Abstract
Background:
The aim is to measure the association between fibromyalgia syndrome (FMS) and post-traumatic stress disorder (PTSD), mood and anxiety disorders using reliable psychiatric diagnoses according to Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV) and with a case–control design.
Methods:
Case–control study with cases (71 consecutive female patients with FMS) and controls (284 subjects without FMS), randomly drawn after a gender- and age-matching technique from the database of an epidemiological survey. Psychiatric diagnoses were conducted according to DSM-IV and carried out by clinical staff using a structured interview (Advanced Neuropsychiatric Tools and Assessment Schedule). QoL was measured by Short Form Health Survey (SF-12).
Results:
The lifetime prevalence of major depressive disorder (MDD; 43.7% vs 8.1%, p < .0001), bipolar disorder (BD; 21.1% vs 0.7%, p < .0001), PTSD (8.4% vs 1.4%, p < .0001) and panic disorder (28.2% vs 5.6%, p < .001) was higher in people with FMS than in controls. People with FMS showed a poorer QoL than controls on the SF-12 (26.43 ± 6.04 vs 37.45 ± 5.80, p < .0001). Those with comorbidity with MDD and BD showed a mean SF-12 score of 24.75 ± 6.31 versus 29.52 ± 4.84 (N = 25) of people with FMS without any mood disorder (p = .002). The attributable burden of FMS in worsening QoL was found comparable to that of serious chronic diseases such as multiple sclerosis.
Conclusion:
FMS is a disorder that ‘in itself’ can have a devastating impact on an individual’s life. The frequency of the association with major depressive and bipolar disorders increases the impact on the QoL of people with FMS. One of the causes of this association appears to be the extreme vulnerability to chronic stress that this disorder involves. The findings have important clinical significance: the physician must interpret in the right dimension and with dignity the suffering of the people with FMS.
Keywords
Background
Fibromyalgia syndrome (FMS) is becoming a relevant public health issue as it affects 0.4%–8.8% of the general population worldwide, with a global mean prevalence of 2.7% (3.1% in the Americas, 2.5% in Europe and 1.7% in Asia; Queiroz, 2013); in addition, many studies have pointed out that this disease greatly compromises the quality of life (QoL) of those who live with it (Hoffman & Dukes, 2007; Sancassiani et al., 2017).
Several reviews (Arnold et al., 2006; Fietta, Fietta, & Manganelli, 2007; Khalid, Simonds, Loukas, & Tubbs, 2018; Palagini et al., 2016) and observational studies (Dell’Osso et al., 2009; Dell’Osso et al., 2011) reported high prevalence rates for mood and post-traumatic stress disorders (PTSDs) in patients with FMS, but with important differences between studies concerning the strength of the association. These differences are not so large in major depressive disorder (MDD), for example, the association in comparison with healthy subjects varies from odds ratio (OR) of 2.92 (95% confidence limits, 2.52–3.33) in a representative sample of the Canadian general population with diagnoses conducted with screening tools (Fuller-Thomson, Nimigon-Young, & Brennenstuhl, 2012) to OR 2.7 (95% confidence interval (CI) = 1.2–6.0, p = .013) in a clinical sample of people with fibromyalgia compared to healthy controls in which the diagnoses were carried out by clinicians according to Diagnostic and Statistical Manual of Mental Disorders (4th ed.; DSM-IV; Arnold et al., 2006). But they were really greater in bipolar disorders (BDs): in clinical samples of patients with fibromyalgia, the frequencies of BD are less than 20% in all the studies conducted with structured interviews or psychiatric evaluation (Aaron et al., 1996; Arnold et al., 2006; Dell’Osso et al., 2009; Hudson, Goldenberg, Pope, Keck, & Schlesinger, 1992; Hudson, Hudson, Pliner, Goldenberg, & Pope, 1985; Wallace & Gotto, 2008) and more than 25% in those carried out with screening tools for BDs (Carta et al., 2006; Wilke, Gota, & Muzina, 2010).
In the case of PTSDs and traumas, the frequencies in clinical samples of patients with fibromyalgia are very high (e.g. 37% in the study of Nelson et al., 2017), but studies were carried out using only rating scales for symptoms rather than reliable diagnoses according to international classification systems (Galek et al., 2013; Raphael, Janal, Nayak, Schwartz, & Gallagher, 2006).
