If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Medicine, School of Medicine, University of California San Francisco, San Francisco, California, USADepartment of Dermatology, University of California San Francisco, San Francisco, California, USA
Correspondence: Wilson Liao, Department of Dermatology University of California San Francisco, 2340 Sutter Street, Box 0808, San Francisco, California 94143, USA.
Department of Dermatology, University of California San Francisco, San Francisco, California, USAInstitute for Human Genetics, University of California San Francisco, San Francisco, California, USA
Psoriasis (PSO) and psoriatic arthritis (PSA) are inflammatory diseases with complex genetic and environmental contributions. Although studies have identified environmental and clinical associations with PSO/PSA, causality is difficult to establish. Mendelian randomization (MR) employs the random assortment of genetic alleles at birth to evaluate the causal impact of exposures. We systematically reviewed 27 MR studies in PSO/PSA examining health behaviors, comorbidities, and biomarkers. Exposures, including smoking, obesity, cardiovascular disease, and Crohn’s disease, were causal for PSO and PSA, whereas PSO was causally associated with several comorbidities. These findings provide insights that can guide preventive counseling and precision medicine.
Psoriasis (PSO) is an immune-mediated skin condition that affects 125 million individuals globally, with an estimated 3.0% prevalence among United States adults (
). Many patients with PSO have multiple comorbidities, including cardiometabolic and mental health conditions, and up to 30% are diagnosed with psoriatic arthritis (PSA) (
). The ability to understand whether behaviors, comorbidities, or biomarkers increase the risk for PSO/PSA or vice versa is central to advancing personalized preventive and therapeutic interventions.
Mendelian randomization (MR) is a research method that can be applied to observational data to investigate causality. MR analyzes whether genetic variants such as SNPs associated with variables of interest affect health outcomes (e.g., development of PSO/PSA) (
). MR allows for evaluation of causation without requiring a randomized controlled trial and overcomes major limitations from observational studies related to unmeasured confounding, ascertainment bias, and small sample sizes (
). For instance, although common epidemiological techniques may uncover an association between smoking and PSO diagnosis, confounders may exist that modify this association, and measurement error may lead to residual confounding even after statistical adjustment, precluding determination of causality (
). In contrast, MR relies on random assortment of genetic variants at birth to determine whether the risk of developing a disease is contingent on genetic liability to a risk factor (
). Because of these advantages, MR-based studies have gained traction over the past 5 years as a method to conduct high-quality investigations identifying risk factors for diseases.
MR uses summary statistics aggregating significance and association data for every genetic variant analyzed in published GWASs, in which SNPs associated with a risk factor of interest are identified from an exposure dataset. These instrumental SNPs are then measured in a separate GWAS dataset for an outcome (e.g., PSO) using two-sample MR to determine the relationship between exposure and disease. MR analyses rest on three assumptions (
) (Supplementary Figure S1): (i) a robust association exists between SNPs and the exposure variable; (ii) SNPs are associated with the outcome only through the exposure variable, minimizing horizontal pleiotropy; and (iii) SNPs are not associated with confounders, evaluated through sensitivity analyses.
This systematic review summarizes areas of consensus and disagreement identified by MR studies regarding risk factors for developing PSO and PSA. A total of 27 MR studies relating to PSO, PSA, or both are discussed. We aim to provide a timely synthesis of MR results, placed within the context of biological insights, to identify the risk factors and biomarkers causally associated with the development of PSO/PSA or comorbidities to which PSO/PSA causally contribute.
Results
Alcohol and smoking
Lifestyle factors, including alcohol and smoking, are implicated as triggers for PSO. Although the effects of alcohol consumption on PSO incidence remain unclear (
), the evidence for smoking and PSO development is better established. A previous meta-analysis suggested that PSO is associated with both current and former smoking (
Association between psoriatic disease and lifestyle factors and comorbidities: cross-sectional analysis and Mendelian randomization [e-pub ahead of print].
