{"paper_id":"11831ffb-4daf-4f03-92cd-5965a0e88d2b","body_text":"In recent years, interest in addressing chronic pelvic pain (CPP) and its complexity in women has increased. CPP is a collective term that encompasses conditions such as vulvodynia, dyspareunia, vaginismus, endometriosis, interstitial cystitis, irritable bowel syndrome, nerve impingements in the pelvic region, visceral sensitivity and pelvic floor myalgia. \n 1 \n  According to the International Association of the Study of Pain (IASP), CPP is pain lasting for at least 6 months, perceived in structures related to the anatomical pelvis, and is often associated with negative cognitive, behavioral, sexual and emotional outcomes. \n 2 \n ,  \n 3\nPopulation‐based studies conducted among women have shown that CPP prevalence ranges widely from 6% to 25%. \n 4 \n  In the only survey conducted in the Nordic countries, 11% of 1200 randomly selected women reported CPP. \n 3 \n  The prevalence of CPP also seems to vary across age groups; although most studies have revealed that CPP is most prevalent during reproductive age, \n 5 \n ,  \n 6 \n ,  \n 7 \n ,  \n 8 \n ,  \n 9 \n  another peak has been found at approximately 46–55 years of age. \n 3\nOften associated with fatigue and mental health conditions, CPP can be an incapacitating condition that significantly impacts quality of life and work efficiency. \n 4 \n ,  \n 10 \n ,  \n 11 \n  In a population‐based study in Denmark, \n 3 \n  half of participating women reported CPP involving at least moderate average pain intensity, and 90% of them reported that the pain interfered with their daily lives.\nDue to the often multifactorial etiology of CPP and its heterogeneous presentation, it is a challenging condition to manage. Overlap with other pain‐related conditions such as chronic migraine or tension‐type headaches, fibromyalgia and temporomandibular disorders is frequently reported. \n 12 \n  In a study among patients with CPP referred for group‐based physiotherapy, a substantial proportion of the patients reported other pain‐related conditions, including neck pain, low back pain and headache. \n 13 \n  Additional pain‐related conditions (eg bladder pain, abdominal and pelvic pain, menstrual pain and irritable bowel symptom) were found among more than 50% of women with vulvodynia who consulted a gynecological outpatient clinic, although only 26% had symptoms of fibromyalgia. \n 14 \n  The presence of overlapping pain‐related conditions may be explained by central sensitization and other shared biological pathways.\nCPP also impacts sexual function, eg in terms of reduced sexual desire and dyspareunia. \n 13 \n ,  \n 15 \n  Beyond that, cross‐sectional self‐reported data on exposure to different forms of sexual violence have revealed that such exposure increases the odds for CPP by 2–3. \n 16\nNevertheless, little is known about the symptom burden of CPP and more complex pelvic pain among different groups of women. However, a questionnaire, the Amsterdam Complex Pelvic Pain Symptom Scale (ACPPS), which covers many symptoms associated with complex pelvic pain, was recently validated. \n 17 \n ,  \n 18 \n  In our study, we therefore aimed to use the validated 16‐item Norwegian version of the ACPPS to describe and compare the symptom severity of complex pelvic pain in three cohorts of women in Norway. We also assessed associations between various demographic and gynecological characteristics and the severity of the condition.\n\nIn our descriptive cross‐sectional study, data were collected as part of a study seeking to validate the Norwegian version of the ACPPS. \n 17 \n  Three cohorts of women were included in the study: (1) patients referred to gynecological outpatient clinics (ie general gynecology, urogynecology or vulvar team) or to a pelvic floor physiotherapist at St. Olavs hospital, Trondheim University Hospital; (2) members of vulvodynia and endometriosis patient associations; and (3) healthy volunteers, ie female students in medicine, nursing or physiotherapy and female employees at the Norwegian University of Science and Technology. To be included in any of the three cohorts, individuals had to be at least 18 years old, not pregnant, able to consent to participate, and able to read, write and understand Norwegian.\nPatients referred to St. Olavs hospital were invited to answer the questionnaire in a short text message as part of the hospital's call‐in routine. They were also encouraged to participate in the study by hospital staff upon arriving for their appointments, in addition to being invited by posters and on‐screen calls for participants in the waiting area. Members of the patient associations were recruited via the social media groups and web pages for their members. Last, healthy volunteers were recruited via social media groups, web pages on the university's intranet, and email invitations. Respondents received a web link to the online questionnaire or downloaded it using a QR code, which gave them access to the questionnaire by logging in with a secure digital signature. The response rate was 42% among individuals who received a secure email or text message (ie mostly patients) and 62% among individuals who downloaded the link (ie mostly healthy volunteers and members of patient associations). The recruitment and distribution of participants in the three cohorts is depicted in a flow chart in Figure  1 . All data were collected between March 5 and May 3, 2022.