Risk of thromboembolism and fatal stroke in patients with psoriasis and nonvalvular atrial fibrillation: a Danish nationwide cohort study

Authors

  • O. Ahlehoff,

    Corresponding author
    1. Department of Cardiology, Copenhagen University Hospital Roskilde, Roskilde, Denmark
    2. Department of Cardiology, Copenhagen University Hospital Gentofte, Hellerup, Denmark
    • Correspondence: Ole Ahlehoff, MD, PhD, Department of Cardiology, Copenhagen University Hospital Roskilde, DK-4000 Roskilde, Denmark. (fax: +45-47-32-61-31; e-mail: olahha01geh.regionh.dk).

    Search for more papers by this author
  • G. Gislason,

    1. Department of Cardiology, Copenhagen University Hospital Gentofte, Hellerup, Denmark
    2. The Danish National Institute of Public Health, University of Southern Denmark, Copenhagen, Denmark
    Search for more papers by this author
  • M. Lamberts,

    1. Department of Cardiology, Copenhagen University Hospital Gentofte, Hellerup, Denmark
    Search for more papers by this author
  • F. Folke,

    1. Department of Cardiology, Copenhagen University Hospital Rigshospitalet, Copenhagen, Denmark
    Search for more papers by this author
  • J. Lindhardsen,

    1. Department of Cardiology, Copenhagen University Hospital Gentofte, Hellerup, Denmark
    2. Department of Rheumatology, Copenhagen University Hospital Rigshospitalet, Copenhagen, Denmark
    Search for more papers by this author
  • C. T. Larsen,

    1. Department of Cardiology, Copenhagen University Hospital Roskilde, Roskilde, Denmark
    Search for more papers by this author
  • C. Torp-Pedersen,

    1. Department of Health Science and Technology, Aalborg University, Aalborg, Denmark
    Search for more papers by this author
  • P. R. Hansen

    1. Department of Cardiology, Copenhagen University Hospital Gentofte, Hellerup, Denmark
    Search for more papers by this author

Abstract

Objectives

Psoriasis is a chronic inflammatory disease that is associated with a prothrombotic state and cardiovascular disease, including atrial fibrillation and thromboembolism. We therefore evaluated the impact of psoriasis in patients with atrial fibrillation and the performance of the CHA2DS2VASc score in these patients.

Design, setting and participants

The study comprised all Danish patients hospitalized with nonvalvular atrial fibrillation in the period 1997–2011 (n = 99 357). Follow-up started 7 days from discharge and excluded subjects treated with anticoagulation. Poisson regression adjusted for CHA2DS2VASc score was used to estimate the incidence rate ratios and 95% confidence intervals.

Main outcome measure

Hospitalization or death from thromboembolism.

Results

Mean follow-up was 3.5, 3.1, and 2.8 years for patients with no psoriasis, mild psoriasis and severe psoriasis, respectively. Patients with psoriasis were younger compared to patients without psoriasis, but CHA2DS2VASc score did not differ between the three groups. Thromboembolism rates per 100 patient-years (95% confidence intervals) were 4.8 (4.7–4.9), 4.8 (4.2–5.4) and 6.1 (5.0–7.5) for patients with no psoriasis, mild psoriasis and severe psoriasis, respectively. Importantly, the observed thromboembolism rates in patients with severe psoriasis were markedly higher (2.6- to3.4-fold) than predicted by the CHA2DS2VASc score. Relative to no psoriasis, incidence rate ratios were 0.99 (0.87–1.11) and 1.27 (1.02–1.57) for mild and severe psoriasis, respectively. Correspondingly, incidence rate ratios for fatal stroke were 0.97 (0.80–1.12) and 1.51 (1.12–2.05).

Conclusions

In patients with nonvalvular atrial fibrillation not treated with oral anticoagulation, severe psoriasis was associated with increased risk of thromboembolism. In these patients, CHA2DS2VASc underestimated the risk of thromboembolism.

