Incidence of Venous Thromboembolism in Patients With Dermatologist-Diagnosed Chronic Inflammatory Skin Diseases | Dermatology | JAMA Dermatology | JAMA Network
[Skip to Navigation]
Sign In
Figure 1.  Estimates of Developing Venous Thromboembolism (VTE) in Patients With vs Without a Chronic Inflammatory Skin Disease (CISD)
Estimates of Developing Venous Thromboembolism (VTE)  in Patients With vs Without a Chronic Inflammatory Skin Disease (CISD)

Chronic inflammatory skin disease was defined as a composite of the individual conditions: psoriasis, atopic dermatitis, alopecia areata, vitiligo, and hidradenitis suppurativa. Non-CISD was defined as 2 dermatologist visits with no diagnosis of a CISD and matched to the exposure (CISD) patients’ cohort entry date ±60 days. Inpatient VTE was defined as a hospitalization with a primary discharge diagnosis of VTE. Outpatient VTE was defined as an outpatient visit with a diagnosis of VTE followed by initiation of anticoagulant therapy within 7 days or hospitalization for VTE within 7 days. Immunomodulator users included patients receiving treatment with nonbiologic immunomodulatory agents (methotrexate, cyclosporine, mycophenolate mofetil, azathioprine, sulfasalazine, and leflunomide), biologic immunomodulatory agents (dupilumab, risankizumab-rzaa, tildrakizumab, ixekizumab, secukinumab, guselkumab, ustekinumab, abatacept, adalimumab, etanercept, golimumab, certolizumab, infliximab, rituximab, anakinra, and tocilizumab), and targeted synthetic immunomodulatory agents (tofacitinib and baricitinib). PS indicates propensity score.

Figure 2.  Time to Venous Thromboembolism (VTE) in Patients With vs Without a Chronic Inflammatory Skin Disease (CISD)
Time to Venous Thromboembolism (VTE) in Patients With vs Without a Chronic Inflammatory Skin Disease (CISD)

The VTE event in patients with psoriasis (A), atopic dermatitis (B), alopecia areata (C), vitiligo (D), hidradenitis suppurativa (E), and overall CISD (F) was defined as inpatient- or outpatient-diagnosed VTE with evidence of treatment after 1:1 propensity-score matching. Dark lines within the graphs represent 95% CIs.

Table 1.  Baseline Patient Characteristics Comparing Patients With vs Without a CISD, Before 1:1 Propensity-Score Matching
Baseline Patient Characteristics Comparing Patients With vs Without a CISD, Before 1:1 Propensity-Score Matching
Table 2.  Incidence Rate and Hazard Ratio of Venous Thromboembolism Before and After 1:1 Propensity-Score Matching
Incidence Rate and Hazard Ratio of Venous Thromboembolism Before and After 1:1 Propensity-Score Matching
Supplement.

eFigure. Cohort Study Design Diagram

eTable 1. All Baseline Characteristics, Comparing Patients With vs Without a CISD, Before 1:1 Propensity-Score Matching

eTable 2. Baseline Patients Characteristics and Standardized Differences, Comparing Patients With vs Without a CISD, Before and After 1:1 Propensity-Score Matching

eTable 3. International Classification of Diseases Codes Used to Identify CISD and Venous Thromboembolism

eTable 4. Follow-up Time and Reasons for Censoring, After Propensity-Score Matching

eTable 5. Exploratory Analysis of Systemic Immunomodulator Use During Follow-up, After 1:1 Propensity-Score Matching

eTable 6. Incidence Rate and Hazard Ratio of Inpatient or Outpatient VTE, Stratified by Subgroups, Before and After 1:1 Propensity-Score Matching

eTable 7. Incidence Rate and Hazard Ratio of Inpatient VTE (Primary Position), Stratified by Subgroups, Before and After 1:1 Propensity-Score Matching

eTable 8. Incidence Rate and Hazard Ratio of Inpatient VTE (Any Position), Stratified by Subgroups, Before and After 1:1 Propensity-Score Matching

eTable 9. Exploratory Analysis of Dermatologic Patient Characteristics on the Visit Day That Defined Cohort Entry

eTable 10. Incidence of VTE in Patients With Rheumatoid Arthritis Versus Without Rheumatoid Arthritis, Before 1:1 Propensity-Score Matching

eTable 11. Incidence Rate and Hazard Ratio of Venous Thromboembolism, With Additional Adjustment for Non-Melanoma Skin Cancer, Before and After 1:1 Propensity-Score Matching

eTable 12. Exploratory Analysis to Confirm Referent Patients Had No CISD Diagnosis Code Recorded by Any Physician Before Cohort Entry or on the Day of Cohort Entry