\nFlow chart showing the recruitment and distribution of participants from three cohorts of women who completed an online questionnaire survey on symptoms of complex pelvic pain from March 5 to May 3, 2022. Q, questionnaire.\nFor sociodemographic data, we collected the participants’ self‐reported age in years, weight in kg and height in cm. Body mass index (BMI) was calculated and divided into four categories according to the World Health Organization's classification: underweight (ie BMI <18.5), normal weight (ie BMI 18.5–24.9), overweight (ie BMI 25.0–29.9) and obesity (ie BMI ≥30.0). The women's self‐reported mother tongue, country of origin, level of education, occupational status, and marital status were also collected. We divided level of education into primary (ie completed 9 years of compulsory school or less), secondary (ie completed high school or equivalent) and higher (ie completed a university degree program). Occupational status was divided into employed (ie full‐time or part‐time), student (ie in school or university) and unemployed (ie redundant, “laid off”, retired or disabled). Last, marital status was categorized as single or in a relationship, married or cohabitating.\nWe also collected self‐reported gynecological data about pregnancy history, menopausal status, current use of contraceptives, injury to the genital area including birth injury, pelvic surgery (eg gynecological surgery or radiation therapy), history of CPP (ie not pregnancy‐related and lasting at least 6 months), current pelvic pain and intensity of pelvic pain in the past 4 weeks (ie mild, moderate or severe). The criterion we used for vulvodynia was the presence of at least one of four symptoms (ie itching, burning sensation, sharp pain or pain when touching the vulva) for at least 3 months. \n 19 \n  Data about the women's general satisfaction with their sex life and whether they had experienced sexual assault as a child and/or an adult were also collected. In addition, referred patients were asked to give reasons for their referral.\nSelf‐reported symptoms of complex pelvic pain were assessed with the Norwegian version of the ACPPS \n 18 \n  and three other questionnaires. The Norwegian version of the ACPPS, with 16 items (the “ACPPS‐16”) is divided into five subdomains: (1) provoked vulvodynia symptoms (PVD), (2) abdominal pain and irritable bowel symptoms (IBS), (3) lower urinary tract symptoms (LUTS), (4) rectal symptoms and (5) physiological symptoms of general stress and/or tension (stress). \n 17 \n  Each item was answered on a 5‐point Likert scale ranging from 1 (never or almost never) to 5 (almost always or always). Total scores range from 5 to 25 points.\nAs for the three other questionnaires, the Female Sexual Function Index (FSFI) is a 19‐item self‐report questionnaire measuring sexual function in six subdomains: (1) desire, (2) arousal, (3) lubrication, (4) orgasm, (5) satisfaction and (6) pain. Taken together, the subdomains provide a total score ranging from 2 (ie poor sexual function) to 36 points (ie good sexual function), with a clinically significant cutoff score of 26.6, such that scores <26.6 indicate low sexual function. \n 20\nNext, the Fibromyalgia Survey Criteria 2016, concerning the disorder of fibromyalgia or widespread pain, consists of two parts: a Symptom Severity Score (SSS), indicating the severity of fibromyalgia and related symptoms, and the Widespread Pain Index (WPI), indicating the number of areas of the body (ie of 19 possible) with pain. SSS gives scores from 0 to 12, and the WPI gives scores from 0 to 19. \n 21 \n ,  \n 22\nLast, the Mental Health Inventory‐5 (MHI‐5) of the SF‐36 (Short Form Health Survey) is a questionnaire with five items used to assess mental health in terms of mood, anxiety symptoms and depressive symptoms, etc. Total scores range from 0 (ie poor mental health) to 100 (ie optimal mental health). \n 23\nDescriptive statistics were subjected to frequency analyses. Means with standard deviations (SD) and 95% confidence intervals (CI) were recorded for the total and subdomain scores on the ACPPS, as well as for other continuous variables. For categorical variables, frequencies and percentages were recorded. We administered Pearson's chi‐square test, Fisher's exact test and analysis of variance to compare demographic and gynecological variables between the three cohorts and in relation to the mean total scores on the ACPPS. Statistical significance was assumed when  P  < 0.01 following multiple comparisons.