Abstract

BACKGROUND

Combined oral contraceptives (COCs) are widely advocated as treatment for primary dysmenorrhoea, but their efficacy has been questioned in a Cochrane review. The aim of this study was to evaluate COCs and the influence of age on the severity of dysmenorrhoea.

METHODS

Postal questionnaires regarding weight/height, contraception, pregnancy history and other reproductive health factors were sent to random samples of 19-year-old women born in 1962 (n = 656), 1972 (n = 780) and 1982 (n = 666) resident in the city of Gothenburg in 1981, 1991 and 2001. The responders were assessed again 5 years later at the age of 24 years. Current severity of dysmenorrhoea was measured on each occasion by a verbal multidimensional scoring system (VMS) and by a visual analogue scale (VAS).

RESULTS

The severity of dysmenorrhoea was lower (P< 0.0001) in COC users compared with non-users. In a longitudinal analysis of the severity of dysmenorrhoea, COC use and increasing age, independently of each other, were associated with the severity of dysmenorrhoea (COC use, VMS score: a reduction of 0.3 units/VAS: a reduction of 9 mm, both P< 0.0001; increasing age, VMS score: a reduction of 0.1 units per 5 years, P< 0.0001/VAS: a reduction of 5 mm per 5 years, P< 0.0001). Childbirth also reduced the severity of dysmenorrhoea (VAS, P< 0.01 with a reduction of 7 mm). Women from the 82-cohort reported a greater severity of dysmenorrhoea compared with the 62 and 72 cohorts at both 19 and 24 years of age.

CONCLUSIONS

In this longitudinal case–control study, COC use and increasing age, independent of each other, reduced the severity of dysmenorrhoea. COC use reduced the severity of dysmenorrhoea more than increasing age and childbirth. There was a trend over time regarding the severity of dysmenorrhoea where women from the 82-cohort reported a greater severity of dysmenorrhoea compared with the 62 and 72 cohorts.

Introduction

Dysmenorrhoea is a frequently occurring condition affecting a large proportion of young women. The highest prevalence has been reported in adolescent women, where as many as 50–75% suffer from dysmenorrhoea (Andersch and Milsom, 1982a; Robinson et al., 1992; Harlow and Park, 1996; Campell and McGrath, 1997; Davis and Westhoff, 2001; Harlow and Campell, 2004). About 15% of adolescent and young women describe their dysmenorrhoea as severe, causing regular absenteeism from school and work (Sundell et al., 1990; Harlow and Ephross, 1995; Burnett et al., 2005; Dawood, 2006). This monthly disability, which interferes with daily work several days each cycle, has been estimated to account for 600 million lost working hours and two billion dollars in lost productivity annually in the USA (Coco, 1999). Effective management of dysmenorrhoea is beneficial for both the afflicted individual and society.

Combined oral contraceptive pills (COC) have been reported to relieve dysmenorrhoea (Hendrix and Alexander, 2002; Winkler et al., 2004; Davis et al., 2005). The study by Davis et al. (2005) specifically addressed the issue of dysmenorrhoea in adolescents using a randomized placebo-controlled design and showed a positive effect on dysmenorrhoea. The efficacy of COCs and other forms of treatment for dysmenorrhoea has been analysed in two systematic reviews (Proctor and Farquhar, 2007; Wong et al., 2009). A Cochrane review (Wong et al., 2009) included 10 studies: six compared COCs with placebo and four compared different dosages of COCs. The Cochrane review (Wong et al., 2009) concluded that there was limited evidence for pain improvement and they found no difference between different COC preparations but the results were limited because of the variable quality of the randomized controlled trials included in the analysis.

In 1981, a prospective longitudinal study (Andersch and Milsom, 1982a,b) was initiated regarding contraceptive use, pregnancies and reproductive health in a random sample of 19-year-old women resident in the city of Gothenburg. The women have been followed prospectively and new cohorts of 19-year-old women were recruited in 1991 and 2001. The same women were re-assessed 5 years later at 24 years of age. The aim of this study was to evaluate COC's and the influence of age on the severity of dysmenorrhoea in this population of young women.

Materials and Methods

Study population

The women initially included were born in 1962 (Andersch and Milsom, 1982a,b) and were 19 years of age. This age was chosen in order to be able to deal with women of legal age, enabling information to be collected via the individuals themselves without the necessity of consent from their parents. Two later cohorts of 19-year-old women (born 1972, and 1982) resident in the city of Gothenburg in 1991 and 2001, respectively were also invited to participate. The study was approved by the Ethics Committee, Faculty of Medicine, University of Gothenburg, and the National Data Inspection Board approved the study design and informed consent was obtained from each participant.

There were 2621 women aged 19 years living in the city of Gothenburg in 1981 and a random sample of every fourth woman was obtained (n = 656) from the population register. The women from the 1962 birth cohort are referred to in the text as the ‘62-cohort’. Ten years later, in 1991 a one in three sample of the 2342 women aged 19 years (n = 780) resident in the city of Gothenburg was obtained from the population register and in 2001 (Lindh et al., 2009) a similar one in three sample of the 1998 women aged 19 years (n = 666) resident in the city of Gothenburg in 2001 was obtained from the population register. The 19-year-old women from the 1972 and 1982 birth cohorts are referred to as the ‘72-cohort’ and ‘82’-cohort. The women were contacted by letter and requested to complete and return an enclosed questionnaire. If no reply was received, reminders were sent out after 2 and 4 weeks. Women who returned the questionnaire in 1981, 1991 and 2001 at the age of 19 years of age were contacted again 5 years later at the age of 24 years (1986, 1996 and 2006) and requested to answer and return a similar questionnaire.

