Turner Syndrome Increases the Risk of Psoriasis: A Systematic Review and Meta-Analysis

Abstract

Aims: Patients with Turner syndrome (TS) may have a higher risk of psoriasis as suggested by some reports. Data on this association are still limited. We investigated the association between TS and the risk of prevalent and incident psoriasis by combining results from available studies using systematic reviews and meta-analysis techniques. Methods: Potentially eligible studies were identified from Medline and EMBASE databases from inception to December 2023 using a search strategy that comprised of terms for “Turner syndrome” and “psoriasis”. An eligible cohort study must comprise of two groups of participants—those with and without TS. It must report our outcome of interest—incidence and/or prevalence of psoriasis in each group. The pooled effect estimates were generated using the generic inverse variance method, which assigns weight to each study in reversal to its variance. Meta-analyses of the prevalent and incident psoriasis were conducted separately. Results: A total of 4919 articles were retrieved. After two rounds of independent review by two investigators, five cohort studies (two incident studies and three prevalent studies) met the eligibility criteria and were included in the meta-analyses. The meta-analyses found a significantly elevated risk of both incident and prevalent psoriasis in patients with TS compared to individuals without TS, with the pooled risk ratio of 5.58 (95% CI, 3.73–8.35; I² 0%) and 5.66 (95% CI, 1.52–21.03; I² 19%), respectively. Conclusions: An increased risk of both incident and prevalent psoriasis among patients with TS was demonstrated in this study.

1. Introduction

Turner syndrome (TS) is the most common sex chromosome aneuploidy, affecting 25–50 per 100,000 females [1]. It is cytogenetically characterized by one intact X chromosome and a missing or structurally abnormal second sex chromosome. Approximately half of individuals with TS result from monosomy X (45,X), whereas the rest are caused by heterogeneous abnormalities of the X chromosome, including mosaicism (as a result of sex chromosome nondisjunction during postzygotic cell division) and isochromosomes. Patients with TS present with multisystem health issues such as short stature, hypergonadotropic hypogonadism, hearing loss, congenital anomalies of the lymphatic, renal, cardiovascular, and skeletal system, atherosclerotic disease, osteoporosis, and neurodevelopmental differences. Autoimmune diseases, particularly Hashimoto’s thyroiditis, diabetes mellitus, and celiac disease, are known comorbidities of TS and are diagnosed in 20–50% of patients [1,2]. Diagnosis of TS is usually made postnatally based on characteristic clinical features plus karyotype analysis but is sometimes made prenatally during chorionic-villus sampling or amniocentesis. Treatment of TS is primarily focused on the management of their cardiovascular disease (which is a leading cause of morbidity and mortality) and an estradiol–progestin replacement to prevent complications of premature ovarian failure.

Psoriasis is a chronic autoimmune dermatologic disorder, affecting 3.2% of adults (both males and females) and 0.1% of children (both boys and girls) in the United States [3]. Complex interactions between dendritic cells, T cells, keratinocytes, neutrophils, and several cytokines, such as interferon-alpha, tumor necrosis factor alpha, interleukin-12, and interleukin-23, play pivotal roles in the initiation and perpetuation of the cutaneous inflammation that eventually leads to keratinocyte hyperproliferation. The association between psoriasis and TS was first described by Dacou-Voutetakis et al. in 1996 [4]. Since then, there have been case reports and observational studies providing insights on this topic; some publications, however, did not find an association between the two conditions [5,6]. The aim of our study is to conduct a systematic review and meta-analysis to evaluate whether the prevalence and incidence of psoriasis is increased in patients with TS [7,8,9].

2. Results

2.1. Search Results

A total of 3262 articles and 1657 articles were retrieved from the EMBASE and MEDLINE databases, respectively. There were 673 duplicated articles that were indexed in both databases. They were eliminated, resulting in an abstract review of 4246 articles. After the abstract review, we found that 3242 articles were clearly not eligible based on the type of article, methodology, and participants. This elimination left a total of 29 articles for the full-length article review. After a critical full-length review of the 29 articles, we found that 22 of them did not report the outcome of interest, leaving seven cohort studies, three incident studies [6,10,11], and four prevalent studies [2,4,5,12] that met the eligibility criteria. However, during a second round of full-length review, we found that both the study by Jorgensen et al. [6] and Viuff et al. [11] utilized the same database (Danish Cytogenetic Central Register) and, thus, present a risk of patient duplication. The study by Jorgensen et al. [6] was excluded because the study by Viuff et al. [11] was more robust, with a higher number of participants as well as a better selection of comparators. In addition, the study by Bakalov et al. [5] used patients with premature ovarian failure as a comparator, not otherwise healthy controls. Thus, it was excluded as well. Finally, five studies were included in the meta-analyses. Figure 1 demonstrates the process of the study review and identification.

