BJD British Journal of Dermatology. Summary. What s already known about this topic? What does this study add? EPIDEMIOLOGY

Transcription

1 EPIDEMIOLOGY BJD British Journal of Dermatology Cohort study of malignancies and hospitalized infectious events in treated and untreated patients with psoriasis and a general population in the United States* A.B. Kimball, 1 J. Schenfeld, 2 N.A. Accortt, 3 M.S. Anthony, 3 K.J. Rothman 4 and D. Pariser 5 1 Department of Dermatology, Massachusetts General Hospital, 50 Staniford Street #240, Boston, MA 02114, U.S.A. 2 DOCS Global, Inc., North Wales, PA, U.S.A. 3 Amgen Inc., Thousand Oaks, CA, U.S.A. 4 RTI International, Research Triangle Park, NC, U.S.A. 5 Eastern Virginia Medical School and Virginia Clinical Research, Inc., Norfolk, VA, U.S.A. Summary Correspondence Alexa B. Kimball. harvardskinstudies@partners.org Accepted for publication 26 July 2015 Funding sources This study was funded by Immunex, a wholly owned subsidiary of Amgen Inc. Amgen was involved in data collection, data analysis, and medical writing assistance. Conflicts of interest See Appendix 1. *Plain language summary available online DOI /bjd Background Psoriasis is associated with risk of malignancy. Some psoriasis treatments may increase the risk of hospitalized infectious events (HIEs). Objectives To evaluate rates of malignancies and HIEs in patients with psoriasis. Methods This retrospective cohort study utilized data from MarketScan â databases. Cohorts included adult general population (GP), patients with psoriasis, and patients with psoriasis treated with nonbiologics, adalimumab, etanercept, infliximab or phototherapy. Outcomes included incidence rates (IRs) per person-years observation (PYO) for all malignancies excluding nonmelanoma skin cancer (NMSC), lymphoma, NMSC, and per person-years of exposure (PYE) for HIEs. Results Incidence rates [95% confidence interval (CI)] for all malignancies except NMSC were 129 ( ) and 142 ( ) for GP (PYO = ) and psoriasis (PYO = ) cohorts, respectively; 109 ( ) and 129 ( ) for lymphoma; and 145 ( ) and 180 ( ) for NMSC. Rates for all malignancies excluding NMSC were similar among treatments but variable for lymphoma and NMSC. IRs (95% CI) for HIEs were 332 ( ) for the nonbiologic cohort (PYE = 3528); 288 ( ) for etanercept (PYE = 6563); 325 ( ) for adalimumab (PYE = 2772); 521 ( ) for infliximab (PYE = 1058); and 334 ( ) for phototherapy (PYE = 1797). IRs for HIEs were lowest for etanercept and higher in patients on baseline systemic corticosteroids across treatment cohorts. Conclusions Malignancy rates were higher in patients with psoriasis than the GP, but these treatments did not appear to increase malignancy risk. What s already known about this topic? Patients with psoriasis appear to have a higher risk of some types of malignancies than the general population, which may be due to underlying disease pathology. Many treatments for psoriasis have immunosuppressive properties, which may contribute to higher risk of infections. What does this study add? This study provides up to 5 years of follow-up on the risk of malignancy for patients with psoriasis treated with systemic therapies. Most systemic therapies evaluated in this study did not appear to be associated with increased malignancy rates This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

2 1184 Malignancies and hospitalized infections in psoriasis, A.B. Kimball et al. This study provides rates of hospitalized infectious events across the treatment spectrum, and adds information that may help differentiate the influence of disease vs. treatments for these outcomes. Psoriasis is a systemic, inflammatory, autoimmune disease that manifests in skin. Psoriasis occurrence varies with age and geographic region. 1 In the U.S., the prevalence of psoriasis in adults has been estimated to be approximately 1 3% and the incidence has been estimated to be 79 cases per person-years. 1 Treatment for moderate to severe plaque psoriasis includes phototherapy [ultraviolet (UV) A radiation, UVB and psoralen plus UVA (PUVA)], systemic nonbiologic agents (methotrexate and ciclosporin), and biologic agents that inhibit specific mediators of psoriasis immunopathology such as tumour necrosis factor inhibitors (TNFi), including adalimumab, etanercept and infliximab. Patients with psoriasis appear to have increased risk of some cancers, 2,3 which may be linked to chronic inflammation. 