The questionnaire

The questionnaire contained approximately 40 questions concerning contraception, reproductive history, menstrual pattern, duration and severity of menstrual pain, need for medical attention, ability to work during menstruation and factors such as height, weight and smoking. The severity of dysmenorrhoea was assessed by a verbal multidimensional scoring system (VMS) described previously (Andersch and Milsom, 1982a) and by a visual analogue scale (VAS) (Melzack and Katz, 1994). The VMS grades pain as none, mild, moderate or severe, and takes also into account the effect on daily activity, systemic symptoms and whether analgesics are required (Table I). VAS is a technique where a 10 cm line on a paper represents the continuum of the woman's opinion of the degree of pain (It was explained that one extremity of the line represented ‘no pain at all’ and the other extremity ‘unbearable pain’). The ability to understand the questionnaire and reliability was evaluated in a sub-sample of 30 women and were found to be satisfactory. Reliability was tested by letting 30 women answer the questionnaire twice within a 3-week interval, and the agreement of the answers to five specific questions was tested. Ten questions in the inquiry were then selected, and the same 30 women were interviewed to see whether they understood the meaning of the questions.

Table I

The VMS for assessment of severity of dysmenorrhoea (Andersch and Milsom, 1982a).

Grade Working ability Systemic symptoms Analgesica 
Grade 0: Menstruation is not painful and daily activity is unaffected Unaffected None Not required 
Grade 1: Menstruation is painful but seldom inhibits the woman's normal activity. Analgesics are seldom required. Mild pain Rarely affected None Rarely required 
Grade 2: Daily activity affected. Analgesics required and give relief so that absence from work or school is unusual. Moderate pain Moderately affected Few Required 
Grade 3: Activity clearly inhibited. Poor effect of analgesics. Vegetative symptoms, e.g. headache, tiredness, nausea, vomiting and diarrhoea. Severe pain Clearly inhibited Apparent Poor effect 
Grade Working ability Systemic symptoms Analgesica 
Grade 0: Menstruation is not painful and daily activity is unaffected Unaffected None Not required 
Grade 1: Menstruation is painful but seldom inhibits the woman's normal activity. Analgesics are seldom required. Mild pain Rarely affected None Rarely required 
Grade 2: Daily activity affected. Analgesics required and give relief so that absence from work or school is unusual. Moderate pain Moderately affected Few Required 
Grade 3: Activity clearly inhibited. Poor effect of analgesics. Vegetative symptoms, e.g. headache, tiredness, nausea, vomiting and diarrhoea. Severe pain Clearly inhibited Apparent Poor effect 

Characteristics of the samples and analysis of the non-responders

The questionnaire was completed and returned by 594 (91%) of the 656 women from the 62 cohort, by 641 (82%) of the 780 women in the 72 cohort and by 514 (77%) of the 666 women in the 82 cohort. The population register contains information regarding civil status, nationality and home address linked to the individual's personal identification number. The socio-economic status of the district (SES) where the woman was resident was classified according to a 3-point socio-economic index (low, medium and high socio-economic status) based on the mean level of education, income and profession/social group for all the inhabitants resident in each district (Statistics Sweden, 2001). There was no significant difference in marital status, nationality or SES of the area where the women were resident between responders and non-responders in any of the three cohorts with one exception: there was a greater (P< 0.01) number of non-Swedish nationals among non-responders (14%) compared with responders (6%) in the 82 cohort.

The questionnaire was completed and returned at both 19 and 24 years of age by 489 (75%) of the women born in 1962, by 523 (67%) of the women born in 1972 and by 392 (59%) of the women born in 1982. There were no significant difference regarding smoking, pregnancies, live births, contraceptive use, weight and SES of the area where the woman was resident between the responders at both 19 and 24 years of age and non-responders at 24 years of age in the three cohorts.

Data analysis and statistical methods

Data description of groups includes n = sample size, mean, standard deviation and 95% confidence limits. Group differences were tested with an ordinary two sample t-test, which throughout was confirmed with the Wilcoxon rank sum test. Individual changes within groups over time were tested with a matched paired t-test. Analysis of variance was used to investigate the influence of various factors on dysmenorrhoea and a nested design was performed to assess the interaction between the factor cohort and the predictors. To further establish possible differences between cohorts, post hoc tests of multiple comparisons due to Tukey were performed. A Pearson correlation test was used to compare the subjective measurement with the VMS and the VAS. Analyses were done using SAS 9.1 software (SAS Institute Inc., Cary, NC, USA).

Results

A comparison of basic characteristics between the three cohorts of women assessed at 19 and 24 years of age is shown in Table II. There was a successive increase in bodyweight between the three cohorts and the frequency of smoking decreased (P< 0.001) over time. The proportion of women using COCs was lower (P< 0.001) at the age of 24 in the 82 cohort compared with the 62 and 72 cohorts. The type of oral contraceptive used at 19 and 24 years of age varied between cohorts (Table III). In the 82 cohort, there were a greater number of women using a progestogen-only method, and COC use was lower in this cohort at the age of 24 years (Table III). The 62 cohort reported no use at all of progestogen-only methods, and in the 72 cohort, there were no use of progestogen-only methods at the age of 19 years but at 24 years of age 5% of the women used these methods. In the –82 cohort 7% reported use of these methods at the age of 19 years and 14% at 24 years of age.

Table II

Comparison of basic characteristics (mean/%) in the same women at 19 and 24 years of age from the cohorts born in 1962 (n = 489), 1972 (n = 523) and 1982 (n = 392).