1.
Silvestri  E, Scalera  A, Emmi  G,  et al.  Thrombosis in autoimmune diseases: a role for immunosuppressive treatments?   Semin Thromb Hemost. 2016;42(6):650-661. doi:10.1055/s-0036-1579642 PubMedGoogle Scholar
2.
Johannesdottir  SA, Schmidt  M, Horváth-Puhó  E, Sørensen  HT.  Autoimmune skin and connective tissue diseases and risk of venous thromboembolism: a population-based case-control study.   J Thromb Haemost. 2012;10(5):815-821. doi:10.1111/j.1538-7836.2012.04666.x PubMedGoogle ScholarCrossref
3.
Kim  SC, Schneeweiss  S, Liu  J, Solomon  DH.  Risk of venous thromboembolism in patients with rheumatoid arthritis.   Arthritis Care Res (Hoboken). 2013;65(10):1600-1607. doi:10.1002/acr.22039 PubMedGoogle Scholar
4.
Shaheen  MS, Silverberg  JI.  Association of inflammatory skin diseases with venous thromboembolism in US adults.   Arch Dermatol Res. 2021;313(4):281-289. PubMedGoogle Scholar
5.
Ogdie  A, Kay McGill  N, Shin  DB,  et al.  Risk of venous thromboembolism in patients with psoriatic arthritis, psoriasis and rheumatoid arthritis: a general population-based cohort study.   Eur Heart J. 2018;39(39):3608-3614. doi:10.1093/eurheartj/ehx145 PubMedGoogle ScholarCrossref
6.
Ahlehoff  O, Gislason  GH, Lindhardsen  J,  et al.  Psoriasis carries an increased risk of venous thromboembolism: a Danish nationwide cohort study.   PLoS One. 2011;6(3):e18125. doi:10.1371/journal.pone.0018125 PubMedGoogle Scholar
7.
Chung  WS, Lin  CL, Sung  FC, Lu  CC, Kao  CH.  Increased risk of venous thromboembolism in patients with dermatomyositis/polymyositis: a nationwide cohort study.   Thromb Res. 2014;134(3):622-626. doi:10.1016/j.thromres.2014.06.021 PubMedGoogle ScholarCrossref
8.
Rogers  MAM, Lin  P, Nallamothu  BK, Kim  C, Waljee  AK.  Longitudinal study of short-term corticosteroid use by working-age adults with diabetes mellitus: risks and mitigating factors.   J Diabetes. 2018;10(7):546-555. doi:10.1111/1753-0407.12631 PubMedGoogle ScholarCrossref
9.
Coelho  MC, Santos  CV, Vieira Neto  L, Gadelha  MR.  Adverse effects of glucocorticoids: coagulopathy.   Eur J Endocrinol. 2015;173(4):M11-M21. doi:10.1530/EJE-15-0198 PubMedGoogle ScholarCrossref
10.
Mease  P, Charles-Schoeman  C, Cohen  S,  et al.  Incidence of venous and arterial thromboembolic events reported in the tofacitinib rheumatoid arthritis, psoriasis and psoriatic arthritis development programmes and from real-world data.   Ann Rheum Dis. 2020;79(11):1400-1413. doi:10.1136/annrheumdis-2019-216761 PubMedGoogle ScholarCrossref
11.
Solimani  F, Meier  K, Ghoreschi  K.  Emerging topical and systemic JAK inhibitors in dermatology.   Front Immunol. 2019;10:2847. doi:10.3389/fimmu.2019.02847 PubMedGoogle ScholarCrossref
12.
Rudy  SF, Li  K, Moubayed  SP, Most  SP.  Risk of venous thromboembolism in patients with keratinocyte carcinoma.   JAMA Facial Plast Surg. 2018;20(6):453-459. doi:10.1001/jamafacial.2018.0331 PubMedGoogle ScholarCrossref
13.
Alotaibi  GS, Wu  C, Senthilselvan  A, McMurtry  MS.  The validity of ICD codes coupled with imaging procedure codes for identifying acute venous thromboembolism using administrative data.   Vasc Med. 2015;20(4):364-368. doi:10.1177/1358863X15573839 PubMedGoogle ScholarCrossref
14.
Walker  RF, Zakai  NA, MacLehose  RF,  et al.  Association of testosterone therapy with risk of venous thromboembolism among men with and without hypogonadism.   JAMA Intern Med. 2020;180(2):190-197. doi:10.1001/jamainternmed.2019.5135 PubMedGoogle ScholarCrossref
15.
Sanfilippo  KM, Wang  TF, Gage  BF, Liu  W, Carson  KR.  Improving accuracy of International Classification of Diseases codes for venous thromboembolism in administrative data.   Thromb Res. 2015;135(4):616-620. doi:10.1016/j.thromres.2015.01.012 PubMedGoogle ScholarCrossref
16.
Ammann  EM, Cuker  A, Carnahan  RM,  et al.  Chart validation of inpatient International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) administrative diagnosis codes for venous thromboembolism (VTE) among intravenous immune globulin (IGIV) users in the Sentinel Distributed Database.   Medicine (Baltimore). 2018;97(8):e9960. doi:10.1097/MD.0000000000009960 PubMedGoogle Scholar
17.
Tamariz  L, Harkins  T, Nair  V.  A systematic review of validated methods for identifying venous thromboembolism using administrative and claims data.   Pharmacoepidemiol Drug Saf. 2012;21(suppl 1):154-162. doi:10.1002/pds.2341 PubMedGoogle ScholarCrossref
18.
Schneeweiss  S.  A basic study design for expedited safety signal evaluation based on electronic healthcare data.   Pharmacoepidemiol Drug Saf. 2010;19(8):858-868. doi:10.1002/pds.1926 PubMedGoogle ScholarCrossref
19.
Anderson  FA  Jr, Spencer  FA.  Risk factors for venous thromboembolism.   Circulation. 2003;107(23)(suppl 1):I9-I16.PubMedGoogle Scholar
20.
Brookhart  MA, Schneeweiss  S, Rothman  KJ, Glynn  RJ, Avorn  J, Stürmer  T.  Variable selection for propensity score models.   Am J Epidemiol. 2006;163(12):1149-1156. doi:10.1093/aje/kwj149 PubMedGoogle ScholarCrossref
21.
Rosenbaum  PR, Rubin DB.  The central role of the propensity score in observational studies for causal effects.   Biometrika. 1983;70(1):41-55. doi:10.1093/biomet/70.1.41 Google ScholarCrossref
22.
Rassen  JA, Shelat  AA, Myers  J, Glynn  RJ, Rothman  KJ, Schneeweiss  S.  One-to-many propensity score matching in cohort studies.   Pharmacoepidemiol Drug Saf. 2012;21(suppl 2):69-80. doi:10.1002/pds.3263 PubMedGoogle ScholarCrossref
23.
Franklin  JM, Rassen  JA, Ackermann  D, Bartels  DB, Schneeweiss  S.  Metrics for covariate balance in cohort studies of causal effects.   Stat Med. 2014;33(10):1685-1699. doi:10.1002/sim.6058 PubMedGoogle ScholarCrossref
24.
Austin  PC.  Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples.   Stat Med. 2009;28(25):3083-3107. doi:10.1002/sim.3697 PubMedGoogle ScholarCrossref
25.
Cox  DR.  Regression models and life tables.   J Royal Stat Society Series B. 1972;34(2):187-220.Google Scholar
26.
Aetion Inc. Aetion evidence platform. Software for real-world data analysis. 2020. Accessed February 1, 2021. http://aetion.com
27.
Wang  SV, Verpillat  P, Rassen  JA, Patrick  A, Garry  EM, Bartels  DB.  Transparency and reproducibility of observational cohort studies using large healthcare Databases.   Clin Pharmacol Ther. 2016;99(3):325-332. doi:10.1002/cpt.329 PubMedGoogle ScholarCrossref
28.
Kim  SC, Solomon  DH, Rogers  JR,  et al.  Cardiovascular safety of tocilizumab versus tumor necrosis factor inhibitors in patients with rheumatoid arthritis: a multi-database cohort study.   Arthritis Rheumatol. 2017;69(6):1154-1164. doi:10.1002/art.40084 PubMedGoogle ScholarCrossref
29.
Patorno  E, Schneeweiss  S, Gopalakrishnan  C, Martin  D, Franklin  JM.  Using real-world data to predict findings of an ongoing phase IV cardiovascular outcome trial: Cardiovascular Safety of Linagliptin Versus Glimepiride.   Diabetes Care. 2019;42(12):2204-2210. doi:10.2337/dc19-0069 PubMedGoogle ScholarCrossref
30.
Sidbury  R, Davis  DM, Cohen  DE,  et al; American Academy of Dermatology.  Guidelines of care for the management of atopic dermatitis: section 3—management and treatment with phototherapy and systemic agents.   J Am Acad Dermatol. 2014;71(2):327-349. doi:10.1016/j.jaad.2014.03.030 PubMedGoogle ScholarCrossref
31.
Menter  A, Strober  BE, Kaplan  DH,  et al.  Joint AAD-NPF guidelines of care for the management and treatment of psoriasis with biologics.   J Am Acad Dermatol. 2019;80(4):1029-1072. doi:10.1016/j.jaad.2018.11.057 PubMedGoogle ScholarCrossref
32.
Alikhan  A, Sayed  C, Alavi  A,  et al.  North American clinical management guidelines for hidradenitis suppurativa: a publication from the United States and Canadian Hidradenitis Suppurativa Foundations: part II—topical, intralesional, and systemic medical management.   J Am Acad Dermatol. 2019;81(1):91-101. doi:10.1016/j.jaad.2019.02.068 PubMedGoogle ScholarCrossref
33.
Messenger  AG, McKillop  J, Farrant  P, McDonagh  AJ, Sladden  M.  British Association of Dermatologists’ guidelines for the management of alopecia areata: 2012.   Br J Dermatol. 2012;166(5):916-926. doi:10.1111/j.1365-2133.2012.10955.x PubMedGoogle ScholarCrossref
34.
Ingram  JR.  The epidemiology of hidradenitis suppurativa.   Br J Dermatol. 2020;183(6):990-998. doi:10.1111/bjd.19435 PubMedGoogle ScholarCrossref
35.
Lee  HH, Gwillim  E, Patel  KR,  et al.  Epidemiology of alopecia areata, ophiasis, totalis, and universalis: a systematic review and meta-analysis.   J Am Acad Dermatol. 2020;82(3):675-682. doi:10.1016/j.jaad.2019.08.032 PubMedGoogle ScholarCrossref
36.
Langan  SM, Irvine  AD, Weidinger  S.  Atopic dermatitis.   Lancet. 2020;396(10247):345-360. doi:10.1016/S0140-6736(20)31286-1 PubMedGoogle ScholarCrossref
37.
Armstrong  AW, Read  C.  Pathophysiology, clinical presentation, and treatment of psoriasis: a review.   JAMA. 2020;323(19):1945-1960. doi:10.1001/jama.2020.4006 PubMedGoogle ScholarCrossref
38.
Rodrigues  M, Ezzedine  K, Hamzavi  I, Pandya  AG, Harris  JE; Vitiligo Working Group.  Current and emerging treatments for vitiligo.   J Am Acad Dermatol. 2017;77(1):17-29. doi:10.1016/j.jaad.2016.11.010 PubMedGoogle ScholarCrossref
39.
Ezzedine  K, Eleftheriadou  V, Whitton  M, van Geel  N.  Vitiligo.   Lancet. 2015;386(9988):74-84. doi:10.1016/S0140-6736(14)60763-7 PubMedGoogle ScholarCrossref
40.
Heit  JA.  Epidemiology of venous thromboembolism.   Nat Rev Cardiol. 2015;12(8):464-474. doi:10.1038/nrcardio.2015.83 PubMedGoogle ScholarCrossref
41.
Silverstein  MD, Heit  JA, Mohr  DN, Petterson  TM, O’Fallon  WM, Melton  LJ  III.  Trends in the incidence of deep vein thrombosis and pulmonary embolism: a 25-year population-based study.   Arch Intern Med. 1998;158(6):585-593. doi:10.1001/archinte.158.6.585 PubMedGoogle ScholarCrossref
42.
Cushman  M, Tsai  AW, White  RH,  et al.  Deep vein thrombosis and pulmonary embolism in two cohorts: the longitudinal investigation of thromboembolism etiology.   Am J Med. 2004;117(1):19-25. doi:10.1016/j.amjmed.2004.01.018 PubMedGoogle ScholarCrossref
43.
Tagalakis  V, Patenaude  V, Kahn  SR, Suissa  S.  Incidence of and mortality from venous thromboembolism in a real-world population: the Q-VTE Study Cohort.   Am J Med. 2013;126(9):832.e13-832.e21. doi:10.1016/j.amjmed.2013.02.024PubMedGoogle ScholarCrossref
44.
Fox  EA, Kahn  SR.  The relationship between inflammation and venous thrombosis: a systematic review of clinical studies.   Thromb Haemost. 2005;94(2):362-365. doi:10.1160/TH05-04-0266 PubMedGoogle Scholar
45.
Branchford  BR, Carpenter  SL.  The role of inflammation in venous thromboembolism.   Front Pediatr. 2018;6:142. doi:10.3389/fped.2018.00142 PubMedGoogle ScholarCrossref
46.
van der Poll  T, Büller  HR, ten Cate  H,  et al.  Activation of coagulation after administration of tumor necrosis factor to normal subjects.   N Engl J Med. 1990;322(23):1622-1627. doi:10.1056/NEJM199006073222302 PubMedGoogle ScholarCrossref
47.
Sandborn  WJ, Panés  J, Sands  BE,  et al.  Venous thromboembolic events in the tofacitinib ulcerative colitis clinical development programme.   Aliment Pharmacol Ther. 2019;50(10):1068-1076. doi:10.1111/apt.15514 PubMedGoogle ScholarCrossref
48.
Newsom  M, Bashyam  AM, Balogh  EA, Feldman  SR, Strowd  LC.  New and emerging systemic treatments for atopic dermatitis.   Drugs. 2020;80(11):1041-1052. doi:10.1007/s40265-020-01335-7 PubMedGoogle ScholarCrossref
49.
Asgari  MM, Wu  JJ, Gelfand  JM,  et al.  Validity of diagnostic codes and prevalence of psoriasis and psoriatic arthritis in a managed care population, 1996-2009.   Pharmacoepidemiol Drug Saf. 2013;22(8):842-849. doi:10.1002/pds.3447 PubMedGoogle ScholarCrossref
50.
Löfvendahl  S, Theander  E, Svensson  Å, Carlsson  KS, Englund  M, Petersson  IF.  Validity of diagnostic codes and prevalence of physician-diagnosed psoriasis and psoriatic arthritis in southern Sweden—a population-based register study.   PLoS One. 2014;9(5):e98024. doi:10.1371/journal.pone.0098024 PubMedGoogle Scholar
51.
Kim  GE, Shlyankevich  J, Kimball  AB.  The validity of the diagnostic code for hidradenitis suppurativa in an electronic database.   Br J Dermatol. 2014;171(2):338-342. doi:10.1111/bjd.13041 PubMedGoogle ScholarCrossref
52.
Lavian  J, Li  SJ, Lee  EY,  et al.  Validation of case identification for alopecia areata using International Classification of Diseases coding.   Int J Trichology. 2020;12(5):234-237. doi:10.4103/ijt.ijt_67_20 PubMedGoogle Scholar
53.
Hsu  DY, Dalal  P, Sable  KA,  et al.  Validation of International Classification of Disease Ninth Revision codes for atopic dermatitis.   Allergy. 2017;72(7):1091-1095. doi:10.1111/all.13113 PubMedGoogle ScholarCrossref
54.
Bolognia  J, Jorizzo JL, Schaffer JV, eds . Dermatology. 4th ed. Elsevier; 2018.
55.
Schneeweiss  S, Glynn  RJ, Tsai  EH, Avorn  J, Solomon  DH.  Adjusting for unmeasured confounders in pharmacoepidemiologic claims data using external information: the example of COX2 inhibitors and myocardial infarction.   Epidemiology. 2005;16(1):17-24. doi:10.1097/01.ede.0000147164.11879.b5 PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Original Investigation
    May 26, 2021