Abbreviations
ATC

Anatomical therapeutic chemical

CHA2DS2VASc

congestive heart failure, hypertension, age ≥75 years (doubled risk weight), diabetes mellitus, history of stroke or thromboembolism (doubled risk weight), vascular disease, age 65–74 years, sex category (female)

ICD

International Classification of Diseases

NCSP

Nordic Medical Statistics Committees Classification of Surgical Procedures

STROBE

Strengthening the Reporting of Observational Studies in Epidemiology

TIA

transient ischaemic attack

Introduction

Atrial fibrillation is the most common type of cardiac arrhythmia and is associated with an approximately sixfold increase in stroke risk [1]. Psoriasis is a T helper (Th)-type 1/Th17 cell chronic inflammatory disorder that affects approximately 2% of the global population [2]. Of interest, it has been suggested that patients with psoriasis and other chronic inflammatory diseases have a prothrombotic disposition [3-6] and that those with severe psoriasis have an increased risk of cardiovascular diseases including atrial fibrillation, venous thromboembolism, myocardial infarction and stroke [7-11]. The risk of stroke or systemic thromboembolism in patients with atrial fibrillation can be markedly attenuated by oral anticoagulation. Treatment decision regarding anticoagulant use is guided by thromboembolic risk prediction schemes, preferably the CHA2DS2VASc [congestive heart failure, hypertension and age ≥75 years (doubled risk weight), diabetes mellitus, history of stroke or thromboembolism (doubled risk weight), vascular disease, age 65–74 years, sex category (female)] score, balanced against the perceived bleeding risk and net clinical benefit for individual patients [12, 13]. However, established risk prediction tools do not include information on psoriasis or other chronic inflammatory diseases and may consequently underestimate the risk of stroke in patients with these conditions. We therefore hypothesised that severe psoriasis may be a novel risk factor for thromboembolic events in subjects with atrial fibrillation and that in this case the risk of thromboembolic events is independent of the CHA2DS2VASc score when compared to patients with atrial fibrillation but without psoriasis. Therefore, the objective of this study was to examine the observed versus CHA2DS2VASc-predicted event rates and determine the risk of stroke and systemic thromboembolism associated with psoriasis amongst patients with nonvalvular atrial fibrillation.

Methods

Ethics

The study was approved by the Danish Data Protection Agency (reference no. 2007-58-0015). Approval by an ethics committee is not required for registry studies in Denmark. The lead author guarantees the integrity of the data and the accuracy of the data analysis.

Data sources

The study was conducted and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) recommendations [14]. In this nationwide Danish cohort study, using the unique personal registration number provided to all Danish residents at birth or immigration, we examined individual-level data from four prospectively recorded registries: (i) the Central Population Registry that records information on vital status for all citizens, (ii) the National Causes of Death Registry that registers all primary and contributing causes of death, (iii) the National Patient Registry that contains information on all inpatient and outpatient visits to Danish hospitals since 1978 [according to the International Classification of Diseases, eighth revision (ICD8) until 1994 and tenth revision (ICD10) thereafter] and (iv) the Registry of Medicinal Product Statistics that includes all prescriptions dispensed from Danish pharmacies since 1995. All dispensed drugs are classified according to the Anatomical Therapeutic Chemical (ATC) classification. Pharmacies in Denmark are required to register all dispensed prescriptions. Since 1996, invasive therapeutic procedures have been coded according to the Nordic Medical Statistics Committees Classification of Surgical Procedures (NCSP).

