Advertisement

Depression Is Associated with an Increased Risk of Psoriatic Arthritis among Patients with Psoriasis: A Population-Based Study

Open AccessPublished:February 22, 2017DOI:https://doi.org/10.1016/j.jid.2016.11.032
      The factors that contribute to the development of psoriatic arthritis (PsA) among patients with psoriasis are not well known; however, systemic inflammation is believed to be important. On the basis of recent laboratory work demonstrating that major depressive disorder (MDD) is associated with increased systemic inflammation, we hypothesized that patients with psoriasis who develop MDD are at increased risk of subsequently developing PsA. We utilized The Health Improvement Network, a primary care medical records database, to identify 73,447 individuals with psoriasis. Patients were followed up to 25 years until the development of the primary outcome of PsA or the censor date. The exposure of interest was the development of MDD. Cox proportional-hazards models showed that patients with psoriasis who developed MDD were at significantly increased risk of subsequently developing PsA compared with patients who did not develop MDD, even after accounting for numerous covariates (hazard ratio 1.37, 95% confidence interval 1.05–1.80, P = 0.021). This result was maintained through numerous sensitivity analyses. These data support the hypothesis that MDD increases the risk of developing PsA among patients with psoriasis, suggesting a need for heightened prevention and management of MDD in patients with psoriasis.

      Abbreviations:

      CI (confidence interval), HR (hazard ratio), IQR (interquartile range), MDD (major depressive disorder), PsA (psoriatic arthritis), THIN (The Health Improvement Network)

      Introduction

      Psoriasis is an inflammatory skin disease characterized by pruritic, erythematous, scaling papules, and plaques (
      • Nestle F.O.
      • Kaplan D.H.
      • Barker J.
      Psoriasis.
      ). Approximately 8.5% of patients with psoriasis have psoriatic arthritis (PsA) (
      • Ogdie A.
      • Langan S.
      • Love T.
      • Haynes K.
      • Shin D.
      • Seminara N.
      • et al.
      Prevalence and treatment patterns of psoriatic arthritis in the UK.
      ), which is characterized by psoriasis plus inflammation of joints and other extra-articular manifestations. Although many studies have highlighted potential risk factors for PsA development, such as psoriasis severity (
      • Reich K.
      • Kruger K.
      • Mossner R.
      • Augustin M.
      Epidemiology and clinical pattern of psoriatic arthritis in Germany: a prospective interdisciplinary epidemiological study of 1511 patients with plaque-type psoriasis.
      ), obesity (
      • Love T.J.
      • Zhu Y.
      • Zhang Y.
      • Wall-Burns L.
      • Ogdie A.
      • Gelfand J.M.
      • et al.
      Obesity and the risk of psoriatic arthritis: a population-based study.
      ), alcohol use (
      • Wu S.
      • Cho E.
      • Li W.Q.
      • Han J.
      • Qureshi A.A.
      Alcohol intake and risk of incident psoriatic arthritis in women.
      ), and smoking (
      • Li W.
      • Han J.
      • Qureshi A.A.
      Smoking and risk of incident psoriatic arthritis in US women.
      ), the pathophysiologic mechanisms that contribute to the development of PsA among patients with psoriasis are not well known. Given the similar genetic and inflammatory associations between psoriasis and PsA (
      • Huffmeier U.
      • Uebe S.
      • Ekici A.B.
      • Bowes J.
      • Giardina E.
      • Korendowych E.
      • et al.
      Common variants at TRAF3IP2 are associated with susceptibility to psoriatic arthritis and psoriasis.
      ,
      • Veale D.J.
      • Ritchlin C.
      • FitzGerald O.
      Immunopathology of psoriasis and psoriatic arthritis.
      ), immune activation to external stressors is a proposed mechanism (
      • Barnas J.L.
      • Ritchlin C.T.
      Etiology and pathogenesis of psoriatic arthritis.
      ). It is believed that elevated systemic inflammation contributes to more severe psoriatic disease (
      • Dowlatshahi E.A.
      • van der Voort E.A.
      • Arends L.R.
      • Nijsten T.
      Markers of systemic inflammation in psoriasis: a systematic review and meta-analysis.
      ), suggesting the possibility of a particular “threshold” that is crossed leading to manifestations beyond skin disease.
      Recently, it has been demonstrated that individuals with major depressive disorder (MDD) have elevated serum inflammatory markers such as tumor necrosis factor-α (
      • Dowlati Y.
      • Herrmann N.
      • Swardfager W.
      • Liu H.
      • Sham L.
      • Reim E.K.
      • et al.
      A meta-analysis of cytokines in major depression.
      ), and that increased levels of tumor necrosis factor-α are associated with reduced activity of serotonin transporters (
      • Krishnadas R.
      • Nicol A.
      • Sassarini J.
      • Puri N.
      • Burden A.D.
      • Leman J.
      • et al.
      Circulating tumor necrosis factor is highly correlated with brainstem serotonin transporter availability in humans.
      ) relevant to MDD pathophysiology. Also, both psoriasis and PsA are associated with MDD (
      • Dalgard F.J.
      • Gieler U.
      • Tomas-Aragones L.
      • Lien L.
      • Poot F.
      • Jemec G.B.
      • et al.
      The psychological burden of skin diseases: a cross-sectional multicenter study among dermatological out-patients in 13 European countries.
      ,
      • Dommasch E.D.
      • Li T.
      • Okereke O.I.
      • Li Y.
      • Qureshi A.A.
      • Cho E.
      Risk of depression in women with psoriasis: a cohort study.
      ,
      • Kurd S.K.
      • Troxel A.B.
      • Crits-Christoph P.
      • Gelfand J.M.
      The risk of depression, anxiety, and suicidality in patients with psoriasis: a population-based cohort study.
      ), commonly attributed to the physical and/or cosmetic disability incurred by psoriatic disease. The concept of high levels of systemic inflammation driving psoriatic disease severity (
      • Dowlatshahi E.A.
      • van der Voort E.A.
      • Arends L.R.
      • Nijsten T.
      Markers of systemic inflammation in psoriasis: a systematic review and meta-analysis.
      ) warrants an investigation on whether the presence of MDD may in fact predispose to PsA in patients with psoriasis. In addition, MDD is known to be associated with poor health behaviors such as unhealthy diet and physical inactivity (
      • Trudel-Fitzgerald C.
      • Tworoger S.S.
      • Poole E.M.
      • Williams D.R.
      • Kubzansky L.D.
      Prospective changes in healthy lifestyle among midlife women: when psychological symptoms get in the way.
      ), which could also contribute toward the development of PsA. Population-based studies are ideal to address the role of MDD in PsA development, given their large and nationally representative samples, lengthy follow-up times, and ability to determine risk of incident PsA given an MDD exposure in patients with psoriasis.
      Our objective was to evaluate clinical data to elucidate the possible role of MDD in the progression of inflammatory disease from one organ system (the skin) to involving multiple organ systems, and thereby highlight a possible risk factor for PsA development among patients with psoriasis. We hypothesized that patients with psoriasis and comorbid incident MDD would be at a greater risk of developing PsA compared with those who have psoriasis without comorbid incident MDD.

