Use of combined oral contraceptives and headaches.

OBJECTIVE
The study was designed to examine the certain patterns of combined oral contraceptive use in women of childbearing potential and evaluate the relationship between the use of combined oral contraceptives and headaches, bad habits, type of work, and concomitant diseases.


MATERIAL AND METHODS
In total, 194 randomly selected women aged 18 to 40 years who visited a gynecologist for preventive gynecological examination were surveyed. Respondents were categorized as combined oral contraceptive users (n=116; study group) and nonusers (n=78; control group). An anonymous questionnaire developed by the authors of this study and a standardized scale called the Migraine Disability Assessment Scale (MIDAS) were used for the survey.


RESULTS
A multivariate logistic regression analysis demonstrated a significantly higher prevalence of combined oral contraceptive use in women older than 20 years (odds ratio, 6.0; 95% CI, 2.6-14), better educated women (odds ratio, 5.7; 95% CI, 2.1-15.2), and women reporting a steady sexual partner (odds ratio, 4.0; 95% CI, 1.5-11.0). Relationship between headaches and use of combined oral contraceptives as well as other factors were analyzed in a group of 178 respondents; the rest 16 respondents reported not having headaches at all. The prevalence of reported minimal-to-mild and moderate-to-severe impact of headaches on daily activities did not differ significantly between the study and control groups, women with and without bad habits, and white-collar and blue-collar groups (P>0.05). However, women with concomitant diseases significantly more often reported moderate-to-severe impact on daily activities due to headaches (P<0.01). Differences in impact of headaches on daily activities between women using combined oral contraceptives containing 20 or less μg of ethinylestradiol and 30 or more μg of ethinylestradiol did not differ significantly.


CONCLUSIONS
The prevalence of combined oral contraceptive use was higher in women older than 20 years, better educated women, and women reporting a steady sexual partner. The impact of headaches on daily activities did not differ significantly between the combined oral contraceptive users and nonusers.


Introduction
Combined oral contraceptives (COCs) were fi rst approved more than four decades ago. Due to effectiveness and comfort of use, they made a breakthrough in the fi eld of family planning (1). Currently, more than 8 million women use COCs worldwide (2). Though COCs are safe and highly effective in preventing unwanted pregnancy, headaches are among the most common adverse effects of COC use, which quite often lead to discontinuation of use by women (3)(4)(5). Many previous studies, which assessed the impact of COC use on the development of migraine headaches, were performed in the 1970s when COCs contained higher levels of estrogen (from 50 to 100 μg of ethinylestradiol). Clinical studies on COCs containing high levels of estrogens demonstrated a general tendency toward the development of more severe headaches in COC users (5). Data on the impact of modern COCs containing low levels of estrogens (15-35 μg) on headaches are limited and rather controversial.
This study was designed to examine certain patterns of COC use in the group of surveyed women and the relationship between COC use and headaches as well as other factors (bad habits, type of work, and concomitant diseases) that can have an impact on the prevalence and manifestation of headaches.

Material and Methods
Randomly selected respondents of childbearing potential (aged 18 to 40) who visited Šančiai, Šilainiai, and Dainava Maternity Centers for preventive gynecological examination were asked to complete an anonymous questionnaire. The study took 23 months from April 2008 through Decem-ber 2009. Questionnaires were given to 198 women; completed questionnaires were returned by 194 respondents (response rate, 97.9%). Of these women, 116 used contraceptive pills (study group), and 78 were nonusers (control group). Exclusion criteria were as follows: 1) women younger than 18 and older than 40 years; 2) women suffering from migraine; 3) women who did not consent to take part in the study. The study was approved by Kaunas Regional Ethics Committee for Biomedical Research (approval No. BC-MF-32). All women gave a written consent to participate in the study.
A two-part questionnaire was used for the survey. It included a 16-item questionnaire developed by the authors and a standardized questionnaire called the Migraine Disability Assessment Scale (MIDAS). The questionnaire developed by the authors contained questions related to sociodemographic characteristics, COC use, bad habits, and diseases suffered by respondents that can have an impact on headaches.
The MIDAS questionnaire contains 5 questions describing the most common types of daily activities for a respondent aged 18 to 40. The patient reports the number of days in the last 3 months he/she missed from work or school (fi rst question), did not do household work (third question), and missed leisure activities (fi fth question) because of headaches. The other two questions assess a severe reduction in productivity. The total score is obtained by summing the scores for all fi ve questions on the impact of headaches on daily activities: Grade I (score 0-5, minimal or no impact on activities); Grade II (score 6-10, mild impact on activities); Grade III (score 11-20, moderate impact on activities); and Grade IV (score 21+, severe impact on activities) (6)(7)(8). For analysis purposes, the impact of headaches on daily activities was grouped into two categories of minimal-to-mild (Grades I and II) and moderateto-severe (Grades III and IV) impact.
For statistical data analysis purposes, the statistical software package SPSS 17.0 for Windows was used. A statistical relationship between qualitative variables was assessed using the Pearson χ 2 test or exact Kruskal-Wallis test (for three and more independent samples) and the Mann-Whitney U test (for two independent samples). Differences in data were considered statistically signifi cant when P<0.05. A multivariate backward stepwise logistic regression method was used to predict the use of combined oral contraceptives. Age, level of education, and marital status were included as independent factors into the logistic regression model.

