Risk factors for the development of systemic sclerosis: a systematic review of the literature

Abstract Objectives Although numerous studies have investigated the roles of various genetic, epigenetic and environmental factors that may impact its aetiology, SSc is still regarded as an idiopathic disease. Given that there is significant heterogeneity in what has been proposed to influence the development of SSc, this systematic review was conducted to assess the impacts of different factors on the aetiology of scleroderma. Methods The search was performed in the PubMed, CINAHL and SCOPUS databases on 17 May 2017. Any study that made explicit reference to scleroderma or SSc that had information about the risk factors or epidemiology of the disease was included. The extracted outcome variables were prevalence, gender preponderance, geographical distribution, family history and various proposed environmental risk factors. Results One thousand five hundred and seventy-four articles were screened for eligibility. Thirty-four articles were eligible for the systematic literature review. Conclusion Age between 45 and 64 years, female sex, positive family history and exposure to silica were found to be risk factors. There were conflicting findings regarding the impact of exposure to organic solvents and microchimerism. No relationship between infectious agents, alcohol consumption or cigarette smoking and the development of SSc was identified.

environmental exposures that predispose to SSc in certain parts of the world [9]. Studies agree on a 7:1 female preponderance, and the onset in most patients is between the ages of 30 and 50 years, which suggests that hormonal factors, pregnancy and age-related influences might play a role [9][10][11][12][13]. Paradoxically, there is a 13-to 19-fold increase in the risk of developing SSc if one has a sibling with the disease, although SSc very rarely runs in families [10][11][12]. Nonetheless, we still know very little about the underlying causes of SSc and its pathogenesis. The aim of this systematic review is to synthesize what has been discovered regarding the risk factors for the development of SSc, in an attempt to enhance our understanding of the causes of this debilitating disease.

Methods
This study was carried out according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines and is registered at the International Prospective Register of Systematic Reviews [14,15]. The registration number for this review on the International Prospective Register of Systematic Reviews website is CRD42017067304, and a published protocol for this study can be found there.

Eligibility criteria
In order to be included in this systematic review, the articles had to be studies written in English that involved humans and made explicit reference to scleroderma or SSc. There was no restriction on publication date or study type. Articles were excluded if they used the same data published in another study, in order to reduce the effect of publication bias. The studies had to examine adult-onset scleroderma, rather than juvenile scleroderma (which is a distinct entity from SSc and may have a different aetiology). Articles were excluded if they focused on the proposed effects of silicone breast implants on SSc aetiology, because many studies have found no relationship [4,6,7,13,[16][17][18].

Search strategy
The systematic search was done on 17 May 2017, using the following databases: PubMed, CINAHL and SCOPUS. The search strategy for PubMed was constructed by the first author, and was as follows: (((risk factors OR risk factor OR risk OR probability OR epidemiolog* OR aetiolog* OR etiolog*)) AND (SSc OR scleroderma)) AND (develop*).
The search strategy was modified to fit CINAHL and SCOPUS. Duplicates among the databases were excluded. The titles and abstracts of the articles were then screened independently by two authors (S.A. and A.H.-H.). If the title and/or abstract met the eligibility criteria, full texts of the articles were obtained. Any discrepancies in the selection process between the two authors were resolved by discussion. Three additional studies were identified from reference lists of the included articles [9,10,19].

Data collection and data items
The articles included in the systematic review were analysed for risk factors that have been suggested for the subsequent development of SSc; these were: patient demographics; positive family history for SSc; occupational exposures to noxious substances (e.g. silica); exposure to infectious agents; microchimerism, oestrogen and pregnancy-related events; cigarette smoking and alcohol consumption; low birthweight and small-for-gestational age; and vitamin D exposure.
The study data collected from the articles were study type, sample size, methodology, author, country and year of publication.

Risk of bias within studies and quality assessment
The quality of each of the articles was assessed according to their study type. A modified version of the Newcastle-Ottawa Scale was used to assess the quality and risk of bias within the case-control and cohort studies [20]. The Newcastle-Ottawa Scale grades articles according to their case selection, comparability and ascertainment of exposure. For systematic reviews and meta-analyses, a standardized quality assessment scale constructed by the National Heart, Lung and Blood Institute was used, and for narrative reviews, a scale proposed by La Torre et al. [21,22], called the International Narrative Systematic Assessment score, was used.

