2.1. Participants and procedures:
Twenty-nine adults (older than 18 years of age) with PWS were enrolled in 2015 and 2016 either during routine visits to the Department of Endocrinology, Karolinska University Hospital, Stockholm, or after responding to advertisements in the journal of the Swedish PWS Association.
Data for controls were retrieved from the Lifelines Cohort Study. Lifelines is a multi-disciplinary prospective population-based cohort study examining in a unique three-generation design, the health and health-related behaviors of 167,729 persons living in the North of The Netherlands. It employs a broad range of investigative procedures in assessing the biomedical, socio-demographic, behavioral, physical and psychological factors, which contribute to the health and disease of the general population, with a special focus on multi-morbidity and complex genetics (13, 14, 15). Data on hair cortisol was available for 266 individuals, who were used as controls in the present study (15).
2.2. Hair processing and analysis
A sample of approximately 100–150 hairs was cut from the posterior vertex of the scalp, as close to the scalp as possible. Hair samples were taped to a piece of paper and stored in envelopes in the dark at room temperature prior to processing.
Hair sample analysis was performed as described previously (13). In summary, from each hair sample, approximately 10 mg of the proximal 3 cm was weighed and cut into 1 cm segments. Hair samples were washed in isopropanol for 2 minutes, and then left to dry. The hair was then extracted with methanol and purified using solid phase extraction. Cortisol was quantified by liquid chromatography − tandem mass spectrometry (LC–MS/MS) (Waters, Milford, MA) (16). Two standard curves of ten calibration standards were prepared on each day of analysis. The functional sensitivity (lower limit of quantification, LLoQ) was determined in hair extracts of male and female adults by serial dilution of hair extract, starting with 40 mg/ml to 0.3125 mg/ml. The LLoQ was calculated in pg/ml hair (16). The geometric mean of hair cortisol analyses in The Lifelines Cohort study (N=266) was 2.7 pg/mg (13).
2.3 Questionnaire
The participants with PWS and/or their caretakers, filled out a questionnaire about their age, sex and anthropometric measurements. This questionnaire was not developed for this study, but has been used for multiple studies in the past
(15). The stress was assessed by the question: did any stressful events occur during the last three months? In case there did, what happened?
In addition, the participants were asked in a standardized manner whether they used any product containing corticosteroids in the past three months and the route of administration (i.e., oral, intravenous, nasal, topical, inhaled, joint injections or others).
Data retrieved from the Lifelines Study for the controls included information on the subjects’ age, sex, anthropometric and stressful life events. Life events were evaluated using the Dutch version of the List of Threatening Experiences (LTE), including occurrence of twelve major life events in the past twelve months (13).
2.4 Matching process and statistical analysis
The PWS patients were categorized into four age-groups [(18-25 years), (26-35 years), (36-45 years) and (46-60 years)], resulting in a fairly balanced age-sex groups for each of the approximately 10 years intervals. Subjects from each group of PWS patients were matched for age and sex to a total of 105 subjects from the corresponding control groups, in a 1:5 ratio. However, due to low number of controls for males in the 18-25 age-group (9 adults with PWS, 23 matched-controls) matching was made in a ratio of 1:2.5 for this particular sex-age group. The matching was programmed using the “seed” command in STATA 15.0 to ensure a random selection of subjects.
Results are presented as mean±SD. For comparison between the groups, Student’s t-test was used. Mean hair cortisol was considered to best capture the clinical reality with large variations in hair cortisol and a generalized linear regression model was used to examine the association of crude levels and levels adjusted for BMI and stress as potential confounders. The Akaike Information Criteria (AIC) test was used to confirm best fit model where by both BMI and reported stress were retained in the model, as they predicted the model well. Statistical significance was set at p<0.05.