Autoimmune thyroid disease (AITD) and psoriatic disease share auto-immunological components. Few studies have investigated the link between both, yielding inconclusive results.
We assessed the association of AITD with psoriatic disease in a prospective cohort study and performed a systematic review and meta-analysis.
8214 participants of the Rotterdam Study (RS) with thyroid peroxidase antibodies (TPO-Abs), thyroid-stimulating hormone (TSH) and/or free thyroxine (FT4) measurements and information on psoriatic disease were included. We performed logistic and Cox regression analyses and a systematic literature search in several electronic databases on AITD and psoriatic disease. We pooled odds ratios (ORs) of included studies using the Mantel-Haenszel method, while adding RS data on prevalent psoriatic disease.
Within the RS, we found no association between TPO-Ab positivity and psoriatic disease. There was a positive trend between TSH and prevalent psoriatic disease, and between FT4 and incident psoriatic disease, although not significant. Out of 1850 articles identified, seven were included in the systematic review and four in the meta-analysis. The risk of psoriatic disease (pooled OR) was 1.71 (confidence interval (CI): 1.27–2.31) for TPO-Ab positivity, 1.25 (CI: 1.14–1.37) for AITD and 1.34 (CI: 1.16–1.54) respectively, and 1.17 (CI: 1.03–1.32) for hypothyroidism and hyperthyroidism.
Our meta-analysis suggests that TPO-Ab positivity, hypothyroidism and hyperthyroidism might be associated with prevalent psoriatic disease. However, there are only few studies with large heterogeneity regarding psoriatic disease definition and indication of publication bias. Additional prospective data are needed to assess the association of AITD with incident psoriatic disease.
Autoimmune thyroid disease (AITD) is the most common cause of both hyper- and hypothyroidism (i.e. Graves’ and Hashimoto’s disease respectively) in iodine-replete areas (1). Anti-thyroid peroxidase antibodies (TPO-Abs) are the most prevalent auto-antibodies in patients with AITD, being present in 90% of the patients with Hashimoto thyroiditis (1, 2). AITD is a reflection of complex gene–environment interactions and involves infiltration of cytokine-producing T-lymphocytes that activate various inflammatory pathways (3). Thyroid disease has a wide variety of clinical presentations, including various cutaneous manifestations. This can vary from a dry, doughy skin in hypothyroidism to erythema and diffuse pigmentation in hyperthyroidism (4, 5, 6).
Psoriasis is a common chronic immune-mediated skin disease that is characterized by rough, well-demarcated scaly plaques. Psoriasis is known for its significant impact on health-related quality of life (7). Psoriatic arthritis (PsA) is a chronic inflammation of the skin and synovium that occurs in approximately 3.2% of the patients with psoriasis (8, 9).
Several inflammatory pathways have been described in AITD and psoriatic disease (psoriasis and/or PsA), suggesting an interconnected pathophysiology. Interleukin 17 (IL17) is an important cytokine that stimulates keratinocyte expression of various chemokines leading to sustained inflammation in the skin, epidermal hyperproliferation, and skin barrier disruption (10). IL17 is found to play an important role in both Graves’ disease (11) and Hashimoto thyroiditis (12) as well. Expression of several other chemokines, including CXCL10 (interferon(IFN)-inducible protein-10; IP10), CXCL9 (monokine induced by IFN-γ; Mig), CCL2 (monocyte chemoattractant protein-1; MCP1) and CCL22 (macrophage-derived chemokine; MDC) has also been found in both AITD and psoriatic disease (13, 14, 15, 16, 17, 18). CCL2 and CCL22 mark the shift from a Th1 immune response to Th2 immunity, which has been shown in an advanced disease course of both PsA and Graves’ disease (15, 17, 18). Furthermore, a disruption in the nuclear factor-κB (NF-κB) signaling pathway, an important feature of several autoimmune diseases, has been described in both AITD and psoriasis (19).
A higher risk of AITD in psoriatic patients could prompt physicians to screen for thyroid dysfunction in psoriatic patients with unexplained symptoms. However, the association between thyroid function and psoriatic disease remains unclear and has been assessed by only a few, mostly retrospective patient population-based studies (20, 21, 22, 23, 24).
Hence, we aim to (1) assess the association of TPO-Ab positivity, thyroid-stimulating hormone (TSH) and free thyroxine (FT4) with psoriatic disease in the Rotterdam Study (RS), a large prospective population-based cohort study and (2) conduct a systematic review and meta-analysis of available epidemiologic studies evaluating the relation between AITD/thyroid function and psoriatic disease including the yet unpublished data from the RS.
Subjects and methods
We assessed the association of TPO-Ab positivity, TSH and FT4 with prevalent and incident psoriatic disease, and the association of AITD (with the subgroups of hypothyroidism and hyperthyroidism) with psoriatic disease cross-sectionally in the RS. Furthermore, we conducted a systematic literature search and meta-analysis on studies assessing the association between AITD/thyroid function and psoriatic disease, also incorporating the cross-sectional data from the RS.