\nRegarding missing data on the ACPPS, in 18 cases when some answers for PVD were missing and in two cases when answers about rectal symptoms were missing, we replaced the missing values with the mean of the remaining items on the scale, as suggested by the developers. Even so, in 28 cases, information about PVD was missing for all three PVD‐related items due to the “No sexual activity” response option. Thus, the scores of those 28 cases were regarded as missing in the total ACPPS scores as well. Last, we conducted linear regression analysis to assess predictors of ACPPS scores in two multivariable models. The normality of data distribution was checked by visually inspecting frequency histograms, and the normality of residuals in the linear regression models was checked by visually inspecting Q–Q plots of the regression standardized residuals. \n 24 \n  Statistical analyses were performed with IBM SPSS Statistics for Windows version 28.0.1.0.\nOur study was approved by the Regional Committee for Medical and Health Research Ethics on June 21, 2021 (Ref. No. 245815, REK‐Midt); by St. Olavs hospital, Department of Obstetrics and Gynecology, R&D, on December 14, 2021 (Ref. No. 2021/14758), and by the Norwegian Center for Research Data on March 8, 2022 (Ref. No. 607016). We performed a Data Protection Impact Assessment and a risk assessment analysis on November 12, 2021. Permission to use the ACPPS was also granted by van Lunsen and van Laan, the instrument's developers, on May 4, 2020. The Mapi Research Trust permitted the use of a modified Norwegian version of the FSFI on November 12, 2021.\nWomen participating in our study received information about the study electronically and consented to participate by pushing the consent button on the online questionnaire. All women were informed that participation in the study was voluntary and that they could withdraw their data from the study at any time.\n\nOf the 397 women who participated in our study, 173 (44%) were patients referred to the outpatient clinics of St. Olavs hospital (ie “referred patients”), 70 (18%) were members of patient associations and 154 (39%) were healthy volunteers. Of the 173 referred patients, 168 (97%) gave reasons for their referral: 56 (32%) for pelvic pain; 30 (17%) for urogynecology; 24 (14%) for abnormal uterine bleeding; 23 (13%) for pelvic mass; 12 (7%) for vulvar problems; 11 (6%) for contraceptive counseling, screening or testing; seven (4%) for hormonal disturbances; and five (3%) for pelvic floor physiotherapy.\nThe distribution of demographic and gynecological characteristics in the three cohorts is shown in Table  1 . Mean age of participants was 38.4 years (SD = 14.4, range: 19–82) and BMI 25.8 kg/m 2  (SD = 5.7, range: 15–58). Referred patients were significantly older and had significantly higher BMI than members of the patient associations and healthy volunteers. Most women had higher education ( n  = 243, 62%) and were employed ( n  = 202, 54%); however, referred patients had relatively a low level of education and were often unemployed compared with the other groups. In all three groups, being in a relationship ( n =  297, 76%) was the most common.\nDemographic and gynecological characteristics of 397 women who completed an online questionnaire survey on symptoms of complex pelvic pain in spring 2022.\nAbbreviation: ACPPS, Amsterdam Complex Pelvic Pain Symptom Scale.\nMembers of the Vulvodynia or the Endometriosis Patient Associations.\nANOVA (analysis of variance) test.\nFisher's Exact test.\nPearson Chi‐Square test of heterogeneity.\nA total of 217 women (55%) had given birth, and 91 (23%) had reached menopause, with a higher proportion among referred patients (67% and 31%, respectively); and 231 (76%) of the 305 premenopausal women reported using contraception. Many women reported a history of CPP ( n =  140, 35%), with a significantly higher frequency among the members of the patient associations ( n  = 52, 74%) than among referred patients ( n  = 70, 41%) or healthy volunteers ( n  = 18, 12%). A similar distribution was found for vulvodynia symptoms. Most participants ( n =  252, 64%) reported general satisfaction with their sex life; however, among members of the patient associations, only 25 (36%) were satisfied. Another 120 women (31%) reported a history of sexual assault, which was rather evenly distributed among the three cohorts.\nMean total score on the ACPPS was 11.6 (SD = 3.7); the distribution of total and subdomain scores on the ACPPS in the three cohorts is shown in Table  2 . Members of the patient associations had a significantly higher total mean score (14.6, SD = 3.4) than referred patients (11.8, SD = 3.8) or healthy volunteers (10.0, SD = 2.8), and the same pattern emerged for all subdomain scores on the ACPPS as well.