It is necessary to point out that previous studies on the association between fibromyalgia and psychiatric disorders did not use semi-structured interviews conducted by psychiatrists which is the most correct methodological approach (Carta & Angst, 2016). In fact, the aforementioned studies were based on screening questionnaires or adopted highly structured interviews, or they used clinical diagnoses without interviews as diagnostic psychiatric tools.
The issue of comorbidity with psychiatric disorders is of major relevance for the therapeutic approach because some antidepressant drugs are commonly used in FMS and have entered therapeutic guidelines for the treatment of FMS (Carta et al., 2013). If the lifetime high risk of BD in FMS is confirmed by a robust methodology, this will have implications for treatment because using an antidepressant without a mood stabilizer in BD the risk of switching to mania is high (Carta et al., 2013).
As concerns PTSD, it would also be advisable to verify whether the association previously found through rating scales can be confirmed by surveys conducted with structured psychiatric interviews.
It is also to be understood that any increased risk of PTSD in people with FMS might be attributed to a higher frequency of mood disorders that are known to be associated with PTSD (Carta et al., 2006) or are rather due to FMS independently of the association with mood disorders. To clarify whether there is an association between PTSD and FMS independent of FMS risk and mood disorders, useful elements can in fact be provided by developing and verifying etiological hypotheses dealing with the questions concerning whether PTSD risk is higher in a subgroup of patients having BD and who are thus exposed to a higher risk of traumatic experiences; or if PTSD may lead to FMS because of some neurobiological mechanisms common to mood disorder or BD. Finally, it can tell if PTSD and FMS are linked by a shared relationship with sensitivity to stress via emotional reaction.
Another important aspect for the clinician is to understand to what extent the impairment of QoL (Mantovani et al., 1996) in FMS may be secondary to a psychiatric disorder. People suffering from FMS often complain that their disorders are interpreted by physicians as ‘psychogenic’ ones, where this term is considered as reductive by the patient as it seems to identify a disorder of low gravity, supposing the term ‘psychogenic’ opposed to ‘organic’ neglects an ‘objective’ origin out of ‘weaknesses or subjective interpretations’ (Sancassiani et al., 2017). For these reasons, the diagnosis of a psychiatric disorder found in FMS often leads to refusal by the affected person as it could be interpreted as a kind of certification of the ‘psychogenicity’ of the disease (Sancassiani et al., 2017). Thus, the maximum scientific clarity has to emphasize these aspects.
The objectives of this study are to measure the association of anxiety and PTSD and mood disorders in FMS by comparing those with the disease with healthy people; to establish, with the same methodology, to what extent FMS compromises the QoL of the affected persons; and to define the attributable burden due to comorbidity with psychiatric disorders in worsening the QoL.
Methods
Design
This is a case–control study.
Sample
The cases are 71 female patients with FMS, consecutively attending (from March 2015 to July 2015) the Rheumatology Outpatient Service of the University Hospital of Cagliari (‘Azienda Ospedaliero Universitaria di Cagliari’). Controls include 284 women without FMS, who were randomly drawn from a database of an epidemiological survey on health in Italy, but also constructed with the objective of being a depository for providing controls in case–control studies of this type (Carta et al., 2015). The choice of sex-, age- and residence-matched controls was performed by means of a randomized block design. A total of 250 people in the database who declared in the epidemiology survey that they suffered from FMS were primarily excluded from the depository to generate possible controls. For each case, a block including all eligible age- (±1 year), gender- and residence- (Italian region of residence) matched controls in the database was created. Four individuals per block were drawn for each case, thus excluding them automatically from the remaining blocks. Age and gender were homogeneous in cases and controls thanks to the methodology of matching.
Demographic data, psychiatric diagnosis interview and QoL assessment
The same standard form for recording basic demographic data was used in cases and controls. The psychiatric diagnoses (lifetime) both in cases and controls were carried out by means of the Advanced Neuropsychiatric Tools and Assessment Schedule (ANTAS). ANTAS is a semi-structured clinical interview administered by clinical staff (psychiatrists, psychologists, nurses and occupational therapists) after a training course (Carta et al., 2010). This tool is partially inspired by the non-patient version of the DSM-IV Structured Clinical Interview (SCIDI/NP-DSM-IV) (First et al., 2002). Like SCIDI/NP, ANTAS can produce psychiatric diagnoses according to the DSM-IV system by means of an algorithm. The diagnosis derived from SCIDI/NP and ANTAS showed good reliability in a study; the results of which were published previously (Carta et al., 2010).