Study population

We identified all patients discharged from hospital following a first time diagnosis of nonvalvular atrial fibrillation or atrial flutter [ICD10: I48; absence of previous mitral or aortic valve disease (ICD8: 394–396, 4240, 4241; ICD10: I05, I06, I34, I35) or surgery (NCSP: KFK, KFM)] during the study period from 1997 to 2011. The diagnosis of atrial fibrillation has previously been validated [15]. Patients with psoriasis were identified by treatment with topical vitamin D (ATC: D05AX) and classified as having severe psoriasis according to inpatient or outpatient hospital-based treatment (ICD10: L40) as reported previously [10, 11]. Treatment with biological drugs approved for severe psoriasis (ATC: L04AB, L04AC), methotrexate (ATC: L01BA01, L04AX03), retinoids (ATC: D05BB), cyclosporine (ATC: L04AD01) and phototherapy (NSCP: BNG) was also recorded. Registration of biological drugs was only complete from 2007. Patients who died or experienced a study outcome within 7 days of discharge were excluded as were those who received oral anticoagulation (ATC B01AA) from 180 days prior to study start or within 7 days of discharge. Patients were censored on 31 December 2011 or during follow-up if they died or experienced a study outcome, whichever came first.

Pharmacological treatment

Pharmacotherapy was determined by means of filled prescriptions and, because treatment may be changed or intensified during or immediately after hospitalization, the baseline assessment and follow-up period started 7 days after discharge. Baseline pharmacological treatment was determined from prescriptions filled from 180 days before to 7 days after discharge: aspirin (ATC: B01AC06, N02BA01), ADP receptor antagonists (B01AC), oral anticoagulants (B01AA), amiodarone (C01BD), digoxin (C01AA), flecainide (C01BC), beta-blockers (C07), calcium antagonists (C08), angiotensin-converting enzyme inhibitors/angiotensin II receptor antagonists (C09), loop diuretics (C03C), thiazide diuretics (C03AB), antidepressants (N06A), antithyroid drugs (H03B), cholesterol-lowering agents (C10A) and glucose-lowering medication (A10). Information on oral anticoagulation was updated during follow-up.

Thromboembolism risk assessment

The predicted risk of stroke or systemic thromboembolism was assessed using the CHA2DS2VASc score to stratify the risk of stroke amongst patients with atrial fibrillation not receiving anticoagulant therapy; values ranged from 0 to 9 with higher scores indicating greater risk [12, 13, 16]. A score of 0 is generally considered ‘low risk’, 1 is intermediate risk and ≥2 is high risk of thromboembolism [12, 13]. We identified patients with heart failure (ICD8: 425, 4270–4271; ICD10: I110, I42, I50, J819, and use of loop diuretics) [17], hypertension (combination treatment with a least two classes of antihypertensive drugs) [18], diabetes mellitus (use of glucose-lowering medication), previous stroke or systemic thromboembolism (ICD8: 433–438, 444; ICD10: G458–G459, I63–I64, I74) and vascular disease (ICD8: 410, 440; ICD10: I21–I22, I700, I702–I709).

Study outcomes

First, focusing on predicted versus observed event rates in patients with severe psoriasis, the primary outcome was hospitalization or death due to thromboembolism [transient ischaemic attack (TIA), pulmonary embolism, nonhaemorrhagic stroke and peripheral artery embolism (ICD10: G458–G459, I26, I63–I64, I74)]. Secondly, fatal stroke was examined. In addition, all-cause mortality rates were calculated. The diagnosis of nonhaemorrhagic stroke has previously been validated [19].

Statistical analysis

Comparisons of baseline characteristics in patients with mild psoriasis, severe psoriasis or without psoriasis (the reference population) were performed using the chi-squared test for categorical covariates and anova for continuous variables. Event rates per 100 patient-years for the study outcomes were calculated according to psoriasis status. All individuals were included in analyses 7 days after discharge. We examined the event rates associated with individual risk factors of the CHA2DS2VASc score and the rates in predicted low risk (CHA2DS2VASc = 0) and moderate/high risk (CHA2DS2VASc ≥ 1) individuals with and without severe psoriasis. Additionally, Poisson regression models were used to estimate the risk of study outcomes associated with psoriasis. The regression analyses were adjusted for CHA2DS2VASc score and year of inclusion with calendar year (divided into 2-year periods after 1 January 1997) as the time scale. The CHA2DS2VASc score was included as a fixed variable defined at baseline. We repeated all analyses restricted to short-term follow-up (i.e. 180 and 365 days). A sensitivity analysis was performed, with censoring of subjects initiating oral anticoagulation during follow-up (defined as first oral anticoagulation prescription claim following inclusion), to assess the impact of changes in treatment on the primary results. In addition, the robustness of the primary results was tested in sensitivity analyses excluding TIA and pulmonary embolism from the outcomes of interest.