\nDistribution of scores among 397 women from three cohorts who completed an online questionnaire survey on symptoms of complex pelvic pain in spring 2022.\n1.8 ± 0.9 (1.7–1.9)\n2.5 ± 1.3 (2.2–2.8)\n<0.001\n2.5 ± 2.0 (2.0–2.9)\n<0.001\nAbbreviations: ACPPS, Amsterdam Complex Pelvic Pain Symptom Scale; FSFI, Female Sexual Function Index; IBS, irritable bowel syndrome; LUTS, lower urinary tract symptoms; MHI‐5, Mental Health Inventory‐5 of the SF‐36 (Short Form Health Survey); PVD, provoked vestibulodynia; SSS, Symptom Severity Score; UTI, urinary tract infection; WPI, Widespread Pain Index.\nMembers of the Vulvodynia or the Endometriosis Patient Associations.\nANOVA (analysis of variance) test.\nThe distribution of scores among the three cohorts on the other self‐report questionnaires is also shown in Table  2 . Members of the patient associations had significantly higher mean scores on the two fibromyalgia‐focused questionnaires compared with the other two cohorts. Healthy volunteers had the highest scores for sexual function (ie on the FSFI) and mental health (ie on the MHI‐5). Using the recommended cutoff score of 26.6 for sexual function (ie total FSFI score), \n 20 \n  indicating good sexual function, only 12 women (18%) from the patient associations, 39 referred patients (25%) and 63 healthy volunteers (43%) scored above the cutoff ( F  = 9.5,  df  = 2,  P  < 0.001, data not shown).\nTable  3  presents the mean total scores on the ACPPS among participants with different demographic and gynecological characteristics. Older and postmenopausal women scored significantly lower than younger and premenopausal women, and unemployed women scored significantly higher than employed women and students. Regarding current use of contraceptives, no significant differences surfaced in the mean ACPPS scores. However, a history of CPP and at least moderate pelvic pain during the last 4 weeks were both associated with significantly higher mean scores on the ACPPS than otherwise. The highest mean ACPPS score emerged among participants who reported three or four symptoms of vulvodynia (mean = 15.0, SD = 4.3). Participants with low sexual function or a history of sexual assault also had significantly higher mean scores than the ones who reported good sexual function or who had not been assaulted (see Table  3 ).\nDemographic and gynecological characteristics by Amsterdam Complex Pelvic Pain Symptom Scale (ACPPS) scores among 369 \n a \n  women who completed an online questionnaire survey on symptoms of complex pelvic pain in spring 2022.\nAbbreviation: ACPPS, Amsterdam Complex Pelvic Pain Symptom Scale.\nn =28 women had no sexual activity, i.e., missing information on the provoked vulvodynia (PVD) and stress subdomains, and hence excluded from the comparisons.\nAnova (Analysis of variance) test.\nAs shown in Table  4 , in the first multivariable linear regression model (Model 2), high sum scores on the two fibromyalgia‐focused questionnaires were the greatest predictors of high scores on the ACPPS, both with a standardized β‐value of 0.4 ( P  < 0.001). As shown in Table  4 , in Model 3 (that is, without the fibromyalgia‐related variables) having a low score for mental health (ie on the MHI‐5) and a history of sexual assault predicted higher ACPPS scores, with standardized β‐values of −0.4 and 0.2, respectively (both  P  < 0.001).\nPredictors of the Amsterdam Complex Pelvic Pain Symptom Scale (ACPPS) scores among 369 \n a \n  women who completed an online questionnaire survey on symptoms of complex pelvic pain in spring 2022. Three linear regression models are shown.\nAbbreviation: ACPPS, Amsterdam Complex Pelvic Pain Symptom Scale; BMI, Body mass index; FSFI, Female Sexual Function Index; SSS, Symptom Severity Score; WPI, Widespread Pain Index; MHI‐5, Mental Health Inventory‐5 of the SF‐36 (Short Form Health Survey).\nn  = 28 women had no sexual activity, ie missing information on the provoked vulvodynia (PVD) and stress subdomains, and hence excluded from the comparisons.\nModel summary: Adjusted  R  square 0.58,  F  change 43.3,  P  < 0.01.\nModel summary: Adjusted  R  square 0.32,  F  change 18.8,  P  < 0.01.\n\nIn our study, many women reported complex pelvic pain. As expected, especially high symptom burden was reported among women in the vulvodynia and endometriosis patient associations. Scores on the ACPPS were lower among older and postmenopausal women, and unemployed women scored higher than employed women and students. Especially high scores emerged among women with complaints of CPP, at least moderate pelvic pain intensity, and/or chronic vulvar pain. Last, women who reported low sexual function and/or a history of sexual assault also had high symptom burden according to the ACPPS. In multivariable regression, fibromyalgia, low mental health and a history of sexual assault were associated with high scores on the ACPPS.