      Results

      Participants and descriptive data

      We utilized The Health Improvement Network (THIN), a general practice medical records database in the United Kingdom (UK), to identify 73,447 individuals with an incident psoriasis diagnosis who were followed until the development of PsA or the censor date (Figure 1). The median (interquartile range, IQR) age at psoriasis diagnosis was 49.5 years (IQR 19.0 years). The median follow-up time in the study was 5.1 years (IQR 6.9 years).
      Figure 1
      Figure 1Study flow diagram. The figure shows the approach for identifying eligible patients with psoriasis within THIN and then identifying those who subsequently developed major depressive disorder (MDD) and/or psoriatic arthritis (PsA). THIN, The Health Improvement Network.
      Between the study start time at psoriasis diagnosis and the study end point at PsA diagnosis or censor date, we identified 5,216 (7.1%) patients with psoriasis who developed MDD. Among those who developed MDD, the median time from psoriasis diagnosis to MDD diagnosis was 3.1 years (IQR 4.7 years). Those who developed MDD were more likely to be younger, female, current smokers, with at least one comorbidity, socially deprived, and with moderate-severe psoriasis, and less likely to be obese or alcohol users (all P-values <0.0001, Table 1). Among patients with moderate-severe psoriasis, the most commonly used therapies were methotrexate and phototherapy (Table 2).
      Table 1Baseline characteristics of patients with psoriasis with and without depression
      VariableDepression (n = 5,216)No depression (n = 68,231)P-value
      Age<0.0001
       20–44 y2,783 (53.4)29,082 (42.6)
       45–90 y2,433 (46.6)39,149 (57.4)
      Sex<0.0001
       Females3,215 (61.6)33,464 (49.0)
      Obesity status<0.0001
       BMI ≤ 30 kg/m22,926 (56.1)40,631 (59.5)
       BMI > 30 kg/m2897 (17.2)12,514 (18.3)
       Missing1393 (26.7)15,086 (22.1)
      Smoking status<0.0001
       Current1,860 (35.7)18,497 (27.1)
       Ex-smoker833 (16.0)14,450 (21.2)
       Never2,023 (38.8)30,857 (45.2)
       Missing500 (9.6)4,427 (6.5)
      Alcohol use<0.0001
       User3,151 (60.4)44,397 (65.1)
       Nonuser819 (15.7)10,155 (14.9)
       Missing1,246 (23.9)13,679 (20.0)
      Charlson comorbidity index
      Higher = more comorbidities.
      <0.0001
       03,851 (73.8)53,308 (78.1)
       1896 (17.2)8,844 (13.0)
       2280 (5.4)3,615 (5.3)
       3114 (2.2)1,453 (2.1)
       ≥475 (1.4)1011 (1.5)
      Townsend deprivation index
      Higher = more socially deprived.
      <0.0001
       11,039 (19.9)16,694 (24.5)
       2955 (18.3)14,351 (21.0)
       31,005 (19.3)13,596 (19.9)
       41,130 (21.7)11,879 (17.4)
       5842 (16.1)8,276 (12.1)
       Missing245 (4.7)3,435 (5.0)
      Psoriasis severity<0.0001
       Mild4,984 (95.6)66,348 (97.2)
       Moderate-severe232 (4.4)1,883 (2.8)
      Values show the number (percentage) of patients with a given characteristic. Moderate-severe psoriasis defined based on therapy use (see the Materials and Methods section).
      Abbreviation: BMI, body mass index.
      1 Higher = more comorbidities.
      2 Higher = more socially deprived.
      Table 2Summary of all therapies used to define moderate-severe psoriasis
      TherapyDepression (n = 5,216)No depression (n = 68,231)Total
      Acitretin14 (0.27)126 (0.18)140
      Azathioprine15 (0.29)119 (0.17)134
      Cyclosporine30 (0.58)159 (0.23)189
      Etretinate0 (0.00)7 (0.01)7
      Hydroxycarbamide11 (0.21)76 (0.11)87
      Methotrexate102 (1.96)974 (1.43)1076
      Mycophenolic acid0 (0.00)16 (0.02)16
      Phototherapy73 (1.40)471 (0.69)544
      Total245 (4.70)1,948 (2.86)2,193
      Values represent the number of individuals (% of individuals relative to all patients in group). Moderate-severe psoriasis defined based on therapy use (see the Materials and Methods section). Total values for each group do not equate to values in Table 1 for number of patients with moderate-severe psoriasis as some patients had more than one therapy. For the expanded therapy list used in sensitivity analysis, see Supplementary Table S1 online.
      Among all patients with psoriasis, 1,466 (2.0%) developed PsA during the observation period. Among patients with psoriasis with MDD, the median time from MDD to PsA was 2.9 years (IQR 4.2 years). Among patients with psoriasis without MDD, the median time from psoriasis to PsA was 2.6 years (IQR 4.8 years).