4 In a systematic review and meta-analysis of epidemiological studies, patients with psoriasis were shown to have increased risk for some solid cancers (respiratory tract cancer, upper aerodigestive tract cancer, urinary tract cancer, liver cancer), haematological cancers (non-hodgkin lymphoma) and skin cancers (squamous cell carcinoma, basal cell carcinoma). 3 Some treatments for psoriasis also appear to increase the risk of malignancy, including PUVA, ciclosporin and possibly methotrexate. 5 7 Short-term use of TNFi therapies has not been shown to increase the risk of malignancy in patients with psoriasis. 8,9 Evaluation of malignancy risk in patients with psoriasis has also been confounded by a high prevalence of smoking in these patients 10 and perhaps a high prevalence of alcohol consumption. 6 Patients with psoriasis have also been shown to be at increased risk for serious infections 8 and death from infections, 11,12 possibly related to the immunosuppressive properties of many treatments for psoriasis in patients with moderate to severe disease. Patients with psoriasis treated with TNFi therapies appear to be at increased risk of all infections, including serious infections. 8,13 Among TNFi therapies, rates of serious infections were shown to be lower for patients treated with etanercept than for those treated with adalimumab or infliximab in one study. 13 The purpose of this cohort study was to describe the incidence of malignancies and hospitalized infectious events (HIEs) in a general population (GP), a psoriasis population (regardless of treatment) and psoriasis populations prescribed specific treatments in a large U.S. insured population. In the 3-year interim analysis of data from this study, patients with psoriasis appeared to have higher rates of malignancy and HIEs than the GP. 14 We now report rates of malignancy and HIEs in patients with up to 5 years of follow-up, including analyses of outcomes for TNFi therapies approved for the treatment of psoriasis in the U.S. with sufficient person-years of exposure (PYE) to provide reliable estimates. Methods Data source Claims data for this analysis were obtained from the Market- Scan â commercial and Medicare Supplemental claims databases. 15 These databases comprise commercial claims from ~74 million covered lives and Medicare data from ~53 million covered lives from 1 January 1995 through 31 December The commercial database includes data from fee-for-service and managed-care health plans, including exclusive provider organizations, health maintenance organizations, indemnity plans, preferred provider organizations, point of service plans, and consumer-driven health plans. The Medicare database includes data from both Medicare-covered and employer-covered payments. These databases include enrolment data, medical claims with diagnosis codes [International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes] and prescription data. Identification of general population cohort For this analysis, the GP cohort comprised patients 18 years of age as of 2005 or at their enrolment date (whichever came later) with 12 months of continuous enrolment (baseline period) in MarketScan between 1 January 2005 and 31 December Medical history, including diagnoses, procedures and prescriptions, was captured during the 12-month enrolment period. The index date was defined as the day after the required 12-month enrolment period. Identification of psoriasis cohort Patients in the psoriasis cohort were selected from the GP cohort. To be included, patients had to have a diagnosis of psoriasis (ICD-9-CM code 6961) on one inpatient claim or two outpatient claims that were 30 and 365 days apart, with the first outpatient qualifying code occurring on or before 31 December The psoriasis cohort index date was defined as the day after the date of the earliest inpatient claim or the second outpatient claim and after they had met the required 12-month enrolment period. Both prevalent (psoriasis diagnostic claim before the index date) and incident (first psoriasis claim at the index date) patients with psoriasis were included.

