Quality of life and mental health in adolescent users of oral contraceptives. Results from the nationwide, representative German Health Interview and Examination Survey for Children and Adolescents (KiGGS)

The KiGGS study was conducted from May 2003 to May 2006 by the Robert Koch-Institute with the aim of establishing a nationally representative sample to assess physical, psychological, and social health in German children and adolescents according to the WHO health definition. In the total, 17,641 study participants up to 17 years of age were surveyed which corresponded to a response rate of 66.6%. A two-stage probability sampling strategy was applied [15]. In the first stage, a sample of 167 municipalities was drawn according to the structure of the federal states and the sizes of the municipalities of the Federal Republic of Germany. In the second stage, at each sample point, a random sampling from local population registries in proportion to the age and gender structure of the nationwide child population was performed. The local study teams consisted of five trained members, led by a physician experienced in pediatrics. The survey comprised physical and medical examinations including a comprehensive psychometric assessment. The present analysis is based on 1695 female study participants, 15–17 years of age, for whom valid measurements from psychometric testing and data on oral contraceptive use were available.

From all study participants aged over 15 years and their accompanying parents, written informed consent was obtained prior to each interview and examination. The study was approved by the Institutional Review Board of the Charité Universitätsmedizin Berlin (Ethics Committee of the Virchow Hospital, Humboldt University) and the German Federal Office for the Protection of Data [15].

Data collection comprised a computer-assisted personal interview, self-administered psychometric assessment, and standardized physical examinations. Socioeconomic status (SES) was classified into three groups (low, medium, and high) using the Winkler index, which employed information about parents’ education, occupational status, and net household income [16]. Based on information on their nationality, country of birth, and year of immigration of both parents, the study participants were classified as having or not having a family immigration background [17]. Anthropometric measurements of body weight and height were performed using a portable Harpenden stadiometer (Holtain Ltd., Crymych, UK) and a calibrated electronic scale (SECA, Birmingham, UK), respectively. From these data, body-mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters. Systolic and diastolic arterial blood pressure recordings were obtained non-invasively from two independent sequential readings taken after 5 min of rest in a sitting position with the subject’s right arm on the desk, the forearm in supination, and the elbow at the level of the right atrium [18‐20]. The Datascope Accutorr Plus sphygmomanometer equipped with a portable monitor was used to detect oscillometric pulsations in the brachial artery during cuff deflation [21, 22]. With regard to the circumference of the subject’s upper right arm, four different sizes of inflatable cuffs were used, as the cuff had to cover at least two-thirds of the length of the upper arm from the axilla to the antecubital fossa. Serum 25-hydroxyvitamin D [25(OH)D] concentrations were measured from venous blood samples using an automated direct LIAISON chemiluminescence immunoassay (CLIA, DiaSorin), as described [23, 24].

The study physicians performed a standardized face-to-face, computer-assisted personal interview with each participant, in which detailed information on medical and psychiatric history was collected. The drug-use history covered both prescribed and non-prescribed over-the-counter (OTC) drugs which had been consumed during the last 7 days prior to the interview including oral contraceptive preparations. Adolescents were encouraged to add data on the use of medicines themselves. For the purpose of verification, parents and caregivers were asked in advance to bring prescriptions or original packaging to the local examination sites. Detailed information on psychotropic medication, such as the duration of use, was not available. All reported medications were divided into various groups according to the Anatomical Therapeutic Chemical (ATC) classification system, including psychotropic drugs.