      Outcomes data

      Using Cox-proportional hazards models, we found that patients with psoriasis who developed MDD had a significantly increased risk of subsequently developing PsA compared with those who did not develop MDD (unadjusted hazard ratio [HR] 1.56, 95% confidence interval [CI] 1.28–1.90, P < 0.0001). Using a likelihood ratio test simultaneously evaluating the significance of all interaction terms (each covariate by MDD), there was no evidence for effect modification (P = 0.387). Through backward elimination, alcohol use was the only covariate that was identified as having a potential confounding effect on the MDD and PsA association among patients with psoriasis. In the most parsimonious model, adjusted only for alcohol use, MDD remained a significant risk factor for the development of PsA among patients with psoriasis (HR 1.41, 95% CI 1.10–1.80, P = 0.007). To provide a conservative estimate of risk, when adjusted for all covariates including age, sex, obesity status, smoking status, alcohol use, Charlson comorbidity index, Townsend deprivation index, and psoriasis severity, the relationship remained significant (HR 1.37, 95% CI 1.05–1.80, P = 0.021). There was no evidence against the proportional hazards assumption based on the Shoenfeld residuals (P = 0.287). These results are summarized in Table 3 and in Kaplan-Meier failure curves in Figure 2. Together, these findings support our original hypothesis that MDD would significantly increase the risk of developing PsA among individuals with psoriasis.
      Table 3Hazard ratios for the risk of psoriatic arthritis associated with depression among patients with psoriasis
      ModelHazard ratio (95% CI)P-value
      Unadjusted model
       Depression (MDD)1.56 (1.28–1.90)<0.0001
      Alcohol use adjusted model
       Depression (MDD)1.41 (1.10–1.80)0.007
       Alcohol use1.07 (0.91–1.25)0.396
      Multivariable adjusted model
       Depression (MDD)1.37 (1.05–1.80)0.021
       Age0.72 (0.63–0.83)<0.0001
       Sex1.21 (1.06–1.38)0.005
       Obesity status1.61 (1.40–1.86)<0.0001
       Smoking status0.87 (0.80–0.94)<0.0001
       Alcohol use1.01 (0.85–1.20)0.917
       Charlson comorbidity index0.95 (087–1.04)0.265
       Townsend deprivation index1.00 (0.95–1.05)0.958
       Psoriasis severity5.02 (4.18–6.04)<0.0001
      Cox proportional hazards models were used to estimate the hazards ratio of developing psoriatic arthritis (PsA) based on whether psoriasis patients were exposed to major depressive disorder or not. As can be seen from these models, MDD significantly increases the risk of developing PsA among patients with psoriasis when using unadjusted models as well as models accounting for numerous covariates.
      Abbreviation: CI, confidence interval; MDD, major depressive disorder.
      Figure 2
      Figure 2Kaplan-Meier failure curves with development of psoriatic arthritis (PsA) as outcome. Here, it can be seen that study follow-up was up to 25 years, and that a greater proportion of individuals developed PsA in the MDD group (black) compared with the no-MDD group (dotted gray). Beyond 15 years of follow-up time, there were few patients remaining in observation in the MDD group (median time to PsA from MDD diagnosis of 2.9 years), and few developing the PsA outcome. MDD, major depressive disorder.

      Sensitivity analyses

      When restricting case definitions of psoriasis to only those whose diagnosis of psoriasis was at least 2 years after registration in THIN, MDD remained a significant predictor for the development of PsA after adjustment for all covariates (HR 1.41, 95% CI 1.05–1.90, P = 0.023). Similar results were obtained when restricting to those with at least 5 years in THIN before psoriasis diagnosis (HR 1.56, 95% CI 1.04–2.33, P = 0.030). These analyses confirmed that identification of incident psoriasis was achieved. We also conducted a sensitivity analysis to evaluate the use of a more inclusive therapy list for the definition of moderate-severe psoriasis. In this case, similar results were obtained showing that MDD increases the risk of developing PsA (HR 1.35, 95% CI 1.03–1.77, P = 0.030). An expanded therapy list with frequencies of use is shown in Supplementary Table S1 online. Together, these sensitivity analyses show that our main results do not appear to be sensitive to the wash out time used to identify incident psoriasis, nor the therapy choices selected for the definition of moderate-severe psoriasis.