\nNo previous studies have used the validated Norwegian version of the ACPPS (ACPPS‐16), \n 17 \n  and few have used the 30‐item original patient‐reported Amsterdam Hyperactive Pelvic Floor Scale (AHPFS) for women. \n 18 \n  In one study, higher scores on the AHPFS emerged among women with post‐traumatic stress disorder \n 25 \n  and/or a history of sexual assault. \n 26 \n  An unpublished study validating the Dutch version of the questionnaire also revealed higher scores among women with vulvodynia and vaginismus than among healthy controls. \n 18 \n  In our study, members of endometriosis and vulvodynia patient associations had the highest mean total and subdomain scores compared with the other cohorts; that result can be expected because the questionnaire has been developed to target that group. For example, the mixed group of patients referred to St. Olavs hospital's general gynecology outpatient clinic or to a pelvic floor physiotherapist (ie “referred patients”) had a lower total mean score (ie 11.8) compared with women from the patient associations (ie 14.6). As shown by their given self‐reported reasons for referral, referred patients represented a broad range of gynecological disorders, not just conditions involving pelvic pain.\nSomewhat surprisingly, a significant proportion of the healthy volunteers reported a history of CPP (12%) and symptoms of vulvodynia (25%), and only approximately 40% reported good sexual function. The volunteers’ mean total ACPPS score of 10 exceeded the score found in the Dutch validation study, \n 18 \n  in which healthy volunteers had a mean score of only 8 points even on the 30‐item scale; however, comparison with the original study should be undertaken with caution due to different score ranges. The different scores could also be due to differences in the studies’ populations. Our so‐called healthy volunteers were recruited from a university environment of students and employees and were not required to be without symptoms. Moreover, 25% of our volunteers reported a history of sexual assault. Although high, that proportion aligns with recently published findings about the prevalence of self‐reported rape (22%) from a phone interview survey of 2100 women 18–74 years old in the general Norwegian population. \n 27\nSimilar to a previous study in the Netherlands, \n 26 \n  we also found a Significantly higher mean ACPPS score (ie 12.5–14.1) among women who reported a history of sexual assault than among the ones who did not (ie 10.8). Retrospective studies have shown that traumatic experiences (eg history of sexual trauma) are associated with CPP, \n 25 \n ,  \n 28 \n  which may result from chronic mental stress and, in that case, act as an unconscious defense mechanism. A growing body of evidence also shows that exposure to adverse childhood events increases the risk of developing a state of central sensitization. \n 29 \n  However, determining whether the fear‐avoidance model or central sensitization mechanism is most applicable requires further research. \n 28 \n  We found that fibromyalgia symptoms and low mental health were associated with high ACPPS scores, indicating that those factors may also interact in patients reporting complex pelvic pain. The dynamic, multidimensional biopsychosocial model describes how physiological, psychological and social factors influence each other and result in chronic and complex pain syndromes. \n 30 \n  That approach is well established when managing chronic widespread pain, following insights that organ‐specific explanations should be replaced with a more integrated model of pain. \n 31 \n  Results of exploring whether a common pathway exists for chronic widespread pain and CPP could suggest the increased use of the biopsychosocial model in treating patients with CPP as well. Although patients with CPP often fall between different medical specialties (eg gynecology and urology), they should receive multidisciplinary, multimodal treatment, at least when other medical or surgical treatments have failed to alleviate their complaints.\nIncreased tone in the pelvic floor muscles is frequently hypothesized to be involved in the pathophysiology of CPP. A recent systematic review and meta‐analysis showed that women with CPP had higher tone in their pelvic floor muscles than women with no pain. \n 32 \n  Although vulvodynia has also been associated with objective pelvic floor hyperactivity, \n 33 \n ,  \n 34 \n  the true association between the sign “hypertonicity” and the subjective patient‐reported outcomes of pelvic pain and female sexual dysfunction remains unclear; the International Urogynecological Association and the International Continence Society working group has recommended using the term hypertonicity to refer strictly to a sign or clinical finding of the altered structure or function of the pelvic floor muscles. \n 35 \n ,  \n 36 \n  Accordingly, hypertonic pelvic floor should be regarded