Study design and population
This study is embedded in the RS, a prospective population-based cohort study that investigates risk factors for cardiovascular, dermatological, endocrine and locomotor, liver, ophthalmic, psychiatric and respiratory diseases in the elderly. The RS started in 1990 and included 14 926 participants aged 45 years and older from Ommoord, Rotterdam, the Netherlands by 2008. This study was approved by the Erasmus MC Medical Ethics Committee and the Dutch Ministry of Health, Welfare and Sports. Written informed consent was obtained from all participants. Detailed information on the RS is available elsewhere (25, 26). We included participants from three independent consecutive cohorts of the RS (visits I–3, II–1 and III–1) with measurements of TSH and/or FT4 and information on psoriatic disease prevalence and incidence (n = 8214). For the longitudinal analyses, all study participants, free of psoriatic disease at baseline (n = 8044), were followed from date of laboratory measurement until incidence of psoriatic disease, death, loss to follow-up or September 1st 2011, whichever occurred first.
Assessment of thyroid function
Serum samples were collected between 1997 and 2008 and the moment of blood drawing was considered the study baseline. TPO-Abs, TSH and FT4 were all assessed in serum samples stored at −80°C using the same assay in all cohorts (thyroid peroxidase antibodies, thyrotropin and thyroxine electrochemiluminescence immunoassay, ‘ECLIA’, Roche). TPO-Ab levels greater than 35 kU/mL were regarded as positive, according to manufacturer recommendations. The normal range of TSH comprised 0.4–4.0 mU/L and the normal range of FT4 comprised 11–25 pmol/L (=0.85–1.94 ng/dL), as mentioned in guidelines and previous reports (27, 28). Hypothyroidism comprised both overt and subclinical hypothyroidism and was defined as TSH values >4 mU/L and FT4 <25 pmol/L (<1.94 ng/dL). Hyperthyroidism comprised both overt and subclinical hyperthyroidism and was defined as TSH values <0.4 mU/L and FT4 >11 pmol/L (>0.85 ng/dL).
Assessment of psoriasis
A validated algorithm was used to define patients with psoriasis (29). General practitioner (GP) records were screened for a psoriasis code (S91) and for use of antipsoriatic medication between January 1st 1991 and December 31st 2009. Subsequently, a medical record review was performed on GP notes, medical specialist reports and hospital discharge letters. In case of uncertainty about the psoriasis diagnosis, consensus was sought based on medication and/or medical records, or the final diagnosis was based on a skin examination at the research center in Ommoord. The psoriasis diagnosis was considered definite if made by a rheumatologist or dermatologist, if made at least twice by the GP, if based on the use of psoriasis specific medication, or if established at the research center. We only included participants with definite psoriasis in our analyses. The date of psoriasis onset was the date of first diagnosis in the medical records, the date of first use of antipsoriatic medication or the self-reported date of onset, whichever was available or came first (29).
Alcohol consumption and smoking status were determined by questionnaires. Alcohol consumption was reported in grams per day and smoking status was categorized as current, former or never. Body mass index (BMI) was determined through physical examination by dividing the body weight in kilograms by the squared height in meters. Diabetes mellitus was defined as a fasting plasma glucose level of at least 7 mmol/L (measured twice within a period of one year), a non-fasting plasma glucose level of at least 11.1 mmol/L (in absence of a fasting measurement), or as the use of antidiabetic medication. Blood pressure was measured twice in a sitting position on the right upper arm using a random-zero sphygmomanometer and the average of two measurements was taken. Hypertension was defined as a systolic blood pressure of at least 140 mmHg, a diastolic blood pressure of at least 90 mmHg, or as the use of antihypertensive drugs. Serum cholesterol (mmol/L) and C-reactive protein (CRP) (mg/mL) levels were measured in the laboratories of the Erasmus MC using standard techniques. Information on education was derived by questionnaires and the highest attained education was taken as a proxy for socioeconomic status (SES).
The association between TPO-Ab positivity or thyroid function and prevalent psoriatic disease in the RS was assessed using logistic regression models. Potential confounders were selected based on biological confounding plausibility and availability of data. In a first model, we adjusted for age, sex and cohort. In a second model, we additionally adjusted for alcohol consumption, smoking status, BMI, diabetes mellitus, hypertension, serum cholesterol, serum CRP level and SES. The association between thyroid function and incident psoriatic disease was assessed using Cox proportional hazards models. For the Cox models, we used the same covariates as for the cross-sectional analyses.
TSH was log-transformed for all analyses, due to the skewed distribution. To control for missing values of the confounders, we used multiple imputation. We used the Markov Chain Monte Carlo method to create 10 imputed datasets, which were pooled for analyses. The 2-sided significance level was set at P < 0.05. There was no departure from linearity in any of the models, as assessed by using restricted cubic splines (3 knots). Furthermore, there was no violation of the proportional hazards assumption in the longitudinal analyses. All analyses were conducted in IBM SPSS Statistics 21, except for the assessment of linearity which was performed in R (rms package, R-project, Institute for Statistics and Mathematics, R Core Team (2016), Vienna, Austria, version 3.2.5).