Table 1. Main characteristics of the cohort studies included in the meta-analysis of incident psoriasis in Turner syndrome.

	Goldacre et al. [10]	Viuff et al. [11]
Country	United Kingdom	Denmark
Study design	Retrospective cohort	Retrospective cohort
Year of publication	2014	2021
Total number of participants	Cases with TS: 2459 Comparators without TS: 3.8 million	Cases with TS: 1156 Comparators without TS: 115,577
Recruitment of participants	Patients with TS were identified from the Hospital Episode Statistics dataset, which included records for all National Health Service hospital admission, from 1999 to 2011. Females in the dataset without TS and were admitted for minor and surgical condition served as comparators.	Patients with TS were identified from the Danish Cytogenetic Central Register which prospectively collected results from all chromosome examinations in cytogenetic labs in Denmark since 1968. Only patients who were first diagnosed between 1960 and 2014 were included. Age-matched females without TS were randomly identified from the National Patient Registry to serve as comparators. A total of 100 comparators were identified per 1 case.
Diagnosis of TS	Presence of diagnostic code for TS in the database	From cytogenetic analysis
Cytogenetic abnormalities - Monosomy X - Isochromosome Xq - Others	NA	422 (37%) 117 (10%) 617 (53%)
Diagnosis of psoriasis	Presence of diagnostic code for psoriasis in the database	Presence of ICD-8-CM or ICD-10-CM for psoriasis in the Danish National Patient Register which prospectively collected data from all inpatient and outpatient care from all residents of Denmark.
Follow-up	Until 31 December 2011	Until 31 December 2014
Average duration of follow-up (years)	NA	NA
Average age of participants (years)	NA	NA
Number of incident cases of psoriasis	Cases with TS: 26 Comparators without TS: NA	Cases with TS: 15 Comparators without TS: 624
Newcastle-Ottawa score	Selection: 2 Comparability: 0 Outcome: 3	Selection: 4 Comparability: 1 Outcome: 3

Abbreviation: NA, not available; TS, Turner syndrome.

and

Table 2. Main characteristics of the cohort studies included in the meta-analysis of prevalent psoriasis in Turner syndrome.

	Dacou-Voutetakis et al. [4]	Su et al. [12]	Gawlik et al. [2]
Country	Greece	United States	Poland
Study design	Prospective cohort	Retrospective cohort	Prospective cohort
Year of publication	1996	2009	2018
Total number of participants	Cases with TS: 35 Comparators without TS: 304	Cases with TS: 20 Comparators without TS: 10	Cases with TS: 37 Comparators without TS: 11
Recruitment of participants	Patients with TS were recruited from the Department of Pediatric Endocrinology of the Athens University. Comparators without TS were recruited from as Athenian school.	Patients with TS were recruited from the Endocrinology Clinic of the University of California, San Francisco. Comparators without TS were recruited from as Athenian school.	Patients with TS were consecutively recruited from the Department of Pediatric Endocrinology of the Medical University of Silesia. Comparators without TS were recruited from the same institute.
Diagnosis of TS	Based on peripheral blood cell karyotype	Review of medical records	Based on peripheral blood cell karyotype
Cytogenetic abnormalities - Monosomy X - Isochromosome Xq - Other	15 (43%) 2 (6%) 18 (51%)	8/20 (40%) 3/20 (15%) 9/20 (45%)	NA 7 (19%) NA
Diagnosis of psoriasis	Direct interview and physical examination	Review of medical records	Direct interview and review of medical records
Average age of participants (years)	Cases with TS: 13.4 Comparators without TS: 11.8	Cases with TS: 22.8 Comparators without TS: 30.4	Cases with TS: 11.9 Comparators without TS: 12.5
Number of prevalent cases of psoriasis	Cases with TS: 35 (17.1%) Comparators without TS: 5 (1.6%)	Cases with TS: 2 (10.0%) Comparators without TS: 0 (0.0%)	Cases with TS: 1 (2.7%) Comparators without TS: 0 (0.0%)

Abbreviation: NA, not available; TS, Turner syndrome.

describe the study methodology and characteristics of subjects in the included studies for the meta-analyses of incident and prevalent psoriasis, respectively.

