Investigating the association between occupational exposure to asbestos and ovarian carcinoma: results from a pilot study in Germany

Between December 2017 and May 2018, one thousand out of about 16.000 insured woman registered at the health service of a German trade association as having been occupationally exposed to asbestos were invited to participate in a pilot study. They received a postal invitation package containing an information letter on the details of the study along with a consent form as well as a return envelope to send back the informed consent form. Non-respondents were reminded up to two times. Trained professionals via telephone individually interviewed those women who gave written informed consent to participate. During the interview, a study questionnaire that incorporated validated instruments from previous studies was completed.

Since it can be assumed that at the time of the asbestos ban in Germany in 1993, the registered women were at least 16 years of age (as they must have been employed, otherwise they could not appear in the registry), we reckoned that the youngest members of the cohort were now at least 38 years old. Applying the incidence rate for Germany provided by the Robert Koch Institute of 29.9 cases per 100,000 persons in the age range of 35 to 84 years to our total sample of 16.000 registered women, 4.8 cases per year could be expected [21]. Hence, during a 22-year follow up there should be 106 cases of ovarian cancer in the sample due to natural rate of disease. Applying the odds ratio calculated by the above mentioned meta-analyses of 1.77 (95% CI: 1.37–2.28), around 200 cases should be expected [18]. In the pilot study, 5% of the the registered women should be included, i.e., 800 persons. Reckoning with a willingness to participate of 80%, a random sample of 1000 women was invited. Based on the above-mentioned calculations, the expected number of cases among those 800 individuals was 5 cases (95% CI: 2–11 cases).

During the interview, all participants were asked if they ever had a diagnosis of ovarian cancer. In case of an affirmative answer, the interviewers asked them for the year of diagnosis and the clinic in which they were treated. As described below, we then tried to collect medical documentation and histologic materials for tumour validation.

For assessment of asbestos exposure, the study questionnaire asked for the participants’ full occupational history including duration (start and end date), job title, tasks and activities, form of employment (full-time or part-time), and exposure to asbestos (yes/no) for each job that was performed for at least 12 months. In case of jobs containing asbestos exposure, participants answered a job-specific supplementary questionnaire. According to the specific type of job reported, interviewers selected one out of 33 job-specific questionnaires (Table 1) that were specifically tailored for collecting detailed information on various aspects of asbestos exposure such as specific materials, distance to the source of exposure, or use of personal protective equipment. Ahrens and colleagues developed those job-specific questionnaires in the 1990s specifically for retrospective assessment of occupational asbestos exposures in German study populations [22, 23]. For the purpose of the present study, we developed an additional supplementary questionnaire, which the interviewers used if the job reported by the interviewee could not be allocated explicitly to one of the 33 job-specific questionnaires. This additional questionnaire contained those aspects of asbestos exposure which the German Social Accident Insurance considers as the most important ones to calculate exposure levels in asbestos fibre-years [24]. Those were type of activity, start and end date, frequency per week, duration per day, type of premises, distance from the source of asbestos exposure, use of personal protective equipment, type of materials and type of tools used, use of asbestos at neighbouring work places, and ventilation measures. Lastly, an expert (GS) calculated individuals’ exposure levels in asbestos-fibre years using the data provided by the supplementary questionnaires. We chose AFY as our primary measure of exposure as they are the valid and applied measure of asbestos exposure in Germany’s occupational disease legislation. Estimation of AFY is based on a multiplication of duration of employment (in years, assuming that a standard working years has 240 workdays), duration of performing asbestos-related tasks per day (in hours per day), and level of exposure defined as fibre concentration (in fibres per cubic centrimetres (F/cm³)). As explained in the example by Felten et al., a worker performing the task of spraying an asbestos pump with a fibre concentration of 400 F/cm³ for one standard working year of 240 workdays and 8 h per workday would accumulate 400 AFY [25].

The study questionnaire also collected information on potential confounders starting with socio-demographic information (age, schooling, highest educational level) and current occupational status (still working full time, part time, or retirement). In addition, the questionnaire asked for information on the participants’ health and lifestyle with items being based on questionnaires from previous studies such as the European Community Respiratory Health Survey [26]. The participants were also asked if there were cases of cancer in their family (i.e., parents and siblings), and if the answer was affirmative, more details on the type of cancer diagnosed and the dates of diagnosis were collected. Furthermore, they were inquired about personal history of ovarian cancer, any other cancer, and if any surgery was performed on their abdomen. In the lifestyle part of the questionnaire, participants were questioned about the use of talcum powder and duration of its use in genital area or other body parts. The lifestyle part also included questions regarding ever smoking (defined as a minimum of 20 packets of cigarettes during lifetime, or at least one cigarette per day for 1 year), physical activity (not active: mostly sitting throughout work; moderately active: being active for 20 min three times a week over a period of 10 years; highly active: doing sports every week for 30 min at least three times per week), number of pregnancies, birth year of the offspring, age at menarche and menopause, and breastfeeding (yes/no) including duration of breastfeeding. An English translation of the study questionnaire is provided in the supplement (Additional file 1).

We collected medical documentation including records from imaging procedures (CT, MRI and x-ray scans) from all interviewees, where available. The images were then screened for indication of pleural plaques by two independent experienced radiologists/occupational physicians (KGH and TK). In addition, we aimed at retrieving histological materials from the suspected ovarian cancer cases for tumour validation by a pathologist specialised in diagnosis of asbestos-related carcinoma (AT). The documents were obtained from the interviewees themselves, from their treating doctors and hospitals, or from records available at the health service of the cooperating trade association.

We evaluated the feasibility of the planned, large-scale study by using the following, a priori defined criteria:

We considered recall bias as a potentially significant source of bias due to difficulties of the individuals to correctly report asbestos exposure as well as start and end dates of their jobs. Therefore, when making an appointment with the participants the interviewers told them to put out ready any work contracts they had with them at the day of the interview in order to be able to look up for the details during the interview. Interviewer bias was addressed with professional training of the interviewers instructing them to conduct the interviews in a standardised form to reduce the influence on respondents’ answers. The interviewers also underwent asbestos-specific training by an expert at the German Social Accident Insurance.

We assessed the feasibility criteria by calculating the following statistics:

Furthermore, we present summary statictics to describe the socio-demographic, lifestyle, and health-related characteristics of the study participants. Summary statistics were calculated as arithmetical means and standard deviations (SD) for continuous variables while categorical variables were expressed as absolute numbers (n) and percentages (%).

Statistical analyses were performed in R version 1.0.44 and SPSS Version 25.

At the time of the interview, the 163 individuals were on average 63 years old. About half of the participants had medium schooling level, a quarter reported low educational level. When asked for their current occupation, about 50% of the individuals reported to be retired. The total number of self-reports of cancer diagnoses other than ovarian cancer was 28, which included breast (9), skin (4), bowel (6), peritoneal (1), splenic (1), thyroid (1), and cervical cancer (5) as well as brain tumour (1). The reported dates of diagnoses were between 1973 and 2017 with twelve reported diagnoses dating back more than 10 years. More than half (53%) of the participants had a family history of cancer and around 37% had a history of gynaecological problems. 146 subjects (89%) reported having ever worked with asbestos. With regard to gynaecologic aspects, individuals indicated an average of two pregnancies. Average duration of OCP use was 15 years. More than half of the subjects reported having had a lower abdomen surgey; in 36 interviewees (22%) this was a surgery of the ovaries. In all variables, the number of missing values was below 10 %, most of them were 100% complete (Table 4).