    Incidence of Venous Thromboembolism in Patients With Dermatologist-Diagnosed Chronic Inflammatory Skin Diseases

    Author Affiliations
    • 1Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women’s Hospital, Boston, Massachusetts
    • 2Harvard Medical School, Boston, Massachusetts
    • 3Department of Dermatology, Brigham and Women’s Hospital, Boston, Massachusetts
    • 4Division of Rheumatology, Inflammation, and Immunity, Department of Medicine, Brigham and Women’s Hospital, Boston, Massachusetts
    • 5Associate Editor, JAMA Dermatology
    • 6Department of Dermatology, George Washington University School of Medicine and Health Sciences
    JAMA Dermatol. 2021;157(7):805-816. doi:10.1001/jamadermatol.2021.1570
    Key Points

    Question  What is the background rate of venous thromboembolism in patients with vs without a chronic inflammatory skin disease, including psoriasis, atopic dermatitis, alopecia areata, vitiligo, and hidradenitis suppurativa?

    Findings  In this cohort study with analysis of claims data on 158 123 patients with a chronic inflammatory skin disease, no indication of an increased incidence of unprovoked clinical venous thromboembolism events was noted in patients with chronic inflammatory skin diseases compared with similar patients without chronic inflammatory skin diseases.

    Meaning  In this newer era of advanced systemic medications for treatment of chronic inflammatory skin diseases, having a background rate of venous thromboembolism established may help monitor the safety of newly marketed systemic treatments in dermatology.

    Abstract

    Importance  Several studies have linked chronic inflammatory skin diseases (CISDs) with venous thromboembolism (VTE) in a range of data sources with mixed conclusions.

    Objective  To examine the incidence of VTE in patients with vs without CISD.

    Design, Setting, and Participants  A cohort study using commercial insurance claims data from a nationwide US health care database from January 1, 2004, through 2019 was conducted. A total of 158 123 patients with dermatologist-recorded psoriasis, atopic dermatitis, alopecia areata, vitiligo, or hidradenitis suppurativa were included. Risk-set sampling identified patients without a CISD. Patient follow-up lasted until the first of the following occurred: VTE, death, disenrollment, or end of data stream.

    Exposures  Patients with vs without CISD.

    Main Outcomes and Measures  Venous thromboembolism events were identified with validated algorithms. Incidence rates were computed before and after 1:1 propensity-score matching to account for VTE risk factors. Hazard ratios were estimated to compare the incidence of VTE in the CISD vs non-CISD cohorts.

    Results  A total of 158 123 patients were identified with CISD: with psoriasis (n = 96 138), atopic dermatitis (n = 30 418), alopecia areata (n = 17 889), vitiligo (n = 7735), or HS (n = 5934); 9 patients had 2 of these conditions. A total of 1 570 387 patients were without a CISD. The median follow-up time was 1.9 years (interquartile range, 0.8-4.0 years) in patients with CISD. The incidence rate (per 1000 person-years) of outpatient or inpatient VTE was 1.57 in psoriasis, 1.83 in atopic dermatitis, 0.94 in alopecia areata, 0.93 in vitiligo, 1.65 in HS and 1.53 in CISD overall, compared with 1.76 in patients without a CISD. Incidence rates increased in patients aged 50 years or older (2.3 per 1000 person-years) and decreased in those aged 18 to 49 years (0.8 per 1000 person-years). After propensity-score matching to patients without a CISD, the hazard ratio (HR) of VTE was 0.86 (95% CI, 0.75-0.99) in psoriasis, 1.19 (95% CI, 0.95-1.48) in atopic dermatitis, 0.97 (95% CI, 0.65-1.46) in alopecia areata, 0.90 (95% CI, 0.49-1.65) in vitiligo, 1.64 (95% CI, 0.82-3.27) in hidradenitis suppurativa, and 0.94 (95% CI, 0.84-1.05) in CISD overall.

    Conclusions and Relevance  In this large-scale cohort study, CISDs were not associated with an increased incidence of VTE after controlling for relevant VTE risk factors in a representative dermatology patient population.