). In all these studies, alcohol consumption was evaluated by the number of alcoholic drinks consumed per week, where one SD equated to nine drinks per week. Genetic predisposition to increased weekly alcohol intake showed no significant association with disease in three studies. Interestingly, one investigation conducted with a smaller cohort of 7,554 Spaniards found that alcohol as a diet category was causally affiliated with reduced PSO development (OR = 0.87) (
Association between psoriatic disease and lifestyle factors and comorbidities: cross-sectional analysis and Mendelian randomization [e-pub ahead of print].
Variables significantly associated with PSO and/or PSA.
, Type 2 DM, uveitis
UK Biobank (456,426 BMI cases): UK. Published GWAS (246,363 depression cases; 561,190 controls): UK. Published GWAS (1,131,881 education cases): Europe. Published GWAS (47,309 HF cases; 930,014 controls): Europe. Published GWAS (12,160 IBD cases; 13,145 controls): Europe. Published GWAS (1,232,091 lifetime smoking exposure, smoking status): Europe, US
Disease consequences of higher adiposity uncoupled from its adverse metabolic effects using Mendelian randomisation [published correction appears in Elife 2022;11:e80233].
BMI, CAD, HDL, LDL, Type 2 DM, total cholesterol, triglyceride level, waist-hip ratio
GIANT consortium (806,834 BMI cases; 697,734 waist-hip ratio cases): UK. Kaiser Permanente health system (94,674 HDL, LDL, total cholesterol, and triglyceride level cases): US
Childhood body size at age 10 years → PSA: 2.18 (1.43‒3.31). Adult body size → PSA: 1.64 (1.18‒2.29). Childhood body size at age 10 years → PSO: 1.39 (1.06‒1.82). Adult body size → PSO: 2.23 (1.78‒2.80).
Childhood body size at 10 years → PSA: 2.75 × 10−4. Adult body size → PSA: 3×10−3. Childhood body size at 10-years-old → PSO: 1.7 × 10−2. Adult body size → PSO: 4.96 × 10−12.
Investigating the causal relationship of C-reactive protein with 32 complex somatic and psychiatric outcomes: a large-scale cross-consortium Mendelian randomization study.
Abbreviations: 25-OHD, 25-hydroxycholecalciferol; Afib, atrial fibrillation; BMI, body mass index; CAD, coronary artery disease; CD, Crohn’s disease; CI, confidence interval; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; DBP, diastolic blood pressure; DM, diabetes mellitus; GERD, gastroesophageal reflux disease; GLIDE, Gene-Lifestyle Interactions in Dental Endpoints; GSCAN, GWAS and Sequencing Consortium of Alcohol and Nicotine; GSMR, Generalized Summary data-based Mendelian Randomization; HbA1c, hemoglobin A1c; HEIDI, heterogeneity in dependent instruments; HF, heart failure; HDL, high-density lipoprotein; HMGCR, 3-hydroxy-3-methylglutaryl-CoA reductase; HTN, hypertension; HUNT, Nord-Trondelag Health Study (Norway); IBD, irritable bowel disease; IV, instrumental variable; IVW, inverse variance-weighted; MBE, modes-based estimate; MHC, major histocompatibility complex; MI, myocardial infarction; MR, Mendelian randomization; MRAPS, Mendelian randomization-Robust Adjusted Profile Score; MR-PRESSO, Mendelian randomization Pleiotropy RESidual Sum and Outlier; NAFLD, nonalcoholic fatty liver disease; NSPHS, Northern Sweden Population Health Study; PD, Parkinson disease; PSA, psoriatic arthritis; PSO, psoriasis; PV, psoriasis vulgaris; RAPS, robust adjusted profile score; SBP, systolic blood pressure; SHBG, sex hormone-binding globulin; sIL-6R, serum interleukin-6 receptor; TSLS, two-staged least squares; UC, ulcerative colitis; UK, United Kingdom; US, United States.