Systematic review of the literature and meta-analysis
We have performed the systematic review and meta-analysis in accordance with the PRISMA guidelines for transparent reporting and have provided the checklist in Table 1.
PRISMA 2009 checklist.
|Section/topic||#||Checklist item||Reported on page#|
|Title||1||Identify the report as a systematic review, meta-analysis, or both||1|
|Structured summary||2||Provide a structured summary including, as applicable: background; objectives; data sources; study eligibility criteria, participants, and interventions; study appraisal and synthesis methods; results; limitations; conclusions and implications of key findings; systematic review registration number||1|
|Rationale||3||Describe the rationale for the review in the context of what is already known||2|
|Objectives||4||Provide an explicit statement of questions being addressed with reference to participants, interventions, comparisons, outcomes, and study design (PICOS)||2|
|Protocol and registration||5||Indicate if a review protocol exists, if and where it can be accessed (e.g., Web address), and, if available, provide registration information including registration number||×|
|Eligibility criteria||6||Specify study characteristics (e.g., PICOS, length of follow-up) and report characteristics (e.g., years considered, language, publication status) used as criteria for eligibility, giving rationale||5|
|Information sources||7||Describe all information sources (e.g., databases with dates of coverage, contact with study authors to identify additional studies) in the search and date last searched||3, 5|
|Search||8||Present full electronic search strategy for at least one database, including any limits used, such that it could be repeated||5|
|Study selection||9||State the process for selecting studies (i.e., screening, eligibility, included in systematic review, and, if applicable, included in the meta-analysis)||5|
|Data collection process||10||Describe method of data extraction from reports (e.g., piloted forms, independently, in duplicate) and any processes for obtaining and confirming data from investigators||5|
|Data items||11||List and define all variables for which data were sought (e.g., PICOS, funding sources) and any assumptions and simplifications made||Table 2|
|Risk of bias in individual studies||12||Describe methods used for assessing risk of bias of individual studies (including specification of whether this was done at the study or outcome level), and how this information is to be used in any data synthesis||5|
|Summary measures||13||State the principal summary measures (e.g., risk ratio, difference in means)||5, 7|
|Synthesis of results||14||Describe the methods of handling data and combining results of studies, if done, including measures of consistency (e.g., I2) for each meta-analysis||5, 7|
|Risk of bias across studies||15||Specify any assessment of risk of bias that may affect the cumulative evidence (e.g., publication bias, selective reporting within studies)||7|
|Additional analyses||16||Describe methods of additional analyses (e.g., sensitivity or subgroup analyses, meta-regression), if done, indicating which were pre-specified||7|
|Study selection||17||Give numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at each stage, ideally with a flow diagram||Fig. 1|
|Study characteristics||18||For each study, present characteristics for which data were extracted (e.g., study size, PICOS, follow-up period) and provide the citations||Table 2|
|Risk of bias within studies||19||Present data on risk of bias of each study and, if available, any outcome level assessment (see item 12)||Table 6|
|Results of individual studies||20||For all outcomes considered (benefits or harms), present, for each study: (a) simple summary data for each intervention group (b) effect estimates and confidence intervals, ideally with a forest plot||Fig. 2|
|Synthesis of results||21||Present results of each meta-analysis done, including confidence intervals and measures of consistency||Fig. 2|
|Risk of bias across studies||22||Present results of any assessment of risk of bias across studies (see Item 15)||10, Fig. 3|
|Additional analysis||23||Give results of additional analyses, if done (e.g., sensitivity or subgroup analyses, meta-regression (see Item 16))||Table 5|
|Summary of evidence||24||Summarize the main findings including the strength of evidence for each main outcome; consider their relevance to key groups (e.g., healthcare providers, users, and policy makers)||10, 11|
|Limitations||25||Discuss limitations at study and outcome level (e.g., risk of bias), and at review-level (e.g., incomplete retrieval of identified research, reporting bias)||11, 12|
|Conclusions||26||Provide a general interpretation of the results in the context of other evidence, and implications for future research||12|
|Funding||27||Describe sources of funding for the systematic review and other support (e.g., supply of data); role of funders for the systematic review||12|
Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 6(6): e1000097. doi:10.1371/journal.pmed1000097.
Study search and identification
We systematically searched online medical databases (Embase, Medline Ovid, Web of Science, Scopus, Cochrane, PubMed Publisher, Cinahl Ebsco, ProQuest and Google Scholar) up to March 4th 2016 to find articles in English assessing the association between thyroid function/disease and psoriatic disease with the help of an experienced librarian (WMB). The performed search strategy yielded 1850 references, after excluding duplicates.