    Introduction

    Chronic inflammatory conditions, such as rheumatoid arthritis, are known to increase the risk of venous thromboembolism (VTE).1-3 Several studies linked chronic inflammatory skin diseases (CISDs) with VTE in a range of data sources. Based on cross-sectional data from the US National Inpatient Sample, a 22% increase in VTE was found among patients with vs without atopic dermatitis (AD), which was slightly higher among adults aged 50 years or older (25%).4 Among patients with severe psoriasis, one study observed a 13% increase in VTE risk in a UK general practice network not excluding patients with cancer, and a Scandinavian claims data analysis found a 2- to 3-fold increase.5,6 In a Taiwan national claims database, patients with dermatomyositis had an 11-fold increase in VTE risk in an analysis that included patients with cancer—a known risk factor for VTE.7

    It is important to establish a robust estimate of the background VTE risk in patients with CISDs, including psoriasis (without psoriatic arthritis), AD, alopecia areata (AA), vitiligo, and hidradenitis suppurativa (HS). These CISDs vary in pathogenesis and, as such, the underlying baseline VTE risk may be different among these individual conditions. Some systemic immunomodulatory agents that are used in the treatment of CISDs are suspected or known to modify the VTE risk, including systemic corticosteroids, high-dose Janus kinase inhibitors, and possibly other agents.8-10 In addition to treatments, some comorbidities associated with CISD, such as inflammatory bowel disease and rheumatoid arthritis, are also associated with VTE risk. It is important to isolate the CISDs’ effect on VTE from already established VTE risk associated with comorbidities and various treatments. The need to establish this background VTE risk has become more pressing as a number of new systemic immunomodulatory agents for treatment of dermatologic conditions, including Janus kinase inhibitors for AD, AA, and vitiligo, will likely be marketed in the near future.11

    There are different approaches to determining the risk for VTE in the CISD population. The variability of findings from previous studies depends on the exclusion criteria. Studies with strict exclusion criteria aim to identify the VTE risk specific to the dermatologic condition of interest independent of patient characteristics, comorbidities, and treatments that may influence the incidence of VTE. Studies with more relaxed exclusion criteria identify the VTE risk of the condition plus its associated comorbidities, and another group of studies identifies the VTE risk of the condition plus the associated comorbidities and their treatments. Although none of these approaches are wrong, they have different utility. To establish a background rate of VTE specific to CISDs and isolated from the noise of comorbidities and various treatments, it is necessary to apply strict exclusions and adjustments. We sought to establish a background incidence of VTE in patients with dermatologist-diagnosed CISDs independent of existing risk factors, comorbidities, and treatments.

    Methods
    Data Source

    We used real-world data from a national commercial insurance claims database (Optum’s deidentified Clinformatics Data Mart Database) from January 1, 2004, through 2019. The database is composed of beneficiaries of commercial and Medicare Advantage health plans. The database contains dated information on plan enrollment, health care use, demographics, and integrated records for inpatient and outpatient encounters and pharmacy dispensing. All patient information was deidentified. The Brigham and Women’s Hospital’s Institutional Review Board approved this study, and a signed data licensing agreement was in place.

    Study Cohort

    This study sought to compare patients with vs without CISD within the dermatology clinic population. We identified patients of all ages who were diagnosed with a CISD, including psoriasis (without psoriatic arthritis), AD, AA, vitiligo, and HS, according to the following criteria: at least 2 visits to a dermatologist within 1 year and a diagnosis of the same CISD at each visit (eTable 3 in the Supplement provides specific codes). We required that the skin conditions be mutually exclusive, meaning that a patient who had 2 encounters with a dermatologist within a year but received a diagnosis of AD at 1 visit and psoriasis at the next would not qualify. The cohort entry date for the CISD group was the second CISD diagnosis after at least 365 days of continuous enrollment. For the non-CISD group, we risk-set sampled 10 patients without CISD from all plan enrollees who had at least 2 visits to a dermatologist (cohort entry date was the second dermatologist visit) within a year but were not diagnosed with a CISD at any time, ensuring a reliable diagnosis separation of the 2 groups. The patients without CISD were matched on the cohort entry date ±60 days of the patient with CISD and were required to have at least 365 days of continuous enrollment before the matched cohort entry date (eTable 9 and eTable 12 in the Supplement provide more information on the non-CISD group). Patients from either the CISD or non-CISD cohort were excluded if they had any of the following conditions before their cohort entry: VTE, use of an anticoagulant, diagnosis of cancer (excluding nonmelanoma skin cancer),12 prothrombotic condition (factor V Leiden, antiphospholipid antibody syndrome, prothrombin deficiency, and factor C and S deficiency), sickle cell disease, HIV/AIDS, or chronic inflammatory conditions associated with VTE (rheumatoid arthritis, inflammatory bowel disease, and lupus) (eFigure in the Supplement). Including these conditions could raise concern as to whether the VTE risk was due to CISD or to the comorbid inflammatory bowel disease, rheumatoid arthritis, or cancer.

    Outcomes

    We identified incident VTE, deep venous thrombosis, or pulmonary embolism events during all available follow-up times. Venous thromboembolism was defined as either a hospitalization with a primary discharge diagnosis of VTE or an outpatient visit with a diagnosis of VTE followed by initiation of an anticoagulant within 7 days or hospitalization for VTE within 7 days (positive predictive value, 72%-90%).13-16 For a more-specific but less-sensitive definition, we limited our secondary outcome to include only inpatient events, defined as hospitalization with a primary discharge diagnosis of deep venous thrombosis or pulmonary embolism (positive predictive value, 83%-90%).13,16 We included a third VTE outcome to capture probable VTE events defined as an inpatient VTE recorded on hospital claims as a discharge diagnosis in any position (positive predictive value, 72%).17 For each definition, the event date was the date on which all criteria were fulfilled. Detailed International Classification of Diseases, Version 9 (ICD-9), and International Statistical Classification of Disease, Version 10 (ICD-10), codes are listed in eTable 3 in the Supplement. Follow-up began the day after the cohort entry day and lasted until the first of the following events occurred: VTE, death, disenrollment, end of enrollment, and end of data stream (eFigure in the Supplement).18

    Patient Characteristics

    All patient characteristics were assessed during the 365 days before cohort entry (eFigure in the Supplement). We considered age at cohort entry, sex, race, year of cohort entry, history of fracture, surgery, tobacco use, hospitalization, hemiplegia or paraplegia, earlier use of oral or nonoral contraceptives, hormone replacement therapy, systemic glucocorticoid use, cumulative systemic prednisone equivalencies over the previous 180 days, systemic immunomodulatory medication use (nonbiologic immunomodulators: methotrexate, cyclosporine, mycophenolate mofetil, azathioprine, sulfasalazine, and leflunomide; biologic immunomodulators: dupilumab, risankizumab-rzaa, tildrakizumab, ixekizumab, secukinumab, guselkumab, ustekinumab, abatacept, adalimumab, etanercept, golimumab, certolizumab, infliximab, rituximab, anakinra, and tocilizumab; and targeted synthetic immunomodulators: tofacitinib and baricitinib), number of physician visits and unique prescription medications filled, and a range of comorbidities that may be risk factors for VTE (Table 1; eTable 1 in the Supplement).19

    Statistical Analysis

    We tabulated baseline patient characteristics for patients with and without CISD and separately for each CISD component condition: psoriasis, AD, AA, vitiligo, and HS. We recorded the duration of follow-up, including reasons for censoring, and computed incidence rates with 95% CIs.