2.2. Risk of Incident Psoriasis in Patients with TS

The meta-analysis found a significantly higher risk of incident psoriasis in patients with TS compared to the control group, with a pooled risk ratio of 5.58 (95% CI, 3.73–8.35). This meta-analysis had non-significant between-study statistical heterogeneity with I² of 0% (Figure 2).

2.3. Risk of Prevalent Psoriasis in Patients with TS

The meta-analysis also found a significantly elevated risk of prevalent psoriasis in patients with TS compared to individuals without TS, with a pooled risk ratio of 5.66 (95% CI, 1.52–21.03). This meta-analysis had low between-study statistical heterogeneity with I² of 19% (Figure 3).

2.4. Evaluation for Publication Bias

Visualization for publication bias using a funnel plot was not feasible and was not performed due to the limited number of included studies in both meta-analyses.

3. Material and Methods

3.1. Search Strategy

The current study was conducted by searching for relevant published studies from two sources, the Medline and Embase databases. The search was conducted for studies indexed in the databases until December 2023. Search terms and keywords used to identify relevant studies comprised of terms related to the two diseases of interest (i.e., TS and psoriasis). Information on the search terms, syntax, and operation used for this systematic review is provided as Supplemental Material S1. We included studies published in any language. This process of study identification was independently conducted by two investigators (B.P. and P.U.).

3.2. Inclusion Criteria

We only included cohort studies that comprised two groups of participants—one group with subjects with TS and another group of subjects without TS. It must report our outcome of interest—incidence or prevalence of psoriasis in both group—or report a ratio comparing the incidence or prevalence between the cohort of participants with and without TS. The ratio can be either a relative risk (RR) or hazard risk ratio (HR). It must also report the associated 95% confidence interval (CI) for the calculation of the pooled risk ratio in meta-analysis.

The inclusion criteria were applied to each retrieved article to determine its eligibility for meta-analysis. The assessment of the study’s eligibility was independently performed by the same two investigators and was cross-checked to avoid any human errors. If determination of the study’s eligibility was different between the two investigators, the study would be reviewed together by the two investigators to find the final consensus. The quality of the included incident studies was rated using the Newcastle–Ottawa quality assessment scale for cohort studies [7].

3.3. Data Extraction

The following information were collected from the included studies: last name of the first author, country where the study was conducted, year(s) when the study was conducted, publication year, study size, how the participants were recruited or identified, how the diagnosis of TS was made, how the diagnosis of psoriasis was made, average follow-up duration (for incident studies), baseline characteristics and demographics of the participants, and variables that were adjusted in multivariate analysis.

3.4. Statistical Analysis

Review Manager 5.3 software from the Cochrane Collaboration was used for all of the analyses. RR and/or HR from all incident studies were combined together to calculate the pooled risk ratio. Each study was weighted according to its precision, which was estimated using its variance (a study with less precision with higher variance will receive a lower weight in the pooled analysis and vice versa). This method was originally described by DerSimonian and Laird and has been widely used in the literature [8]. We decided to run the random-effect model instead of the fixed-effect model, as the basic assumption of the fixed-effect model that all studies should yield the same outcome is almost universally not true in the real world due to differences in study design, method of diagnosis, measurement off outcome, and background population studied. Meta-analyses of the prevalent and incident psoriasis were conducted separately. Assessment of between-study heterogeneity was performed using the Cochran’s Q test and I² statistic. We divided the value of I² into four quarters—0–25%, 26–50%, 51–75%, and 75–100%—to represent insignificant, low, moderate, and high heterogeneity, respectively [9]. We planned to use visualization of a funnel plot to look for asymmetry of distribution of included studies in relation to their RR and/or HR and precision (defined by variance) to assess for evidence of potential publication bias if enough studies were eligible and included in the meta-analyses.

4. Discussion

We conducted a systematic review and meta-analysis of 3707 individuals with TS from seven cohort studies and found higher incidence and prevalence of psoriasis in patients with TS when compared to control groups, with a pooled risk ratio of 5.58 (95% CI, 3.73–8.35) and a pooled risk ratio of 5.66 (95% CI, 1.52–21.03), respectively. The results suggest that TS is a risk factor for psoriasis, in addition to previously known dermatologic diagnoses such as melanocytic nevi, pilomatricomas, dermatitis/eczema, alopecia areata, vitiligo, and halo nevi [1,11]. Our findings are supported by a population-based study which has shown that anti-psoriatic medications were prescribed more frequently in TS patients when compared to the general population (hazard ratio of 2.5, 95% CI 1.9–3.2) [11]. This study had 37% of patients with monosomy and 10% of patients with isochromosome.