  Cohort 1962
 
Cohort 1972
 
Cohort 1982
 
 19 years 24 years 19 years 24 years 19 years 24 years 
Height, cm, mean, (95% CI) 167.1 (166.6–167.6) 167.3 (166.8–167.8) 167.2 (166.7–167.7) 167.4 (166.8–167.9) 167.3 (166.7–167.9) 167.4 (166.8–168.0) 
Weight, kg, mean, (95% CI) 58.2 (57.6–58.9) 59.7 (59.0–60.5) 59.6 (58.7–60.4) 62.6 (61.6–63.5) 61.5 (60.6–62.5) 62.8 (61.9–63.8) 
BMI, mean, (95% CI) 20.8 (20.6–21.0) 21.3 (21.1–21.5) 21.2 (21.0–21.5) 22.3 (22.0–22.6) 21.9 (21.7–22.2) 22.4 (22.1–22.7) 
Menarche age year, mean, (95% CI) 12.9 (12.8–13.0)  12.9 (12.8–13.0)  12.8 (12.7–13.0)  
Smoker, %, (95% CI) 41 (36–45) 43 (38–47) 33 (29–38) 30 (26–34) 27 (22–31) 17 (13–21) 
Number of pregnancies, mean, (95% CI) 0.19 (0.15–0.24) 0.80 (0.69–0.91) 0.15 (0.10–0.19) 0.65 (0.56–0.74) 0.16 (0.10–0.21) 0.43 (0.35–0.52) 
Number of children, mean, (95% CI) 0.06 (0.03–0.08) 0.38 (0.31–0.44) 0.02 (0.01–0.04) 0.34 (0.27–0.41) 0.02 (0.00–0.04) 0.12 (0.08–0.15) 
COC use, %, (95% CI) 46 (42–51) 51 (47–56) 46 (42–51) 49 (44–53) 46 (41–51) 34 (30–39) 
  Cohort 1962
 
Cohort 1972
 
Cohort 1982
 
 19 years 24 years 19 years 24 years 19 years 24 years 
Height, cm, mean, (95% CI) 167.1 (166.6–167.6) 167.3 (166.8–167.8) 167.2 (166.7–167.7) 167.4 (166.8–167.9) 167.3 (166.7–167.9) 167.4 (166.8–168.0) 
Weight, kg, mean, (95% CI) 58.2 (57.6–58.9) 59.7 (59.0–60.5) 59.6 (58.7–60.4) 62.6 (61.6–63.5) 61.5 (60.6–62.5) 62.8 (61.9–63.8) 
BMI, mean, (95% CI) 20.8 (20.6–21.0) 21.3 (21.1–21.5) 21.2 (21.0–21.5) 22.3 (22.0–22.6) 21.9 (21.7–22.2) 22.4 (22.1–22.7) 
Menarche age year, mean, (95% CI) 12.9 (12.8–13.0)  12.9 (12.8–13.0)  12.8 (12.7–13.0)  
Smoker, %, (95% CI) 41 (36–45) 43 (38–47) 33 (29–38) 30 (26–34) 27 (22–31) 17 (13–21) 
Number of pregnancies, mean, (95% CI) 0.19 (0.15–0.24) 0.80 (0.69–0.91) 0.15 (0.10–0.19) 0.65 (0.56–0.74) 0.16 (0.10–0.21) 0.43 (0.35–0.52) 
Number of children, mean, (95% CI) 0.06 (0.03–0.08) 0.38 (0.31–0.44) 0.02 (0.01–0.04) 0.34 (0.27–0.41) 0.02 (0.00–0.04) 0.12 (0.08–0.15) 
COC use, %, (95% CI) 46 (42–51) 51 (47–56) 46 (42–51) 49 (44–53) 46 (41–51) 34 (30–39) 

COC, combined oral contraceptive.

Table III

Comparison of the type of COC used at 19 and 24 years of age in the three cohorts of women born in 1962 (n = 489), 1972 (n = 523) and 1982 (n = 392).

COC type Cohort 1962
 
Cohort 1972
 
Cohort 1982
 
19 years 24 years 19 years 24 years 19 years 24 years 
All COC users, n 225 249 244 254 183 135 
COC with 50 μg EE 46 (20%) 26 (10%) 5 (2%) 1 (0.4%) 0 (0%) 0 (0%) 
COC with 30 μg EE + 150 μg LNG 157 (70%) 85 (34%) 7 (3%) 15 (6%) 15 (8%) 23 (17%) 
Triphasic COC 0 (0%) 104 (42%) 90 (37%) 112 (44%) 83 (45%) 44 (33%) 
Other COCs 4 (2%) 11 (4%) 123 (50%) 100 (39%) 68 (37%) 56 (42%) 
Name unknown 18 (8%) 23 (9%) 19 (8%) 26 (10%) 17 (9%) 12 (9%) 
COC type Cohort 1962
 
Cohort 1972
 
Cohort 1982
 
19 years 24 years 19 years 24 years 19 years 24 years 
All COC users, n 225 249 244 254 183 135 
COC with 50 μg EE 46 (20%) 26 (10%) 5 (2%) 1 (0.4%) 0 (0%) 0 (0%) 
COC with 30 μg EE + 150 μg LNG 157 (70%) 85 (34%) 7 (3%) 15 (6%) 15 (8%) 23 (17%) 
Triphasic COC 0 (0%) 104 (42%) 90 (37%) 112 (44%) 83 (45%) 44 (33%) 
Other COCs 4 (2%) 11 (4%) 123 (50%) 100 (39%) 68 (37%) 56 (42%) 
Name unknown 18 (8%) 23 (9%) 19 (8%) 26 (10%) 17 (9%) 12 (9%) 

Data are n (% of all COC users).

EE, ethinyl estradiol; LNG, levonorgestrel.

Other COCs = 30 µg or 20 µg ethinyl estradiol in combination with desorgestrel/drospirenone etc.

The severity of dysmenorrhoea

A correlation analysis was performed to compare VMS and VAS in all women who completed the VMS and VAS (n = 2642) at both assessment points. There was a significant correlation (r = 0.83, P< 0.0001) between the assessment of the severity of dysmenorrhoea by the two methods. The results of the VMS are therefore presented as mean values.