      Discussion

      Associations between MDD and psoriasis or PsA have been documented previously and largely attributed to the physical or cosmetic disability incurred by psoriatic disease (
      • Dalgard F.J.
      • Gieler U.
      • Tomas-Aragones L.
      • Lien L.
      • Poot F.
      • Jemec G.B.
      • et al.
      The psychological burden of skin diseases: a cross-sectional multicenter study among dermatological out-patients in 13 European countries.
      ,
      • Kurd S.K.
      • Troxel A.B.
      • Crits-Christoph P.
      • Gelfand J.M.
      The risk of depression, anxiety, and suicidality in patients with psoriasis: a population-based cohort study.
      ,
      • McDonough E.
      • Ayearst R.
      • Eder L.
      • Chandran V.
      • Rosen C.F.
      • Thavaneswaran A.
      • et al.
      Depression and anxiety in psoriatic disease: prevalence and associated factors.
      ,
      • Schmitt J.M.
      • Ford D.E.
      Role of depression in quality of life for patients with psoriasis.
      ). The present study has expanded our knowledge in this area by identifying a temporal relationship between psoriasis, MDD, and PsA and conservatively demonstrates that MDD increases the risk of developing PsA among patients with psoriasis by as much as 37%. As such, this study highlights important considerations for clinicians, who should exercise heightened awareness and primary prevention strategies for MDD among patients with psoriasis, as MDD appears to significantly increase the risk of developing PsA.
      Because we based our definition of psoriasis severity on the use of systemic therapies (
      • Gelfand J.M.
      • Neimann A.L.
      • Shin D.B.
      • Wang X.
      • Margolis D.J.
      • Troxel A.B.
      Risk of myocardial infarction in patients with psoriasis.
      ,
      • Kurd S.K.
      • Troxel A.B.
      • Crits-Christoph P.
      • Gelfand J.M.
      The risk of depression, anxiety, and suicidality in patients with psoriasis: a population-based cohort study.
      ), some of which are also used to treat PsA, it could have been suspected that patients with moderate-severe psoriasis are protected from PsA as they are being treated for PsA before it is diagnosed. However, psoriasis severity was not found to be an effect modifier or confounder in this study, therefore suggesting that our definition of psoriasis severity was not positively or negatively associated with the outcome of PsA. Similarly, although smoking and obesity have been documented as risk factors for PsA previously (
      • Li W.
      • Han J.
      • Qureshi A.A.
      Smoking and risk of incident psoriatic arthritis in US women.
      ,
      • Love T.J.
      • Zhu Y.
      • Zhang Y.
      • Wall-Burns L.
      • Ogdie A.
      • Gelfand J.M.
      • et al.
      Obesity and the risk of psoriatic arthritis: a population-based study.
      ), these were not found to be effect modifiers or confounders in the present study. Conversely, only alcohol demonstrated a confounding effect that attenuated the risk of developing PsA among those with MDD, but the association remained significant.
      This study draws into question the role of MDD in the development of PsA. MDD is known to be associated with poor health behaviors such as unhealthy diet and physical inactivity (
      • Trudel-Fitzgerald C.
      • Tworoger S.S.
      • Poole E.M.
      • Williams D.R.
      • Kubzansky L.D.
      Prospective changes in healthy lifestyle among midlife women: when psychological symptoms get in the way.
      ), and thus it is possible that these negative health behaviors, indirectly captured in our data as MDD, contribute toward the development of PsA. Supporting this, recent studies have demonstrated that regular exercise (
      • Naldi L.
      • Conti A.
      • Cazzaniga S.
      • Patrizi A.
      • Pazzaglia M.
      • Lanzoni A.
      Diet and physical exercise in psoriasis: a randomized controlled trial.
      ), weight loss (
      • Gisondi P.
      • Del Giglio M.
      • Di Francesco V.
      • Zamboni M.
      • Girolomoni G.
      Weight loss improves the response of obese patients with moderate-to-severe chronic plaque psoriasis to low-dose cyclosporine therapy: a randomized, controlled, investigator-blinded clinical trial.
      ), and low-energy diets (
      • Jensen P.
      • Zachariae C.
      • Christensen R.
      • Geiker N.R.
      • Schaadt B.K.
      • Stender S.
      Effect of weight loss on the severity of psoriasis: a randomized clinical study.
      ) can all improve psoriasis symptoms and severity. Similarly, perceived lack of social support may lead to increased susceptibility to psoriasis exacerbations (
      • Picardi A.
      • Mazzotti E.
      • Gaetano P.
      • Cattaruzza M.S.
      • Baliva G.
      • Melchi C.F.
      Stress, social support, emotional regulation, and exacerbation of diffuse plaque psoriasis.
      ). Taken together, it is possible that detrimental health behaviors associated with MDD may be important factors that contribute to increased risk of PsA and should be explored in future studies.
      It is well known that both psoriasis and PsA are inflammatory diseases with similar underlying serum markers and genetic links (
      • Huffmeier U.
      • Uebe S.
      • Ekici A.B.
      • Bowes J.
      • Giardina E.
      • Korendowych E.
      • et al.
      Common variants at TRAF3IP2 are associated with susceptibility to psoriatic arthritis and psoriasis.
      ,
      • Veale D.J.
      • Ritchlin C.
      • FitzGerald O.
      Immunopathology of psoriasis and psoriatic arthritis.
      ); however, it is relatively recent that MDD is also believed to contribute to systemic inflammation (
      • Dowlati Y.
      • Herrmann N.
      • Swardfager W.
      • Liu H.
      • Sham L.
      • Reim E.K.
      • et al.
      A meta-analysis of cytokines in major depression.
      ). From our results, it remains possible that MDD superimposed on psoriasis may alter an individual’s inflammation so as to increase the risk of PsA development. This hypothesis is in agreement with basic science research showing that MDD is associated with increased systemic inflammation (
      • Iwata M.
      • Ota K.T.
      • Duman R.S.
      The inflammasome: pathways linking psychological stress, depression, and systemic illnesses.
      ,
      • Krishnadas R.
      • Nicol A.
      • Sassarini J.
      • Puri N.
      • Burden A.D.
      • Leman J.
      • et al.
      Circulating tumor necrosis factor is highly correlated with brainstem serotonin transporter availability in humans.
      ). In light of this, further study of MDD in the context of inflammatory disease seems crucial. In addition, these results raise the possibility that although MDD is a distinct medical diagnosis, it could be a potential manifestation of PsA for select patients, similar to anterior uveitis, enthesitis, or onycholysis. It should be highlighted that although our results support the concept of MDD and systemic inflammation, systemic inflammation was not objectively captured in our data, and thus it is not possible to confirm that this mechanism occurred in the present study.
      At present, it is not possible to discern whether MDD increased the risk of PsA by means of negative health behaviors or systemic inflammation, and this should be a focus of future research. Importantly, our data show that MDD contributed to an increased risk of PsA, but this relationship was not causal, that is, patients with psoriasis without MDD also developed PsA and not all patients with psoriasis with MDD developed PsA. Thus, other risk factors remain to be discovered that also contribute to the development of PsA.

      Strengths and limitations

      This study is strengthened by the use of a population-based design, a large and nationally representative sample, and long follow-up time. In addition, the study utilized medical records from a primary care database, and thus results are less likely to be affected by a specialty bias. Given the large sample size, caution must be exercised in determining the clinical relevance of statistically significant results.
      Patient misclassification was likely infrequent given previous validation studies on psoriasis and PsA in THIN (
      • Ogdie A.
      • Alehashemi S.
      • Love T.J.
      • Jiang Y.
      • Haynes K.
      • Hennessy S.
      • et al.
      Validity of psoriatic arthritis and capture of disease modifying antirheumatic drugs in the health improvement network.
      ,
      • Seminara N.M.
      • Abuabara K.
      • Shin D.B.
      • Langan S.M.
      • Kimmel S.E.
      • Margolis D.
      • et al.
      Validity of The Health Improvement Network (THIN) for the study of psoriasis.
      ); however, in cases where misclassification did occur, it is not expected that misclassification would occur differentially in favor or against diagnosis of psoriasis, MDD, PsA, or psoriasis severity because data were recorded in primary care offices where hypotheses of our study were not known.
      Although we did not assess health care utilization directly as a confounder in the association between exposure and outcome, which could introduce surveillance bias for the outcome, we did include the Charlson comorbidity index as an indirect measure. We also note that UK NICE guidelines (

      National Institute for Health and Care Excellence. Psoriasis: Assessment and Management, Clinical Guidelines. http://nice.org.uk/guidance/cg153 (accessed 24 October 2012).

      ) recommend annual screening of all patients with psoriasis for PsA using validated tools and referral to a rheumatologist for complete workup when PsA is suspected. As a result, both the referent and MDD groups should have been actively screened for PsA, and therefore the risk of surveillance bias should be minimal. In addition, given that 99% of the UK population is registered with a family physician and 5% of the UK population is captured in THIN, it is not likely that patients with psoriasis selected into our study would differ from other patients with psoriasis in the UK regarding their risk of developing MDD and PsA.
      The design used in this study allowed us to study a large number of patients to identify cases of PsA among patients with psoriasis—a relatively infrequent outcome. As a result, MDD was identified as a risk factor for PsA development. However, a limitation of this design is that we are unable to consider all possible confounding or mediating variables and thus are unable to develop a causal model for why MDD increases the risk of PsA. Future research is needed to identify potential mechanisms by which MDD increases the risk of PsA, and may include factors such as inflammation, lifestyle, or therapies associated with MDD.
      Lastly, although our analysis assessed MDD as a risk factor for PsA among patients with psoriasis, it is possible that this association may be explained in part by mediating effects, which were not assessed directly in this study. Similarly, serum markers of systemic inflammation were not evaluated in this study.