It is unclear why individuals with TS develop psoriasis. They have different immunologic profiles, with lower immunoglobulin levels, lower CD4+/CD8+ T-lymphocyte ratios, increased pro-inflammatory cytokines (interleukin [IL]-6, transforming growth factor [TGF] β1), and lower anti-inflammatory cytokines (IL-10, TGFβ2) [2,13]. One of the widely referenced theories is haploinsufficiency (deficient expression) of immune system genes on the X chromosome, particularly FOXP3. The gene is crucial for the development of regulatory T cells, of which, their impaired function is a known pathogenesis of psoriasis [14]. Pathogenic loss-of-function variants in the FOXP3 gene result in severe immunodeficiency, autoimmunity, and, in a few patients, psoriasiform rashes [15]. Furthermore, TS patients, especially those with autoimmunity, were found to have quantitative and qualitative defects in regulatory T cells [2,16]. Other plausible explanations for an increased risk of psoriasis in TS patients include aneuploidy itself (as demonstrated by prevalent autoimmunity in individuals with Down syndrome and Klinefelter syndrome), aberrations in other immune system genes, both on X chromosomes [17] and autosomes [18], epigenetic dysfunction [19], ovarian insufficiency and androgen deficiency [5], and defective thymic negative selection of self-reactive T cells [12].

The diagnosis of psoriasis impacts healthcare utilization, as the risk of hospital admission related to this diagnosis was increased in female TS patients compared with background females (incidence rate ratio 6.9, 95% CI 2.3–20.5) [11]. Identification of TS individuals with psoriasis is thus important from clinical and public health perspectives. The clinical practice guidelines for TS, published by the International Turner Syndrome Consensus Group, do not provide specific recommendations regarding psoriasis, but they recommend annual skin exams in adult patients to assess for keloid and changes in pigmented nevi [1]. We observe that among the seven included studies, the highest frequency of psoriasis (6/35, 17%) was reported when Dacou-Voutetakis et al. specifically looked for psoriasis during a routine clinic visit [4]. Notably, all six psoriasis patients in that study had exceptionally early onset, ranging from infancy to five years. We hypothesize that psoriasis is underdiagnosed in TS and that it is not uncommon in pediatric population. It is therefore prudent that physicians perform annual skin exams beginning at a young age for early detection and referral, if indicated.

Our study has limitations. First, there were only two and three studies for final analyses of incident and prevalent psoriasis, respectively. Therefore, the pooled effect estimates are relatively imprecise, as reflected by the wide 95% CIs. Second, the included studies did not provide data on two factors that are potentially associated with autoimmunity in TS, i.e., type of cytogenetic abnormalities and parental origin of the X chromosome [6]. Further studies on genotype–phenotype correlation are needed. Third, the data on type of psoriasis (e.g., plaque psoriasis, pustular psoriasis), natural history, and management of psoriasis in TS are lacking. For example, in contrast to juvenile rheumatoid arthritis, which is a known comorbidity of TS [1], psoriatic arthritis has been rarely reported [20]. Prospective, longitudinal studies in a larger cohort are warranted. Fourth, the apparent increased risk could be a function of surveillance bias, as patients with TS tend to have more exposure to medical providers (as a result of their chronic illness) and, thus, more frequent skin exams. Finally, we could not perform a formal evaluation for publication bias as the number of included studies was too small (only two and three studies for each meta-analysis). Therefore, publication bias in favor of studies that showed positive associations may have been present.

5. Conclusions

Patients with TS have a higher incidence and prevalence of psoriasis when compared to control groups. Aberrant genetic, epigenetic, immunologic, and hormonal factors have been postulated as the underlying pathophysiology of this association. It is possible that psoriasis is underdiagnosed in TS and is not uncommon in pediatric population. Annual skin exams beginning at a young age could be considered, similar to surveillance to autoimmune hypothyroidism, in which the elevated risk is already well known [21] and the surveillance is widely adopted in clinical practice. Further studies are needed to evaluate the risk factors and natural history of psoriasis in individuals with TS.