Example of electronic search strategy (EMBASE):
(‘thyroid gland’/exp OR ‘thyroid disease’/exp OR ‘thyroid hormone’/de OR ‘thyrotropin’/de OR ‘thyroxine’/de OR ‘thyroid function’/exp OR ‘thyroid hormone blood level’/exp OR ‘thyroid peroxidase antibody’/de OR ‘thyroid gland examination’/exp OR (thyroid* OR hyperthyro* OR hypothyro* OR tsh OR ft4 OR thyroxin* OR Hashimoto OR Graves OR Thyronine*):ab,ti) AND (‘psoriasis’/exp OR ‘psoriatic arthritis’/de OR (psoria* OR (pustulo* NEAR/3 palmoplantar*)):ab,ti)
We searched for published studies that satisfied the following criteria: (1) Studies comparing patients with psoriatic disease (defined as psoriasis, PsA or pustulosis palmoplantaris (PPP)) to healthy controls and (2) studies reporting effect estimates (risk ratio, odds ratio or hazard ratio) for the association of hyperthyroidism, hypothyroidism or thyroid autoimmunity (Graves’, Hashimoto and TPO-Ab positivity) with psoriatic disease or (3) studies providing mean values with standard deviations (s.d.’s) for thyroid function measurements (T3, FT3, T4, FT4, TSH). We excluded case reports, case series, animal studies or studies, in subjects younger than 18 years. We also excluded studies that only looked at the effect of thyroid function altering medication on psoriatic disease or studies that compared thyroid hormone levels in patients before and after treatment for psoriatic disease. We did include studies in which, from the provided information, risk estimates or other measures could be calculated.
Two independent reviewers (SRK, AB) screened the titles and abstracts of the eligible articles. Full-text articles were retrieved for studies that met the inclusion criteria. The same eligibility criteria were applied to the full-text articles. The inter-reviewer agreement was determined by calculating the kappa-statistic (κ). The agreement was very good for abstract and title (κ = 0.80) and good for full-text screening (κ = 0.77). In case of discrepancy, consensus was sought or a third independent reviewer (LC) was consulted.
Data collection and quality assessment process
Standardized data extraction forms were used to extract relevant information on article source, study methods, demographics of included patients and controls, and the outcomes and conclusion of the included articles, among others (Table 2). SRK assessed the quality of the included studies using the Newcastle–Ottawa Scale (NOS) for non-randomized studies in meta-analyses (30). AB cross-checked a random proportion of the data and there were no disagreements.
Description, characteristics and results of included studies on the association between thyroid (dys)function and psoriatic disease.
|Reference||Country||Totalna||Age (mean)||% Female||n; type of cases||Odds ratio (95% CI)||TSH (mU/L), mean (s.d.)||(F)T4 (ng/dL), mean (s.d.)||(F)T3 (pg/mL), mean (s.d.)|
|TPO-Ab positivity||Hypothyroidism||Hyperthyroidism||Covariates adjusted for||Cases||Controls||Cases||Controls||Cases||Controls|
|(31)||Italy||426||39.6b||48.4b||108; PsA||2.89 (1.61–5.19)c||2.92 (1.67–5.11)c||NA||Age||NA||NA||NA||NA||NA||NA|
|(24)||Italy||480||57.8b||45b||80; PsA||3.60 (1.82–7.10)c||2.24 (1.03–4.88)b,c||NA||Age and sex||1.79 (2.08)||1.38 (1.41)||FT4: 0.96 (0.25)||FT4: 1.00 (0.20)||FT3: 2.46 (0.60)||FT3: 2.76 (0.92)|
|(21)||Iran||60||NR||NR||30; psoriasis||NA||NA||NA||Age and sex||2.82 (1.96)||2.65 (1.63)||T4: 9.48 (1.65) mg/dLe||T4: 8.93 (1.19) mg/dLe||T3: 105.30 (19.99) ng/dLf||T3: 106.90 (23.84) ng/dLf|
|(34)||Italy||94||56.1b,d||52.2b||42; PsA||NA||NA||NA||NR||1.4 (0.9)||1.8 (0.9)||FT4: 1.12 (0.32)g,h||FT4: 0.95 (0.23)g,h||FT3: 3.71 (0.91)i||FT3: 3.39 (0.98)i|
|(33)||USA||5553||48.9b||50.7b||162; psoriasis||NA||1.16 (NR)||1.29 (NR)||Age, sex, ethnicity, alcohol, smoking, BMI||1.76 (NR)||1.88 (NR)||FT4: 0.864 (NR)||FT4: 0.845 (NR)||FT3: 3.18 (NR)||FT3: 3.14 (NR)|
|(32)||Taiwan||2,59,000||46.4b||38.4b||51 800; PsA/psoriasis||NA||1.28 (1.02–1.60)b,c||1.24 (1.05–1.46)b,c||Age, sex, urbanization level of residential area||NA||NA||NA||NA||NA||NA|
|(22)||USA||1,52,046||48.9b||48.4b||25 341;PsA/psoriasis||NA||1.22 (0.97–1.53)b||1.09 (0.90–1.32)b||Age, sex and length of enrollment||NA||NA||NA||NA||NA||NA|
Total number of cases and controls of interest for this particular research question; bCalculated from provided study data; cP-value <0.05; dMedian age; e1 mg/dL = 1 × 106 ng/dL; f1 ng/dL = 10 pg/mL; gP-value for difference <0.01; hFT4 converted from pmol/L: 1 pmol/L = 0.0777 ng/dL; iFT3 converted from pmol/L: 1 pmol/L = 0.6510 pg/mL.