    To account for potential differences in VTE risk factors, we used propensity score (PS) matching.20,21 We included all pre-exposure patient characteristics without further variable selection as independent variables for PS matching. The PS predicting the presence or absence of CISD was estimated with a multivariable logistic regression for each CISD component condition. Propensity score matching was performed using 1:1 nearest neighbor matching with a maximum caliper of 0.02 on the PS scale.22 We computed standardized differences in the covariates between the 2 groups after PS matching to quantify the balance achieved at baseline.23,24 Kaplan-Meier curves were plotted for CISD and each component condition after PS matching. Hazard ratios (HRs) comparing the incidence of VTE in patients with vs without CISD were estimated by fitting Cox proportional hazards regression models.25

    To study effect modification, we separately analyzed 3 subgroups of patients: those (1) with no previous use of systemic immunomodulators, (2) aged 18 to 49 years, and (3) aged 50 years or older. To demonstrate the robustness of our approach, we included a positive control (rheumatoid arthritis vs non–rheumatoid arthritis) and a negative tracer outcome (skull or face fracture). In a sensitivity analysis, we included additional adjustment for nonmelanoma skin cancer requiring excision. All analyses were conducted using the Aetion Evidence Platform, version 4.12,26 which has been validated.27-29

    Results

    We identified 158 123 qualifying patients with a diagnosis of a CISD, which comprised patients with psoriasis (n = 96 138), AD (n = 30 418), AA (n = 17 889), vitiligo (n = 7735), and HS (n = 5934). We identified 1 570 387 patients without a CISD through 10:1 risk-set sampling. The median follow-up time was 1.9 years (interquartile range, 0.8-4.0 years) in patients with CISD and 2.0 years (interquartile range, 0.8-4.2) in those without CISD; 25% had 4.2 years of follow-up or longer. The reasons for censoring were similar between groups (eTable 4 in the Supplement).

    Patients with a CISD were overall younger than those without CISD (mean [SD], 47.83 [16.40] vs 51.58 [18.42] years) (Table 1). The youngest patients were those with HS (mean [SD], 37.68 [13.22] years) and alopecia (mean [SD], 41.78 [14.35] years). There was a slight female predominance in the non-CISD (922 155 [58.7%]), AD (18 493 [60.8%]), and alopecia (10 555 [59.0%]) cohorts and a female predominance in the HS (4544 [76.6%]). Across all patients, White race was most common, with a slightly lower percentage in the vitiligo (4546 [58.8%]), alopecia (10 614 [59.3%]), and HS (3931 [66.2%]) cohorts. Use of systemic immunomodulators was more prevalent in patients with a CISD (nonbiologic immunomodulator, 8404 [5.3%] and biologic immunomodulator, 15 589 [9.9%]) compared with those without a CISD (nonbiologic immunomodulator, 99 [1.3%] and biologic immunomodulator, 11 [0.1%]). During follow-up, use of systemic immunomodulators increased in patients with CISD (eTable 5 in the Supplement). Across CISDs, we observed that, before matching, the demographics, comorbidity patterns, and systemic immunomodulatory treatments were aligned with the known epidemiologic factors and treatment guidelines.30-39 As an example, patients with HS were younger (37.68 vs 51.58 years in patients with non-CISD), more commonly female (76.6% vs 58.7%), Black (1062 [17.9%] vs 95 035 [6.1%]), and obese (1462 [24.6%] vs 128 263 [8.2%]), with increased use of tobacco (694 [11.7%] vs 73 151 [4.7%]), isotretinoin (237 [4.0%] vs 24 507 [1.6%]), oral contraceptives (1168 [19.7%] vs 211 249 [13.5%]) and tumor necrosis factor inhibitors (329 [5.5%] vs 588 [0%) as anticipated (Table 1).34 After 1:1 PS matching, all exposure patients (ie, those with CISDs) were matched with 1 of 10 non-CISD candidates. The distribution of all patient characteristics was well balanced between groups after 1:1 PS matching (eTable 2 in the Supplement).

    Incidence of VTE

    Before PS matching, the incidence rate per 1000 person-years of inpatient or outpatient VTE with evidence of treatment was 1.57 in psoriasis, 1.83 in AD, 0.94 in AA, 0.93 in vitiligo, and 1.65 in HS compared with 1.76 in patients without CISD (Table 2). Overall, the incidence rate of VTE was 1.53 per 1000 person-years in CISD vs 1.76 in non-CISD. The incidence rate was almost unchanged when excluding patients with systemic immunomodulator use (1.5) (Figure 1).

    After PS matching, the risk of inpatient or outpatient VTE was not significantly increased in patients with psoriasis (HR, 0.86; 95% CI, 0.75-0.99), AD (HR, 1.19; 95% CI, 0.95-1.48), AA (HR, 0.97; 95% CI, 0.65-1.46), vitiligo (HR, 0.90; 95% CI, 0.49-1.65), or CISD overall (HR, 0.94; 95% CI, 0.84-1.05), when compared with those without a CISD. We observed a numerically increased risk in patients with HS (HR, 1.64; 95% CI, 0.82-3.27), which was based on 34 events among 13 120 patient pairs resulting in wide 95% CIs (Figure 1). Kaplan-Meier plots reflected these results, showing no increase in VTE risk for the CISD vs non-CISD groups over up to 13 years of follow-up (Figure 2).

    Incidence of Inpatient VTE Only

    When limiting our outcome definition to inpatient VTE, the incidence rate (per 1000 person-years) was 0.59 in psoriasis, 0.69 in AD, 0.35 in AA, 0.42 in vitiligo, and 0.60 in HS compared with 0.75 in non-CISD (Table 2). Overall, the incidence rate of inpatient VTE was 0.58 in CISD vs 0.75 in non-CISD. After PS matching, the risk of inpatient VTE was not increased in patients with psoriasis (HR, 0.73; 95% CI, 0.59-0.91), AD (HR, 1.04; 95% CI, 0.73-1.48), AA (HR, 0.89; 95% CI, 0.46-1.70), HS (HR, 0.79; 95% CI, 0.29-2.11), or CISD overall (HR, 0.85; 95% CI, 0.72-1.01), compared with patients without a CISD. We observed a numerically increased risk in patients with vitiligo (HR, 1.48; 95% CI, 0.53-4.17), based on only 16 inpatient events (Table 2).

    Sensitivity Analyses

    The incidence rate remained almost unchanged when excluding patients using systemic immunomodulators in CISD vs non-CISD. This was also observed for the component conditions (Figure 2). Another sensitivity analysis that included eczema not otherwise specified in our definition of AD did not change the HR estimate meaningfully. Additional adjustment for nonmelanoma skin cancer with excision also did not change the effect estimates meaningfully (eTable 11 in the Supplement).

    Stratified by age, the incidence (per 1000 person-years) of inpatient and outpatient VTE events in patients aged 18 to 49 years was 0.8 vs 0.7 in patients without a CISD and, in those aged 50 years or older, was 2.3 vs 2.5 in patients without a CISD. The increased incidence among patients aged 50 years or older was consistently observed for the component conditions as well (Figure 2; eTable 6, eTable 7, and eTable 8 in the Supplement provide more details). After PS matching, there was no significant difference in VTE risk in patients with vs without CISD who were otherwise comparable in those aged 18 to 49 years (HR, 1.00; 95% CI, 0.81-1.25) or 50 years or older (HR, 1.01; 95% CI, 0.88-1.14) (Figure 2).

    To examine the sensitivity of our approach to detect elevated VTE risk, we replicated earlier findings of Kim et al3 of a 2-fold increase in VTE risk among patients with rheumatoid arthritis (HR, 2.01; 95% CI, 1.83-2.21) (eTable 10 in the Supplement). To further test the robustness of our study approach, we estimated the incidence of skull or face fractures as a negative tracer outcome that is not possibly associated with CISD. We found no association with our negative tracer for CISD (HR, 0.96; 95% CI, 0.86-1.07) or across all component conditions (Table 2).

    Discussion

    In this large-scale study in a commercially insured US population, patients with CISD diagnosed by a dermatologist and without known risk factors for VTE had an incidence rate of VTE of 1.53 per 1000 person-years, in line with those reported by population-representative surveys.40-43 None of the 5 component CISD conditions (psoriasis, AD, AA, vitiligo, and HS) showed elevated VTE risks compared with risks in individuals without CISD. The VTE incidence is mostly associated with patients aged 50 years or older as expected, given older age is a risk factor for VTE.40 Findings may generalize to the entire spectrum of patients treated by dermatologists capturing the patients seen in routine practice, but may have limited generalizability to underinsured patients.

    Active systemic inflammation modulates thrombosis and can affect hypercoagulability.44 Inflammatory mediators and cytokines associated with VTE, including tumor necrosis factor and interleukin-6, are elevated in chronic inflammatory diseases, such as inflammatory bowel disease and rheumatoid arthritis.45 Although these and other inflammatory mediators are involved in some CISDs, the overall lower inflammatory burden in patients with such dermatologic conditions may explain our study findings of no increased background risk of VTE in CISDs.

    To our knowledge, this is the first large-scale study to examine the background VTE risk in CISD that specifically compared patients with CISD against those who are similar in all characteristics except they do not have a CISD. The study further separated 5 CISD conditions from each other to avoid contamination by patients with multiple conditions and estimate an interpretable background rate. For example, patients with psoriatic arthritis were excluded from the psoriasis group.