The perception of QoL in cases and controls was measured by means of the Short Form Health Survey (SF-12; Ware, Kosinski, & Keller, 1996). The SF-12 evaluated the following seven dimensions: physical activity, limitations on activities due to health conditions, emotional life, pain, general health, vitality, social network and mental health. The period of reference for the measure was the month prior to evaluation. The higher the score of SF-12, the better the QoL.
FMS diagnosis in cases
All 71 patients with FMS were examined by an expert rheumatologist and met the criteria for fibromyalgia (FM) diagnosis of the American College of Rheumatology (Wolfe et al., 2016).
A patient meets the criteria for diagnosis when the following three conditions are met:
Widespread Pain Index (WPI) ⩾ 7 and score to Symptom Severity Scale (SS) ⩾ 9.
Symptoms have been present at a constant intensity for at least 3 months.
The patient does not have a disorder that would otherwise explain the pain.
FMS screening in controls
During the epidemiological survey on controls, subjects were asked about their well-being, the presence of current or past illness, the use of health facilities and consultation with physicians. Available medical tests and documents were requested whether they were conducted routinely (e.g. by periodic controls at work/school or for driver’s licence eligibility) or for health issues. Diagnoses of ascertained illness as FMS were reported using a structured form.
Data analysis
Lifetime prevalence of MDD, BD, including Bipolar I and Bipolar II, PTSD, panic disorder (PD) and generalized anxiety disorder (GAD) according to the DSM-IV system was calculated in the cases and control groups. The OR for DSM-IV diagnosis (dependent variable) was calculated using the control group as ‘pivot’ with univariate analysis (owing to the matching method, the groups were balanced by age and gender).
The analysis of variance (ANOVA one-way) was used for the comparison of parametric variables; the χ2 test was used for nonparametric variables. The OR and 95% CI were calculated adopting Miettinen’s simplified method. The ‘Attributable Burden’ of FMS in worsening the QoL was measured by subtracting the score at the SF-12 of the sample with FMS from the mean score achieved at the SF-12 of a balanced sample by age and gender without FMS (Carta et al., 2015; Carta et al., 2014).
Ethical aspects
People recruited in case and control groups gave their informed consent for the use of anonymous data. The ethics committee of the Italian National Health Institute (Rome) approved the epidemiological survey from which data bank controls had been drawn, and the approved project planned the conduction of a series of case–control studies using the data bank of the study. The ethics committee of the Azienda Ospedaliero Universitaria di Cagliari approved the study presented in this article. Subject data have been de-identified; a coded number identifies each subject.
Results
The study sample consisted of 71 female outpatients with FMS. The control group consisted of 284 females without a diagnosis of FMS; thanks to the matching method, controls were perfectly homogeneous with cases by age. The demographic characteristics of cases and controls are summarized in Table 1.
Demographic characteristics of cases and controls.
FMS: fibromyalgia syndrome; SD: standard deviation.
Table 2 shows the lifetime prevalence of psychiatric disorders (PTSD, BD, MDD, PD and GAD) in cases and controls. All psychiatric disorders evaluated in this study (PTSD, MDD, BD, PD), except for GAD, showed in cases a statistically significant association with FMS. MDD shows the higher frequency in cases (43.66% vs 8.07% in controls); BD shows the higher OR for cases (37.90; 95% CI, 7.95–247.28). Among the cases, the six FMS patients with PTSD all had comorbidity with a mood disorder (three MDDs and three BDs), while among the four controls with PTSD, comorbidity was present in only one subject (one MDD). This difference is statistically significant (Fisher exact test p = .033).
Lifetime prevalence of psychiatric disorders in cases and controls.
CI: confidence interval; GAD: generalized anxiety disorder; MDD: major depressive disorder; OR: odds ratio; PTSD: post-traumatic stress disorder.
The mean ± standard deviation of SF-12 scores in cases was 26.43 ± 6.04 against 37.45 ± 5.80 in controls (ANOVA 1-way df = 1, 158, 159; F = 62.245; p < .0001); people with FM and diagnosis of DSM-IV mood disorder (N = 46, 15 BD; 31 MDD) showed a mean SF-12 score of 24.75 ± 6.31 versus 29.52 ± 4.84 (N = 25) of people with FMS without mood disorder (ANOVA 1-way, df = 1, 69, 70, F = 10.80, p = .002). Both mood disorders and PTSD appear to be factors that significantly increase the QoL impairment in FMS (Table 3).