Model assumptions, including the absence of interactions between model covariates, were tested and found to be valid (unless otherwise indicated). A two-sided P value of <0.05 was considered to indicate statistical significance. All analyses were performed using SAS (version 9.2; SAS Institute, Cary, NC, USA) and Stata software (version 11.0; StataCorp, College St., TX, USA).

Results

A total of 99 357 patients were included in the study (Fig. 1). The maximum follow-up period was 15 years, with a mean follow-up of 3.5, 3.1 and 2.8 years for patients without psoriasis and with mild and severe psoriasis, respectively. Patients with psoriasis were significantly younger than those without the condition; there were no significant differences in sex or CHA2DS2VASc score. Patients with severe psoriasis were treated with tumour necrosis factor inhibitors (1.1%), methotrexate (28.4%), cyclosporine (4.9%), retinoids (14.4%) or other antipsoriatic therapies, including phototherapy [ultraviolet (UV) A, UVB and psoralen plus UVA (PUVA)] and climate therapy (51.2%). Baseline characteristics of the study population are presented in Table 1. Patients with severe psoriasis had the highest death rate: 14.6, 13.4 and 18.8 per 100 patient-years for patients without psoriasis and with mild and severe psoriasis, respectively.

Table 1. Baseline characteristics of the study population
 No psoriasis = 97 115Mild psoriasis = 1693Severe psoriasis = 549P-value
  1. CHA2DS2VASc, congestive heart failure, hypertension, age ≥75 years (doubled risk weight), diabetes mellitus, history of stroke or thromboembolism (doubled risk weight), vascular disease, age 65–74 years, sex category (female). Data are presented as counts and percentages unless otherwise stated.

Male, n (%)46 891 (48.3)864 (51.0)264 (48.1)0.08
Mean (SD) age, years74.4 (14.2)72.5 (12.7)71.7 (13.0)<0.001
Total patient-years339 58351391513 
Mean (SD) follow-up, years3.5 (3.6)3.1 (3.1)2.8 (2.9) 
Mean (SD) CHA2DS2VASc score3.0 (1.7)2.9 (1.8)3.0 (1.8)0.69
Age ≥ 75 years55 900 (57.6)817 (48.3)263 (47.9)<0.001
Age 65–74 years19 435 (20.0)466 (27.5)136 (24.8)<0.001
Congestive heart failure14 249 (14.7)239 (14.1)92 (16.8)0.31
Hypertension35 059 (36.1)734 (43.4)207 (37.7)<0.001
Stroke or thromboembolism16 746 (17.2)293 (17.3)107 (19.5)0.38
Vascular disease16 163 (16.6)303 (17.9)104 (18.9)0.14
Diabetes8608 (8.9)204 (12.1)95 (17.3)<0.001
Beta-blocker30 876 (31.8)633 (37.4)138 (25.1)<0.001
Calcium antagonist12 631 (13.0)233 (13.8)72 (13.1)0.66
Amiodarone2785 (2.9)69 (4.1)13 (2.4)0.01
Flecainide994 (1.0)26 (1.5)2 (0.4)0.04
Digoxin28 084 (28.9)424 (25.0)162 (29.5)0.002
Antidepressant8485 (8.7)167 (9.9)46 (8.4)0.25
Antithyroid drug1933 (2.0)29 (1.7)5 (0.9)0.14
Thiazide diuretic6478 (6.7)106 (6.3)34 (6.2)0.73
Loop diuretic23 331 (24.0)373 (22.0)117 (21.3)0.06
Cholesterol-lowering agent8104 (8.3)217 (12.8)60 (10.9)<0.001
Aspirin27 826 (28.7)469 (27.7)148 (27.0)0.48
ADP receptor antagonist3301 (3.4)78 (4.6)18 (3.3)0.03
Figure 1.