    Establishing a background rate of VTE is overdue for 2 reasons. First, nonskin chronic inflammatory diseases, such as rheumatoid arthritis, have long been linked to an elevated VTE risk owing to their substantially increased inflammatory state3,44,46; therefore, ruling out an association between CISDs and an increased VTE risk is useful. Second, several systemic immunomodulatory agents that have quickly gained popularity in dermatology, in particular, Janus kinase inhibitors, were associated with an increased risk for VTE in high doses among a cohort of patients with rheumatoid arthritis3 and a cohort of those with inflammatory bowel disease.47 Some of these Janus kinase inhibitors have recently received a secondary indication for the treatment of dermatologic diseases, and many new Janus kinase inhibitors that are currently in clinical trials have shown efficacy in treating dermatologic conditions including, but not limited to, psoriasis, AA, vitiligo, and AD.48 In this new era of effective systemic medications for CISD, having a background VTE risk established will help monitor the safety of newly marketed systemic treatments in dermatology.

    Limitations

    Establishing population-based incidence rates of serious medical events, such as VTE, in dermatologic conditions is subject to several considerations and limitations. The first challenge was a clear separation of homogeneous patient groups. A diagnosis of a CISD by a dermatologist is reliable, particularly if this diagnosis is confirmed over more than 1 visit.49-52 Although, to our knowledge, no ICD-10 code validation study has been conducted for AD, a US validation study investigating ICD-9 codes for AD found that 1 code for AD by a dermatologist had a positive predictive value of 67.7% (for definite or probable AD).53 We additionally required a second dermatologist visit with an AD code, which should have further increased the positive predictive value; however, we cannot completely rule out any misclassification of the underlying condition. Establishing a group of patients who have no CISD with high certainty is more difficult because subtle disease presentation or lack of access to a specialist may leave CISD unrecorded. We therefore required that our non-CISD comparison group equally have 2 dermatologist visits, but that did not result in a CISD recording. Assuming that a dermatologist would, with highest likelihood, diagnose and record an existing CISD may result in a group free of CISD. Patients with past CISD (eg, childhood AD), although sometimes recorded, may have not been detected through our screening; the VTE risk in those with lower inflammatory burden associated with nonactive disease is likely similar to those without CISD. Similarly, we used the fact that the CISD group had dermatologist visits to categorize patients into 5 mutually exclusive groups of well-characterized CISDs. The patient characteristics in each of the 5 conditions were as expected from clinical practice, which supports that we were able to capture the intended populations in claims data.54

    The second consideration was outcome surveillance. We used a broad outcome definition to include inpatient and outpatient VTE events, resulting in a representative incidence rate estimate; consequently, the incidence rate of VTE in our non-CISD population (0.75-1.76 per 1000 person-years) is in line with other population-based surveys (1.04-1.83 per 1000 person-years).40-43 For comparative analyses of incidence rates, it is desirable to have high-specificity end point definitions to result in unbiased HR estimates.55 For this reason, we included a secondary outcome limited to inpatient VTE events recorded as the primary discharge diagnosis.

    Separating the disease from its treatment and comorbidities presented the third important consideration of this study. To be able to draw conclusions on the VTE risk attributable to the underlying skin disease rather than its treatment or comorbidities, we balanced all patients with pertinent VTE risk factors; treatments, including use of systemic immunomodulators; and comorbidities via exclusions or 1:1 PS matching. This allowed us to make comparative statements on the VTE risk attributable to CISD or its component disease relative to patients who have equal risk factor distributions but not the skin disease of interest. Differences in results between our work and others are attributable to the above, particularly, excluding patients with cancer from our analysis, adding confounder adjustment with 1:1 PS matching including matching on systemic immunomodulator use, not limiting the population to an inpatient setting, and using a cohort study design rather than a cross-sectional design.

    Insensitive methods may lead to null findings, which is the fourth consideration in our study. Other studies had shown an increase in VTE risk, for example, in patients with severe psoriasis, which we cannot reproduce. This is explained by our focus on the disease independent of its treatment, comorbidities, and VTE risk factors (eg, cancer). Nevertheless, negative findings need special scrutiny of the study methods to avoid false-negative results. First, we used the same study method to replicate the known finding of an increased VTE risk in patients with rheumatoid arthritis. Second, we used a negative tracer outcome (skull or face fractures that are unrelated to skin diseases) and, as expected, found no association.

    The fifth consideration involved residual confounding. Claims data are known to underrecord several risk factors for VTE, such as smoking and obesity. Although these are weaker risk factors than hormone therapy, trauma, or coagulopathy, we attempted to indirectly identify those condition via diagnoses related to smoking (eg, chronic obstructive pulmonary disease and chronic bronchitis), and obesity (diabetes, varicose veins, statin use, and hyperlipidemia).

    Effect modification by CISD severity was the sixth consideration. Disease severity and inflammatory burden may modify the risk of VTE but are difficult to assess in claims data. This study population captured patients who sought specialty care (2 dermatologist visits in the past year), of whom 25% used systemic immunomodulators or systemic corticosteroids; these characteristics may suggest that the population comprises more patients with moderate to severe disease. The subgroup analysis without systemic immunomodulator use may represent patients with milder disease.

    Conclusions

    In this large-scale analysis of claims data, we found no indication of an increased background rate of clinical VTE events in a representative population of patients with dermatologist-diagnosed CISD.

    Back to top
    Article Information

    Accepted for Publication: April 2, 2021.

    Published Online: May 26, 2021. doi:10.1001/jamadermatol.2021.1570

    Corresponding Author: Maria Schneeweiss, MD, Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women’s Hospital, One Brigham Circle, Ste 3030, Boston, MA 02120 (mschneeweiss@bwh.harvard.edu).

    Author Contributions: Dr M. Schneeweiss had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: M. Schneeweiss, Kim, Wyss, Merola, S. Schneeweiss.

    Acquisition, analysis, or interpretation of data: M. Schneeweiss, Kim, Jin, Chin, Mostaghimi, Silverberg, S. Schneeweiss.

    Drafting of the manuscript: M. Schneeweiss, Wyss, Chin, Merola.

    Critical revision of the manuscript for important intellectual content: Kim, Jin, Mostaghimi, Silverberg, S. Schneeweiss.

    Statistical analysis: M. Schneeweiss, Jin, Chin.

    Obtained funding: S. Schneeweiss.

    Administrative, technical, or material support: M. Schneeweiss, Merola.

    Supervision: M. Schneeweiss, Kim, Merola, Mostaghimi, S. Schneeweiss.

    Conflict of Interest Disclosures: Dr Kim reported receiving grants from Pfizer, AbbVie, Roche, and Bristol-Myers Squibb outside the submitted work. Dr Wyss reported receiving funding from Aetion for serving as a consultant for work unrelated to the topic of this study. Dr Merola reported receiving personal fees from AbbVie, Amgen, Bayer, Lilly, Novartis, Janssen, Celgene, Biogen, Pfizer, and BMS outside the submitted work; and receives funding from Dermavant, UCB, Sanofi, and Leo for serving as a principal investigator. Dr Mostaghimi reported receiving personal fees from Pfizer, Concert, Lilly, Digital Diagnostics, AbbVie, and hims and hers equity outside the submitted work. Dr Silverberg reported receiving personal fees from Eli Lilly and Company during the conduct of the study and personal fees from AbbVie, Afyx, Arena, Asana, BioMX, Bluefin, Bodewell, Boehringer Ingelheim, Celgene, Dermavant, Dermira, Eli Lilly, Galderma, GlaxoSmithKline, Incyte, Kiniksa, Leo, Luna, Menlo, Novartis, Pfizer, RAPT, Regeneron, and Sanofi-Genzyme outside the submitted work. Dr S. Schneeweiss reported receiving grants from Boehringer Ingelheim and personal fees from Aetion Inc as a consultant and having equity outside the submitted work. No other disclosures were reported.

    Funding/Support: This study was supported by the Division of Pharmacoepidemiology and Pharmacoeconomics, Brigham and Women’s Hospital.

    Role of the Funder/Sponsor: The funding organization had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    Disclaimer: Dr Mostaghimi is Associate Editor for JAMA Dermatology, but he was not involved in any of the decisions regarding review of the manuscript or its acceptance.

