure PMS’) or during the entire cycle with premenstrual aggravation (‘PMA’), were recruited to participate in a randomized study investigating the effect of three different oral contraceptives (OCs) on mood symptoms; 32 out of the 37 women completed the study. The monophasic ethinylestradiol (EE)/desogestrel (DSG) OC was compared, in a single (doctor)-blind cross-over design, with a monophasic and a triphasic levonorgestrel (LNG)-containing combined OC. The women kept a record of their symptoms and complaints by noting daily ratings using a validated visual analogue scale. One pretreatment cycle was followed by four treatment cycles, two cycles on each OC. All OCs had a beneficial effect on the PMS symptoms compared to the pre-treatment period. There were no changes between consecutive cycles. Cyclical symptom changes were noted during all OC treatment. The monophasic desogestrel pill provoked less changes in mood parameters than the monophasic and triphasic levonorgestrel OCs. However, physical complaints were less frequently reported during the use of the triphasic preparation as compared to the monophasic desogestrel preparation. Women with ‘pure’ PMS (premenstrual syndrome) were more consistent in their reactions on OCs compared to women with PMA. Address correspondence to: T. Backstrom. MD, Ph.D. Dept Obstetrics and Gynaecology Academic Hospital S-757 85 Uppsala Fax no. Sweden Submitted for publication Accepted for publication
Copyright
+46-l 8559775
October 16, July 2, 1992
0 1992 Butterworth-Heinemann
1991
254
Contraception
Introduction It is well known that certain women suffer from severe mood changes when using oral contraception, and that this has been an important factor in the arguments for the discontinuation of the Pill. Earlier studies have shown that mood changes as side effects of OC mainly occur in women suffering from premenstrual syndrome (PMS) (1). Cullberg noted that only women suffering from PMS reacted negatively to the Pill. This suggests that women with PMS are more susceptible to hormonal provocation than women without PMS. A high percentage of women with PMS report negative mood changes while using an OC (2,3,4). Others have noted that premenstrual syndromelike mood changes are less frequently reported in women on the Pill (4,5). This can be interpreted to mean that women with PMS would be less likely to continue the use of an OC. There are few prospective studies into the effect of OCs on mood in women with PMS. One study showed a greater beneficial effect from a monophasic OC containing norgestrel than a triphasic preparation with the same progestogen (6). Another placebo-controlled prospective study showed no beneficial effect on premenstrual symptoms when taking a triphasic OC (7). One of the main causes for women to discontinue the use of an OC is the occurrence of undesirable changes in mood. More knowledge is necessary on whether different types of progestogens have different effects in provoking ‘negative’ mood changes, and also whether the quantity of the progestogen is of importance. That the progestogen is of interest has been shown in postmenopausal women using a sequential estrogen/progestogen treatment. They develop cyclical mood changes similar to that seen during the menstrual cycle in women with PMS (8). The effect of OCs on mood is best studied in women who have hormone-related cyclical mood changes, as differences between the preparations are probably more easily detected in this group. Women with PMS are probably more susceptible to developing undesirable changes in mood during OC use, than women in general (1,2,3). Therefore, we have chosen to study, in a randomized design, the effects of three different OC preparations in a group
Contraception
of women suffering from PMS. These women noted daily ratings in order to record cyclical mood changes during treatment cycle and the four following treatment cycles.