A comparison of menstrual bleeding pattern and the severity of dysmenorrhoea were made both within and between the three cohorts at the ages of 19 and 24 years (Table IV). Fewer women in the 82 cohort reported no dysmenorrhoea at all according to the VMS system. Women in the 82 cohort had at 19 and 24 years a higher VMS score (P< 0.05; P< 0.001) compared with the –62 cohort and at the age of 24 years a higher VMS score compared with the 72 cohort (P< 0.01). Similarly, there was a greater severity of dysmenorrhoea as assessed by the VAS in the 82 cohort compared with the 62 cohort (P< 0.001) and the 72 cohort (P< 0.01) at 19 and 24 years old, respectively. Absenteeism due to dysmenorrhoea was reported between 31 and 33% at the age of 19 years and decreased in all three cohorts at the age of 24 years (20–23%).

Table IV

Comparison of menstrual bleeding pattern and severity of dysmenorrhoea in the same women at 19 and 24 years of age from the cohorts born in 1962 (n = 489), 1972 (n = 523) and 1982 (n = 392).

  Cohort 1962
 
Cohort 1972
 
Cohort 1982
 
19 years 24 years 19 years 24 years 19 years 24 years 
Menstrual cycle 
 Length of cycle, days, (95% CI) 28.0 (27.6–28.3) 27.2 (26.8–27.5) 28.1 (27.7–28.6) 27.7 (27.2–28.2) 27.2 (26.5–28.0) 27.0 (26.3–27.6) 
 Duration of menses, days, (95% CI) 5.3 (5.2–5.4) 5.2 (5.1–5.3) 5.4 (5.2–5.5) 5.2 (5.0–5.3) 5.3 (5.1–5.4) 5.2 (5.1–5.4) 
Dysmenorrhoea 
 VAS, mean, (95% CI) 41.2 (38.3–44.0) 34.2 (31.7–36.6) 42.9 (40.1–45.7) 38.7 (36–41.2) 49.6 (46.8–52.5) 45.1 (42.2–48.0) 
 VMS n (%)       
  0 134 (27.9) 157 (33.0) 137 (27.2) 137 (28.3) 64 (16.5) 72 (18.9) 
  1 170 (35.3) 167 (35.1) 156 (31.0) 173 (35.7) 142 (36.7) 153 (40.1) 
  2 109 (22.7) 106 (22.3) 142 (28.2) 147 (30.4) 137 (35.4) 121 (31.7) 
  3 68 (14.1) 46 (9.7) 69 (13.7) 27 (5.6) 44 (11.4) 36 (9.4) 
 VMS, mean, (95% CI) 1.23 (1.14–1.32) 1.1 (1.0–1.17) 1.28 (1.20–1.37) 1.13 (1.05–1.21) 1.42 (1.33–1.51) 1.32 (1.23–1.41) 
 Absenteeism due to dysmenorrhoea, %, (95% CI) 32 (28–36) 21 (18–25) 33 (29–37) 20 (16–23) 31 (27–36) 23 (19–28) 
 Analgesic required, %, (95% CI) 26 (22–30) 22 (18–25) 32 (28–36) 30 (26–34) 29 (25–34) 28 (24–33) 
  Cohort 1962
 
Cohort 1972
 
Cohort 1982
 
19 years 24 years 19 years 24 years 19 years 24 years 
Menstrual cycle 
 Length of cycle, days, (95% CI) 28.0 (27.6–28.3) 27.2 (26.8–27.5) 28.1 (27.7–28.6) 27.7 (27.2–28.2) 27.2 (26.5–28.0) 27.0 (26.3–27.6) 
 Duration of menses, days, (95% CI) 5.3 (5.2–5.4) 5.2 (5.1–5.3) 5.4 (5.2–5.5) 5.2 (5.0–5.3) 5.3 (5.1–5.4) 5.2 (5.1–5.4) 
Dysmenorrhoea 
 VAS, mean, (95% CI) 41.2 (38.3–44.0) 34.2 (31.7–36.6) 42.9 (40.1–45.7) 38.7 (36–41.2) 49.6 (46.8–52.5) 45.1 (42.2–48.0) 
 VMS n (%)       
  0 134 (27.9) 157 (33.0) 137 (27.2) 137 (28.3) 64 (16.5) 72 (18.9) 
  1 170 (35.3) 167 (35.1) 156 (31.0) 173 (35.7) 142 (36.7) 153 (40.1) 
  2 109 (22.7) 106 (22.3) 142 (28.2) 147 (30.4) 137 (35.4) 121 (31.7) 
  3 68 (14.1) 46 (9.7) 69 (13.7) 27 (5.6) 44 (11.4) 36 (9.4) 
 VMS, mean, (95% CI) 1.23 (1.14–1.32) 1.1 (1.0–1.17) 1.28 (1.20–1.37) 1.13 (1.05–1.21) 1.42 (1.33–1.51) 1.32 (1.23–1.41) 
 Absenteeism due to dysmenorrhoea, %, (95% CI) 32 (28–36) 21 (18–25) 33 (29–37) 20 (16–23) 31 (27–36) 23 (19–28) 
 Analgesic required, %, (95% CI) 26 (22–30) 22 (18–25) 32 (28–36) 30 (26–34) 29 (25–34) 28 (24–33) 

VAS, visual analogue scale, VMS, verbal multidimensional scoring system.