The generic, German-language KINDL-R (Kinder-Lebensqualitätsfragebogen) questionnaire was employed in the KiGGS study as a screening instrument for health-related quality of life [25, 26]. This questionnaire distinguishes six dimensions of health-related quality of life with reference to the past week: physical well-being, emotional well-being, self-esteem, family, friends, and every day functioning. It contains 4 items per each subscale which were measured on a 5-point Likert scale (never, seldom, sometimes, often, all the time). The subscale scores were transformed to a range of 0 to 100, with a higher total sum score representing better health-related quality of life. The second instrument used to estimate prevalence rates for psychopathological problems with respect to oral contraceptive use was the Strengths and Difficulties Questionnaire (SDQ). This well-evaluated psychometric screening tool for children and adolescents contains five different subscales measuring emotional symptoms, conduct problems, hyperactivity-inattention, peer relationship problems, and prosocial behavior [27‐29]. Each of the 25 items of the SDQ is scored on a 3-point Likert scale with 0 = not true, 1 = somewhat true, or 2 = certainly true, with higher scores indicating greater problems, except for prosocial behavior, where a higher score indicates more positive behavior. A total difficulties score ranging from 0 to 40 was obtained by summing the scores of all subscales, except prosocial behavior.

Anonymized data obtained from the Robert Koch-Institute were analyzed using the Statistical Package for the Social Sciences (SPSS, IBM, New York, USA) software, version 22. To ensure that the survey cohort was representative of the national population, sampling weights were included in the analysis, which accounted for unequal sampling probabilities by adjusting for deviations in demographic characteristics (age, sex, residence in West or East Germany, and level of urbanicity). Descriptive statistics for users and non-users of oral contraceptives were expressed as means and standard deviations or as frequencies and percentages in the case of categorical data. Significant differences between the two groups were tested using Student’s t test for numerical data and Chi-square test for categorical variables, respectively. Fisher’s exact probability test was performed to test whether psychotropic drug use was more common in hormonal contraceptive users. Pearson’s correlation test was used to assess relationships between blood pressure and dimensions of mental well-being. In order to assess the relationships between quality of life and psychopathological problems, respectively, and oral contraceptive use, a series of linear regression models were calculated using components of mental well-being (self- and parent-rated KINDL-R or SDQ sum score) as the dependent variable and hormonal contraception as independent variable. Based on the univariate comparisons, these models were adjusted for age, body-mass index, migration status, log-transformed values for 25(OH)D, and mean arterial blood pressure. In addition, psychotropic medication use was added to the list of confounders, since from the existing literature on adolescents, it is known that hormonal contraception increases the risk of psychotropic drug use [30]. Migration status was included as a confounder in these models, since cultural diversity is known to influence contraceptive use. Furthermore, we computed mediation models using the SPSS macro PROCESS v2.10 [31] to test for possible interaction effects of (1) serum 25(OH)D on the relationship between oral contraceptive use and mean arterial blood pressure and (2) of blood pressure mediating indirect effects of oral conception on serum 25(OH)D. In order to keep the results comparable to the linear regression models, the covariates entered into the mediation models were kept the same. The mediator variable in the models was considered to be significant, when the 95th percentile-based confidence interval for the indirect effect did not exceed zero. In all tests, statistical significance was defined as p < 0.05.

Approximately one-third of all female adolescents aged 15–17 years (n = 522, 30.8%) reported current medication with oral contraceptive pills. The percentage of study participants exposed to hormonal contraceptives increased steadily with age, as only 85 out of 560 girls (15.2%) in the youngest age stratum of 15 years were taking these drugs as compared to 184 out of 571 (32.2%) participants who were 16 years old. Among the age group of 17-year-old girls, nearly half used hormonal forms of contraception (254 out of 564, 45.0%). Study participants taking hormonal contraceptives were older as compared to non-users (16.8 ± 0.8 vs. 16.3 ± 0.9 years, p < 0.001) and had a non-significantly higher body-mass index (22.4 ± 3.5 vs. 22.2 ± 4.3 kg/m², p = 0.327). Participants with an immigration background were less likely to be exposed to hormonal contraceptives (6.7% vs. 20.7%, p < 0.001). However, the two groups did not differ with respect to their SES (p = 0.304). A detailed epidemiological characterization of the total study cohort dichotomized along the use of oral contraceptives is given in Table 1.