3 Malignancies and hospitalized infections in psoriasis, A.B. Kimball et al Identification of treated subcohorts Patients within the psoriasis cohort who received a qualifying treatment after their psoriasis index date were identified and included in treatment subcohorts. These treatment subcohorts separated patients during the time they were treated with nonbiologic systemic therapies (methotrexate, ciclosporin), phototherapy (PUVA, UVB), etanercept, adalimumab and/or infliximab. Exposure to each therapy was defined as having one prescription claim for that therapy on or after the psoriasis index date and before the end of the continuous enrolment cycle. The treatment index date was defined as the earliest prescription fill date or earliest psoriasis procedure date. Patients could belong to more than one treatment subcohort. Biologic and nonbiologic therapies were identified in the database using procedure codes [Healthcare Common Procedure Coding system (HCPC) codes or Current Procedural Terminology (CPT) codes] or drug names (generic or trade names) and corresponding National Drug Codes. Phototherapy was identified by CPT and HCPC codes. The first continuous exposure period for each treatment was analysed. Study outcomes Outcomes included all malignancies excluding nonmelanoma skin cancer (NMSC), lymphoma, NMSC and HIEs. Algorithms to identify cases of all malignancies excluding NMSC (including lymphoma), 19 NMSC 20 and HIEs 21 were selected based on published algorithms. Using the algorithms for malignancies, rates of specific cancer types were compared with age-standardized rates from the Surveillance, Epidemiology, and End Results database. 22 An incident case of malignancy (excluding NMSC) or lymphoma was identified based on either one inpatient discharge diagnosis or two outpatient claims occurring 30 and 365 days apart with an ICD-9-CM code for a specific cancer and no confirmed diagnosis (one inpatient or two outpatient) of the identified ICD-9-CM code in the patient s claim record in the 12-month enrolment period before the index date. The malignancy or lymphoma diagnosis date was defined as the earliest inpatient diagnosis or the second outpatient diagnosis (for patients with two outpatient diagnoses). An incident case of NMSC was defined as one inpatient discharge diagnosis or one outpatient diagnosis claim with an accompanying inpatient or outpatient ICD-9-CM procedure code occurring on the same date for lesion removal, and no presence of the ICD-9-CM diagnosis code in the previous 3 months. A new HIE was defined as one overnight hospitalization in conjunction with an ICD-9-CM infection code listed as either primary or secondary discharge diagnosis and no hospitalization with that specific code as a cause for hospitalization in the previous 3 months. Statistical considerations Incidence rates (IRs) with 95% confidence intervals (CIs) were calculated for all malignancies excluding NMSC, lymphoma, NMSC and HIEs. Malignancy analyses were based on persontime at risk of cancer. Follow-up time was stopped at the earliest of the first cancer event, disenrolment from the healthcare plan, or completion of 5 years of follow-up after the index date. HIE analyses were based on person-time experienced by patients while being treated. Therefore, HIE analyses were limited to the treatment cohorts and not conducted among the GP and psoriasis cohorts. Follow-up was stopped at the earliest of the first HIE, time of disenrolment, 5 years of follow-up, 30 days after last treatment, or 30 days after switching treatment. IRs per person-years of observation (PYO) or exposure (PYE) were standardized to the psoriasis population 18 years of age for each outcome. IRs were standardized for age (grouped into five categories: 18 39, 40 49, 50 59, and 70 years), presence of rheumatoid arthritis or psoriatic arthritis during the 12-month enrolment period, and presence of systemic psoriasis medication exposure during the 12-month enrolment period. Standardized IR ratios with 95% CIs used the nonbiologic group as a reference category for analyses of patients with unique treatment exposures. Standardized IR ratios with 95% CIs are also provided for patients with and without systemic (oral or intravenous) corticosteroid exposure at baseline. Results Patients More than 18 million patients were included in the GP cohort; of these, were eligible for inclusion in the psoriasis cohort (Table 1). Across all cohorts, the median age was similar (range years), approximately half the patients were male, and most psoriasis cases (625%) were diagnosed before the start of follow-up (i.e. they were prevalent cases at the start of follow-up). Results for the GP and psoriasis cohorts were similar to what was previously reported in the 3-year analysis. 14 Malignancies The psoriasis cohort had higher IRs than the GP cohort for all malignancies excluding NMSC, lymphoma and NMSC (Table 2). The rates for all malignancies excluding NMSC in the treatment subcohorts ranged from 78 per PYO for adalimumab to 230 per PYO for infliximab. Rates for lymphoma were lowest in the adalimumab and etanercept subcohorts and highest in the infliximab subcohort, although the overall number of cases was very low among all the treatment groups (Table 2). Incidence rates for NMSC were comparable across the treatment subcohorts and similar to the overall psoriasis population. Table 3 displays the rates for patients who were unique to their particular treatment subcohort (i.e. not treated with another qualifying psoriasis treatment during the study period); rates did not seem to vary widely when analyses were limited to those not exposed to multiple qualifying therapies.