      Conclusions

      We utilized a large, nationally representative medical records database with up to 25 years of follow-up to identify that MDD significantly increased the risk of developing PsA among patients with psoriasis. Consequently, heightened attention for preventing and managing MDD in patients with psoriasis may be warranted. Future research is necessary to elucidate the possible mechanisms by which MDD contributes to the increased risk of PsA.

      Materials and Methods

      Data source, setting, and study design

      THIN was used as a data source. THIN is an electronic database that contains general practice medical records from more than 12 million individuals in the UK. In the UK, 99% of individuals are registered with a general practice physician, and THIN represents approximately 5% of the UK population. When a patient joins a THIN-registered practice, this date serves as their start date in THIN. Any medical history before joining THIN is not available; however, if an individual has a chronic disease, codes for these diseases are expected to appear after joining in THIN. In the UK, all specialist referrals originate from patients’ general practice physician, and data from specialists’ offices and hospitals are returned to the general practice for recording in THIN. Thus once in THIN, the patient’s full medical history is documented from the date of joining the practice, until death or transfer out of practice. Patients registered in THIN have similar age and sex distributions to the general UK population (
      • Wallace H.
      • Shorvon S.
      • Tallis R.
      Age-specific incidence and prevalence rates of treated epilepsy in an unselected population of 2,052,922 and age-specific fertility rates of women with epilepsy.
      ). Therefore, THIN is a large, representative database with longitudinal follow-up and detailed medical history. Data for this study were collected prospectively from the year 1987 to 15 May 2012, representing up to 25 years of follow-up. THIN has been used previously for the study of psoriasis (
      • Chiesa Fuxench Z.C.
      • Shin D.B.
      • Ogdie Beatty A.
      • Gelfand J.M.
      The Risk of cancer in patients with psoriasis: a population-based cohort study in the health improvement network.
      ,
      • Gelfand J.M.
      • Neimann A.L.
      • Shin D.B.
      • Wang X.
      • Margolis D.J.
      • Troxel A.B.
      Risk of myocardial infarction in patients with psoriasis.
      ,
      • Yeung H.
      • Takeshita J.
      • Mehta N.N.
      • Kimmel S.E.
      • Ogdie A.
      • Margolis D.J.
      • et al.
      Psoriasis severity and the prevalence of major medical comorbidity: a population-based study.
      ), PsA (
      • Dubreuil M.
      • Rho Y.H.
      • Man A.
      • Zhu Y.
      • Zhang Y.
      • Love T.J.
      • et al.
      Diabetes incidence in psoriatic arthritis, psoriasis and rheumatoid arthritis: a UK population-based cohort study.
      ,
      • Ogdie A.
      • Yu Y.
      • Haynes K.
      • Love T.J.
      • Maliha S.
      • Jiang Y.
      • et al.
      Risk of major cardiovascular events in patients with psoriatic arthritis, psoriasis and rheumatoid arthritis: a population-based cohort study.
      ), and MDD (
      • Kurd S.K.
      • Troxel A.B.
      • Crits-Christoph P.
      • Gelfand J.M.
      The risk of depression, anxiety, and suicidality in patients with psoriasis: a population-based cohort study.
      ,
      • Meier C.R.
      • Wilcock K.
      • Jick S.S.
      The risk of severe depression, psychosis or panic attacks with prophylactic antimalarials.
      ,
      • Rait G.
      • Walters K.
      • Griffin M.
      • Buszewicz M.
      • Petersen I.
      • Nazareth I.
      Recent trends in the incidence of recorded depression in primary care.
      ).

      Study population and sample size

      THIN was used to identify adults aged 20–90 years with new-onset (incident) psoriasis. Psoriasis was defined by the presence of any one of the Read codes for psoriasis validated previously (
      • Seminara N.M.
      • Abuabara K.
      • Shin D.B.
      • Langan S.M.
      • Kimmel S.E.
      • Margolis D.
      • et al.
      Validity of The Health Improvement Network (THIN) for the study of psoriasis.
      ). Read codes are the standard clinical codes used in UK general practice, based on International Classifications of Disease (ICD) codes. The approach by
      • Seminara N.M.
      • Abuabara K.
      • Shin D.B.
      • Langan S.M.
      • Kimmel S.E.
      • Margolis D.
      • et al.
      Validity of The Health Improvement Network (THIN) for the study of psoriasis.
      demonstrated a 90% positive predictive value for psoriasis when validated against a general practice physician’s clinical assessment. To be included in the study, the patient must have been enrolled in THIN for at least 1 year without a psoriasis code to ensure that cases of psoriasis were incident. This has been done previously when defining incident psoriasis in THIN (
      • Dubreuil M.
      • Rho Y.H.
      • Man A.
      • Zhu Y.
      • Zhang Y.
      • Love T.J.
      • et al.
      Diabetes incidence in psoriatic arthritis, psoriasis and rheumatoid arthritis: a UK population-based cohort study.
      ). By using a 1-year washout for inclusion in our study, prevalent cases of psoriasis who could have been treated before joining a THIN practice were excluded. In addition, any patients with diagnoses of MDD or PsA before psoriasis diagnosis were excluded. Together, these factors contribute to clear discernment of the temporal sequencing of events in the population data. Follow-up time began at first service date with a psoriasis code (after the patient’s practice reached acceptable mortality reporting). Any patients meeting these criteria were included in the study, yielding 73,447 individuals with psoriasis.

      Outcome

      The primary outcome was the development of PsA. Censoring occurred when the patient died, transferred out of the practice, or reached the end of the study period. The time of outcome was defined as the first date on which either the outcome of PsA was observed or the patient became censored. The development of PsA was defined by the presence of any one of the PsA Read codes previously validated (
      • Ogdie A.
      • Alehashemi S.
      • Love T.J.
      • Jiang Y.
      • Haynes K.
      • Hennessy S.
      • et al.
      Validity of psoriatic arthritis and capture of disease modifying antirheumatic drugs in the health improvement network.
      ). This method demonstrated an 85% positive predictive value for PsA when validated against a physician’s clinical assessment (
      • Ogdie A.
      • Alehashemi S.
      • Love T.J.
      • Jiang Y.
      • Haynes K.
      • Hennessy S.
      • et al.
      Validity of psoriatic arthritis and capture of disease modifying antirheumatic drugs in the health improvement network.
      ), thus contributing to a low risk of misclassification bias of PsA. Although exclusion of cases with Read codes for rheumatoid arthritis or osteoarthritis was considered, previous PsA validation research in THIN (
      • Ogdie A.
      • Alehashemi S.
      • Love T.J.
      • Jiang Y.
      • Haynes K.
      • Hennessy S.
      • et al.
      Validity of psoriatic arthritis and capture of disease modifying antirheumatic drugs in the health improvement network.
      ) has shown that this would not enhance our positive predictive value for selecting true cases of PsA, and so the recommended and validated approach of utilizing a single PsA code without any further exclusions was performed. Because diagnosis of PsA was based on the presence or absence of a PsA Read code, there were no missing data for this outcome. As this study aimed to estimate the risk of developing PsA among those with psoriasis, any cases of PsA preceding the psoriasis diagnosis were excluded.