This table summarizes the exposure variable(s), outcome variable(s), description of the dataset(s) used, number of SNPs used as instrumental variables, MR analysis techniques, ORs, and P-values used in all the 27 MR studies. Studies are organized into sections (alcohol and smoking; obesity, metabolic syndrome, dietary intake; other medical conditions; serum biomarkers) and ordered alphabetically within sections by the first author. Only exposure‒outcome pairings with significant ORs and P-value results are explicitly listed in the rightmost two columns; nonsignificant exposure/outcome pairings are not listed. Only results from the primary mode of MR analysis technique (bolded text, e.g., IVW) are described in the OR and P-value columns.
1 Inputted as the number of SNPs for the exposure and outcome variable(s), when reported by the study authors.
2 Indicates genome-wide significance P-value.
3 Variables significantly associated with PSO and/or PSA.
Two MR studies evaluated multiple aspects of smoking behavior, including initiation (ever smoking), quantity (cigarettes per day), cessation (current or former smoker), and lifetime use (in relation to population-based averages) (
Association between psoriatic disease and lifestyle factors and comorbidities: cross-sectional analysis and Mendelian randomization [e-pub ahead of print].
Association between psoriatic disease and lifestyle factors and comorbidities: cross-sectional analysis and Mendelian randomization [e-pub ahead of print].
). With PSO, smoking initiation showed an OR ranging from 1.34 to 1.48, and lifetime smoking exposure showed an OR ranging from 2.14 to 2.56; smoking quantity and smoking cessation were associated with ORs of 1.63 and 1.46, respectively. Thus, smoking appears to consistently predict PSO incidence, and certain features—including smoking more than one SD above that of the average population—are associated with over twofold odds of developing PSO. However, smoking for any period does not appear to be a predictor of PSA development.
Obesity, metabolic syndrome, atherosclerosis, and dietary intake
Obesity and diet have been previously associated with PSO incidence. Patients diagnosed with PSO have a higher prevalence of obesity; furthermore, individuals with severe PSO phenotypes have greater odds of obesity than those with milder PSO symptoms (
). More recent studies have focused on the impact of specific dietary intake factors such as saturated fatty acids and exposure to high-sugar diets or other inducers of gut dysbiosis in causing PSO/PSA (
). These modifiable risk factors fall under the larger umbrella of metabolic dysregulation and metabolic syndrome, including hypertension, obesity, insulin resistance, and dyslipidemia (
Ten MR studies investigated whether various aspects of metabolic dysregulation were causally associated with the development of PSO or PSA (Table 1). Studies were conducted using outcome datasets involving a mean (SD) of 6,051 (5,366) patients with PSO/PSA. Five studies investigated the relationship of body mass index (BMI) with PSO, which was found to be causally associated in all the five studies, with an OR ranging from 1.09 to 1.59 (
Association between psoriatic disease and lifestyle factors and comorbidities: cross-sectional analysis and Mendelian randomization [e-pub ahead of print].
). Findings were validated across multiple regional populations, including the well-established UK Biobank, a Norwegian GWAS dataset, and a Japanese biobank (
). Two additional studies investigated the role of excess fat distribution as a risk factor for the development of psoriatic disease. One study of PSO examined favorable and unfavorable adiposity, where fat distribution in subcutaneous and visceral adipose tissue as well as ectopic liver and pancreatic fat were taken into consideration (
Disease consequences of higher adiposity uncoupled from its adverse metabolic effects using Mendelian randomisation [published correction appears in Elife 2022;11:e80233].