Results
The mean age of respondents in this study was 23.6 years (SD, 5.2). The mean age of COC us-ers and nonusers was 25.6 years (SD, 5.2) and 20.5 years (SD, 3.4), respectively. The receiver operating characteristic (ROC) curve analysis was performed to select an optimal cutoff point for assessing COC use. In our study, the cutoff point was 20 years of age (AUC 83.9%, P<0.001) with a sensitivity of 0.9 and a specifi city of 0.7 (Fig. 1).
For further analysis, women were assigned to two groups according to their age: women aged 20 and younger and women older than 20 years. It appeared that signifi cantly more respondents in the study group (COC users) were older than 20 years of age (81.8%) as compared with the control group (i.e., nonusers), which contained signifi cantly more women aged 20 and younger (76.7%) (P<0.001).
The data analysis of the level of education and marital status also revealed signifi cant differences between the COC user and nonuser groups. COC users signifi cantly more often had better education and lived with a steady sexual partner (P<0.001).
A binary multivariate backward stepwise logistic regression method was used to determine the odds ratios for individual characteristics of the COC users and select optimal combinations (Table 1).
Based on the available data, higher prevalence of COC use in women older than 20 years (odds ratio, 6.0; 95% CI, 2.6-14), better educated women (odds ratio, 5.7; 95% CI, 2.1-15.2), and women reporting a steady sexual partner (odds ratio, 4.0; 95% CI, 1.5-11.0) can be predicted. The classifi cation table of the logistic regression analysis shows that 83.2% of predictions were correctly classifi ed. A rather high coeffi cient of determination (r 2 =0.4) shows that the model was selected correctly. The majority of women, i.e., 91.8% (n=178), in both groups reported having experienced headaches at some point in their life; 8.2% (n=16) of respondents reported that they had never had headaches. The distribution of answers to this question did not differ signifi cantly between the study and control groups.
All 178 women who reported having experienced headaches at some point in their life also completed the MIDAS questionnaire.
Furthermore, 78.4% of COC users reported mini mal-to-mild impact of headaches on their daily activities, while 21.6% of COC users assessed the impact of headaches on their daily activities as moderate-to-severe (Fig. 2). The nonusers less frequently reported moderate-to-severe impact of headaches on daily activities, but the difference was not significant (Table 2).
Other factors that can be related to headaches, such as the level of ethinylestradiol in the COC pills used, smoking, alcohol consumption, and concomitant di seases, were also assessed in the study and control groups.
Fifteen women (13.6%) using COC containing 30 μg of ethinylestradiol reported a minimal-tomild impact of headaches on daily activities, while 5 women (4.5%) in this group rated the impact as moderate-to-severe. The impact of headaches on daily activities in the users of COCs containing 20 μg of ethinylestradiol was minimal-to-mild in 34 cases (30.9%) and moderate-to-severe in 8 cases (7.3%). The impact of headaches on daily activities of COC users who used formulations with 20 and 30 μg of ethinylestradiol did not differ signifi cantly.
The impact of headaches on daily activities did not differ signifi cantly between smokers and nonsmokers as well as between moderate alcohol users and nonusers (P>0.05 in both cases). The prevalence of moderate-to-major impact of headaches on daily activities was not signifi cantly higher in COC users with bad habits (P>0.05) as compared with nonusers with bad habits.
Women with concomitant diseases (renal diseases, hypertension, diabetes mellitus, etc.) significantly more often reported a moderate-to-severe impact on daily activities due to headaches as compared with women having no concomitant diseases (P<0.05). A univariate regression method used for data evaluation allowed for prediction that women with concomitant diseases had a 2.7-fold (95% CI, 1.1-6.6) higher risk of headaches with a moderateto-severe impact on their daily activities as compared with healthy women.