Impairment of quality of life in FMS, the burden of comorbidity: comparison of impaired quality of life among people with FM without psychiatric comorbidity versus controls and people with FMS and comorbidity (PTSD, MDD and BD).
BD: bipolar disorder; FMS: fibromyalgia syndrome; MDD: major depressive disorder; PTSD: post-traumatic stress disorder; SF-12: Short Form Health Survey.
There can be no comorbidity between MDD and BD; all those who have PTSD diagnosis have either MDD or BD.
Table 4 shows the ‘Attributable Burden’ (and due to FMS without mood disorders) in worsening QoL and the comparison versus the ‘Attributable Burden’ of other conditions on the basis of the same database for psychiatric diseases (MDDs, PDs and eating disorders) or on the basis of case–control studies where cases were consecutive patients selected at the same University Hospital of Cagliari, and control groups were extracted with a similar matching technique after stratification by sex and age from the same database (Carta et al., 2015; Carta et al., 2014). FMS leads to a serious impairment of QoL whose burden is comparable to that of severe chronic conditions such as multiple sclerosis and MDD; it is higher than the burden attributable to the other conditions considered (such as eating disorders and PDs). If FMS is associated with a mood disorder (which occurs in 65% of cases), the disorder results in impairment of the perception of the QoL even higher than that of multiple sclerosis.
Attributable burden in worsening quality of life to FMS on comparison with other chronic diseases.
FMS: fibromyalgia syndrome; SF-12: Short Form Health Survey; SD: standard deviation; PTSD: post-traumatic stress disorder.
Discussion
This study has confirmed that FMS is closely associated with several psychiatric disorders and specifically mood disorders: around 65% of people with FMS have a lifetime diagnosis of BD (21.12%) or MDD (43.66%).
These results were reached by adopting a case–control design and psychiatric diagnosis carried out by psychiatrists according to DSM-IV by means of psychiatric semi-structured interviews. With this methodology, the frequencies of MDD found are higher and almost double those of previous clinical studies carried out with simple clinical evaluation without interviews (Arnold et al., 2006; Marangell et al., 2011) and also superior to the studies adopting screening tools and community samples (Fuller-Thomson et al., 2012). In this case, however, it should be considered that a bias determined by the fact that the screening tools provide the prevalence of current depressive episodes, but the diagnosis of MDD is a long-term condition with a lifetime prevalence about three times higher than the point prevalence (Weissman et al., 1996). Fibromyalgia is a chronic condition several years lasting, so these studies probably not only lower rates of depression but also underestimate the association between fibromyalgia and MDD. On the other hand, ‘… specific life events predict increases in particular depression symptoms, and there is evidence for direct causal links among symptoms …’ (Fried & Nesse, 2015). Thus, the authors suggest ‘that the pervasive use of sum-scores to estimate depression severity has obfuscated crucial insights and contributed to the lack of progress in key research areas such as identifying biomarkers …’ (Fried & Nesse, 2015).
The frequency of BD found in our sample (lifetime prevalence 21% BD in people with fibromyalgia) was on average higher than those in studies adopting a diagnostic approach based on rigidly structured interviews with yes/no answers, such as those of Aaron et al. (1996; lifetime prevalence 18%), Wallace and Gotto (2008; lifetime prevalence 10%) and Dell’Osso et al. (2009; 3.5%). However, the approach with structured tools does not appear to guarantee good diagnostic accuracy of BDs because of the difficulty of identifying previous episodes of hypomania with such a methodology (Carta & Angst, 2016).
In contrast, the frequencies in patients with fibromyalgia in our sample are lower than those found by adopting screening tools such as the Mood Disorder Questionnaire (frequency 59%, Carta et al., 2006) or Hypomania Checklist 32 (frequency 86%, Alciati et al., 2013), but it is known that screening tools identify an area of ‘spectrum’ which is not superimposable on the cases identified with a clinical diagnosis of BD (Carta & Angst, 2016), with respect to which the screening tools for BDs cause numerous false positives (Zimmerman, Galione, Chelminski, Young, & Dalrymple, 2011). Although our results are intermediate in the gap produced by the studies that adopted different methodologies, at present the approach we follow appears to be the most reliable one.