Study selection
The search retrieved a total of 2347 results. Seven hundred and seventy duplicates were identified; therefore, 1577 titles were screened. Of these articles, 1268 were excluded because of an irrelevant title, and a further 254 were rejected when the abstract was deemed irrelevant. Fifty-five articles underwent full-text review, which included 5 meta-analyses, 2 systematic reviews, 27 narrative reviews, 15 case-control studies, 3 cohort studies and 3 case reports. Three articles were excluded because they had an inappropriate study design [23][24][25]. A further eight were excluded because they were not written in English [26][27][28][29][30][31][32][33]. Four articles were excluded because of their focus on the proposed effects of silicone breast implants on SSc aetiology [7,16,34,35]. Another two articles were excluded because they used the data of another study [19,36]. One article was excluded because it did not make specific reference to SSc [37]. Finally, three articles were excluded because they did not focus on risk factors that lead to the development of SSc [38][39][40][41]. A total of 34 articles were finally included in the systematic review [4-6, 8-10, 12, 13, 17, 18, 41-64]. These are outlined in the PRISMA flow diagram shown in Fig. 1. Upon analysis of the included texts, the methods, risk factors examined and outcomes of each text were noted, and these are listed in Table 1.

Risk of bias across studies
Supplementary Tables S1 and S2, available at Rheumatology online, depict the quality assessment and risk of bias of the included articles. The International Narrative Systematic Assessment scoring system that was used to assess the quality of the narrative reviews considered factors such as the objective of the review, selection of studies and presentation of results. A score of 5/7 or higher is considered to be good, and all of the included reviews scored a grade of 5 or higher. Both systematic reviews struggled to fulfil the criteria of the quality assessment scale that was used in this review. They are, however, comprehensive articles that form a coherent argument and do not represent any conflict of interest or publication bias and were thus deemed to be eligible for inclusion. The quality scale used for the meta-analyses was a modified version of the scale used for the assessment of the systematic reviews, which included an additional criterion regarding whether heterogeneity was assessed [21].

Risk factors for the development of scleroderma
A summary of the findings for risk factors for SSc derived from the studies is shown in Table 2.

Female sex
Female sex as a risk factor for SSc was assessed in seven of the studies, all of which made note of the marked female preponderance, with female:male prevalence ratios ranging from 3:1 to 8:1 [4,8,9,12,13,53].

Age
Patient age was investigated as a risk factor in four of the studies [5,8,13,47]. The common finding between all of the studies was that the risk of SSc peaks between the ages of 45 and 64 years [9,10,12].

Geographical and ethnic group distribution
Nine of the studies examined the effects of geographical location and ethnicity [5, 8-10, 12, 13, 47, 49, 52]. PRISMA: preferred reporting items for systematic reviews and meta-analysis.    SSc occurs much more frequently in women than in men. SSc is five times more common in the USA than in Britain and Japan. African-American (continued) Seven of these articles discovered a higher prevalence in Australia and North America compared with Continental Europe, the UK and Japan [5,8,10,12,13,49,52]. SSc was found to be more prevalent in African-Americans than Caucasian Americans in six of the studies [8,12,13,36,49,52]. One study, on the contrary, did not find an association with ethnicity [9]. A northsouth gradient in Europe was reported in two of the studies, with the incidence being lower in northern Europe [10,49].

Positive family history
Seven of the studies discussed the impact of a positive family history on SSc susceptibility [4, 8-10, 12, 13, 47]. All of these studies reported a positive family history as a risk factor for SSc, and four of them suggested that it is the strongest risk factor for development of SSc, with odds ratios (ORs) varying between 10 and 16 [8][9][10]47]. In contrast, the other three studies argued that it is unlikely that genetic factors are a major cause of SSc [4,12,13]. Relative Risk (of developing SSc with being exposed vs non-exposed to organic solvents).
Exposure to silica Exposure to silica as a risk factor for SSc was examined in 14 of the included articles and was reported to be a significant risk factor in nine of them (including one meta-analysis), with ORs ranging between 3.20 and 25 [6,10,13,18,47,49,52,56,58]. In contrast, two articles concluded that occupational exposure to silica may be a significant risk factor for men, but not for women [57,59]. The final three articles did not find a significant correlation between exposure to silica and SSc [4,8,17].

Infections
Exposure to infections as a risk factor was studied in eight of the included articles, again with considerable heterogeneity [4,8,18,44,53,59,62]. Four of the articles reported a positive correlation between infectious agents with SSc [8,44,59,62]. In contrast, one of the studies reported that there have been conflicting results in the literature regarding infections and SSc but concluded that their involvement in the disease cannot be ruled out [53]. The final three studies stated that there is insufficient evidence in the literature to implicate either bacterial or viral infection as a risk factor for SSc [4,18,49].