). This investigation found that unfavorable adiposity was associated with 2.11 greater odds of developing PSO. Another study investigating childhood and adult body sizes found that increased body size at age 10 years, in addition to body size at maturity, was a highly significant predictor of both PSO and PSA development (
). The study was strengthened by the validation of the body size genetic instruments in three independent populations. Intriguingly, the OR of childhood body size was 1.39 and 2.18 for predicting PSO and PSA, respectively, whereas the OR of adult body size was 2.23 and 1.64 for predicting PSO and PSA, respectively. These results suggest that adult body size confers one of the highest risks for PSO development among all modifiable risk factors investigated using MR, whereas childhood body size is the most important modifiable risk factor for PSA occurrence (Figure 1). Furthermore, the risk of developing rheumatic and skin manifestations in PSO/PSA persists after adjustment for adult body size, suggesting that risk because of childhood body size may not be fully reversible (
Figure 1Summary of MR studies in psoriasis and psoriatic arthritis. Among published MR studies, psoriasis and psoriatic arthritis were evaluated as both outcome variables and exposure variables. Outcome variables are ranked by increasing OR, with green arrows indicating OR < 1 (protective association), red arrows indicating OR > 1 (increased association), and gray arrows indicating uncertainty between multiple MR studies. Exposure or outcome variables found to have no significance in relation to psoriasis or psoriatic arthritis are listed under each respective section. Numbers in parentheses indicate the number of MR analyses that were conducted for each respective exposure; an asterisk indicates that bidirectional MR analysis was performed for that exposure variable. BMI, body mass index; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; GERD, gastroesophageal reflux disease; HbA1c, hemoglobin A1c; HDL, high-density lipoprotein; IBD, inflammatory bowel disease; LDL, low-density lipoprotein; MR, Mendelian randomization; NAFLD, nonalcoholic fatty liver disease; PD, Parkinson disease.
noted a positive relationship between type 2 diabetes mellitus (T2DM) and PSO, which remained after adjusting for BMI (adjusted OR = 1.35). A separate study identified an association between blood sugar and PSO in European but not Japanese subpopulations (
). Three studies investigating dyslipidemia, including high-density lipoprotein (HDL), low-density lipoprotein (LDL), total cholesterol, triglyceride levels, and 3-hydroxy-3-methylglutaryl-coA reductase inhibition as a proxy for statin use, presented results with lack of concordance (
). Finally, one study found that coronary atherosclerosis was consistently associated with an increased risk of developing PSO (OR = 1.14), a significant association that remained after adjustment for possible metabolic confounders (BMI, waist‒hip ratio, and cholesterol levels) (
With diet, 14 food groups were investigated in a smaller Spanish population, for which fruit was found to be significantly protective for PSO (OR = 0.89) but not for PSA development (
). An additional study investigating the role of vitamin D (25-hydroxycholecalciferol [25-OHD]) in PSO showed that a genetically predicted one SD increment in circulating 25-OHD level was associated with a 24% decreased risk of PSO (OR = 0.76) (
In summary, MR studies strongly indicate that increased body mass and adiposity are causally associated with increased risk for PSO and PSA. The presence of coronary artery disease and T2DM may confer a greater risk for PSO, whereas increased circulating 25-OHD and consumption of fruit may be protective.
Other medical comorbidities
There is interest in using MR to understand whether the risk of developing PSO/PSA can be predicted by the presence of other medical conditions and whether PSO/PSA increases the risk of developing other diseases.
Two MR analyses investigated heart failure (HF) and its association with the development of PSO and PSA. In one cohort, HF of any etiology was significantly associated with susceptibility to PSA (OR = 1.79) but not with PSO development (
Association between psoriatic disease and lifestyle factors and comorbidities: cross-sectional analysis and Mendelian randomization [e-pub ahead of print].
). Interestingly, whereas this study found no significant reverse causal relationship between PSO/PSA and HF, a second MR study did find that PSO increased the susceptibility for multiple cardiac comorbidities, including HF (OR = 1.04), atrial fibrillation (OR = 1.04), myocardial infarction (MI) (OR = 1.07), and large artery stroke (OR = 1.11) (
Similar to PSO and PSA, inflammatory bowel disease (IBD), which includes Crohn’s disease (CD) and ulcerative colitis (UC) subtypes, is a chronic inflammatory disorder. IBD correlates with PSO/PSA prevalence, and common genetic susceptibility loci and shared immunologic features have been described (
Consideration of confounders, accuracy of diagnosis, and disease severity in assessing the risk of inflammatory bowel disease in patients with psoriasis and psoriatic arthritis/ankylosing spondylitis beginning interleukin-7 inhibitor treatment: comment on the article by Penso et al.