      Exposure

      The exposure in this study was the development of MDD. This exposure was defined by the presence of any one of the MDD Read codes selected based on expert consensus from psychiatrists and represents a definition of a depressive disorder, not just a limited number of depression symptoms. The date of exposure was taken as the date on which one of these Read codes first appeared on the patient’s record. Patients with codes for bipolar disorder, mania, or hypomania were excluded. This case definition of MDD is likely more specific than previously used definitions in patients with psoriasis in THIN (
      • Kurd S.K.
      • Troxel A.B.
      • Crits-Christoph P.
      • Gelfand J.M.
      The risk of depression, anxiety, and suicidality in patients with psoriasis: a population-based cohort study.
      ). Because diagnosis of MDD was based on the presence or absence of a MDD Read code, there were no missing data for the exposure. Any cases of MDD preceding the psoriasis diagnosis or occurring after a diagnosis of PsA were excluded. MDD was treated as a time-dependent exposure whereby the observation time was partitioned into unexposed time (between the psoriasis diagnosis and the diagnosis of MDD) and exposed time (from the MDD diagnosis until the outcome occurred or the end of the study period). This was done to minimize the threat of survival bias that could result from patients having to survive or be free of PsA by definition until at least the MDD diagnosis in the exposed group, but not in the unexposed group who did not develop MDD (
      • Zhou Z.
      • Rahme E.
      • Abrahamowicz M.
      • Pilote L.
      Survival bias associated with time-to-treatment initiation in drug effectiveness evaluation: a comparison of methods.
      ). Once exposed to MDD, there was no mechanism to revert a patient back to the unexposed group, and thus our analysis accounts for ever having a diagnosis of MDD and not just a current episode of major depression. This is consistent with the diagnostic definition of MDD, as contained in the DSM-5 (

      American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 5th ed (DSM-5). Arlington, VA; 2013.

      ), in which the diagnosis of MDD is characterized by one or more episodes of major depression—the diagnosis of MDD remains when patients’ episodes remit. For patients who died or transferred out of practice before MDD or PsA development, this data was considered in our analysis as either exposed time (if MDD was present) or unexposed time (if no MDD).

      Variables

      Age was taken as the age at the time of psoriasis diagnosis. The sample demonstrated a bimodal age distribution at psoriasis diagnosis (Supplementary Figure S1 online). Based on this, the trough between the two modes can be seen at approximately 45 years of age. Therefore, age was stratified based on age < 45, or age ≥ 45, representing early versus late onset psoriasis.
      Sex was taken as the biological sex at the time of psoriasis diagnosis and stratified as male or female.
      Obesity status was based on body mass and height records in closest proximity to the date of psoriasis diagnosis, from which body mass index was calculated and stratified on the World Health Organization definition of whether obesity was present (body mass index ≥ 30 kg/m2), absent (body mass index < 30 kg/m2), or missing.
      Smoking status was based on records in closest proximity to psoriasis diagnosis and stratified based on current smoker, ex-smoker, never smoker, or missing.
      Alcohol use was based on records in closest proximity to psoriasis diagnosis and stratified based on user (any amount), nonuser (none), or missing.
      The Charlson comorbidity index, a validated index of a patient’s relative burden of disease based on Read codes (
      • Khan N.F.
      • Perera R.
      • Harper S.
      • Rose P.W.
      Adaptation and validation of the Charlson Index for Read/OXMIS coded databases.
      ), and used previously in psoriasis (
      • Yeung H.
      • Takeshita J.
      • Mehta N.N.
      • Kimmel S.E.
      • Ogdie A.
      • Margolis D.J.
      • et al.
      Psoriasis severity and the prevalence of major medical comorbidity: a population-based study.
      ), was determined from available Read codes within 3 years of the index date (
      • St Germaine-Smith C.
      • Liu M.
      • Quan H.
      • Wiebe S.
      • Jette N.
      Development of an epilepsy-specific risk adjustment comorbidity index.
      ).
      The Townsend deprivation index, a validated index on socioeconomic status from geographical location (
      • Townsend P.
      • Phillimore P.
      • Beattie A.
      Inequalities in health in the Northern region: an interim report.
      ), was based on the participant’s postal address closest to the index date.
      The presence of mild versus moderate-severe psoriasis was defined from medication usage—an approach that has been used in other psoriasis THIN studies (
      • Gelfand J.M.
      • Neimann A.L.
      • Shin D.B.
      • Wang X.
      • Margolis D.J.
      • Troxel A.B.
      Risk of myocardial infarction in patients with psoriasis.
      ,
      • Kurd S.K.
      • Troxel A.B.
      • Crits-Christoph P.
      • Gelfand J.M.
      The risk of depression, anxiety, and suicidality in patients with psoriasis: a population-based cohort study.
      ). Specifically, individuals were considered to have moderate-severe psoriasis if they had a drug code for any of the following therapies between the time of psoriasis diagnosis and study censoring date: acitretin, azathioprine, cyclosporine, etretinate, hydroxycarbamide, methotrexate, mycophenolic acid, and phototherapy with or without psoralen. Given that each of these medications may alter systemic inflammation, the psoriasis severity variable also accounted for medication use that could affect the development of PsA. Patients without codes for any of the above therapies were classified as mild psoriasis. Because severity was based on the presence or absence of codes, there were no missing data for this covariate.
      All covariates, except psoriasis severity, were determined in as close proximity to the index date of psoriasis diagnosis as possible. Thus, the characteristics of the study population may be taken as the presenting features of the patients in clinical practice.