Insufficient vitamin D exposure
Two of the included studies investigated the effects of hypovitaminosis D on SSc, and they both concurred that vitamin D deficiency can be a risk factor for SSc [41,49].

Cigarette smoking and alcohol consumption
All of the six studies that investigated smoking and alcohol consumption as a risk factor for SSc found neither of them to be significant risk factors [6,8,13,18,46,49].
Exposure to heavy metals Exposure to heavy metals was examined in one study, which found a significant correlation between SSc and exposure to a number of heavy metals, including palladium, cadmium, zinc and antimony [55].

Microchimerism and pregnancy-related events
Microchimerism refers to the persistence of the retained cells from the fetus of a previous pregnancy in a mother's peripheral bloodstream. The possible effects of microchimerism as a risk factor for SSc were investigated in eight of the articles included in this review [12,47,48,53,59,61,63,64]. Five of these articles reported microchimerism as being more common in SSc patients than in controls and found a positive relationship between SSc and a history of pregnancy (especially a history of having had a son) [47,48,53,59,64]. The other three studies failed to show a relationship and argued against microchimerism as a risk factor [12,61,63].

Low birthweight
One of the included studies examined low birthweight as a risk factor and found a statistically significant

Discussion
This review found complete concordance with the fact that SSc is primarily a disease of middle-aged women, as expected. Likewise, there was strong concordance regarding the geographical distribution of patients with SSc. In contrast, the finding of a north-south gradient across Europe, with SSc being less frequent in northern Europe, is out of keeping with the findings of an Italian systematic review by Antico et al. [10,41,49], which found that patients with SSc have lower vitamin D levels than healthy controls, in percentages varying from 46 to 84%. However, this finding was based on a modest total of 313 subjects, whereas the epidemiological studies that have noted the higher frequency of SSc in southern Europe have been based on significantly larger, betterpowered registries. Therefore, even if vitamin D deficiency is a risk factor for SSc, it seems that there must be other environmental exposures in southern Europe that predispose to SSc that can still contribute to the development of SSc in patients with normal vitamin D levels.
A positive family history of SSc has repeatedly been reported as the strongest risk factor for the development of SSc, but in spite of this, numerous studies have noted that monozygotic twins with SSc are very rare and that 98% of SSc patients do not have another affected family member of any degree of blood relation [9,12,47]. The logical explanation for this would be that there is a modest genetic predisposition for SSc and that the environmental exposures necessary for the development of SSc in a predisposed person are uncommon. Therefore, although the brother of a patient with SSc may be at a 14-fold increased risk of developing SSc throughout his lifetime, the absolute risk of him developing SSc is low, owing to the rarity of the putative environmental exposures necessary for SSc to occur [10].
Perhaps the most contentious issue regarding environmental risk factors for SSc is the exposure to organic solvents. This review included three meta-analyses that examined organic solvents as a risk factor for SSc, all of which found a significant correlation between SSc and exposure to organic solvents, with combined OR values ranging from 2.07 to 2.91 [5,42,43]. Although these results appear compelling, there is the possibility of a significant publication bias, either owing to researchers not submitting negative results or owing to journals declining to publish the findings of studies that have not found organic solvents to be a significant risk factor for SSc. Furthermore, all of the four case-control studies included in this review that reported a significant correlation between SSc and organic exposure were retrospectively based on self-reported exposure and were therefore subject to significant recall bias [51,54,56,60]. Finally, as noted by Dospinescu et al. [6], exposure to organic solvents has usually been ascertained on the basis of the subjects' occupations, and it is likely that these subjects (e.g miners) had concomitant, confounding exposure to other environmental agents. Therefore, although the consensus in the literature favours organic solvents as being a significant risk factor for SSc, the evidence for this is still arguably dubious.
Of all of the environmental exposures that have been investigated as risk factors for SSc, occupational exposure to silica appears to be the most convincing. A welldesigned meta-analysis by McCormic et al. [58] found a significant relationship between occupational silica exposure and SSc, with a combined estimator or relative risk of 3.20 (95% CI: 1.24-7.35). In contrast, a large case-control study by Burns et al. [17] did not detect any effect of occupational exposure to silica on the risk of SSc, but this study only included female subjects, who are less likely to experience occupational exposures to silica than men (e.g. as in abrasive grinding, mining and sandblasting). This is in keeping with the finding of Mora et al. [59], who argued that occupational exposure to silica significantly increases the risk of SSc in men, but not in women, primarily because the exposure to silica of the male cohort is far greater.