4 1186 Malignancies and hospitalized infections in psoriasis, A.B. Kimball et al. Table 1 Demographics and clinical characteristics at baseline Characteristic GP n = PSO n = NB n = 5857 ETN n = 6856 ADA n = 3314 INF n = 1044 PT n = 5156 Age, median years (Q1, Q3) 46 (35, 57) 52 (41, 60) 53 (43, 60) 49 (39, 57) 49 (40, 56) 49 (40, 56) 51 (41, 59) Age category, % < 65 years years Sex, % male PSO diagnosis, % Incident NA Prevalent NA Comorbidities, % Diabetes Hypertension Rheumatoid arthritis Psoriatic arthritis Hyperlipidaemia/ hypercholesterolaemia Prior therapies, % ADA ETN INF Methotrexate Corticosteroids a ADA, adalimumab; ETN, etanercept; GP, general population; INF, infliximab; NB, nonbiologics; PSO, psoriasis; PT, phototherapy; Q1, first quartile; Q3, third quartile; NA, not applicable. a Includes oral or intramuscular administration. Table 2 Cases, exposure and IRs for malignancies for all cohorts and subcohorts GP PSO NB ETN ADA INF PT All malignancies except NMSC No. of cases PYO a IR b % CI Lymphoma No. of cases PYO a IR b % CI to NMSC No. of cases PYO a IR b % CI ADA, adalimumab; CI, confidence interval; ETN, etanercept; GP, general population; INF, infliximab; IR, incidence rate; NB, nonbiologics; NMSC, nonmelanoma skin cancer; PSO, psoriasis; PT, phototherapy; PYO, person-years of observation. a Patients were excluded from an analysis if they had a history of the condition. b IRs were per PYO, standardized for age categories (18 39, 40 49, 50 59, 60 69, 70 years), presence of rheumatoid arthritis or psoriatic arthritis at baseline, and systemic medication exposure at baseline. Table 4 shows the rates of all malignancies by year of follow-up for the GP. Rates were highest in the first year of follow-up and then stabilized in years 3 through 5. Hospitalized infectious events HIEs were examined in the nonbiologic, etanercept, adalimumab, infliximab and phototherapy cohorts. Among patients treated with TNFi therapies, the highest IR of HIEs was seen in the infliximab subcohort and the lowest IR in the etanercept subcohort (Table 5). Analyses stratified by baseline systemic corticosteroid use showed lower rates of HIEs for all treatment groups among patients not exposed to systemic corticosteroids at baseline (Table 5). Rates among the treatment subcohorts did not differ substantially when analyses were limited to patients who were not exposed to other qualifying

5 Malignancies and hospitalized infections in psoriasis, A.B. Kimball et al Table 3 Cases, exposure, IRs and rate ratios for malignancies for patients with unique exposures a NB ETN ADA INF PT All malignancies except NMSC No. of cases PYO IR b % CI Rate ratio Reference % CI NA Lymphoma No. of cases PYO IR b % CI to to Rate ratio Reference % CI NA NMSC No. of cases PYO IR b % CI Rate ratio Reference % CI NA ADA, adalimumab; CI, confidence interval; ETN, etanercept; INF, infliximab; IR, incidence rate; NB, nonbiologics; NMSC, nonmelanoma skin cancer; PT, phototherapy; PYO, person-years of observation; NA, not applicable. a Patients were excluded from the analysis if they had a history of the condition. b IRs were per person-years of observation, standardized for age categories (18 39, 40 49, 50 59, 60 69, 70 years), presence of rheumatoid arthritis or psoriatic arthritis at baseline, and systemic medication exposure at baseline. Table 4 Cases, exposure, and IRs for all malignancies excluding NMSC for the MarketScan general population stratified by 12-month intervals of follow-up Total PYO Cases Standardized IR a 0 12 months months months months months IR, incidence rate; NMSC, nonmelanoma skin cancer; PYO, person-years of observation. a IRs were per PYO, standardized for age categories (18 39, 40 49, 50 59, 60 69, 70 years), presence of rheumatoid arthritis or psoriatic arthritis at baseline, systemic medication exposure at baseline, prior corticosteroid exposure rank, and comorbidity rank score, and standardized to the psoriasis population. treatments, except that the rate for the infliximab subcohort was more similar to that for the other treatment subcohorts (Table 6). Discussion How long do I need to stay on my medication? is