as a condition in which the pelvic floor muscles have signs of elevated contractile activity and/or an increased stiffness that will resist passive movements of the muscles by, for instance, a physiotherapist or a gynecologist during a digital palpation. More objective measures to assess hypertonicity may include ultrasound measurements of the anterior–posterior diameter of the levator hiatus, manometry or myoelectrical activity. \n 32\nAs per that updated definition, the use of “hyperactive pelvic floor” in the original Dutch questionnaire's title is somewhat misleading. Our research group used the ACPPS as a proxy for symptoms on complex pelvic pain in a recent validation study and chose to rename it to the more appropriate “Amsterdam Complex Pelvic Pain Symptom Scale”. \n 17 \n  The association between objective signs of hyperactivity in the pelvic floor muscles in physical examinations and symptom scores on the ACPPS has never been assessed, or at least not published. Future research should therefore focus on assessing the association between scores on the validated questionnaire and the results of clinical examinations of the pelvic floor muscles. Beyond that, a new clinical cutoff score should be established. Even so, we support the validated 16‐item ACPPS questionnaire as a screening tool for symptom burden in assessing, for example, gynecological patients with CPP and pelvic floor complaints. However, as in any clinical setting, symptom scores on a questionnaire should be supplemented by taking patients’ medical history and clinically examining them. Thus, in consulting patients presenting with symptoms of complex pelvic pain, gynecologists and pelvic floor physiotherapists should assess the pelvic floor muscles by at least performing a digital palpation.\nOur study has many strengths. It was the first to use the validated 16‐item version of the ACPPS among Norwegian‐speaking women, which has contributed to the study of complex pelvic pain. We used other relevant, validated questionnaires with good psychometric properties, had an adequate sample size of both volunteers and symptomatic women, and there was little missing information on the individual questionnaires, which allowed for sufficient analyses and statistical comparisons. Nevertheless, some sub‐analyses could have involved type II errors. We did not perform a separate power calculation for our descriptive, comparative study.\nAmong the study's limitations, self‐report data are subjected to exaggeration and under‐reporting, especially about intimate health‐related issues involving sexual function and/or sexual assault, as well as to recall bias. Although few data were missing, probably due to the electronic distribution of the questionnaire, approximately 9% of data about the PVD factor of the ACPPS and about the FSFI were missing because the participants, mostly referred patients, did not report sexual activity. Moreover, as mentioned, our study was subject to selection bias, as most of the women were relatively young, highly educated, and employed or students. Older adult women without internet access could face difficulties in answering an online survey such as ours and would therefore become excluded from participation. Moreover, volunteers recruited to participate may have been members of the general population with the most severe pelvic pain symptoms, which would have increased their mean score on the ACPPS. Thus, our study does not represent the general population of women in Norway, and further research should be done with a more heterogeneous sample. In addition, the ACPPS does not indicate a period for the complaints that it asks about, meaning that our participants could report symptoms both at present and at any time in their lives. Finally, our study's population included women only, because the questionnaire was developed to assess complex pelvic pain specifically among women. However, future studies should aim to assess similar pelvic pain in men as well.\n\nAlthough complex pelvic pain was reported by many women across the cohorts, multivariable regression revealed that fibromyalgia, low mental health and a history of sexual assault were associated with high symptom scores on the ACPPS, indicating high symptom burden from complex pelvic pain. Those findings support taking a biopsychosocial approach in treating women who present with such complaints.\n\nDesign and conception of the study: CTH, SNS, and SSaga. Data collection: CTH, IB, VD, SSpetalen, SSaga and SNS. Data curation and analysis: CTH, SSpetalen, SSaga, SNS, IB and VD. Interpretation of results: CTH, SSpetalen, SNS, SSaga, IB and VD. Drafting the paper: CTH, SSpetalen, SNS, SSaga, IB and VD. Revising the manuscript critically and approving the final version: CTH, SSaga, SNS, SSpetalen, IB and VD.\n\nThe authors have no conflicts of interest to declare.","source_license":"CC0","license_restricted":false}