Although the theory of infection leading to molecular mimicry and causing autoimmune disease is biologically plausible, there is currently insufficient evidence to support the notion of infection being a significant risk factor for SSc. In a case-control study involving a modest total of 30 patients with SSc, Bilgin et al. [44] found that a higher proportion of SSc patients had antibodies against Helicobacter pylori, CNV, EBV and parvovirus B19. However, that study did not account for the impact of other environmental factors on the development of SSc (i.e. did not adjust the results with multivariate logistic regression), which limits the validity of the conclusions. Owing to the poor quality of this evidence, infections are not regarded as a risk factor for SSc [4,18]. Likewise, there is insufficient evidence for cigarette smoking as a risk factor for the development of SSc, although it might exacerbate the severity and prognosis, once the disease process has begun [6,46,49].
Microchimerism as a risk factor for SSc is an interesting theory, which stemmed from the observation of an obvious female preponderance, with females being affected preferentially in the post-childbearing years [47]. It has since been discovered that microchimeric cells are more commonly detected in women with SSc than in healthy controls [47,53,59]. These observations motivated many of the studies included in this review, which yielded widely conflicting results. Firstly, Cockrill et al. [48] found, in their highly powered case-control study involving 987 patients with SSc and 3088 sibling controls, that a history of one or more pregnancies increased the risk of SSc by an OR of 2.8. In direct contrast to this, Pisa et al. [61], in their case-control study, found that a history of pregnancy reduces the risk of SSc (OR: 0.3; 95% CI: 0.1-0.8). It would seem that microchimerism is only one of a multitude of factors that have been implicated in the pathogenesis of SSc but are yet to be proven aetiological factors. This is evidenced by the observation that microchimerism occurs in many healthy, normal women, and SSc often occurs in nulliparous women [47,61].
Although the articles that examined exposure to heavy metals and low birthweight in this review were strong studies, there are currently insufficient data in the literature to confirm whether or not these factors predispose to the development of SSc [18,50,55]. Further research in these areas would be valuable to determine the role of these agents in SSc pathogenesis.
Ultimately, ascertaining the risk factors that predispose to SSc is an ongoing dilemma for three reasons. Firstly, SSc is a very rare disease, making it difficult to obtain cohorts that are adequately powered to provide meaningful data. Secondly, the majority of environmental exposures that are currently under investigation are often experienced concomitantly, and therefore it may be impossible to attribute the development of SSc to one exposure alone. Thirdly, in order to implicate an exposure in SSc aetiology fully, prospective studies are required, because retrospective studies are subject to significant recall bias. The feasibility of such studies is doubtful, because the likelihood of a subject developing SSc as a result of being exposed to a given agent is extremely low, and there are obvious ethical implications involved with risking the development of such a debilitating disease in a subject. Although we have discovered that occupational exposure to silica is a significant risk factor in SSc, in addition to a number of aforementioned patient demographics, the absolute and attributable risks of these factors to SSc aetiology are clearly low, because the vast majority of individuals with these risk factors do not develop SSc. Unfortunately, performing a meta-analysis for the various risk factors for SSc was not possible for the present review, owing to there being excess heterogeneity among the literature. Therefore, as noted numerous times by Roberts-Thomson et al. [9,19], SSc is likely to result from a number of stochastic events that are impossible to identify, thus SSc may continue to be seen as an idiopathic disease.

Conclusion
In conclusion, this review found age between 45 and 64 years, female sex, living in the USA and Australia, positive family history and occupational exposure to silica to be significant risk factors for the development of SSc. There were conflicting findings regarding the impact of exposure of organic solvents and microchimerism on the development of SSc. There are currently insufficient data to implicate infectious agents as a risk factor. Alcohol consumption and cigarette smoking were not found to be risk factors. Given that we currently have only empirical treatment for SSc, the possible prevention of this disease is of the utmost importance. The results of this review suggest that industries involving exposure to silica should ensure that their staff use protective measures and have regular health checks to limit the possibility of their occupational exposure leading to the development of SSc. Further research is required to ascertain the role of organic solvents, microchimerism and infectious agents as risk factors for SSc.
Funding: No specific funding was received from any bodies in the public, commercial or not-for-profit sectors to carry out the work described in this manuscript.
Disclosure statement: The authors have declared no conflicts of interest.