    References
    1.
    Silvestri  E, Scalera  A, Emmi  G,  et al.  Thrombosis in autoimmune diseases: a role for immunosuppressive treatments?   Semin Thromb Hemost. 2016;42(6):650-661. doi:10.1055/s-0036-1579642 PubMedGoogle Scholar
    2.
    Johannesdottir  SA, Schmidt  M, Horváth-Puhó  E, Sørensen  HT.  Autoimmune skin and connective tissue diseases and risk of venous thromboembolism: a population-based case-control study.   J Thromb Haemost. 2012;10(5):815-821. doi:10.1111/j.1538-7836.2012.04666.x PubMedGoogle ScholarCrossref
    3.
    Kim  SC, Schneeweiss  S, Liu  J, Solomon  DH.  Risk of venous thromboembolism in patients with rheumatoid arthritis.   Arthritis Care Res (Hoboken). 2013;65(10):1600-1607. doi:10.1002/acr.22039 PubMedGoogle Scholar
    4.
    Shaheen  MS, Silverberg  JI.  Association of inflammatory skin diseases with venous thromboembolism in US adults.   Arch Dermatol Res. 2021;313(4):281-289. PubMedGoogle Scholar
    5.
    Ogdie  A, Kay McGill  N, Shin  DB,  et al.  Risk of venous thromboembolism in patients with psoriatic arthritis, psoriasis and rheumatoid arthritis: a general population-based cohort study.   Eur Heart J. 2018;39(39):3608-3614. doi:10.1093/eurheartj/ehx145 PubMedGoogle ScholarCrossref
    6.
    Ahlehoff  O, Gislason  GH, Lindhardsen  J,  et al.  Psoriasis carries an increased risk of venous thromboembolism: a Danish nationwide cohort study.   PLoS One. 2011;6(3):e18125. doi:10.1371/journal.pone.0018125 PubMedGoogle Scholar
    7.
    Chung  WS, Lin  CL, Sung  FC, Lu  CC, Kao  CH.  Increased risk of venous thromboembolism in patients with dermatomyositis/polymyositis: a nationwide cohort study.   Thromb Res. 2014;134(3):622-626. doi:10.1016/j.thromres.2014.06.021 PubMedGoogle ScholarCrossref
    8.
    Rogers  MAM, Lin  P, Nallamothu  BK, Kim  C, Waljee  AK.  Longitudinal study of short-term corticosteroid use by working-age adults with diabetes mellitus: risks and mitigating factors.   J Diabetes. 2018;10(7):546-555. doi:10.1111/1753-0407.12631 PubMedGoogle ScholarCrossref
    9.
    Coelho  MC, Santos  CV, Vieira Neto  L, Gadelha  MR.  Adverse effects of glucocorticoids: coagulopathy.   Eur J Endocrinol. 2015;173(4):M11-M21. doi:10.1530/EJE-15-0198 PubMedGoogle ScholarCrossref
    10.
    Mease  P, Charles-Schoeman  C, Cohen  S,  et al.  Incidence of venous and arterial thromboembolic events reported in the tofacitinib rheumatoid arthritis, psoriasis and psoriatic arthritis development programmes and from real-world data.   Ann Rheum Dis. 2020;79(11):1400-1413. doi:10.1136/annrheumdis-2019-216761 PubMedGoogle ScholarCrossref
    11.
    Solimani  F, Meier  K, Ghoreschi  K.  Emerging topical and systemic JAK inhibitors in dermatology.   Front Immunol. 2019;10:2847. doi:10.3389/fimmu.2019.02847 PubMedGoogle ScholarCrossref
    12.
    Rudy  SF, Li  K, Moubayed  SP, Most  SP.  Risk of venous thromboembolism in patients with keratinocyte carcinoma.   JAMA Facial Plast Surg. 2018;20(6):453-459. doi:10.1001/jamafacial.2018.0331 PubMedGoogle ScholarCrossref
    13.
    Alotaibi  GS, Wu  C, Senthilselvan  A, McMurtry  MS.  The validity of ICD codes coupled with imaging procedure codes for identifying acute venous thromboembolism using administrative data.   Vasc Med. 2015;20(4):364-368. doi:10.1177/1358863X15573839 PubMedGoogle ScholarCrossref
    14.
    Walker  RF, Zakai  NA, MacLehose  RF,  et al.  Association of testosterone therapy with risk of venous thromboembolism among men with and without hypogonadism.   JAMA Intern Med. 2020;180(2):190-197. doi:10.1001/jamainternmed.2019.5135 PubMedGoogle ScholarCrossref
    15.
    Sanfilippo  KM, Wang  TF, Gage  BF, Liu  W, Carson  KR.  Improving accuracy of International Classification of Diseases codes for venous thromboembolism in administrative data.   Thromb Res. 2015;135(4):616-620. doi:10.1016/j.thromres.2015.01.012 PubMedGoogle ScholarCrossref
    16.
    Ammann  EM, Cuker  A, Carnahan  RM,  et al.  Chart validation of inpatient International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) administrative diagnosis codes for venous thromboembolism (VTE) among intravenous immune globulin (IGIV) users in the Sentinel Distributed Database.   Medicine (Baltimore). 2018;97(8):e9960. doi:10.1097/MD.0000000000009960 PubMedGoogle Scholar
    17.
    Tamariz  L, Harkins  T, Nair  V.  A systematic review of validated methods for identifying venous thromboembolism using administrative and claims data.   Pharmacoepidemiol Drug Saf. 2012;21(suppl 1):154-162. doi:10.1002/pds.2341 PubMedGoogle ScholarCrossref
    18.
    Schneeweiss  S.  A basic study design for expedited safety signal evaluation based on electronic healthcare data.   Pharmacoepidemiol Drug Saf. 2010;19(8):858-868. doi:10.1002/pds.1926 PubMedGoogle ScholarCrossref
    19.
    Anderson  FA  Jr, Spencer  FA.  Risk factors for venous thromboembolism.   Circulation. 2003;107(23)(suppl 1):I9-I16.PubMedGoogle Scholar
    20.
    Brookhart  MA, Schneeweiss  S, Rothman  KJ, Glynn  RJ, Avorn  J, Stürmer  T.  Variable selection for propensity score models.   Am J Epidemiol. 2006;163(12):1149-1156. doi:10.1093/aje/kwj149 PubMedGoogle ScholarCrossref
    21.
    Rosenbaum  PR, Rubin DB.  The central role of the propensity score in observational studies for causal effects.   Biometrika. 1983;70(1):41-55. doi:10.1093/biomet/70.1.41 Google ScholarCrossref
    22.
    Rassen  JA, Shelat  AA, Myers  J, Glynn  RJ, Rothman  KJ, Schneeweiss  S.  One-to-many propensity score matching in cohort studies.   Pharmacoepidemiol Drug Saf. 2012;21(suppl 2):69-80. doi:10.1002/pds.3263 PubMedGoogle ScholarCrossref
    23.
    Franklin  JM, Rassen  JA, Ackermann  D, Bartels  DB, Schneeweiss  S.  Metrics for covariate balance in cohort studies of causal effects.   Stat Med. 2014;33(10):1685-1699. doi:10.1002/sim.6058 PubMedGoogle ScholarCrossref
    24.
    Austin  PC.  Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples.   Stat Med. 2009;28(25):3083-3107. doi:10.1002/sim.3697 PubMedGoogle ScholarCrossref
    25.
    Cox  DR.  Regression models and life tables.   J Royal Stat Society Series B. 1972;34(2):187-220.Google Scholar
    26.
    Aetion Inc. Aetion evidence platform. Software for real-world data analysis. 2020. Accessed February 1, 2021. http://aetion.com
    27.
    Wang  SV, Verpillat  P, Rassen  JA, Patrick  A, Garry  EM, Bartels  DB.  Transparency and reproducibility of observational cohort studies using large healthcare Databases.   Clin Pharmacol Ther. 2016;99(3):325-332. doi:10.1002/cpt.329 PubMedGoogle ScholarCrossref
    28.
    Kim  SC, Solomon  DH, Rogers  JR,  et al.  Cardiovascular safety of tocilizumab versus tumor necrosis factor inhibitors in patients with rheumatoid arthritis: a multi-database cohort study.   Arthritis Rheumatol. 2017;69(6):1154-1164. doi:10.1002/art.40084 PubMedGoogle ScholarCrossref