Flow chart showing selection of the study population.

Individual risk factors of the CHA2DS2VASc score

We evaluated the event rates associated with either the individual risk factors of the CHA2DS2VASc score or severe psoriasis as a single risk factor for thromboembolism. The thromboembolism rates associated with severe psoriasis were comparable to the rates associated with several of the established risk factors, such as congestive heart failure, hypertension, diabetes mellitus and vascular disease (Table 2). Of note, the observed thromboembolism rate in patients with severe psoriasis was markedly higher (2.6- to 3.4-fold) than predicted by the CHA2DS2VASc score in individuals predicted to be at low risk (i.e. CHA2DS2VASc = 0). By contrast, thromboembolism rates in patients with mild psoriasis and without psoriasis were similar to each other (Table 3).

Table 2. Thromboembolism rates (total follow-up period) per 100 patient-years associated with individual risk factors of the CHA2DS2VASc score
Risk factorEvent rate
  1. CHA2DS2VASc, congestive heart failure, hypertension, age ≥75 years (doubled risk weight), diabetes mellitus, history of stroke or thromboembolism (doubled risk weight), vascular disease, age 65–74 years, sex category (female). Event rates were calculated as (i) CHA2DS2VASc = 1 and presence of one of the following: congestive heart failure, hypertension, diabetes mellitus, vascular disease, age 65–74 years or female sex; (ii) CHA2DS2VASc = 2 and the presence of age ≥75 years or a history of stroke or thromboembolism; or (iii) CHA2DS2VASc = 0 and the presence of mild or severe psoriasis. Data are presented as event rates (95% confidence intervals).

CHA2DS2VASc
CHA2DS2VASc = 00.9 (0.8–1.0)
Congestive heart failure2.3 (1.4–3.7)
Hypertension1.5 (1.3–1.8)
Age ≥75 years5.2 (4.8–5.5)
Diabetes mellitus2.1 (1.5–3.0)
History of stroke/thromboembolism9.8 (8.2–11.7)
Vascular disease1.7 (1.3–2.2)
Age 65–75 years2.7 (2.4–2.9)
Female sex1.0 (0.9–1.2)
Non-CHA2DS2VASc
Mild psoriasis1.0 (0.5–2.1)
Severe psoriasis2.3 (0.9–6.1)
Table 3. Predicted versus observed thromboembolism rate per 100 patient-years in patients with severe psoriasis according to CHA2DS2VASc score (predicted low risk versus moderate/high risk)
CHA2DS2VAScPredicted event rateObserved event rate
  1. CHA2DS2VASc, congestive heart failure, hypertension, age ≥75 years (doubled risk weight), diabetes mellitus, history of stroke or thromboembolism (doubled risk weight), vascular disease, age 65–74 years, sex category (female). Data are presented as event rates (95% confidence intervals).

180 days of follow-up
00.7 (0.5–0.9)2.4 (0.3–16.9)
≥16.9 (6.7–7.1)9.2 (6.6–12.8)
365 days of follow-up
01.1 (0.9–1.4)2.5 (0.3–17.6)
≥110.0 (9.7–10.2)11.7 (8.7–15.8)
Total follow-up period
00.9 (0.8–1.0)2.3 (0.9–6.1)
≥15.5 (5.5–5.6)6.6 (5.3–8.1)

Thromboembolism and fatal stroke

Patients with severe psoriasis had higher thromboembolism and fatal stroke incidence rates and adjusted incidence rate ratios compared to patients without psoriasis (Tables 4 and 5).

Table 4. Event rates per 100 patient-years
 No psoriasisMild psoriasisSevere psoriasis
  1. Data are presented as event rates (95% confidence intervals).