a frequent question posed by patients as they commence systemic treatment for psoriasis. The natural history of psoriasis, as currently understood, suggests that complete remission is rare and that most patients will require treatment for much of their lives. Withdrawal portions of dosing registration studies similarly suggest that while dosing changes may be appropriate for some patients, psoriasis is relentless for most and predictably returns. Given this sobering reality, long-term safety data are of paramount importance. Data from the 5-year OBSERVE-5 registry 23 showed that there are patients who remain on therapy for years, and that while there are identifiable risks with treatment, there are few surprises with long-term psoriasis treatment. Comparison of rates of malignancies and HIEs from the 3-year analysis 14 and the final 5-year analysis of the study reported here also indicate that these risks remain flat rather than increasing over extended periods. In addition to extending follow-up for an additional 2 years, some methodological refinements were made to the study since the 3-year analysis was conducted. 14 Updated versions of the MarketScan database were used in the 5-year analysis, with the latest data available for the enrolled individuals, start dates, end dates, diagnosis dates and diagnosis codes. Enrolment gaps of 30 days were allowed in the 3-year analysis, whereas continuous enrolment was required for the 5-year analysis. In the 3-year analysis, IRs were standardized to the GP for sex and age, whereas in the current analysis rates were standardized to the psoriasis cohort for age and for comorbidities and systemic medication use during enrolment. Standardization for sex was dropped because sex was not a confounding variable. In the 3-year analysis, patients were excluded from malignancy analyses if they had any single

6 1188 Malignancies and hospitalized infections in psoriasis, A.B. Kimball et al. Table 5 Cases, exposure, IRs a and rate ratios for HIEs NB ETN ADA INF PT All patients No. of cases PYE IR a % CI Patients without systemic CS exposure No. of cases PYE IR a % CI Patients with systemic CS exposure No. of cases PYE IR a % CI to RR ( CS : +CS) % CI ADA, adalimumab; CI, confidence interval; CS, corticosteroid; ETN, etanercept; HIE, hospitalized infectious events; INF, infliximab; IR, incidence rate; NB, nonbiologics; NMSC, nonmelanoma skin cancer; PT, phototherapy; PYE, person-years of exposure; RR, rate ratio. a IRs are per PYE, standardized for age categories (18 39, 40 49, 50 59, 60 69, 70 years), presence of rheumatoid arthritis or psoriatic arthritis at baseline, and systemic medication exposure at baseline. Table 6 Cases, exposure, IRs and rate ratios for HIEs in patients with unique exposures NB ETN ADA INF PT No. of cases PYE IR a % CI RR Reference % CI NA ADA, adalimumab; CI, confidence interval; ETN, etanercept, HIEs, hospitalized infectious events; INF, infliximab; IR, incidence rate; NB, nonbiologics; PT, phototherapy; PYE, personyears of exposure; RR, rate ratio; NA, not applicable. a IRs were per person-years of observation, standardized for age categories (18 39, 40 49, 50 59, 60 69, 70 years), presence of rheumatoid arthritis or psoriatic arthritis at baseline, and systemic medication exposure at baseline. cancer diagnosis (one inpatient or one outpatient) during the 12-month baseline period; in the 5-year analysis, this criterion was changed to a confirmed cancer diagnosis (one inpatient or two outpatient) so that all eligible patients (including unconfirmed cases) could be eligible to be considered an incident cancer case at the beginning of the follow-up period. This change expanded the number of patients who qualified as a malignancy case and subsequently increased the number of malignancy cases, but should provide a more complete capture and accurate assessment of malignancy cases. These changes to the statistical analyses resulted in higher observed IRs overall, primarily because we now standardized to patients with psoriasis, who were older and had more comorbidities than the GP, but this change affected all cohorts and the trends were