255
symptom one pre-
Methods Thirty-seven women with either an established PMS or PMA (premenstrual aggravation of pre-existing symptoms) were recruited for this study, to be run at four centres in Sweden. Nineteen were randomly assigned to a treatment with a monophasic preparation containing 30 mcg ethinylestradiol (EE) combined with either 150 mcg desogestrel (DSG) or 150 mcg levonorgestrel (LNG) (monophasic DSG trial). The other seventeen women were randomized to a treatment with either the monophasic DSG OC or a triphasic preparation containing EE (30/40/30 mcg) and levonorgestrel (LNG, 50/75/125 mcg) (triphasic LNG trial). After two treatment cycles, the cross-over was made, and the women were treated another two cycles with the other preparation. The drugs were foreign and the brand names were not available in Sweden. The trial medication was blinded not only for the doctor, but also for the patient as the brand names were not used in Sweden. Trial medication was supplied and randomized by Organon Sweden AD. The inclusion criteria were: women asking for oral contraception aged between 20 and 40 years with established PMS or PMA. Willingness to note daily symptom ratings. Exclusion criteria: contraindications to the use of OCs, as specified in the OCs’ “Method of Use”; hormonal treatment during the last two months; alcohol or drug abuse; a present or recent history of psychiatric disease. The study was approved by the Ethical Committee, University of Umea, Sweden. Of the 37 women who started, 32 completed the study and 5 had dropped out by the end of the study. Two women dropped out because of an increase of PMS symptoms during treatment. One woman was on monophasic EE/DSG treatment, the other on triphasic EE/LNG. Three women dropped out because of other reasons: one became pregnant before starting on the OC, the
256
Contraception
second became menopausal at the time that she was treatment, and the third woman lost her rating cards. Procedure
for
establishing
PMS
and
to
start
PMA
Before randomization, all women had daily rated their symptoms and complaints during one or two menstrual cycles, using a previously validated visual analogue scale (2). Diagnoses were made with the help of a procedure described elsewhere in detail (9). In principle, symptom scores during premenstrual and preovulatory phases were compared and the number of symptomfree preovulatory days calculated. In order to be eligible for entry into this study, the women had to meet the following criteria, apart from the ones already stated above: 1) a history of recurrent cyclical mood changes with at with maximum severity during least one psychological symptom, the premenstrual phase; 2) a significant difference in at least 3 out of 12 rated symptoms between the 9 pre-menstrual days and the 9 pre-ovulatory days of the pre-treatment cycle. This is taken as indication of a cyclical pattern of the symptoms and complaints. estimated according to earlier The severity of PMS was published definitions (lo), taking into account the effect of symptoms on family and social life as well as on the ability to work. All women had a moderateto- severe PMS, PMA. Subsequently, the women were sub-divided into two groups depending on the pattern of symptom distribution during the menstrual cycle. Women with only symptoms and complaints during the luteal phase of the cycle were diagnosed as ‘pure PMS’. Women with symptoms and complaints both in the preovulatory and luteal phase of the menstrual cycle, but with a significant aggravation of symptoms during the latter, were diagnosed as ‘PMA’. Of the 32 women who completed the study, 14 were diagnosed as having pure PMS (8 in the monophasic DSG trial, 6 in the triphasic LNG trial), and 18 with PMA (8 in DSG trial, IO in triphasic LNG trial).
257
Contraception
Rating
scale
Every evening during the 4 consecutive treatment cycles, the women filled out the daily rating scales, the same as used in the pre-treatment cycle. In total, 4 positive and 4 negative psychological mood parameters, 3 physical subjective symptoms, and 3 sexual symptoms were rated (see Tables I and II). The severity of withdrawal bleedings was also rated daily. For each item, the woman had to mark on a 10 cm long line how she had experienced a particular symptom or complaint during that day. On the scale, 0 stood for a complete absence of symptoms, whereas 10 meant maximum severity of a symptom or complaint, as compared to the woman’s personal experience in the pretreatment cycle. They also had to mark on the scale whether the symptoms or complaints had influenced their family or social life and their ability to work. Apart from this, the women were asked to describe events during the day that might have influenced their mood, as well as to record all concomitant medication taken.
Statistical
analysis
Statistical analysis was made using one- or two-way analysis of ANOVA, followed ‘ad hoc’ with the method of least variance, significant difference (LSDS) to find the differences between individual groups participating in the ANOVA. One-way ANOVA was performed first with the diagnostic cycle in the ANOVA test, thereafter without it. The other tests were made using two-way ANOVA, with the differences between treatments as the discriminating factors. The symptom scores of the 8 premenstrual days in the pretreatment cycle, and of the last 8 days in the treatment cycle were used as test variables. In order to get an overall impression of all symptoms and complaints tested, the daily scores were summarized to four summary scores per day of the pre-treatment and treatment cycles. In the summarized sexual score, the score for sexual interest and the score of positive sexual thought were while the score for negative sexual thought was added, subtracted from the summary score. The ranges for the summary scores were -10 to +20 for the sexual score, 0 to 40 for the negative mood score, 0 to 30 for the negative physical complaints and 0 to 40 for the positve score.
258
Contraception
In order to show the mean symptom and complaint scores graphically, these have been calculated for each cycle day, related to the onset of menstrual bleeding in the pre-treatment cycle, as well as to the onset of OC use in the treatment cycles. OC use started on the first day of menstrual bleeding, while the next cycles were started after a one-week pill-free interval. Because of the shorter period of the first treatment cycle, we have made one figure from the day when OC use started and for the next 4 weeks; this was done in the treatment cycles. The pretreatment cycle was compared with treatment cycles 2, 3, and 4, and figures were calculated from the onset of withdrawal bleeding with a one-week break at the beginning of the figure. RESULTS
An
analysis
of
time-related
changes
First, an analysis of the influence of time was made. The ANOVA over all cycles, pre-treatment cycle included, showed significant heterogeneity between the cycles for all symptoms and complaints (p