The use of COC and dysmenorrhoea

The severity and changes in severity of dysmenorrhoea were assessed using both the VMS and VAS in a longitudinal, individual analysis of the same women in the three cohorts at 19 and 24 years of age grouped according to the use or non-use of COC (Table V). The women were assessed in two separate groups using a case control model; one group where the women were using COC at 19 years of age and no COC at 24 years of age; and the second group consisted of women who used no COC at 19 years of age and who then used a COC at 24 years of age. These two separate groups were then compared with each other to investigate possible differences in severity of dysmenorrhoea when using COC. Parous women and women using an intrauterine device (IUD) or progestogen-only methods were excluded from these analyses.

Table V

A longitudinal, individual analysis of the severity of dysmenorrhoea assessed by the VMS and the VAS in the same women at 19 and 24 years of age, related to change from use to non-use of COCs or the reverse between the two ages.

Age, year   Mean
 
Difference 19–24 (95% CI)   Mean
 
 Difference 19–24 (95% CI) Total benefit associated with COC use (95% CI) 
19
 
24
 
19
 
24
 
COC No COC No COC COC 
VMS 
 Cohort-62 33 1.39 1.58 0.18 (−0.12 to 0.48) 84 1.32 0.75 −0.57 (−0.81 to −0.33) 0.75 (0.34–1.17), P< 0.001 
 Cohort-72 34 1.15 1.29 0.15 (−0.16 to 0.46) 52 1.29 0.92 −0.37 (−0.66 to −0.07) 0.51 (0.07–0.95), P< 0.05 
 Cohort-82 39 1.36 1.46 0.10 (−0.14 to 0.35) 31 1.26 1.03 −0.23 (−0.49 to 0.04) 0.33 (−0.03 to 0.68) 
 All women 106 1.30 1.44 0.14 (−0.02 to 0.30) 167 1.30 0.86 −0.44 (−0.60 to −0.29) 0.58 (0.35–0.81), P< 0.0001 
VAS 
 Cohort-62 36 45.0 49.9 4.9 (−5.7 to 15.2) 84 44.3 24.5 −19.8 (−27.1 to −12.6) 24.7 (11.7–37.7), P< 0.001 
 Cohort-72 34 42.8 44.2 1.5 (−6.4 to 9.4) 50 44.9 31.5 −13.4 (−22.3 to −4.5) 14.9 (2.4–27.4) P< 0.05 
 Cohort-82 38 49.1 48.3 −0.8 (−8.5 to 6.9) 27 54.6 40.0 −14.6 (−24.5 to −4.7) 13.8 (1.74–25.9), P< 0.05 
 All women 108 45.7 47.5 1.8 (−3.1 to 6.8) 161 46.2 29.2 −17.0 (−21.9 to −12.1) 18.8 (11.6–26.0), P< 0.0001 
Age, year   Mean
 
Difference 19–24 (95% CI)   Mean
 
 Difference 19–24 (95% CI) Total benefit associated with COC use (95% CI) 
19
 
24
 
19
 
24
 
COC No COC No COC COC 
VMS 
 Cohort-62 33 1.39 1.58 0.18 (−0.12 to 0.48) 84 1.32 0.75 −0.57 (−0.81 to −0.33) 0.75 (0.34–1.17), P< 0.001 
 Cohort-72 34 1.15 1.29 0.15 (−0.16 to 0.46) 52 1.29 0.92 −0.37 (−0.66 to −0.07) 0.51 (0.07–0.95), P< 0.05 
 Cohort-82 39 1.36 1.46 0.10 (−0.14 to 0.35) 31 1.26 1.03 −0.23 (−0.49 to 0.04) 0.33 (−0.03 to 0.68) 
 All women 106 1.30 1.44 0.14 (−0.02 to 0.30) 167 1.30 0.86 −0.44 (−0.60 to −0.29) 0.58 (0.35–0.81), P< 0.0001 
VAS 
 Cohort-62 36 45.0 49.9 4.9 (−5.7 to 15.2) 84 44.3 24.5 −19.8 (−27.1 to −12.6) 24.7 (11.7–37.7), P< 0.001 
 Cohort-72 34 42.8 44.2 1.5 (−6.4 to 9.4) 50 44.9 31.5 −13.4 (−22.3 to −4.5) 14.9 (2.4–27.4) P< 0.05 
 Cohort-82 38 49.1 48.3 −0.8 (−8.5 to 6.9) 27 54.6 40.0 −14.6 (−24.5 to −4.7) 13.8 (1.74–25.9), P< 0.05 
 All women 108 45.7 47.5 1.8 (−3.1 to 6.8) 161 46.2 29.2 −17.0 (−21.9 to −12.1) 18.8 (11.6–26.0), P< 0.0001 

Parous women and women fitted with an IUD or using progesterone-only contraception were excluded.

When considering all three cohorts together, the women who were using COC at the age of 19 years and had no use of COC at the age of 24 years reported increased severity of dysmenorrhoea, while women with no use of COC at 19 years reported a decrease in the severity of dysmenorrhoea when using COC at the age of 24 (Table V). The total benefit associated with COC use on the severity of dysmenorrhoea between groups was equally apparent when using both the VMS system [P< 0.0001, 0.58: confidence interval (CI): 0.35–0.81] and the VAS (P< 0.0001, 18.8: CI: 11.6–26.0). Thus in this case–control analysis where each woman was her own control, there was a significant difference in the severity of dysmenorrhoea depending on whether or not the woman used COCs when controlling for age. When analysing the cohorts separately, a corresponding decreased severity of dysmenorrhoea was noted for the 62 and 72 cohorts according to both measurement systems, but in the 82 cohort only when using the VAS.