). Two independent MR studies examined IBD and PSO/PSA. One found that genetic predisposition to IBD was associated with an increased risk of PSO (OR = 1.13) and that CD was causally associated with both PSO (OR = 1.16) and PSA (OR= 1.14) (
). Similarly, a bidirectional MR analysis found that genetically predicted IBD was associated with a higher risk of PSO (OR = 1.10) and PSA (OR = 1.10) (
). This study also found that CD—but not UC—was causally associated with both PSO (OR = 1.16) and PSA (OR = 1.13). All sensitivity analyses assessing the independence of instrumental variables (IVs) as one of the three underlying assumptions of strong MR analyses found no notable directional pleiotropy or reverse directional causality (Table 1).
In addition, understanding the relationship between PSO and susceptibility to COVID-19 is particularly important for assessing COVID-19 risk in patients with PSO receiving immunosuppressive treatment. A bidirectional MR study found that COVID-19 was not significantly associated with the development of PSO but that the genetic risk of PSO was associated with increased susceptibility to COVID-19 (βinverse variance-weighted = 2.94, P = 0.01) (
). Five MR studies investigated how PSO/PSA may lead to increased susceptibility to depression, Parkinson disease (PD), Alzheimer’s disease, osteoporosis, and lung cancer (
Association between psoriatic disease and lifestyle factors and comorbidities: cross-sectional analysis and Mendelian randomization [e-pub ahead of print].
) or vice versa. One study found both clinically diagnosed and self-reported major depression to increase the risk of PSO development (OR = 1.41) but not that of PSA (
Association between psoriatic disease and lifestyle factors and comorbidities: cross-sectional analysis and Mendelian randomization [e-pub ahead of print].
). A separate analysis found PSO to increase the risk of susceptibility to PD, including the rate of PD progression as measured by Hoehn‒Yahr stage (OR = 1.05), depression (OR = 1.06), and dementia (OR = 1.07) (
). Similarly, neither PSO nor PSA was significantly associated with increased susceptibility to measures of osteoporosis, including estimated bone mineral density and fracture risk (
In summary, MR studies have identified several exposures that contribute to increased susceptibility for psoriatic disease—with depression mildly increasing the risk for PSO and HF for PSA development and IBD and CD (but not UC) strongly contributing to both PSO and PSA. Several comorbidities previously reported as psoriatic complications were more closely investigated, with MR analyses finding that PSO conferred a mildly increased risk for cardiovascular complications, COVID-19, PD, and lung cancer.
Serum biomarkers
No validated biomarkers are routinely used for the confirmatory diagnosis of PSO and PSA; current diagnostic processes focus on clinical presentation, supplemented by possible imaging or biopsy (
). Although promising-omics and metabolic markers have been identified, none are regularly used to predict disease progression and therapeutic response (
). The diagnostic need is particularly urgent for PSA, which presents heterogeneously; genetic markers proposed to date often exhibit high accuracy but low sensitivity (
Investigating the causal relationship of C-reactive protein with 32 complex somatic and psychiatric outcomes: a large-scale cross-consortium Mendelian randomization study.
). Data sources used included aggregated GWAS biobanks that incorporated individuals across Europe (e.g., United Kingdom, Finland, Sweden), similar to most PSO MR studies. In PSO, three circulating biomarkers were causally associated with the development of the disease: SDF-1α (OR = 0.59) (
). Sensitivity analyses using alternative MR methods (e.g., MR-Egger, MR Pleiotropy RESidual Sum and Outlier [MR-PRESSO]) were consistent with results from the primary method of analysis across all papers, showing the same direction of effect but not necessarily significance (Table 1). MR also identified three circulating biomarkers causally associated with the development of PSA: IL-12B (OR = 0.79) (
Investigating the causal relationship of C-reactive protein with 32 complex somatic and psychiatric outcomes: a large-scale cross-consortium Mendelian randomization study.