      Statistical analysis

      All analyses were performed using STATA/MP v13.1 with a two-tailed significance level of 0.05. Chi-square tests were used to compare baseline data between patients with psoriasis who developed MDD and patients with psoriasis who did not develop MDD. Kaplan-Meier failure curves were generated to display the development of PsA over the full follow-up time. Cox proportional hazards models were used to evaluate the risk of developing PsA among patients with psoriasis who had MDD, expressed as a HR with corresponding 95% CIs. We included all interaction terms in the initial model and evaluated for possible effect modification by means of an omnibus likelihood ratio test that compared this model with a model without interactions. As no interactions were found to be significant, we then developed a fully adjusted model that controlled for age, sex, obesity status, smoking status, alcohol use, Charlson comorbidity index, Townsend deprivation index, and psoriasis severity as covariates and used a backward elimination procedure to assess each covariate for possible confounding effects. From this, a more parsimonious model was generated that only included covariates that induced a large (i.e., >10%) change to the resulting HR when removed from the model. Ultimately, alcohol use was the only covariate identified as inducing a change of this magnitude. An unadjusted model was also explored. The proportional hazards assumption was evaluated by testing the Schoenfeld residuals.
      Sensitivity analyses evaluated whether the 1-year washout period used to define incident cases of psoriasis differed from results that utilized a 2- or 5-year washout period for defining incident cases. In addition, the definition of moderate-severe psoriasis was altered as a sensitivity analysis to include a broader list of therapies developed based on UK NICE guidelines for psoriasis and PsA, and also based on expert consensus from a dermatologist and rheumatologist. This expanded list included adalimumab, certolizumab pegol, dapsone, dexamethasone, efalizumab, etanercept, gold, golimumab, hydrocortisone, hydroxycarbamide, infliximab, leflunomide, prednisolone, prednisone, sulfasalazine, thalidomide, ustekinumab in addition to the therapies described previously. Steroids were only considered if they were taken orally, intramuscularly, or intravenously. These therapies were used in sensitivity analyses as opposed to the primary analysis as it was anticipated that recording of these highly specialized therapies may not be as well recorded in THIN (
      • Ogdie A.
      • Alehashemi S.
      • Love T.J.
      • Jiang Y.
      • Haynes K.
      • Hennessy S.
      • et al.
      Validity of psoriatic arthritis and capture of disease modifying antirheumatic drugs in the health improvement network.
      ).
      This manuscript was prepared in accordance with the STROBE statement, which includes all relevant data reporting (
      • von Elm E.
      • Altman D.G.
      • Egger M.
      • Pocock S.J.
      • Gøtzsche P.C.
      • Vandenbroucke J.P.
      The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies.
      ).

      Conflict of Interest

      The authors state no conflict of interest.

      Acknowledgments

      RTL was supported by a CIHR-Leo Studentship in Psoriatic Disease through the Canadian Association of Psoriasis Patients. CB is a Canadian Institutes for Health Research New Investigator in Community-Based Primary Healthcare. RTL, ADF, and IAV are supported through an Alberta Innovates Health Solutions MD/PhD Studentships. Funding sources played no role in study design, analysis, or manuscript preparation. Ethics approval was obtained for this study from the Conjoint Health Research Ethics Board at the University of Calgary (ID REB15-0203), and from the United Kingdom’s IMS Health Scientific Review Committee (ID SRC 16THIN024).

      Supplementary Material

      References

      1. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 5th ed (DSM-5). Arlington, VA; 2013.

        • Barnas J.L.
        • Ritchlin C.T.
        Etiology and pathogenesis of psoriatic arthritis.
        Rheum Dis Clin North Am. 2015; 41: 643-663
        • Chiesa Fuxench Z.C.
        • Shin D.B.
        • Ogdie Beatty A.
        • Gelfand J.M.
        The Risk of cancer in patients with psoriasis: a population-based cohort study in the health improvement network.
        JAMA Dermatol. 2016; 152: 282-290
        • Dalgard F.J.
        • Gieler U.
        • Tomas-Aragones L.
        • Lien L.
        • Poot F.
        • Jemec G.B.
        • et al.
        The psychological burden of skin diseases: a cross-sectional multicenter study among dermatological out-patients in 13 European countries.
        J Invest Dermatol. 2015; 135: 984-991
        • Dommasch E.D.
        • Li T.
        • Okereke O.I.
        • Li Y.
        • Qureshi A.A.
        • Cho E.
        Risk of depression in women with psoriasis: a cohort study.
        Br J Dermatol. 2015; 173: 975-980
        • Dowlati Y.
        • Herrmann N.
        • Swardfager W.
        • Liu H.
        • Sham L.
        • Reim E.K.
        • et al.
        A meta-analysis of cytokines in major depression.
        Biol Psychiatry. 2010; 67: 446-457
        • Dowlatshahi E.A.
        • van der Voort E.A.
        • Arends L.R.
        • Nijsten T.
        Markers of systemic inflammation in psoriasis: a systematic review and meta-analysis.
        Br J Dermatol. 2013; 169: 266-282
        • Dubreuil M.
        • Rho Y.H.
        • Man A.
        • Zhu Y.
        • Zhang Y.
        • Love T.J.
        • et al.
        Diabetes incidence in psoriatic arthritis, psoriasis and rheumatoid arthritis: a UK population-based cohort study.
        Rheumatol. 2014; 53: 346-352
        • Gelfand J.M.
        • Neimann A.L.
        • Shin D.B.
        • Wang X.
        • Margolis D.J.
        • Troxel A.B.
        Risk of myocardial infarction in patients with psoriasis.
        JAMA. 2006; 296: 1735-1741
        • Gisondi P.
        • Del Giglio M.
        • Di Francesco V.
        • Zamboni M.
        • Girolomoni G.
        Weight loss improves the response of obese patients with moderate-to-severe chronic plaque psoriasis to low-dose cyclosporine therapy: a randomized, controlled, investigator-blinded clinical trial.
        Am J Clin Nutr. 2008; 88: 1242-1247
        • Huffmeier U.
        • Uebe S.
        • Ekici A.B.
        • Bowes J.
        • Giardina E.
        • Korendowych E.
        • et al.
        Common variants at TRAF3IP2 are associated with susceptibility to psoriatic arthritis and psoriasis.
        Nat Genet. 2010; 42: 996-999
        • Iwata M.
        • Ota K.T.
        • Duman R.S.
        The inflammasome: pathways linking psychological stress, depression, and systemic illnesses.
        Brain Behav Immun. 2013; 31: 105-114
        • Jensen P.
        • Zachariae C.
        • Christensen R.
        • Geiker N.R.
        • Schaadt B.K.
        • Stender S.
        Effect of weight loss on the severity of psoriasis: a randomized clinical study.
        JAMA Dermatol. 2013; 149: 795-801
        • Khan N.F.
        • Perera R.
        • Harper S.
        • Rose P.W.
        Adaptation and validation of the Charlson Index for Read/OXMIS coded databases.
        BMC Fam Pract. 2010; 11: 1
        • Krishnadas R.
        • Nicol A.
        • Sassarini J.
        • Puri N.
        • Burden A.D.
        • Leman J.
        • et al.
        Circulating tumor necrosis factor is highly correlated with brainstem serotonin transporter availability in humans.
        Brain Behav Immun. 2016; 51: 29-38
        • Kurd S.K.
        • Troxel A.B.
        • Crits-Christoph P.
        • Gelfand J.M.
        The risk of depression, anxiety, and suicidality in patients with psoriasis: a population-based cohort study.
        Arch Dermatol. 2010; 146: 891-895
        • Li W.
        • Han J.
        • Qureshi A.A.
        Smoking and risk of incident psoriatic arthritis in US women.
        Ann Rheum Dis. 2012; 71: 804-808
        • Love T.J.
        • Zhu Y.
        • Zhang Y.
        • Wall-Burns L.
        • Ogdie A.
        • Gelfand J.M.
        • et al.
        Obesity and the risk of psoriatic arthritis: a population-based study.
        Ann Rheum Dis. 2012; 71: 1273-1277
        • McDonough E.
        • Ayearst R.
        • Eder L.
        • Chandran V.
        • Rosen C.F.
        • Thavaneswaran A.
        • et al.
        Depression and anxiety in psoriatic disease: prevalence and associated factors.
        J Rheumatol. 2014; 41: 887-896
        • Meier C.R.
        • Wilcock K.
        • Jick S.S.
        The risk of severe depression, psychosis or panic attacks with prophylactic antimalarials.
        Drug Saf. 2004; 27: 203-213
        • Naldi L.
        • Conti A.
        • Cazzaniga S.
        • Patrizi A.
        • Pazzaglia M.
        • Lanzoni A.
        Diet and physical exercise in psoriasis: a randomized controlled trial.
        Br J Dermatol. 2014; 170: 634-642
      2. National Institute for Health and Care Excellence. Psoriasis: Assessment and Management, Clinical Guidelines. http://nice.org.uk/guidance/cg153 (accessed 24 October 2012).