similar in the 3-year and 5-year analyses. When overall rates were stratified by year of follow-up, we noted a higher rate of malignancy in the first 12 months, which then decreased in subsequent years. Sensitivity analyses showed the higher rate in the first 12 months was influenced by the inclusion of patients with a single outpatient diagnosis during their baseline period, which would have led to exclusion as a malignancy case from the 3-year analysis. We believe this to be an artefact of the claims database, and may partially account for differences between the 3-year and 5-year analyses. While classification issues may affect comparisons of IRs between external populations and the study cohorts, the same criteria were applied to all study cohorts and therefore would not be expected to affect internal comparisons. Some of these changes affected the results for infliximab, likely due to small sample size and fewer person-years compared with previous presentations of these results. 24,25 Of note, the malignancy rate in years 3 through 5 was very stable, suggesting little increased risk over an extended follow-up time. As shown previously, patients with psoriasis had higher rates for all of the malignancy outcomes assessed in this analysis compared with a GP A strength of this study was the inclusion of long-term safety information across the TNFi therapies approved for the treatment of psoriasis. Little difference in overall malignancy rates was noted across the treatment subcohorts. Lymphoma and NMSC rates tended to vary more widely across treatment subcohorts; these outcomes were rare and affected by small numbers of cases, particularly lymphoma. For example, the high lymphoma rate for infliximab was based on only four cases, and with standardization an individual case can be unduly influential if it falls in a

7 Malignancies and hospitalized infections in psoriasis, A.B. Kimball et al small stratum. Rates for outcomes with few events are imprecise and should be interpreted with that in mind. The current analysis showed little difference across treatment subcohorts for HIEs. The HIE rate for infliximab was higher than that for the other treatment groups but this was a relatively small group and the CI for this rate was wide. Similar results were seen when analyses were restricted to patients without corticosteroid exposure. We did find that corticosteroid exposure was associated with higher rates across the treatment groups. The effect of corticosteroid exposure did appear to differ between the treatments. It would be interesting to see whether this association was related to dose and/or duration of corticosteroid use, but we lacked the information to investigate it at that level of detail. The study design allowed us to investigate whether having psoriasis was associated with a higher risk of these outcomes, which could be independent of the therapies used to treat psoriasis. The large cohorts allowed us to examine patients who were exposed to multiple treatments, and therefore better understand the real-world implications of this practice. Moreover, we had a sufficient number of patients to examine those who were uniquely exposed to each therapy. When limited to uniquely exposed patients, current TNFi treatments appeared to confer no increased risk compared with nonbiologic treatment for any of the malignancy outcomes or for HIEs. While claims databases provide large cohorts for analysis, they come with limitations, such as lack of medical history prior to enrolment, misclassification stemming from claims coding, and channelling of patients with more severe disease into certain therapies. Because the baseline period to identify prior treatments was only 1 year, any psoriasis treatment that was stopped prior to that 1-year period would not have been identified. We were also not able to identify any treatments not reimbursed by health insurance, such as home phototherapy. Differences in dosing regimens, including single vs. multiple exposures, were not taken into consideration in analyses of treated patients. Notably, cancers that were undetected during the study period would have led to