It is particularly important to assess the validity and independence of proposed serum biomarkers to draw robust conclusions given MR assumptions. For instance, genomic literature has reported an association between IL-12B and PSO/PSA susceptibility (
), which may reduce the validity of IL-12B as an independent variable. With both PSO and PSA, MR sensitivity testing of IL-12B uncovered minimal confounding, no significant pleiotropy in the data, and the maintenance of significance using differing P-value thresholds (
). Horizontal pleiotropy and sensitivity analyses were also performed with the four other circulating proteins; no significant heterogeneity or asymmetry was found. Notably, although higher IL-17 levels were associated with decreased risk of PSA, these results did not hold significance when using secondary MR analysis techniques (
In summary, five circulating biomarkers exhibited causal associations with the increased development of psoriatic disease: SDF-1α, IL-12B, and RANTES with PSO and IL-12B, IL-17, and CRP with PSA.
Education
Previous cross-sectional studies have shown that among patients with PSO, a variety of health literacy levels are found, which can influence engagement with healthcare providers (
). Furthermore, lower educational attainment (e.g., never reading books) has been associated with decreased odds of receiving biologic therapy in Italy, despite uniform access provided by the National Health System (
). A recent MR study sought to further define the relationship between education and PSO/PSA development in European-descent individuals. Education, as measured by age at completion of full-time education, was compared across 5-year increments to determine whether there was an associated change in disease development (
Association between psoriatic disease and lifestyle factors and comorbidities: cross-sectional analysis and Mendelian randomization [e-pub ahead of print].
). Educational attainment was a predictive factor against disease presence (OR for PSO = 0.67; OR for PSA = 0.75) and was the most important predictive risk factor for PSA in our overall analysis (Figure 1). Although these findings are striking, it is important to note that a source of potential bias in sensitivity analyses conducted by the authors is the presence of horizontal pleiotropy, in which SNP variants may influence the disease outcome through pathways separate from the exposure variable. Thus, it remains possible that the protective effects of higher educational attainment are mediated through other confounders, despite the fact that MR is more effective at minimizing such effects than other cross-sectional epidemiological methods.
Discussion
This systematic review examines the risk factors for PSO and PSA development as well as PSO/PSA as risk factors for other diseases derived using MR analyses (Figure 1). The goal of this review was to interpret MR results and highlight areas of agreement and disagreement while placing these findings into the context of previously conducted translational and epidemiological studies. A secondary aim is to identify the next steps in collaborative research and clinical care, with a special emphasis on prevention and early identification of factors that can contribute to PSO and PSA development.
First, we will discuss the exposure variables that were found in MR studies to increase susceptibility to PSO and PSA development. The identification of smoking and obesity but not alcohol as important risk factors for PSO clarifies previous controversies (
). Notably, across four studies utilizing varying GWAS outcome datasets, smoking was not found to be a significant risk factor for PSA development. For obesity, which exhibited a previously unclear association with PSO prevalence, childhood body size was found to have an outsized impact on PSA occurrence, which contrasts with the increased impact of adult body size on PSO development (
). Agreement across multiple MR studies using differing instrumental SNPs and GWAS from varying geographic regions further adds confidence to the obesity findings. These results have important implications for early intervention in preventative health visits. Clinically, counseling for weight loss and reduced smoking quantity or cessation even late in life may have a significant impact on reducing the risk for psoriatic development, whereas reduced alcohol intake may not lead to a significant reduction in PSO/PSA risk. Targeted counseling during pediatric visits regarding weight loss may reduce the odds of developing PSA and related complications by over twofold; further research studies should be conducted to examine how childhood health independently influences the development of cutaneous psoriatic disease.
With diet, markers of lipidemia—including HDL, LDL, and total cholesterol levels—were not found to increase the risk of psoriatic disease. However, MR results do suggest that increased circulating vitamin D levels may be protective against PSO development (
). These findings are intriguing because several studies have investigated the potential impact of systemic vitamin D supplementation in PSO but did not arrive at a consensus (
). Further investigation of 25-OHD’s protective effects may serve as promising avenues for preventative and therapeutic effects.