    29.
    Patorno  E, Schneeweiss  S, Gopalakrishnan  C, Martin  D, Franklin  JM.  Using real-world data to predict findings of an ongoing phase IV cardiovascular outcome trial: Cardiovascular Safety of Linagliptin Versus Glimepiride.   Diabetes Care. 2019;42(12):2204-2210. doi:10.2337/dc19-0069 PubMedGoogle ScholarCrossref
    30.
    Sidbury  R, Davis  DM, Cohen  DE,  et al; American Academy of Dermatology.  Guidelines of care for the management of atopic dermatitis: section 3—management and treatment with phototherapy and systemic agents.   J Am Acad Dermatol. 2014;71(2):327-349. doi:10.1016/j.jaad.2014.03.030 PubMedGoogle ScholarCrossref
    31.
    Menter  A, Strober  BE, Kaplan  DH,  et al.  Joint AAD-NPF guidelines of care for the management and treatment of psoriasis with biologics.   J Am Acad Dermatol. 2019;80(4):1029-1072. doi:10.1016/j.jaad.2018.11.057 PubMedGoogle ScholarCrossref
    32.
    Alikhan  A, Sayed  C, Alavi  A,  et al.  North American clinical management guidelines for hidradenitis suppurativa: a publication from the United States and Canadian Hidradenitis Suppurativa Foundations: part II—topical, intralesional, and systemic medical management.   J Am Acad Dermatol. 2019;81(1):91-101. doi:10.1016/j.jaad.2019.02.068 PubMedGoogle ScholarCrossref
    33.
    Messenger  AG, McKillop  J, Farrant  P, McDonagh  AJ, Sladden  M.  British Association of Dermatologists’ guidelines for the management of alopecia areata: 2012.   Br J Dermatol. 2012;166(5):916-926. doi:10.1111/j.1365-2133.2012.10955.x PubMedGoogle ScholarCrossref
    34.
    Ingram  JR.  The epidemiology of hidradenitis suppurativa.   Br J Dermatol. 2020;183(6):990-998. doi:10.1111/bjd.19435 PubMedGoogle ScholarCrossref
    35.
    Lee  HH, Gwillim  E, Patel  KR,  et al.  Epidemiology of alopecia areata, ophiasis, totalis, and universalis: a systematic review and meta-analysis.   J Am Acad Dermatol. 2020;82(3):675-682. doi:10.1016/j.jaad.2019.08.032 PubMedGoogle ScholarCrossref
    36.
    Langan  SM, Irvine  AD, Weidinger  S.  Atopic dermatitis.   Lancet. 2020;396(10247):345-360. doi:10.1016/S0140-6736(20)31286-1 PubMedGoogle ScholarCrossref
    37.
    Armstrong  AW, Read  C.  Pathophysiology, clinical presentation, and treatment of psoriasis: a review.   JAMA. 2020;323(19):1945-1960. doi:10.1001/jama.2020.4006 PubMedGoogle ScholarCrossref
    38.
    Rodrigues  M, Ezzedine  K, Hamzavi  I, Pandya  AG, Harris  JE; Vitiligo Working Group.  Current and emerging treatments for vitiligo.   J Am Acad Dermatol. 2017;77(1):17-29. doi:10.1016/j.jaad.2016.11.010 PubMedGoogle ScholarCrossref
    39.
    Ezzedine  K, Eleftheriadou  V, Whitton  M, van Geel  N.  Vitiligo.   Lancet. 2015;386(9988):74-84. doi:10.1016/S0140-6736(14)60763-7 PubMedGoogle ScholarCrossref
    40.
    Heit  JA.  Epidemiology of venous thromboembolism.   Nat Rev Cardiol. 2015;12(8):464-474. doi:10.1038/nrcardio.2015.83 PubMedGoogle ScholarCrossref
    41.
    Silverstein  MD, Heit  JA, Mohr  DN, Petterson  TM, O’Fallon  WM, Melton  LJ  III.  Trends in the incidence of deep vein thrombosis and pulmonary embolism: a 25-year population-based study.   Arch Intern Med. 1998;158(6):585-593. doi:10.1001/archinte.158.6.585 PubMedGoogle ScholarCrossref
    42.
    Cushman  M, Tsai  AW, White  RH,  et al.  Deep vein thrombosis and pulmonary embolism in two cohorts: the longitudinal investigation of thromboembolism etiology.   Am J Med. 2004;117(1):19-25. doi:10.1016/j.amjmed.2004.01.018 PubMedGoogle ScholarCrossref
    43.
    Tagalakis  V, Patenaude  V, Kahn  SR, Suissa  S.  Incidence of and mortality from venous thromboembolism in a real-world population: the Q-VTE Study Cohort.   Am J Med. 2013;126(9):832.e13-832.e21. doi:10.1016/j.amjmed.2013.02.024PubMedGoogle ScholarCrossref
    44.
    Fox  EA, Kahn  SR.  The relationship between inflammation and venous thrombosis: a systematic review of clinical studies.   Thromb Haemost. 2005;94(2):362-365. doi:10.1160/TH05-04-0266 PubMedGoogle Scholar
    45.
    Branchford  BR, Carpenter  SL.  The role of inflammation in venous thromboembolism.   Front Pediatr. 2018;6:142. doi:10.3389/fped.2018.00142 PubMedGoogle ScholarCrossref
    46.
    van der Poll  T, Büller  HR, ten Cate  H,  et al.  Activation of coagulation after administration of tumor necrosis factor to normal subjects.   N Engl J Med. 1990;322(23):1622-1627. doi:10.1056/NEJM199006073222302 PubMedGoogle ScholarCrossref
    47.
    Sandborn  WJ, Panés  J, Sands  BE,  et al.  Venous thromboembolic events in the tofacitinib ulcerative colitis clinical development programme.   Aliment Pharmacol Ther. 2019;50(10):1068-1076. doi:10.1111/apt.15514 PubMedGoogle ScholarCrossref
    48.
    Newsom  M, Bashyam  AM, Balogh  EA, Feldman  SR, Strowd  LC.  New and emerging systemic treatments for atopic dermatitis.   Drugs. 2020;80(11):1041-1052. doi:10.1007/s40265-020-01335-7 PubMedGoogle ScholarCrossref
    49.
    Asgari  MM, Wu  JJ, Gelfand  JM,  et al.  Validity of diagnostic codes and prevalence of psoriasis and psoriatic arthritis in a managed care population, 1996-2009.   Pharmacoepidemiol Drug Saf. 2013;22(8):842-849. doi:10.1002/pds.3447 PubMedGoogle ScholarCrossref
    50.
    Löfvendahl  S, Theander  E, Svensson  Å, Carlsson  KS, Englund  M, Petersson  IF.  Validity of diagnostic codes and prevalence of physician-diagnosed psoriasis and psoriatic arthritis in southern Sweden—a population-based register study.   PLoS One. 2014;9(5):e98024. doi:10.1371/journal.pone.0098024 PubMedGoogle Scholar
    51.
    Kim  GE, Shlyankevich  J, Kimball  AB.  The validity of the diagnostic code for hidradenitis suppurativa in an electronic database.   Br J Dermatol. 2014;171(2):338-342. doi:10.1111/bjd.13041 PubMedGoogle ScholarCrossref
    52.
    Lavian  J, Li  SJ, Lee  EY,  et al.  Validation of case identification for alopecia areata using International Classification of Diseases coding.   Int J Trichology. 2020;12(5):234-237. doi:10.4103/ijt.ijt_67_20 PubMedGoogle Scholar
    53.
    Hsu  DY, Dalal  P, Sable  KA,  et al.  Validation of International Classification of Disease Ninth Revision codes for atopic dermatitis.   Allergy. 2017;72(7):1091-1095. doi:10.1111/all.13113 PubMedGoogle ScholarCrossref
    54.
    Bolognia  J, Jorizzo JL, Schaffer JV, eds . Dermatology. 4th ed. Elsevier; 2018.
    55.
    Schneeweiss  S, Glynn  RJ, Tsai  EH, Avorn  J, Solomon  DH.  Adjusting for unmeasured confounders in pharmacoepidemiologic claims data using external information: the example of COX2 inhibitors and myocardial infarction.   Epidemiology. 2005;16(1):17-24. doi:10.1097/01.ede.0000147164.11879.b5 PubMedGoogle ScholarCrossref
    ×