Factors influencing the severity of dysmenorrhoea

A longitudinal variance analysis was performed utilizing all three cohorts together. Women using an IUD or a progestogen-only method were excluded from these analyses. Factors such as COC use, age, child/no child, smoking/no smoking and BMI were included in the model to find predictors for any change in the severity of dysmenorrhoea. Separate analyses were performed using the two different measures of the severity of dysmenorrhoea, i.e. the VMS system and the VAS as separate dependent variables. Predictors for reducing dysmenorrhoea when using the VMS system were COC use (P< 0.0001)—which resulted in a 0.3 unit reduction of dysmenorrhoea—and age (P< 0.0001)—which resulted in a reduction of the severity of dysmenorrhoea of 0.1 units per 5 years. There was no significant influence on the severity of dysmenorrhoea assessed by the VMS system due to the birth of a child (−0.07), smoking (0.1) or BMI (–0.01/BMI unit). When using the VAS as a dependent variable, predictors for reducing dysmenorrhoea were COC use (P< 0.0001) resulting in a decrease of 9 mm, age (P< 0.0001), which resulted in a reduction of 5 mm for the 5-year period and the birth of a child (P< 0.01) with a reduction of 7 mm. Smoking (2.13 mm) and BMI (−0.5 mm/BMI unit) did not produce any significant influence. A comparison of the estimated changes in the severity of dysmenorrhoea and observed changes in the severity of dysmenorrhoea over the 5 years period was made and showed a good fit when using both the VMS system (R2= 0.76, P< 0.0001) and the VAS (R2= 0.76, P< 0.0001). We also performed this longitudinal analysis of variance with ‘cohort’ included in the analysis, as the three cohorts differed in life experiences, which may have influenced the results of this longitudinal analysis. There was no effect of cohort and the identified predictors were unchanged.

Discussion

In this longitudinal cohort study, COC use and increasing age, independent of each other, were associated with a reduced severity of dysmenorrhoea. COC reduced the severity of dysmenorrhoea more than increasing age or childbirth. The decrease in the severity of dysmenorrhoea achieved by COCs was equivalent to the transfer of every third woman one step down on the VMS scale, which—in clinical terms—will result in less pain, improved working ability and a decrease in the need for analgesics.

Two reviews (Proctor and Farquhar, 2007; Wong et al., 2009) have been published assessing the efficacy of COCs in the management of primary dysmenorrhoea. One of these reviews (Proctor and Farquhar, 2007) concluded that it was not possible to say whether COC's reduced the pain of dysmenorrhoea, as the included studies were too small and many of the products assessed were no longer available. A Cochrane review (Wong et al., 2009) that assessed the effectiveness of oral contraceptives for primary dysmenorrhoea concluded that although oral contraceptives are widely advocated as standard treatment, there was only scanty rigorous evidence to support this practice. The authors concluded that it was not possible to conclusively assess the efficacy of oral contraceptives for dysmenorrhoea due to the lack of methodologically sound randomized controlled trials. The Cochrane authors also discussed earlier data from our longitudinal epidemiological study of women in Gothenburg (Milsom et al., 1990) indicating a lower prevalence and severity of dysmenorrhoea in women who were COC users. The Cochrane review authors noted that parity had been controlled for in these analyses but they concluded that the efficacy of COCs could be attributed to the passage of time. Thus in the present study, we have analysed data, taking into account the influence of age as a confounding factor.

One of the strengths of this study design was that the women in the different groups for the purpose of this analysis served as their own individual controls. In the comparisons, we excluded parous women, as studies have shown reduced severity of dysmenorrhoea after childbirth (Sundell et al., 1990). Women fitted with an IUD that may increase the severity of dysmenorrhoea (Grimes et al., 2007) were also excluded as well as women using progesterone methods such as progesterone-only pills, implants, the levonorgestrel-releasing intrauterine system (LNG) etc. as they may also potentially influence the severity of dysmenorrhoea. One of the limitations of this study is that few women had given birth to a child, which limited the ability to test the influence of childbirth on the severity of dysmenorrhoea.

Several factors have been reported to influence the severity of dysmenorrhoea such as COC use, age, childbirth, smoking etc. (Andersch and Milsom, 1982a; Milsom et al., 1990; Sundell et al., 1990; Harlow and Park, 1996; Dawood, 2006). In this study, COC use, age and childbirth were shown to influence the severity of dysmenorrhoea. Childbirth was shown to influence the severity of dysmenorrhoea despite the fact that the reported number of children was low in these three cohorts of women aged 19–24. A longitudinal study (Weismann et al., 2004) in women aged between 19 and 44 years followed for 6 years reported that older age and higher parity were associated with less severe dysmenorrhoea. In the same study, tobacco use was also associated with more severe dysmenorrhoea, which could not be confirmed by the present study.

There was a trend over time regarding the severity of dysmenorrhoea when comparing 19- and 24-year-old women from the 62-, 72- and 82-cohorts where women from the 82-cohort reported a greater severity of dysmenorrhoea compared with the other cohorts. In addition, when performing the longitudinal individual analysis, a change in the severity of dysmenorrhoea was more obvious in all measurements recorded for women in the 62- and 72-cohorts compared with the 82-cohort. During the course of this longitudinal study spread over 20 years where 19-year-old women were recruited in 1981, 1991 and 2001, there have been some changes in the types of COC used. The recorded difference between the 82-cohort and the other two cohorts may reflect the use of COCs, which may have been less effective due to a different composition (changes in estrogen and progestogen type and content) or a different appreciation of pain in the younger cohort. Previous studies found that earlier menarche and longer menstrual periods were associated with a higher incidence of dysmenorrhoea (Andersch and Milsom, 1982a; Harlow and Park, 1996). The menarcheal age in these three cohorts of women did n't differ and neither did the length of cycles.