The relationships between systemic inflammatory diagnoses and psoriatic disease are particularly interesting. PSO and IBD share genetic correlations, and the presence of IBD alongside PSO can guide biologic therapy selection for treatment because certain agents can exacerbate one or the other condition (
). Across multiple GWAS cohorts and using bidirectional MR analyses, genetic predisposition to IBD or CD but not to UC was associated with an increased risk of both PSO and PSA (
). Reverse causation was not found. These findings enable the targeted capture of patients with IBD for personalized counseling and argue for improved disease control as a means of decreasing the risk of psoriatic development.
The MR studies described in this paper also support mild causal associations between PSO/PSA and additional medical conditions. Cardiovascular comorbidities are a well-known potential complication of PSO (
). MR analyses provide further evidence for this causal relationship because both PSO and PSA were associated with an increased risk of numerous cardiovascular complications, including MI and large artery stroke (
)—conditions with previously unclear associations with psoriatic disease, warranting further research.
Additional studies that integrate large clinical sequencing datasets to identify robust serum biomarkers are important in enabling earlier PSO detection and management. Five promising biomarkers associated with the underlying biology of PSO (SDF-1α, IL-12B, RANTES) and PSA (IL-12B, IL-17, CRP) have been identified using MR. In particular, CRP was identified through an exposure GWAS with larger sample sizes (over 194,000 patient cases) (
Investigating the causal relationship of C-reactive protein with 32 complex somatic and psychiatric outcomes: a large-scale cross-consortium Mendelian randomization study.
) and thus may be more resistant to variability in the exposure of interest. Markers such as SDF-1α are also intriguing because inhibition of this signaling axis has previously been shown to decrease the presence of macrophages and inflammatory angiogenesis (
). Further exploration of signaling pathways represented by the biomarkers highlighted in this review is warranted for the development of less invasive diagnostic testing and prognostication for PSO/PSA.
In the few cases of remaining controversy between MR studies (Figure 1), methodological discrepancies between studies can be identified. For instance, the derivation of different instrumental SNPs from smaller GWAS datasets may lead to less reliable results (
Davey Smith GD, Thompson SG, EPIC- InterAct Consortium. Using published data in Mendelian randomization: a blueprint for efficient identification of causal risk factors.
). Strategies such as implementing more stringent thresholds for significant SNP selection and using a positive control SNP when available can be implemented to better detect instrumental SNP performance (
Davey Smith GD, Thompson SG, EPIC- InterAct Consortium. Using published data in Mendelian randomization: a blueprint for efficient identification of causal risk factors.
). In addition, multivariable MR approaches can be explored. Multivariable MR allows for the assessment of SNP variants that are pleiotropic or associated with multiple related exposures (e.g., not only triglyceride levels but also HDL and LDL levels) (
). Using likelihood- and regression-based methods, multivariable MR can generate ORs predicting the causal impact of each individual exposure variable with greater confidence (
). Moving forward, we recommend that researchers include sufficient methodological detail for readers to assess the strategies implemented to ensure that MR assumptions were upheld, which were not standardized across all reviewed studies (Table 1). Additional strategies, including removing IVs associated with outcomes to reduce pleiotropy (e.g., utilizing MR-PRESSO for individual outlier removal and implementing contamination mixture methods to derive valid inferences despite invalid SNPs [
]) should also be considered to increase MR robustness.
The use of more heterogeneous GWAS study cohorts poses an important challenge. We observed that almost all cohorts used in psoriatic MR studies involved participants of European descent; these results might not be extrapolated to other populations and could be an important source of health inequity (
). However, MR statistical measurements can be invalidated by a mismatch between exposure and outcome data sources, especially because instrumental SNPs are derived from population-specific inheritance patterns (
). Studies in other clinical fields have addressed this concern by simultaneously performing trans-ancestry studies in which exposure‒outcome relationships are validated in different populations (
); this may be an important addition to future MR studies.