180 days of follow-up
Thromboembolism6.3 (6.1–6.4)6.2 (5.0–7.7)8.5 (6.2–11.8)
Fatal stroke2.6 (2.5–2.7)1.9 (1.3–2.7)4.4 (2.8–6.8)
365 days of follow-up
Thromboembolism9.0 (8.8–9.2)8.7 (7.3–10.4)10.8 (8.0–14.4)
Fatal stroke4.1 (3.9–4.2)3.5 (2.6–4.5)5.7 (3.9–8.3)
Total follow-up period
Thromboembolism4.8 (4.7–4.9)4.8 (4.2–5.4)6.1 (5.0–7.5)
Fatal stroke2.1 (2.0–2.1)1.8 (1.5–2.2)2.8 (2.1–3.8)
Table 5. CHA2DS2VASc-adjusted incidence rate ratios (95% confidence intervals)
 No psoriasisMild psoriasisSevere psoriasis
  1. Data are presented as adjusted incidence rate ratios (95% confidence intervals). CHA2DS2VASc, congestive heart failure, hypertension, age ≥75 years (doubled risk weight), diabetes mellitus, history of stroke or thromboembolism (doubled risk weight), vascular disease, age 65–74 years, female sex.

180 days of follow-up
ThromboembolismReference0.99 (0.80–1.21)1.41 (1.00–1.99)
Fatal strokeReference0.90 (0.62–1.23)2.13 (1.32–3.45)
365 days of follow-up
ThromboembolismReference1.00 (0.84–1.18)1.27 (0.93–1.74)
Fatal strokeReference0.88 (0.66–1.16)1.88 (1.22–2.89)
Total follow-up period
ThromboembolismReference0.99 (0.87–1.11)1.27 (1.02–1.57)
Fatal strokeReference0.97 (0.80–1.16)1.51 (1.12–2.05)

Analyses with a maximum follow-up of 180 and 365 days yielded similar results to those of the primary analyses, with higher event rates and adjusted incidence rate ratios in patients with severe psoriasis compared to those without psoriasis (Tables 4 and 5).

Sensitivity analyses

In the analyses excluding TIA and pulmonary embolism, adjusted incidence rate ratios were comparable to those observed in the primary analyses (data not shown). Likewise, censoring of subjects initiating oral anticoagulation during follow-up did not essentially alter the results. In these analyses, thromboembolism rates and 95% confidence intervals were 5.2 (5.1–5.3), 5.2 (4.5–5.9) and 6.4 (5.2–8.0) for patients without psoriasis and with mild and severe psoriasis, respectively, and adjusted incidence rate ratios and 95% confidence intervals were 1.02 (0.90–1.16) and 1.26 (1.00–1.58) for those with mild and severe disease, respectively.

Discussion

In this nationwide cohort study comprising all Danish individuals with nonvalvular atrial fibrillation not treated with anticoagulants, severe psoriasis was associated with adverse outcomes including thromboembolic events, fatal stroke and all-cause mortality. In patients with severe psoriasis, the increased risk of adverse outcomes remained significant after adjustment for CHA2DS2VASc score, and these patients had markedly higher event rates than predicted by the score. These results indicate that severe psoriasis is a risk factor for adverse outcomes in patients with atrial fibrillation and that CHA2DS2VASc may underestimate the risk in these patients.

The relationship between adverse outcomes in patients with chronic inflammatory diseases and atrial fibrillation and the relative contribution of chronic inflammatory diseases to thromboembolic risk stratification algorithms in these patients have not, to our knowledge, been investigated previously. Our results, however, are consistent with previous reports of an apparent increased risk of stroke independent of atrial fibrillation in patients with chronic inflammatory diseases such as psoriasis and rheumatoid arthritis [10, 20, 21].

It is noteworthy that our results add to recent evidence of a suboptimal performance of established risk prediction tools (e.g. the Framingham score) in patients with chronic inflammatory diseases, including psoriasis and rheumatoid arthritis [22-25].