BMI, body mass index; NA, not applicable; NR, not reported; PsA, psoriatic arthritis; TPO-Ab positivity, thyroid peroxidase antibody positivity; TSH, thyroid stimulating hormone; (F)T4, (free) thyroxine; (F)T3, (free) triiodothyronine; 95% CI, 95% confidence interval.
We used ORs with their 95% CI as provided by the included articles, or calculated these manually using the available information in the studies. For the calculations of the pooled estimates, we also added our own data on prevalent psoriatic disease to the meta-analysis. We used our minimally adjusted model, adjusting for age, sex and cohort, in line with the adjustments of the included articles (22, 24, 31, 32). Due to the small number of studies obtained, a fixed effect model with the Mantel-Haenszel method was used to obtain a pooled OR with 95% CI for the association of TPO-Ab positivity and AITD (consisting of the subgroups of hypothyroidism and hyperthyroidism) with psoriatic disease. Most studies did not distinguish between different causes of hypothyroidism and hyperthyroidism, but autoimmunity being the most important cause in general, we combined all cases of hypothyroidism and hyperthyroidism to AITD. We also constructed forest plots to show the association and obtained I2 statistic for heterogeneity. We assessed possible publication bias, visually (funnel plots) and statistically (Egger test).
In order to further assess heterogeneity, we conducted sensitivity analyses using a random effect model, and excluding studies in a non-Caucasian population, studies with <100 events, studies with hospital controls, studies that only looked at PsA as outcome, and studies that were outliers in the funnel plots. Furthermore, we stratified by mean age and percentage women of the studies. We also performed a subgroup analysis by gender. Meta-analyses were performed in Review Manager (RevMan), version 5.3. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014, except for the Egger test, which was performed in R (metafor package, R-project, Institute for Statistics and Mathematics, R Core Team (2016), Vienna, Austria, version 3.2.5).
We included 8214 study participants with baseline measurements of thyroid function, outcome data on psoriatic disease and who had given informed consent for follow-up. The mean age was 62.3 years (s.d.: 8.4) and 55.9% of the participants were female. Baseline characteristics are shown in Table 3. There were a total of 173 prevalent cases of psoriatic disease (158 psoriasis, 3 PsA and 12 both) and 124 incident psoriatic cases (psoriasis only) that occurred during a median follow-up duration of 10.2 years (interquartile range (IQR): 4.4–15.4). The incidence rate of psoriasis was 16 per 10 000 person-years.
Baseline characteristics of the 8214 Rotterdam Study participants with TSH and FT4 measurements.
|Start age, years, mean (s.d.)||62.3 (8.4)|
|Sex, female||4591 (55.9)|
|TPO-antibody positivity||1077 (13.1)|
|TSH, U/L, median (IQR)||1.9 (1.3–2.8)|
|Ln(TSH), mean (s.d.)||0.6 (0.8)|
|FT4, ng/dL, mean (s.d.)||1.2 (0.2)|
|Alcohol intake, g/day, median (IQR)||15.0 (1.4–20.0)|
|BMI, kg/m2, mean (s.d.)||27.2 (4.2)|
|Diabetes Mellitus||933 (11.4)|
|Total serum cholesterol, mmol/L, mean (s.d.)||5.7 (1.0)|
|Serum CRP level, mg/mL, median (IQR)||4.3 (1.9–8.2)|
Unless stated otherwise.
BMI, body mass index; CRP, C-reactive protein; FT4, free thyroxine; IQR, interquartile range; Ln(TSH), the natural logarithm of TSH; s.d., standard deviation; TPO, thyroid peroxidase; TSH, thyroid stimulating hormone.
In the cross-sectional analyses, there was no association of TPO-Ab positivity or FT4 with prevalent psoriatic disease. Our results show a positive trend for the association between TSH and prevalent psoriatic disease, but not significantly (Table 4). We did not find an association of hypothyroidism and hyperthyroidism with prevalent psoriatic disease (Table 4). Longitudinally, there was no association of TPO-Ab positivity and TSH with psoriasis incidence. Higher FT4 levels seemed to increase the risk of psoriasis, but this did not reach statistical significance in the multivariable analysis (HR: 2.19; 95% CI: 0.89–5.38) (Table 4).
Association between thyroid function and psoriatic disease within the Rotterdam Study.
|Variable||Events/subjects (N)a||Model 1b||Model 2c|
|TPO-Ab positivity||173/8204||1.04 (0.66–1.62)||1.03 (0.66–1.62)|
|TSH (U/L)||173/8210||1.21 (0.98–1.49)||1.22 (0.99–1.51)|
|FT4 (ng/dL)||173/8210||1.13 (0.49–2.59)||1.11 (0.48–2.57)|
|Hypothyroidisme||173/8214||1.23 (0.76–1.97)||1.26 (0.78–2.02)|
|Hyperthyroidismf||173/8214||1.06 (0.43–2.60)||1.05 (0.42–2.60)|
|TPO-Ab positivity||124/8033||1.15 (0.69–1.91)||1.16 (0.70–1.93)|
|TSH (U/L)||124/8039||1.09 (0.87–1.36)||1.12 (0.90–1.41)|
|FT4 (ng/dL)||124/8039||2.46 (1.01–5.97)g||2.19 (0.89–5.38)|
Subjects are the study participants with baseline TSH and/or FT4 measurements for the cross-sectional analyses and with baseline TSH and/or FT4 measurements and no history of psoriatic disease for the longitudinal analyses.