COC use was shown to reduce dysmenorrhoea severity, and this information is important for health providers when informing women of the non-contraceptive benefits of COC during contraceptive counselling (The ESHRE Capri Workshop Group, 2005). It has previously been shown that women who gain relief from dysmenorrhoea while using COCs have a higher continuation rate of COC use (Robinson et al., 1992). Self-treatment for dysmenorrhoea is common among adolescent girls and young women (O'Connell et al., 2006). Women experiencing dysmenorrhoea on a monthly basis often use non-steroidal anti-inflammatory drugs (NSAIDs) regularly for pain relief and thus the occurrence of adverse events associated with long-term/regular use may be more apparent in this group of women. In a review (Zahradnik et al., 2010) on the use of NSAIDs and hormonal contraceptives for the treatment of dysmenorrhoea the authors concluded that COCs were preferential to NSAIDs for women who wish contraception.

Dysmenorrhoea was common in all three cohorts but women from the youngest cohort, the 82 cohort, reported a greater overall severity of dysmenorrhoea. Absenteeism due to dysmenorrhoea was also common, and the use of analgesics due to dysmenorrhoea was reported by approximately one in four women. COC use and increasing age, independent of each other, were shown to reduce dysmenorrhoea severity. Childbirth was also shown to reduce dysmenorrhoea severity, but COC use reduced dysmenorrhoea more than both childbirth and increasing age. Effective management of dysmenorrhoea is beneficial for both the afflicted individual and society and thus the possibility of a beneficial influence of COCs on dysmenorrhoea should be included in contraceptive counselling. However, it is desirable that the findings from this study providing evidence for a beneficial effect of COCs in dysmenorrhoea be confirmed by a placebo-controlled, randomized trial where the efficacy of COCs in dysmenorrhoea are assessed as a primary outcome measure.

Authors’ roles

Study design: I.L., A.A.E. and I.M.,  execution: I.L. and I.M., analysis: I.L. and I.M., manuscript drafting: I.L., A.A.E. and I.M. and critical discussion: I.L., A.A.E. and I.M.

Funding

The study was supported by grants from The Gothenburg Medical Society, Hjalmar Svenssons Fund and a National LUA/ALF grant no. 11315.

Conflict of interest

I.L. and I.M. have, during the past 5 years, participated in clinical trials sponsored by Organon. I.M. has participated in International Advisory Boards sponsored by Organon/Schering-Plough/MSD and Schering/Bayer Pharma. I.M. has received lecture fees for presentations sponsored by Organon, Schering/Bayer and Pfizer. I.L. has received lecture fees from Organon/MSD.

Acknowledgements

We thank Ms Marianne Sahlén and Ms Anja Andersson for help with data registration and Björn Areskoug MSc for expertise in statistical programming.

References

Andersch
B
Milsom
I
An epidemiologic study of young women with dysmenorrhea
Am J Obstet Gynecol
 , 
1982
, vol. 
144
 (pg. 
655
-
660
)
Andersch
B
Milsom
I
Contraception and pregnancy among young women in an urban Swedish population
Contraception
 , 
1982
, vol. 
26
 (pg. 
211
-
219
)
Burnett
MA
Antao
V
Black
A
Feldman
K
Grenville
A
Lea
R
Lefebvre
G
Pinsonneault
O
Robert
M
Prevalence of primary dysmenorrhea in Canada
J Obstet Gynaecol Can
 , 
2005
, vol. 
27
 (pg. 
765
-
770
)
Campell
MA
McGrath
PJ
Use of medication by adolescents for the management of menstrual discomfort
Arch Pediatr Adolesc Med
 , 
1997
, vol. 
151
 (pg. 
905
-
913
)
Coco
AS
Primary dysmenorrhea
Am Fam Physician
 , 
1999
, vol. 
60
 (pg. 
489
-
498
)
Davis
AR
Westhoff
CL
Primary Dysmenorrhea in Adolescent Girls and Treatment with Oral Contraceptives
J Pedriatr Adolesc Gynecol
 , 
2001
, vol. 
14
 (pg. 
3
-
8
)
Davis
AR
Westhoff
C
O'Connell
K
Callagher
N
Oral contraceptives for dysmenorrhea in adolescent girls: A randomized trial
Obstet Gynecol
 , 
2005
, vol. 
106
 (pg. 
97
-
104
)
Dawood
MY
Primary dysmenorrhea—advances in pathogenesis and management
Obstet Gynecol
 , 
2006
, vol. 
108
 (pg. 
428
-
441
)
Grimes
DA
Lopez
LM
Manion
C
Schultz
KF
Cochrane systematic reviews of IUD trials: lessons learned
Review Article Contraception
 , 
2007
, vol. 
75
 
Supplement 1
(pg. 
S55
-
S59
)
Harlow
SD
Campell
OM
Epidemiology of menstrual disorders in developing countries: a systematic review
BJOG
 , 
2004
, vol. 
111
 (pg. 
6
-
16
)
Harlow
SD
Ephross
SA
Epidemiology of menstruation and its relevance to women's health
Epidemiol Rev
 , 
1995
, vol. 
17
 (pg. 
265
-
286
)
Harlow
SD
Park
M
A longitudinal study of risk factors for the occurrence, duration and severity of menstrual cramps in a cohort of college women
Br J Obstet Gynaecol
 , 
1996
, vol. 
103
 (pg. 
1134
-
1142
)
Hendrix
SL
Alexander
NJ
Primary dysmenorrhea treatment with a desogestrel-containing low-dose oral contraceptive
Contraception
 , 
2002
, vol. 
66
 (pg. 
393
-
399
)
Lindh
I
Blohm
F
Andersson Ellström
A
Milsom
I
Contraceptive use and pregnancy outcome in three generations of Swedish female teenagers from the same urban population
Contraception
 , 
2009
, vol. 
80
 (pg. 
163
-
169
)
Melzack
R
Katz
J
Wall
PD
Melzack
R
Pain measurement in persons in pain
Textbook of Pain
 , 
1994
3rd edn
London
Churchill Livingstone
(pg. 
337
-
351
)
Milsom
I
Sundell
G
Andersch
B
The influence of different combined oral contraceptives on the prevalence and severity of dysmenorrhea
Contraception
 , 
1990
, vol. 
42
 (pg. 
497
-
506
)
O′Connell
KJ
Davis
AR
Westhoff
C
Self treatment patterns among adolescent girls with dysmenorrhea
J Pediatr Adolesc Gynecol
 , 
2006
, vol. 
19
 (pg. 
285
-
289
)
Proctor
ML
Farquhar
CM
Dysmenorrhea
BMJ Clin Evid
 , 
2007
pg. 
pii: 0813
  