Discussion
Irrespective of efforts by family planning centers and education of society, effective contraceptives still gain limited popularity in certain countries, especially in Eastern Europe and Baltic countries (9). Lithuania is also out of step with many other developed European countries according to effective modern contraceptive use (10)(11)(12). Our study revealed that COCs were more often used by slightly older, better educated women reporting a steady sexual partner. Consequently, organizations engaged in educational work and health care providers (family doctors and obstetricians-gynecologists) should take into account that young, unmarried, and less educated women rarely use effective contraception, which results in very high risks of unwanted pregnancy.
Characteristic Analyzed OR (95% CI) Age (in years) 20 and < >20    One of the common reasons for reluctance to use hormonal contraceptives or discontinuation immediately after starting the use is the fear of side effects or actual side effects. Headaches are a common complaint in women of different ages undoubtedly related to hormonal swings in the woman's body (13). Even in our small-scale study, 21.4% of all surveyed women reported a moderate-to-severe impact of headaches on daily activities. Limited data are available in the world's literature on the relationship between COC use and prevalence of headaches. To the knowledge of the authors, no such studies have been performed in Lithuania at all.
The MIDAS questionnaire was chosen for this study. Pain is a rather subjective feeling, and what is of utmost importance is the level to which daily activities are disturbed by pain. A variety of questionnaires are used for evaluation of headaches, such as the Head-HUNT Questionnaire, the Headache Impact Questionnaire (HimQ), etc. All these questionnaires are precise, but rather long, and it is diffi cult to interpret the results obtained. In 1997, American neurologists developed the MIDAS questionnaire to assess the impact of migraine and headaches on the quality of life in general (8). This short self-administered questionnaire was translated into Lithuanian and validated in Lithuania (14). The Head-HUNT questionnaire is often used in the studies to assess the impact of COC use on headaches. Only one study was found in this fi eld to use the MIDAS questionnaire and assess the impact of headaches on daily activities of COC users (15).
The data obtained in our study show that the impact of headaches on daily activities does not differ between COC users and nonusers. Most of the studies performed by other authors also did not confi rm the relationship between COC use and headaches. In 2005, van Vliet et al. conducted a study in women suffering from cluster headache attacks in order to assess the impact of hormonal products on such attacks (16). In this study, 224 women were surveyed. A screening questionnaire based on the International Headache Society (IHS) criteria for cluster headache was used. Of the 169 women who had ever used COCs and suffered cluster headache attacks, 20 (12%) reported an increase in headaches after starting COC use. Seven women (4%) reported an improvement. According to the authors, the study revealed an insignifi cant impact of oral contraceptive use on cluster headache attacks in general population. Loder et al. examined if the prevalence of headaches (not migraine) was higher in those taking monophasic contraceptive pills as compared with the placebo group (17). They found no signifi cant difference in the prevalence of headaches between women using placebo and monophasic contraceptives; however, women with a history of migraine or other head-related problems may be at the increased risks of developing more severe headaches in the background of COCs. In 2007, Westhoff et al. performed a large-scale study (n=1716). The purpose of this study was to fi nd out the impact of side effects of contraceptives, especially headaches, on discontinuation of COC use in a group of women younger than 25 years (18). Nearly 60% of subjects discontinued COC use by 6 months; however, most subjects reported no increase in headaches. Women complaining about the increased headache were more likely to early discontinue COC use. Nonetheless, the discontinuation of COC use is not related to the side effects of contraceptive pills, especially headaches. In the same year, American scientists conducted a research to assess the severity of headaches in COC users: a standard 28-day combined oral contraceptive cycle consisting of 21 hormonecontaining pills (estrogen + progestin) and 7 placebo pills was converted to a 168-day (24 continuous weeks) extended placebo-free COC regimen (15). The study assessed headaches on a severity scale of 0 to 10 (daily) along with the results according to the MIDAS questionnaire (weekly). During the fi rst 24 days, the results did not differ in both study groups (P>0.05). A signifi cant difference was detected on days 25 through 28 (P<0.0001), which demonstrated that the elimination of a 7-day placebo period from the standard COC regimen led to a decrease in headache severity based on the daily assessments according to the pain scale. Comparison of the MI-DAS score in women using the 21/7 COC regimen and the extended placebo-free regimen revealed that women using placebo-free COC regimen had a decrease in the impact of headaches on daily activities (P<0.05) and the duration of headache (P=0.003). Thus, as compared with a standard COC (21/7-day) regimen, a 168-day extended placebo-free COC regimen led to a decrease in headache severity along with an improvement in the productivity related to daily activities. Finally, the results of recent studies performed in 2009 demonstrated no relationship between headache and the composition of COCs (both in the type and amount of hormones), but headaches were more manifested in COC users suffering from menstruation-related migraine (19).
Therefore, it should be noted that a number of studies (different in methodology and samples used) performed to date did not prove that healthy COC users more often suffered from headaches or that headaches suffered by healthy COC users were more manifested as compared with nonusers; however, women suffering from migraine developed more severe headaches. As it was mentioned above, our study, in fact, confi rmed no relationship between COC use and headaches. The study also assessed the impact of headaches on daily activities (the MIDAS questionnaire was used), which makes our study different from others. A sample studied in our study was not very large but uniform in terms of selected population, which is an undoubted advantage of the study. All subjects were white Lithuanian women, and this eliminated potential racial or ethnic predispositions to headaches. According to literature, headaches are more common in white women as compared with general population (20). Moreover, contrary to other studies (15,16,19,21), other factors that can have an impact on the development of headaches and their impact on daily activities were also taken into account.

Conclusions
In summary of the study results, an average com-bined oral contraceptive user in the study group can be described. It is a woman older than 20 years, married or living with a steady sexual partner, and predominantly with higher education. An original methodology (the MIDAS questionnaire) was used to study the relationship between headaches and combined oral contraceptive use. No relationship between the impact of headaches on daily activities and combined oral contraceptive use was established. The fi nal conclusion on the impact of combined oral contraceptives on headaches calls for a larger-scale and more detailed study.