Sexual behavior and factors associated with young age at first intercourse and HPV vaccine uptake among young women in Germany: implications for HPV vaccination policies

This study was nested in a study designed to determine HPV prevalence in Germany via home-based self-sampling [11]. Briefly, we conducted a cross-sectional study among 20 to 25 year old women in Germany between 2010 and 2012. We used a two-step cluster sampling approach based on the protocol of the German Health Interview and Examination Survey for Adults [12]: Within selected sampling points women were randomly selected from local registration offices. An invitation letter, a self-sampling kit, and a questionnaire were mailed to each woman who gave written informed consent. Self-sampling was performed by cervicovaginal lavage (5 ml volume) with the first generation Delphi-Screener (Delphi-Bioscience, Scherpenzeel, The Netherlands) as described previously [13]. The study was approved by the local ethics committee (Charité, EA2/028/10) and registered at Deutsches Register Klinischer Studien (DRKS 00000599).

With a structured, self-administered questionnaire in the German language we obtained information on sociodemographic factors, data on sexual behavior (such as age at first intercourse and number of lifetime and recent (i.e. in the previous 12 months) sexual partners), medical and smoking history, contraceptive use, past pregnancies, past delivery of a child, as well as HPV-vaccine uptake (time of vaccination and number of doses). The questionnaire has been developed by the Robert Koch Institute and was previously used in another study [14]. Details about the questionnaire and definitions are described elsewhere [11]. Data from the questionnaires were self-reported and not validated.

We calculated proportions on young age at first intercourse, contraceptive use and past pregnancies with corresponding 95% confidence intervals (95% CI). We evaluated the lifetime number of sexual partners by estimating the median, range and interquartile bounds (IQB). For comparison of proportions of categorical variables, we used chi-squared test and Fisher’s exact test, and Student’s t-test for numerical variables. Differences between three age groups (20–21, 22–23 and 24–25 years) were assessed by the Kruskal-Wallis-test for categorical variables and by analysis of variance (ANOVA) for continuous variables.

The age cut-off for ‘young’ age at first intercourse was defined as the lower quartile (age ≤ 14) and a ‘high’ number of lifetime sexual partners was defined as the upper quartile (≥ 7 partners) in the respective variable distribution. For assessing factors associated with HPV vaccine uptake, women with at least one HPV vaccine dose were defined as “vaccinated”.

In order to identify factors associated with (i) young age at first intercourse and (ii) HPV vaccine uptake, we conducted univariate analyses for all binary or categorical exposure variables and calculated odds ratios (OR) and 95% confidence intervals (95% CI). We then used multivariate logistic regression analysis and included variables with a p-value <0.1 in univariate analysis. In addition, we included important potential confounding factors on the basis of the current literature. Two-sided hypothesis tests were performed and a p-value of less than 0.05 was considered as statistically significant. Missing data were treated as such and not imputed. We used the statistical software package STATA, version 11 (STATA Corp., College Station, TX, USA).

We used data of the German Federal Statistical Office (as of 31 December 2011 [15]) to compare characteristics of our study population with the general German female population of the same age.

Overall, 823 out of 5805 women with a valid postal address (14.2%) completed the questionnaire and were included in the analysis (Figure 1). Mean age of participants was 22.6 years (standard deviation (SD) = 1.6). Key characteristics of the study population compared to the general female population in Germany of the same age group are presented in Table 1.

Overall, 95.4% (95% CI, 93.7-96.7%) had experienced sexual debut at the time of the study. More than half of the women were 16 years or older at sexual debut, whereas only 4.6% (95% CI, 3.3-6.3%) were younger than 14 years (Table 2). 73.5% (95% CI, 70.3-76.5%) were living in a relationship and 69.6% (95% CI, 66.5-72.8%) reported having had just one sexual partner in the previous 12 months. The proportion of women with more than one sexual partner in the last 12 months was significantly higher among women not living in a relationship as compared to those living in a relationship (35.6% vs. 16.3%, p-value < 0.01).

17.2% (95% CI, 14.7-20.0%) of sexually active women reported that they experienced first intercourse before or at age of 14 years (Table 2). More than half of the women were older than 15 years at sexual debut and one fourth was older than 17 years. In univariate analysis younger age at first intercourse (defined as ≤ 14 years) was independently associated with lower educational status, current smoking, a higher number of recent and lifetime sexual partners, bisexual experience, usage of contraceptive pill, past pregnancy, and genital warts (data not shown).

The final multivariate model indicated that current smoking, a higher number of sexual partners, and past pregnancy were independently associated with young age at first intercourse (Table 3). In addition, women with younger age at sexual debut were more likely to have bisexual experience and to have used the contraceptive pill. Adjusting for age did not substantially affect the model.

A total of 92.7% (95% CI, 90.1-94.5%) women reported usage of any type of contraception, with the contraceptive pill and condoms being the most common ones (Table 2). 10.8% (95% CI, 8.7-12.9%) of the participants had ever used a vaginal ring and 1.5% (95% CI, 0.6-2.3%) an implant. Overall, 80.8% (95% CI, 77.9-83.4%) of women stated that they would always use condoms during a one-night stand. However, the probability of consistent condom use decreased with an increasing number of a woman's lifetime sexual partners (p-value <0.01). The proportion of women with a previous pregnancy and who delivered a child was 8.6% (95% CI, 6.6-10.4%) and 5.6% (95% CI, 4.0-7.2%), respectively.

Two hundred thirty-four out of 772 participants (30.3%, 95% CI, 27.1-33.7%) with information on HPV vaccination status received at least one dose of an HPV vaccine and 206 (26.7%, 95% CI, 23.6-29.9%) received the full course. Statistically, vaccinated women were significantly younger than unvaccinated women (mean age, 21.5 years (SD = 1.5) vs. 23.0 years, (SD = 1.4), p-value <0.01). Fifty-six (30.1%, 95% CI, 23.6-37.2%) out of 186 women received the first dose of vaccination before their first intercourse. Nineteen (3.5%, 95%, 2.1-5.5%) out of 538 unvaccinated women reported that they were still planning to receive the HPV vaccine. All of these women, except one, already had had their sexual debut.

In multivariate analysis HPV vaccine uptake was independently associated with younger age and a higher educational status. Vice versa, active smoking and migrant background were associated with a lower chance of being vaccinated (Table 4).

Some limitations of our study need to be taken into account. First, the response rate was low (14%) with the potential of introducing participation bias. Since this study was implemented in a population based survey to evaluate HPV prevalence using a self-sampling method, the response rate was comparable to a similar study [34]. However, women with a migrant background were underrepresented and those with a higher education were overrepresented. As educational status was independently associated with young age at first intercourse, it is possible that the true proportion of these women is higher than estimated by our study. However, other characteristics of our study sample, such as place of residency, living in a city, and smoking behavior were comparable to the German female population of similar age groups. Second, information on sexual behavior and vaccination status were self-reported and could be subject to misclassification. Third, due to the sensitive nature of the topic on sexual behavior, social desirability bias cannot be excluded. Although anonymity and confidentiality was ensured, women might have had the tendency to give more “desirable” answers with for example a lower number of sexual lifetime partners or an older age at first intercourse. Forth, a long time interval since relevant exposures (e.g., time at first intercourse) may have led to miscalculation by participants and may have influenced the magnitude of relative risk estimates.