Model 1 is adjusted for age, sex and cohort.
Model 2 is adjusted for model 1, alcohol consumption, smoking status, body mass index, diabetes mellitus, hypertension, serum cholesterol, serum C-reactive protein and socioeconomic status.
The shown Odds Ratios and Hazard Ratios are for TPO-Ab positivity compared to TPO-Ab negativity, per unit increase of log-transformed TSH, per unit increase of FT4, hypothyroidism compared to no hypothyroidism, or hyperthyroidism compared to no hyperthyroidism.
Hypothyroidism is defined as TSH values >4 U/L and FT4 <25 pmol/L (<1.94 ng/dL).
Hyperthyroidism is defined as TSH values <0.4 U/L and FT4 >11 pmol/L (>0.85 ng/dL).
95% CI, 95% confidence interval; FT4, free thyroxine; TPO-Ab positivity, thyroid peroxidase antibody positivity; TSH, thyroid stimulating hormone.
Systematic review of the literature and meta-analysis
We identified 1850 articles after removing duplicates, of which we excluded 1778 articles based on title and abstract. We screened the full text of the remaining 72 articles and excluded 65 articles because they either did not meet our eligibility criteria, or because full texts were not available (not even after contact with first and corresponding authors of these particular articles; n = 6). Seven articles were included in the systematic review and four articles could also be included in the meta-analysis (Fig. 1).
Seven case–control studies were included with a total of 230 PsA patients, 192 psoriasis patients, 77 141 patients with PsA and/or psoriasis and 340 096 controls. Three of the included studies were performed in Italy and the others were from Iran, Taiwan or the USA (Table 2). Furthermore, four studies (21, 24, 33, 34) provided means with or without s.d.s for thyroid function measurements. Three studies could not be included in the meta-analysis, because effect estimates (21, 34) or information necessary to obtain CIs for ORs (33) were not provided. We therefore limited these studies to the qualitative synthesis (systematic review).
Of the included studies in the systematic review that provide information on thyroid function serum measurements, only Bianchi et al. (34) report a difference in mean thyroid hormone levels, with FT4 being higher in PsA patients compared to their controls (Table 2). Of the studies reporting effect estimates, two show a positive association of TPO-Ab positivity with psoriatic disease, three show a positive association of hypothyroidism with psoriatic disease and only Tsai et al. report a positive association of hyperthyroidism. Wu et al. show an increased odds of hypothyroidism and hyperthyroidism in psoriatic patients, although not significantly. Lai et al. also report increased odds for hypothyroidism and hyperthyroidism, but do not report a 95% CI (Table 2).
However, both the definition of thyroid autoimmunity as well as that of psoriatic disease was different in each study. While Peluso et al. have a cut-off value of >34 IU/mL for TPO-Ab positivity (31), similar to our cut-off point, Antonelli et al. employ a threshold of >100 IU/mL (24). Furthermore, only Peluso et al. included hospital controls instead of community controls, possibly leading to selection bias (31). Although Peluso et al. and Antonelli et al. included only patients with PsA in a tertiary care setting (24, 31), Tsai et al. performed a nationwide study in a solely Asian population (32). Also, Tsai et al. and Wu et al. did not include thyroid hormone measurements, but used record linkage with ICD-9-CM codes to assess the association of thyroid disease with psoriatic disease (22, 32).
Of the four articles included in the meta-analysis, two provide information on the association between psoriatic disease and TPO-Ab positivity (24, 31). Both report a higher odds for TPO-Ab positivity in patients with PsA compared to their controls. After adding RS data, the pooled OR for TPO-Ab positivity was 1.71 (95% CI: 1.27–2.31) with a substantial amount of heterogeneity (I2: 85%; Fig. 2A).
All included articles report on the association between (subclinical) hypothyroidism and psoriatic disease, whereas only Tsai et al. and Wu et al. report on the association with hyperthyroidism (22, 32). We combined hypothyroidism and hyperthyroidism to AITD. All included articles show a positive association of AITD with psoriatic disease. Although we did not find a significant association within the RS, after pooling, the odds of AITD was increased in subjects with psoriatic disease (OR: 1.25; 95% CI: 1.14–1.37), with 66% heterogeneity (Fig. 2B). The subgroup analyses also showed a positive association between hypothyroidism and psoriatic disease (OR: 1.34; 95% CI: 1.16–1.54) and hyperthyroidism and psoriatic disease (OR: 1.17; 95% CI: 1.03–1.32; Fig. 2B). Overall, effect estimates did not change in the stratified or sensitivity analyses; however, the included number of studies was too small to render meaningful results (Table 5).