MR has immense potential to uncover the impact of health behaviors and other medical risk factors for the development of PSO and PSA. It is important to keep in mind the limitations and best practices for conducting and interpreting such studies. However, it is exciting to see the potential of MR in enabling precision medicine to better allow the recommendation of individualized behavioral, diagnostic, and therapeutic changes.
Materials and Methods
This systematic review was prospectively registered in the International Prospective Register of Systematic Reviews (number CRD42022357554). A literature search was performed on October 4, 2022 according to Preferred Reporting Items for Systematic Reviews and Meta-analyses guidelines (
) using the National Library of Medicine (PubMed), Embase, and Cochrane electronic databases; all studies published from database inception to October 4, 2022 were considered. Two search terms were used: Mendelian randomization and psoriasis and Mendelian Randomization and psoriatic arthritis.
Initial search yielded a total of 134 studies, 27 of which were ultimately included (Supplementary Figure S2). Articles were evaluated by three authors independently with the aid of Covidence systematic review software; discrepancies were resolved after a joint article review and discussion. Studies were included if they were observational case-control studies written in English and utilizing MR methods to investigate exposure variables associated with outcomes of PSO and/or PSA. Conference proceedings and non‒peer-reviewed articles were excluded. The strength of clinical data was evaluated using the Critical Appraisal Skills Program checklist for case-control studies (
TB is a principal investigator for trials sponsored by Abbvie, Castle, CorEvitas, Dermavant, Galderma, Mindera, and Pfizer. She has been an advisor for Abbvie, Arcutis, Boehringer-Ingelheim, Bristol Myers Squibb, Janssen, Leo, Lilly, Novartis, Pfizer, Sun, and UCB. WL has received research grant funding from Abbvie, Amgen, Janssen, Leo, Novartis, Pfizer, Regeneron, and TRex Bio. The remaining authors state no conflict of interest.
Acknowledgments
JQJ has received research grant funding from the National Psoriasis Foundation and the University of California San Francisco School of Medicine (San Francisco, CA). TB has received research grant funding from Novartis and Regeneron. WL has received research grant funding from Abbvie, Amgen, Janssen, Leo, Novartis, Pfizer, Regeneron, and TRex Bio.
Disease consequences of higher adiposity uncoupled from its adverse metabolic effects using Mendelian randomisation [published correction appears in Elife 2022;11:e80233].
Investigating the causal relationship of C-reactive protein with 32 complex somatic and psychiatric outcomes: a large-scale cross-consortium Mendelian randomization study.
Association between psoriatic disease and lifestyle factors and comorbidities: cross-sectional analysis and Mendelian randomization [e-pub ahead of print].
The CASP checklist for case-control studies was used to assess the risk and bias of included studies. Each study was appraised using the checklist and was awarded + for Yes, ‒ for No, and? for Cannot tell for each question on the checklist.
Supplementary Figure S2PRISMA diagram depicting the study selection process. PRISMA, Preferred Reporting Items for Systematic Reviews and Meta-Analyses.
Davey Smith GD, Thompson SG, EPIC- InterAct Consortium. Using published data in Mendelian randomization: a blueprint for efficient identification of causal risk factors.
Disease consequences of higher adiposity uncoupled from its adverse metabolic effects using Mendelian randomisation [published correction appears in Elife 2022;11:e80233].
Investigating the causal relationship of C-reactive protein with 32 complex somatic and psychiatric outcomes: a large-scale cross-consortium Mendelian randomization study.
Consideration of confounders, accuracy of diagnosis, and disease severity in assessing the risk of inflammatory bowel disease in patients with psoriasis and psoriatic arthritis/ankylosing spondylitis beginning interleukin-7 inhibitor treatment: comment on the article by Penso et al.
Association between psoriatic disease and lifestyle factors and comorbidities: cross-sectional analysis and Mendelian randomization [e-pub ahead of print].
02844-5&id=gr1.jpg){kind=link}
02844-5&id=fx1.jpg){kind=link}
02844-5&id=fx2.jpg){kind=link}