The results on BDs strongly support the conclusions of a first survey of our group that showed an association of positivity at a screening test for BDs in people living with FMS (Carta et al., 2006) and those of a case series survey on patients with FMS in which comorbidity with bipolar spectrum disorders was measured using both a categorical DSM-IV and a dimensional approach (Alciati et al., 2013).
As some antidepressants are indicated in the treatment of FMS, the concept that patients with FMS should be systematically evaluated clinically to exclude the risk of a BD prior to treatment with an antidepressant appears to be of extreme importance (Wilke et al., 2010).
This study found, for the first time with a case–control methodology and the use of clinical psychiatric diagnosis carried out by means of semi-structured interviews, a close association between FM and a lifetime diagnosis of PTSD.
Interestingly, in people living with FMS the association between PTSD and mood disorder was stronger than the association between PTSD and mood disorder in controls without FMS, so the increases in PTSD frequency in FMS cannot be considered as a mere consequence of the increases in mood disorders, which, as is well known, are associated with FMS (Barlow, 2000).
The association between FMS and PTSD in this case–control study, though strong, was found to be much lower than in studies in which the diagnoses of PTSD were conducted by means of questionnaires or by screeners on stress symptoms. For example, in a survey (Galek et al., 2013) on 398 consecutive FMS patients of eight different clinical sites (rheumatology; orthopaedic surgery, psychosomatic and pain medicine, physical and integrative medicine), it was found that 65.7% of FMS patients met the criteria of a potential depressive disorder, 67.9% of a potential anxiety disorder and 45.5% of a potential PTSD. Similar high rates of PTSD symptoms were found in other two studies, in which PTSD was diagnosed by means of ‘well-validated self-report questionnaires’ (Coppens et al., 2017), and PTSD symptoms were detected by means of the Trauma and Loss Spectrum Self-Report (TALS-SR) lifetime version (Dell’Osso et al., 2011).
The profile of a close association between FMS and mood disorders and of a not-so-high risk association with PTSD does not contradict the results of a longitudinal study that assessed the vulnerability and long-term influence of traumatic stress on patients with FMS caused by the Great East Japan Disaster, which occurred in March 2011 (Usui et al., 2013). The survey found that the mean score at the Event Scale–Revised (ES-R) 1 month after the disaster in an FMS patient group was significantly higher than in a rheumatoid arthritis control group, but 7 months after the disaster the score was similar to a comparable FMS control group without exposure to the disaster. The risk of depressive symptoms was higher in the FMS-exposed group versus controls at 1 year than at 1 month after the disaster. The authors interpreted this worsening over time of depression-related symptoms as possibly being in response to chronic stress induced by the fear of radiation due to the nuclear power disaster. Thus, in line with their findings, the authors suggested that patients with FMS were vulnerable to chronic stress rather than acute stress. This specific kind of chronic stress sensitivity may also explain the strong association with depressive and bipolar disorders; the risk of mood disorders may in fact be the result of the specular increases in chronic stress due to the lowering of the subjective threshold. If you consider that the effect of lowering the ability to cope with chronic stress can also be a consequence of a mood disorder (other than a characteristic of FM), this may suggest a possible bi-univocal sense of causation of the association. This theory is thus not in contradiction, but may be complementary, to the evidence of family risk factors common to FMS and mood disorders (Raphael et al., 2006). If depression disorders are supposed to increase vulnerability to FMS because they increase sensitivity to stress, then it is obvious that there will be a cross-linked familiarity between FM and depression. The complexity of the interaction of the two different causality sources can be investigated only in prospective studies.
The results of the Japanese study may also suggest a lower risk of PTSD, a disorder that is a consequence of acute rather than chronic stress. However, this interpretation appears simplistic owing to the close link between acute and chronic stress; for example, having experienced an acute stress and having PTSD as a consequence is a well-known condition of the future vulnerability to chronic stress. This was well highlighted in the cohorts of Israeli soldiers with post-war PTSD and Major Depression (Ginzburg & Solomon, 2010). It can thus be hypnotized two different possible pathogenesis of the association PTSD–FMS and mood disorders–FMS: the first due to sensitivity to chronic stress (and therefore to the risk of FMS) due to PTSD; the second to both of sensitivity to chronic stress (and therefore to the risk of FMS) of people with mood disorders and, in addition, to vulnerability to chronic stress (and hence the risk of mood disorders) in people with FMS.