Stratified and sensitivity analyses of the association between TPO-Ab positivity or AITDa and psoriatic diseaseb.
|Pooled OR (95% CI)||No. of studies||P for heterogeneityc||Pooled OR (95% CI)||No. of studies||P for heterogeneityc|
|Random effect model||2.13 (0.93–4.89)||3||0.001||1.38 (1.11–1.71)d||5||0.02|
|≥50||1.41 (0.99–2.01)||2||0.002||1.32 (0.91–1.91)||2||0.14|
|<65||1.71 (1.27–2.31)d||3||0.001||1.25 (1.14–1.37)d||5||0.02|
|In a non-Caucasian population||NA||NA||NA||1.25 (1.10–1.43)d||4||0.008|
|With <100 events||1.45 (1.04–2.04)d||2||0.005||1.24 (1.13–1.37)d||4||0.02|
|With hospital controls||1.41 (0.99–2.01)||2||0.002||1.22 (1.11–1.34)d||4||0.43|
|With only PsA as outcome||NA||NA||NA||1.21 (1.10–1.33)d||3||0.80|
|As outlier in funnel plot||NA||1||NA||1.22 (1.11–1.34)d||4||0.43|
|<50||3.15 (2.02–4.91)d||2||0.63||1.26 (1.14–1.38)d||4||0.009|
|Stratified by gender|
|Male||1.73 (0.92–3.26)||2||0.007||1.15 (0.59–2.23)||2||0.44|
|Female||1.33 (0.85–2.06)||2||0.06||1.44 (0.91–2.27)||2||0.03|
The overall quality of the included studies was moderate to high on the NOS assessment scale. One study scored eight out of nine stars, one scored seven out of nine stars, four studies scored six out of nine stars and the remaining study scored five out of nine stars (Table 6).
Evaluation of publication bias
The funnel plot for the association with TPO-Ab positivity showed two borderline outliers, and the Egger test indicated significant funnel plot asymmetry (P = 0.0004). The funnel plot for the association with AITD showed one clear outlier (30) (Fig. 3). Exclusion of this study did not change the effect estimate, but it did remove heterogeneity (Table 5). The Egger test for AITD showed publication bias (P = 0.0110).
To our knowledge, this is the first prospective population-based cohort study, as well as the first systematic review and meta-analysis of studies on the association between AITD/thyroid function and psoriatic disease. In our prospective cohort study, we did not find an association of AITD and thyroid function parameters with psoriatic disease. Still, there was a positive trend of higher odds of prevalent psoriatic disease with higher TSH levels and higher incidence of psoriatic disease with higher FT4 levels, although not statistically significantly. In our meta-analysis (including the RS data), TPO-Ab positivity and AITD (hypothyroidism and hyperthyroidism) were associated with an increased odds ratio of prevalent psoriatic disease. These results, however, should be interpreted with caution due to the small number of identified studies and the large (statistical and clinical) heterogeneity across the included studies.
There was no association of TPO-Ab positivity and thyroid function measurements with prevalent psoriatic disease in the RS, which may be explained by the fact that the majority of our participants suffered from mild psoriasis. The studies included in the meta-analysis that found the strongest association between AITD and psoriatic disease exclusively looked at PsA as outcome (24, 31). This is in line with a recent study investigating the occurrence of other autoimmune diseases in patients with autoimmune thyroiditis. The authors report a statistically significantly higher prevalence of PsA in patients with autoimmune thyroiditis as compared to controls. There was also a higher prevalence of psoriasis in general in patients with autoimmune thyroiditis, but this did not reach statistical significance (35). A difference between psoriasis and PsA patients has been found for other autoimmune diseases as well. Inflammatory bowel disease and celiac disease were stronger associated with PsA than with psoriasis, and giant cell arteritis and pulmonary fibrosis were only associated with PsA (36). A possible explanation could be stronger systemic inflammation in PsA patients compared to psoriasis alone. This is supported by higher levels of IL6, CD16+ proinflammatory monocytes, osteoprotegerin (a member of the tumor necrosis factor (TNF) receptor family), high-sensitive CRP and vascular endothelial growth factor (VEGF) in patients with PsA (37). However, in our meta-analysis, excluding studies with solely PsA as outcome did not change the pooled effect estimate (Table 5), even though the majority (between 80 and 90%) of the patients included in the remaining articles suffered from mild psoriasis only. In the longitudinal analyses, there was no association of TPO-Ab positivity and TSH with psoriatic disease. There was an association between FT4 and incidence of psoriatic disease, which lost statistical significance in the multivariable model, possibly due to an insufficient number of incident psoriasis events. Larger studies are needed to confirm these findings.
We show a positive trend for the association of TSH and prevalent psoriatic disease, whereas in the longitudinal analyses, higher FT4 levels seem to increase the risk of psoriasis incidence. There are two possible explanations for this discrepancy between the cross-sectional and longitudinal analyses. Firstly, as hypothesized based on autoimmunity, both hypothyroidism and hyperthyroidism could be associated with psoriatic disease. Secondly, the associations found could be explained by the phenomenon of reverse causation. In cross-sectional analyses, the direction of temporality is unclear, making these analyses prone to reverse causation. We do not know for sure if hypothyroidism preceded the psoriatic disease or if psoriatic patients are more likely to develop hypothyroidism, e.g. through non-thyroidal illness (38). In longitudinal analyses, however, this phenomenon is unlikely to occur.