Since previous work from our and other groups has shown that, in adolescent KiGGS study participants, higher systolic blood pressure recordings and elevated vitamin D levels are both linked to better well-being and low distress [20, 32, 33], we next investigated the relationship between hormonal contraception and these two somatic parameters. As demonstrated in Table 1, both systolic (115.6 ± 8.9 vs. 113.9 ± 9.5 mmHg, p < 0.001) and diastolic blood pressure recordings (70.6 ± 6.9 vs. 69.0 ± 7.2 mmHg, p < 0.001) were higher in young women using oral contraception as compared to non-users. In study participants, who took oral contraceptives, the calculated mean arterial blood pressure was on average 1.7 mmHg higher than in non-users (88.2 ± 7.4 vs. 86.5 ± 7.7 mmHg, p < 0.001). In addition, serum 25(OH)D levels were significantly higher in oral contraceptive users than in non-users (59.5 ± 32.9 vs. 46.1 ± 28.0 nmol/L, p < 0.001). Figure 1 shows the mean blood pressure (A) and serum 25(OH)D levels (B) by age groups separately for the users and non-users of hormonal contraceptives.

Univariate comparisons were performed to address whether the hormonal contraception status was linked to changes in health-related quality of life and mental health. Our data demonstrated that the self-rated KINDL-R scores did not differ between the two groups of current users and non-users of oral contraceptives (69.2 ± 11.2 vs. 69.2 ± 11.0, p = 0.943, Table 2). Although the results from self-rating and proxy-rating by parents correlated significantly with each other (r = 0.533, p < 0.001), parents more often reported higher KINDL-R scores than their adolescent daughters. Again, there was no difference in parent-rated KINDL-R scores between users and non-users (72.9 ± 10.6 vs. 72.9 ± 10.5, p = 0.985). The hormone group and the non-hormone group did not significantly differ with respect to self- (10.9 ± 4.4 vs. 10.8 ± 4.6, p = 0.732) and parent-rated SDQ (7.2 ± 4.8 vs. 7.0 ± 4.6, p = 0.390). Similar to the KINDL-R data, the self- and proxy-rated SDQ versions correlated highly with each other (r = 0.483, p < 0.001). Next, we tested whether psychotropic drug prescription was more frequent in participants taking oral contraceptive pills compared with those not using hormonal contraception. In total, 93 participants using oral contraceptives reported additional psychotropic drug medication (17.8%), while only 14.4%, namely 169 out of 1173 non-users were taking psychotropic drugs (Table 2). In contrast, experience with psychotherapeutic treatment was more frequent in non-users (1.4%) as compared to users (1.0%) of oral contraceptives, although this difference was not at all statistically significant (p = 0.478). In summary, in univariate analysis, there was a trend towards a higher rate of psychotropic drug prescription in girls taking oral contraceptive pills compared with those not using hormonal contraception (p = 0.052), but psychometric indicators for mental well-being were similar in the two groups.

A series of linear regression models was performed with either self- or parent-rated KINDL-R or SDQ as the dependent variable and contraceptive drug use as independent variable to test for associations with blood pressure and 25(OH)D (Table 3). Notably, oral contraceptive use failed to be significantly associated with all parameters of mental well-being, irrespective of whether self- or parent-ratings were included in these models. However, in all four models psychotropic medication was a significant and inversely associated predictor of mental well-being (in all models, p < 0.001). In three of the four models tested, blood pressure and serum 25(OH)D were independently and significantly linked to mental well-being (Table 3). While blood pressure was significantly linked to self-rated KINDL-R and both self- and proxy-rated SDQ (in all models, p ≤ 0.001), it failed to reach significance when parent-rated KINDL-R was used as the dependent variable (p = 0.326, Model 2 in Table 3). In all models, higher circulating serum vitamin D concentrations significantly predicted better mental well-being, except for self-rated SDQ, where this relationship was observed only as a trend (p = 0.075).

Using a mediation model (Model 1 in Table 4), we confirmed the direct effect of oral contraceptive use on mean arterial blood pressure (p = 0.021), but did not find any indication of serum 25(OH)D being a significant mediator of this relationship (95% CI =  − 0.254 to 0.094). Similarly, as shown in Model 2 in Table 4, oral contraception was positively associated with serum 25(OH)D levels (p < 0.001), whereas blood pressure did not elicit any significant interacting effect on this relationship (95% CI =  − 0.007 to 0.001).