        • Nestle F.O.
        • Kaplan D.H.
        • Barker J.
        Psoriasis.
        N Engl J Med. 2009; 361: 496-509
        • Ogdie A.
        • Alehashemi S.
        • Love T.J.
        • Jiang Y.
        • Haynes K.
        • Hennessy S.
        • et al.
        Validity of psoriatic arthritis and capture of disease modifying antirheumatic drugs in the health improvement network.
        Pharmacoepidemiol Drug Saf. 2014; 23: 918-922
        • Ogdie A.
        • Langan S.
        • Love T.
        • Haynes K.
        • Shin D.
        • Seminara N.
        • et al.
        Prevalence and treatment patterns of psoriatic arthritis in the UK.
        Rheumatology. 2013; 52: 568-575
        • Ogdie A.
        • Yu Y.
        • Haynes K.
        • Love T.J.
        • Maliha S.
        • Jiang Y.
        • et al.
        Risk of major cardiovascular events in patients with psoriatic arthritis, psoriasis and rheumatoid arthritis: a population-based cohort study.
        Ann Rheum Dis. 2015; 74: 326-332
        • Picardi A.
        • Mazzotti E.
        • Gaetano P.
        • Cattaruzza M.S.
        • Baliva G.
        • Melchi C.F.
        Stress, social support, emotional regulation, and exacerbation of diffuse plaque psoriasis.
        Psychosomatics. 2005; 46: 556-564
        • Rait G.
        • Walters K.
        • Griffin M.
        • Buszewicz M.
        • Petersen I.
        • Nazareth I.
        Recent trends in the incidence of recorded depression in primary care.
        Br J Psychiatry. 2009; 195: 520-524
        • Reich K.
        • Kruger K.
        • Mossner R.
        • Augustin M.
        Epidemiology and clinical pattern of psoriatic arthritis in Germany: a prospective interdisciplinary epidemiological study of 1511 patients with plaque-type psoriasis.
        Br J Dermatol. 2009; 160: 1040-1047
        • Schmitt J.M.
        • Ford D.E.
        Role of depression in quality of life for patients with psoriasis.
        Dermatol. 2007; 215: 17-27
        • Seminara N.M.
        • Abuabara K.
        • Shin D.B.
        • Langan S.M.
        • Kimmel S.E.
        • Margolis D.
        • et al.
        Validity of The Health Improvement Network (THIN) for the study of psoriasis.
        Br J Dermatol. 2011; 164: 602-609
        • St Germaine-Smith C.
        • Liu M.
        • Quan H.
        • Wiebe S.
        • Jette N.
        Development of an epilepsy-specific risk adjustment comorbidity index.
        Epilepsia. 2011; 52: 2161-2167
        • Townsend P.
        • Phillimore P.
        • Beattie A.
        Inequalities in health in the Northern region: an interim report.
        Northern Regional Health Authority/University of Bristol, United Kingdom1986
        • Trudel-Fitzgerald C.
        • Tworoger S.S.
        • Poole E.M.
        • Williams D.R.
        • Kubzansky L.D.
        Prospective changes in healthy lifestyle among midlife women: when psychological symptoms get in the way.
        Am J Prev Med. 2016; 51: 327-335
        • Veale D.J.
        • Ritchlin C.
        • FitzGerald O.
        Immunopathology of psoriasis and psoriatic arthritis.
        Ann Rheum Dis. 2005; 64: ii26-ii29
        • von Elm E.
        • Altman D.G.
        • Egger M.
        • Pocock S.J.
        • Gøtzsche P.C.
        • Vandenbroucke J.P.
        The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies.
        Lancet. 2007; 370: 1453-1457
        • Wallace H.
        • Shorvon S.
        • Tallis R.
        Age-specific incidence and prevalence rates of treated epilepsy in an unselected population of 2,052,922 and age-specific fertility rates of women with epilepsy.
        Lancet. 1998; 352: 1970-1973
        • Wu S.
        • Cho E.
        • Li W.Q.
        • Han J.
        • Qureshi A.A.
        Alcohol intake and risk of incident psoriatic arthritis in women.
        J Rheumatol. 2015; 42: 835-840
        • Yeung H.
        • Takeshita J.
        • Mehta N.N.
        • Kimmel S.E.
        • Ogdie A.
        • Margolis D.J.
        • et al.
        Psoriasis severity and the prevalence of major medical comorbidity: a population-based study.
        JAMA Dermatol. 2013; 149: 1173-1179
        • Zhou Z.
        • Rahme E.
        • Abrahamowicz M.
        • Pilote L.
        Survival bias associated with time-to-treatment initiation in drug effectiveness evaluation: a comparison of methods.
        Am J Epidemiol. 2005; 162: 1016-1023