In contrast, the results of our meta-analysis do suggest that an association exists of TPO-Ab positivity and AITD with prevalent psoriatic disease. Furthermore, separately studying the association of hypothyroidism and hyperthyroidism with psoriatic disease yielded comparable results. However, there were large differences in outcome definition (psoriasis and/or PsA) and other study characteristics, making the included studies hard to compare. The included studies also used different definitions of AITD. Although Peluso et al. looked at Hashimoto thyroiditis based on auto-antibodies and thyroid ultrasound (31), Antonelli et al. looked at the association with subclinical hypothyroidism (24), and Wu et al. and Tsai et al. used ICD-9-CM codes for Hashimoto thyroiditis respectively unspecified acquired hypothyroidism (22, 32). For hyperthyroidism, Wu et al. used an ICD-9-CM code for Graves’ disease (22), while Tsai et al. used an ICD-9-CM code for thyrotoxicosis without mention of goiter or other cause (32). Furthermore, there were indications of small study effects (24, 31) and we were unable to assess the effect of AITD and thyroid function on incidence of psoriatic disease due to the unavailability of longitudinal data in the current literature.
Our study has several strengths. Most importantly, we were able to combine available reported evidence in a systematic review and meta-analysis. Furthermore, we are the first to conduct longitudinal analyses embedded in a large, prospective population-based cohort study. In our analyses within the RS, we were able to take a wide number of potential confounders into account. Our biggest limitation is the low number of studies with relatively few events included in the systematic review and meta-analysis, illustrating the scarcity of evidence on this topic and the unavailability of prospective data. Furthermore, due to the small number of included studies both visual as well as statistical exploration of publication bias was difficult to interpret. Also, there was a substantial amount of heterogeneity, both statistically and in study characteristics. However, all study specific effect estimates pointed in the same direction, and the quality of included studies was generally good. Furthermore, the studies included in the systematic review and meta-analysis adjusted for a limited number of potential confounders. Also, we were unable to longitudinally assess the association of hypothyroidism and hyperthyroidism with psoriatic disease in the RS, due to a small number of incident events in these subgroups. Lastly, we could not determine which patients with hypothyroidism or hyperthyroidism suffered from Hashimoto thyroiditis respectively Graves’ disease within the RS.
Current evidence on the association between thyroid (dys)function and psoriatic disease remains unclear, hindering investigation of the efficiency of screening for AITD in psoriatic patients. Although our meta-analysis suggests an association between AITD and prevalent psoriatic disease, further investigation is warranted. Furthermore, larger prospective studies are needed to assess the association of AITD and thyroid function with incidence of psoriatic disease.
AITD, TPO-Ab positivity, hypothyroidism and hyperthyroidism might be associated with prevalent psoriatic disease. More prospective studies are needed to further assess the association of AITD with incidence of psoriatic disease before a final conclusion can be made.
Declaration of interest
All authors declare no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
The R S is supported by the Erasmus MC and Erasmus University Rotterdam; the Netherlands Organization for Scientific Research (NWO); the Netherlands Organization for Health Research and Development (ZonMw); the Research Institute for Diseases in the Elderly (RIDE); the Netherlands Genomics Initiative (NGI); the Ministry of Education, Culture and Science; the Ministry of Health Welfare and Sports; the European Commission (DG XII); and the Municipality of Rotterdam. The funding sources had no involvement in the collection, analysis, writing, interpretation, or in the decision to submit the paper for publication.
The authors are grateful to the study participants, the staff from the R S, and participating general practitioners and pharmacists. They would like to thank Wichor M Bramer, biomedical information specialist at the Erasmus MC, for his essential contribution to the online literature search.
Garcia-LopezMASanchoDSanchez-MadridFMarazuelaM. Thyrocytes from autoimmune thyroid disorders produce the chemokines IP-10 and Mig and attract CXCR3+ lymphocytes. Journal of Clinical Endocrinology and Metabolism 2001 86 5008–5016. (doi:10.1210/jcem.86.10.7953)
RomagnaniPRotondiMLazzeriELasagniLFrancalanciMBuonamanoAMilaniSVittiPChiovatoLTonaccheraM Expression of IP-10/CXCL10 and MIG/CXCL9 in the thyroid and increased levels of IP-10/CXCL10 in the serum of patients with recent-onset Graves’ disease. American Journal of Pathology 2002 161 195–206. (doi:10.1016/S0002-9440(10)64171-5)
FallahiPFerrariSMRuffilliIEliaGBiricottiMVitaRBenvengaSAntonelliA. The association of other autoimmune diseases in patients with autoimmune thyroiditis: review of the literature and report of a large series of patients. Autoimmunity Reviews 2016 15 1125–1128. (doi:10.1016/j.autrev.2016.09.009)
MakredesMRobinsonDJrBalaMKimballAB. The burden of autoimmune disease: a comparison of prevalence ratios in patients with psoriatic arthritis and psoriasis. Journal of the American Academy of Dermatology 2009 61 405–410. (doi:10.1016/j.jaad.2009.02.015)