From this point of view, our observational study is also a source of hypotheses that only ad hoc studies with appropriate longitudinal designs can verify. In fact, future research must clarify how both physical (Beggs, 2015; Schwaller & Fitzgerald, 2014) and psychological traumas in early life are associated with chronic pain conditions (i.e. FM) in adult life (Afari et al., 2014; Altamura, Carta, Tacchini, Musazzi, & Pioli, 1998; Burke, Finn, Mcguire, & Roche, 2017). Besides key neurobiological substrates, including the hypothalamic–pituitary–adrenal axis, the neurosteroid, monoaminergic, opioidergic, endocannabinoid systems, the immune and auto-immune response and some epigenetic mechanisms, play an important role in this association (Burke et al., 2017).
Having a diagnosis of FMS seriously worsened the perception of QoL with a severity comparable to that of people who live with severe chronic illnesses such as multiple sclerosis. This strong impact is present also in people who suffer from FMS without comorbidity with mood disorders, so this is a consequence of FMS but in that case of comorbidity the impairment of the QoL was found even stronger than in people with multiple sclerosis. These findings have a clinical significance: the physician must interpret in the right dimension and give the right dignity of the suffering of people who live with FMS. The results show that the disorder cannot be regarded as an ‘imaginary’ illness and must be treated in its great complexity.
An evident limitation of this survey is that while the cases had received a diagnosis of FMS in a clinical setting according to standard criteria, in the control sample the diagnoses were based only on medical history and previous investigations. In this framework, a number of people who screened negative for FMS may actually be undiagnosed cases of FMS. But this limit does not detract from the relevance of these findings because, if this hypothesis is true, it will only decrease the measure of association between cases and controls: the limit is that the association might be even stronger than was found.
Conclusion
FMS is a disorder that ‘in itself’ can have a devastating impact on an individual’s life. The frequency of its association with MDD and BD increases the impact of QoL of people living with FMS.
The findings of this study have important clinical significance: the physician must interpret in the right dimension and give the right dignity of the suffering of the people who live with FMS. The results show that the disorder must be dealt with in its overall complexity.
Footnotes
Acknowledgements
We are grateful to the patients of the Rheumatology Outpatient Service of the University Hospital of Cagliari (‘Azienda Ospedaliero Universitaria di Cagliari’) for their kind participation in the study. M.G.C., M.F.M., F.S., E.C., F.R., V.R. and A.E.N. participated in the design and coordination of the study, in the acquisition and analysis of the data and drafted the manuscript. L.M., S.M. and R.C.F. participated in the analysis of the data and drafted the manuscript. F.L.P., G.T. and M.P. collaborated in the design of the study, in the acquisition and analysis of the data and drafted the manuscript. All authors read and approved the final manuscript.
Availability of data and materials
The database is kept at the Chair of Quality of Care, Faculty of Medicine and Surgery, University of Cagliari, Italy. Access is available by arrangement with the Principal Investigator and under authorization by the Ethics Committee of the University Hospital of Cagliari.
Conflict of interest
The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: M.G.C. has received grants from the European Commission, European Social Fund, AIFA (Agenzia Italiana del Farmaco), Fondazione Banco di Sardegna and the Sardinia Region. M.G.C. is advisor for the Economic and Social Committee of the European Union. M.F.M. received grants from Fondazione Banco di Sardegna. L.M. received a Grant from the Italian Ministry of Universities and Research (MURST). A.E.N. has received research grants from Brazilian Council for Scientific Development (CNPq) and Fundação de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ).
Ethics approval and consent to participate
The study was conducted in accordance with the ethical principles contained in the Helsinki Declaration. The ethics committee of the Italian National Health Institute (Rome) approved the epidemiological survey from which data bank controls had been drawn, the approved project planned the conduction of a series of case–control studies using the data bank of the study. The ethics committee of the Azienda Mista Ospedaliero Universitaria di Cagliari approves the study presented in this paper. The participants gave informed written consent to participate and to subsequent publication of the results. Subject data have been de-identified; a coded number identifies each subject.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The research was funded by the University of Cagliari. The fundings were provided in the framework of ‘current research’, that is, research carried out within the framework of its institutional activities in the Cagliari University by Professor Carta. All participants carried out the different phases of survey, that is, collection, analysis, and interpretation of data and writing the manuscript in the framework of their institutional activities and no additional funds were required for staff salaries. A call for ‘Visiting Professor’ by the University of Cagliari was won by E.D., in his visiting period at the University of Cagliari, the last part of the research was jointly carried out.