Online PMID: 19454059
Robinson
JC
Plichta
S
Weisman
CS
Nathanson
CA
Ensminger
M
Dysmenorrhea and the use of oral contraceptives in adolescent women attending a family planning clinic
Am J Obstet Gynecol
 , 
1992
, vol. 
166
 (pg. 
578
-
583
)
Statistics Sweden 2001
Population Statistics
 
Örebro, Sweden
 
Sundell
G
Milsom
I
Andersch
B
Factors influencing the prevalence and severity of dysmenorrhea in young women
Br J Obstet Gynaecol
 , 
1990
, vol. 
97
 (pg. 
588
-
594
)
The ESHRE Capri Workshop Group
Noncontraceptive health benefits of combined oral contraception
Human Reproduction Update
 , 
2005
, vol. 
5
 (pg. 
513
-
525
)
Weismann
AM
Hartz
AJ
Hansen
MD
Johnson
SR
The natural history of primary dysmenorrhea: a longitudinal study
BJOG
 , 
2004
, vol. 
111
 (pg. 
345
-
352
)
Winkler
UH
Ferguson
H
Mulders
JAPA
Cycle control, quality of life and acne with two low-dose oral contraceptives containing 20 μg ethinyl estradiol
Contraception
 , 
2004
, vol. 
69
 (pg. 
469
-
476
)
Wong
CL
Farquhar
C
Roberts
H
Proctor
M
Oral contraceptive pill for primary dysmenorrohoea
Cochrane database of Systematic Reviews
 , 
2009
 
Art. No.: CD002120 DOI: 10.1002/14651858. CD002120.pub3
Zahradnik
HP
Hanjalic-Beck
A
Groth
K
Nonsteroidal anti-inflammatory drugs and hormonal contraceptives for pain relief from dysmenorrhea: a review
Contraception
 , 
2010
, vol. 
81
 (pg. 
185
-
196
)

Comments

1 Comment
"To the editor on, ""The effect of combined oral contraceptives and age on dysmenorrhoea: an epidemiological study. """
22 March 2012
Jhumka Gupta

Dear Editors, I read with great interest the paper by Lindh et al on the effect of combined oral contraceptives on reducing dysmenorrheal pain.1 This longitudinal, case-control study suggests the use of combined oral contraceptives reduced the severity of dysmenorrhea in the study sample; taking into account the influence of age. An important limitation of this study was notably absent--the authors did not determine whether the women in the study sample had primary or secondary dysmenorrhea.

While combined oral contraceptives may be helpful in reducing severity of pain among women with primary dysmenorrhea, the protective role of oral contraceptives in pain arising from secondary dysmenorrhea is far more complex. This is particularly true in the case of endometriosis, including deep infiltrating endometriosis. It takes women an average of 7 -11 years to receive a laparoscopically-confirmed diagnosis of endometriosis.2 Delays in definitive diagnosis are often due to lack of awareness of symptoms on the patients' part, but physician practices such as empiric use of oral contraceptives, can also contribute to delays in diagnostics and surgical treatment, including complete excision. In fact, recent research suggests the use of oral contraceptives during adolescence may be a marker for the development of deep infiltrating endometriosis in adulthood,3-5 thus implicating the inevitable progression of untreated disease and importance of earlier surgical intervention.3-5 Moreover, pain relief from oral contraceptives for endometriosis pain may only be temporary and thus may offer an explanation of the authors' finding that women who used oral contraceptives at age 19, but not at age 24 reported an increase in pain.

Future prospective research on oral contraceptives and dysmenorrheal pain should strive to explicitly examine the underlying etiologies of such pain, and perform sufficiently powered sub-group analysis by such confirmed etiologies as one size is likely not to fit all, particularly in the case of secondary dysmenorrhea involving endometriosis.

Sincerely, Jhumka Gupta, ScD Assistant Professor Division of Chronic Disease Epidemiology Yale School of Public Health New Haven, CT USA

1. Lindh I, Ellstrom AA, Milsom I. The effect of combined oral contraceptives and age on dysmenorrhoea: an epidemiological study. Hum Reprod 2012;27(3):676-82. 2. Hadfield R, Mardon H, Barlow D, Kennedy S. Delay in the diagnosis of endometriosis: a survey of women from the USA and the UK. Hum Reprod 1996;11(4):878-80. 3. Chapron C, Borghese B, Streuli I, de Ziegler D. Markers of adult endometriosis detectable in adolescence. J Pediatr Adolesc Gynecol 2011;24(5 Suppl):S7-12. 4. Chapron C, Lafay-Pillet MC, Monceau E, Borghese B, Ngo C, Souza C, et al. Questioning patients about their adolescent history can identify markers associated with deep infiltrating endometriosis. Fertil Steril 2011;95(3):877-81. 5. Chapron C, Souza C, Borghese B, Lafay-Pillet MC, Santulli P, Bijaoui G, et al. Oral contraceptives and endometriosis: the past use of oral contraceptives for treating severe primary dysmenorrhea is associated with endometriosis, especially deep infiltrating endometriosis. Hum Reprod 2011;26(8):2028-35.

Conflict of Interest:

None declared

Submitted on 22/03/2012 8:00 PM GMT