Psoriasis is a chronic inflammatory disease characterized by Th1- and Th17-driven inflammation with similar inflammatory markers and mediators as in atherosclerosis [2, 26, 27]. Accordingly, markers of subclinical atherosclerosis, for example carotid intima-media thickening, have been observed in patients with chronic inflammatory diseases [28, 29]. These findings may partly explain the mounting evidence linking psoriasis to increased risk of cardiovascular disease, including arterial and venous thrombosis and atrial fibrillation [7-11, 20, 30-32]. Of note, inflammation, as determined by circulating levels of tumour necrosis factor α and interleukin 6, may be associated with increased risk of stroke in patients with atrial fibrillation [33, 34]. In addition to the evidence of a systemic inflammatory response in patients with psoriasis, it is interesting that a psoriasis severity-dependent increase in platelet activation has been demonstrated [3, 4] and the findings of a single study have indicated a shorter platelet activity regeneration time following aspirin intake in patients with psoriasis compared to nonpsoriatic control subjects [5]. Indeed, the importance of interactions between inflammatory responses, platelet activity and thrombosis is increasingly being recognised [35]. Current guidelines from the European League Against Rheumatism (EULAR) on prevention of cardiovascular disease in patients with inflammatory arthritis recommend intensified treatment goals in these patients [24] and those from the European Society of Cardiology emphasize the increased risk of cardiovascular disease in patients with chronic inflammatory diseases, including psoriasis [36]. Similarly, attention to novel risk factors (e.g. chronic inflammatory diseases) may prove beneficial in patients with atrial fibrillation. Furthermore, the dose–response relationship between psoriasis severity and adverse outcomes observed in the present study clearly strengthens the likelihood of a causal relationship between inflammation and thromboembolism in patients with psoriasis. Nevertheless, we cannot draw definite conclusions about the underlying mechanisms leading to the observed increased risk of thromboembolic events in patients with severe psoriasis independent of traditional risk factors, and further studies are needed to determine the clinical implications including the value of potential inclusion of chronic inflammatory diseases in thromboembolism risk prediction schemes.

Several strengths and limitations of our study should be considered. In Denmark, health care is readily accessible and essentially free of charge, and by including the entire Danish population we avoided selection bias related to, for example, sex, age, socio-economic status, health care insurance status and labour market participation. The use of validated exposure and outcome measures and the completeness of follow-up also strengthen the study conclusions. We were unable to identify patients with psoriasis either treated with topical corticosteroids alone (due to the broad nonpsoriatic indications for these agents) or not treated, allowing for misclassification of some patients with psoriasis as the reference population. However, this would most likely have led to underestimation of event rates in patients with psoriasis and attenuation of the true psoriasis-related risk of adverse events. Of importance, the very large sample size attenuates the potential impact of any misclassification. Because we relied on registry data from hospitalizations to determine the presence of atrial fibrillation, our results cannot be applied to patients with atrial fibrillation treated in primary care. In addition, the application of diagnoses and prescription drug use to identify the individual risk factors of the CHA2DS2VASc score may have underestimated the true prevalence of these conditions. Our databases lacked information on several traditional cardiovascular risk factors, including body mass index, lipid levels, smoking and level of physical activity, and therefore, unmeasured confounding may have influenced our findings. Finally, the Danish population is primarily Caucasian, and therefore, these results should only be extrapolated to other ethnicities with caution.

Conclusion

In a nationwide cohort of patients with nonvalvular atrial fibrillation not treated with oral anticoagulation, severe psoriasis was associated with increased risk of thromboembolism and fatal stroke. A CHA2DS2VASc of 0 markedly underestimated the risk of thromboembolism and therefore may be misleading when used to support deferral of oral anticoagulation in these patients.

Conflict of interest statement

OA has received honoraria as a consultant and/or speaker for AbbVie, Pfizer, MSD and Janssen Pharmaceuticals and is supported by unrestricted grants from The Danish Psoriasis Foundation, The Region Zealand Research Foundation, Dermbio and The Axel Muusfeldt Foundation. GG is supported by an unrestricted clinical research scholarship from the Novo Nordisk Foundation. The sponsors had no role in the conduct of the study or the decision to submit the manuscript for publication.

Ancillary