an underestimation of malignancy rates, particularly if a treatment was associated with a specific malignancy. For patients aged 65 years or older, only those with supplemental insurance are represented in the MarketScan database, and results in that age group may not be generalizable to all Medicare patients. The only biologic therapies presented here are the established TNFi therapies. As ustekinumab, an interleukin (IL)-12/IL-23 blocker, was not marketed until 2009, the number of exposed patients and duration of exposure in this data set were limited and led to calculations for malignancy and HIE rates that were highly unstable when subjected to different assumptions. Several limitations of the real-world study design should be noted. Patients with single exposures to treatments (i.e. only one prescription filled) were included, although single exposures are not likely to contribute substantially to cancer risk. However, we felt that this was the most conservative approach for safety outcomes. The psoriasis cohort included both incident and prevalent cases of psoriasis, even though treatment patterns are likely to be different between these two groups. Incident cases are likely to be receiving step-up therapy (sequential use of phototherapy, nonbiologic therapy and biologic therapy), whereas prevalent cases are more likely to be receiving biologics. In summary, the results of this analysis demonstrate that the rates of these potential safety concerns remain stable for up to 5 years of follow-up, and for some outcomes might be more influenced by the disease rather than the treatment, extending what is known about the long-term safety profile of these treatments. This information should assist clinicians as they respond to patient questions about the balance of risks and benefits of long-term psoriasis treatment. Acknowledgments Julia R. Gage PhD (on behalf of Amgen Inc.) and Dikran Toroser PhD (Amgen Inc.) assisted with writing and editing the manuscript. References 1 Parisi R, Symmons DP, Griffiths CE et al. Global epidemiology of psoriasis: a systematic review of incidence and prevalence. J Invest Dermatol 2013; 133: Naldi L, Mercuri SR. Epidemiology of comorbidities in psoriasis. Dermatol Ther 2010; 23: Pouplard C, Brenaut E, Horreau C et al. Risk of cancer in psoriasis: a systematic review and meta-analysis of epidemiological studies. J Eur Acad Dermatol Venereol 2013; 27 (Suppl. 3): Alexandrescu DT, Riordan NH, Ichim TE et al. On the missing link between inflammation and cancer. Dermatol Online J 2011; 17:10. 5 Naldi L. Malignancy concerns with psoriasis treatments using phototherapy, methotrexate, cyclosporin, and biologics: facts and controversies. Clin Dermatol 2010; 28: Richard MA, Barnetche T, Horreau C et al. Psoriasis, cardiovascular events, cancer risk and alcohol use: evidence-based recommendations based on systematic review and expert opinion. J Eur Acad Dermatol Venereol 2013; 27 (Suppl. 3): Archier E, Devaux S, Castela E et al. Carcinogenic risks of psoralen UV-A therapy and narrowband UV-B therapy in chronic plaque psoriasis: a systematic literature review. J Eur Acad Dermatol Venereol 2012; 26 (Suppl. 3): Dommasch ED, Abuabara K, Shin DB et al. The risk of infection and malignancy with tumor necrosis factor antagonists in adults with psoriatic disease: a systematic review and meta-analysis of randomized controlled trials. J Am Acad Dermatol 2011; 64: Kamangar F, Neuhaus IM, Koo JY. An evidence-based review of skin cancer rates on biologic therapies. J Dermatolog Treat 2012; 23: Armstrong AW, Harskamp CT, Dhillon JS et al. Psoriasis and smoking: a systematic review and meta-analysis. Br J Dermatol 2014; 170: Abuabara K, Azfar RS, Shin DB et al. Cause-specific mortality in patients with severe psoriasis: a population-based cohort study in the U.K. Br J Dermatol 2010; 163: Gelfand JM, Troxel AB, Lewis JD et al. The risk of mortality in patients with psoriasis: results from a population-based study. Arch Dermatol 2007; 143:

8 1190 Malignancies and hospitalized infections in psoriasis, A.B. Kimball et al. 13 Girolomoni G, Altomare G, Ayala F et al. Safety of anti-tnfalpha agents in the treatment of psoriasis and psoriatic arthritis. Immunopharmacol Immunotoxicol 2012; 34: Kimball AB, Schenfeld J, Accortt NA et al. Incidence rates of malignancies and hospitalized infectious events in patients with psoriasis with or without treatment and a general population in the U.S.A.: Br J Dermatol 2014; 170: Truven Health Analytics Inc. MarketScan Research Databases. Available at: MarketScanResearch_SS_Web.pdf (last accessed 30 September 2015). 16 Anderson LA, Gadalla S, Morton LM et al. Population-based study of autoimmune conditions and the risk of specific lymphoid malignancies. Int J Cancer 2009; 125: Javitz HS, Ward MM, Farber E et al. The direct cost of care for psoriasis and psoriatic arthritis in the United States. J Am Acad Dermatol 2002; 46: Fowler JF, Duh MS, Rovba L et al. The impact of psoriasis on health care costs and patient work loss. J Am Acad Dermatol 2008; 59: Brackley ME, Penning MJ, Lesperance ML. In the absence of cancer registry data, is it sensible to assess incidence using hospital separation records? Int J Equity Health 2006; 5: Eide MJ, Krajenta R, Johnson D et al. Identification of patients with nonmelanoma skin cancer using health maintenance organization claims data. Am J Epidemiol 2010; 171: Schneeweiss S, Robicsek A, Scranton R et al. Veteran s affairs hospital discharge databases coded serious bacterial infections accurately. J Clin Epidemiol 2007; 60: National Cancer Institute. Surveillance, Epidemiology, and End Results Program. Available at: (last accessed 30 September). 23 Kimball AB, Rothman KJ, Kricorian G et al. OBSERVE-5: observational postmarketing safety surveillance registry of etanercept for the treatment of psoriasis final 5-year results. J Am Acad Dermatol 2015; 72: Kimball AB, Schenfeld J, Accortt NA et al. Incidence rates of all malignancies excluding nonmelanoma skin cancer (NMSC), lymphoma, and NMSC in psoriasis patients with and without treatment compared with the general US population. Presented at the 23rd European Academy of Dermatology and Venereology Congress, Amsterdam, the Netherlands; 8 12 October Kimball AB, Schenfeld J, Accortt NA et al. Incidence rates of hospitalized infectious events (HIEs) in patients treated for psoriasis. Presented at the 23rd European Academy of Dermatology and Venereology Congress, Amsterdam, the Netherlands; 8 12 October Boffetta P, Gridley G, Lindelof B. Cancer risk in a population-based cohort of patients hospitalized for psoriasis in Sweden. J Invest Dermatol 2001; 117: Margolis D, Bilker W, Hennessy S et al. The risk of malignancy associated with psoriasis. Arch Dermatol 2001; 137: Gelfand JM, Berlin J, Van Voorhees A et al. Lymphoma rates are low but increased in patients with psoriasis: results from a population-based cohort study in the United Kingdom. Arch Dermatol 2003; 139: Gelfand JM, Shin DB, Neimann AL et al. The risk of lymphoma in patients with psoriasis. J Invest Dermatol 2006; 126: Appendix 1 Conflicts of interest A.B.K. has received research funding from Abbott Laboratories, Amgen Inc., Janssen Biotech Inc., Merck & Co., Inc. and Pfizer Inc.; honoraria from AbbVie, Lilly, Merck & Co., Inc., Pfizer Inc. and Vascular Biogenics; and fellowship funding from Janssen-Biotech Inc. J.S. is an employee of DOCS Global, Inc., which received funding for the study from Amgen Inc. N.A.A. is an employee and shareholder of Amgen Inc. M.S.A. was an employee and shareholder of Amgen Inc. at the time of the study and is currently an employee of RTI Health Solutions and shareholder of Amgen Inc. K.J.R. is an employee of RTI Health Solutions, an independent nonprofit research organization that does work for government agencies and pharmaceutical companies. D.P. has served as a consultant, principal investigator, or participated on advisory boards for Abbott Laboratories, Amgen Inc., Astellas Pharma US, Inc., Asubio Pharmaceuticals, Inc., Basilea, Bickel Biotechnology, Celgene Corporation, Dermira, Dow Pharmaceutical Sciences, Inc., DUSA Pharmaceuticals, Inc., Eli Lilly and Company, Galderma Laboratories, L.P., Genentech, Inc., Graceway Pharmaceuticals, LLC, Intendis, Inc., Janssen-Ortho Inc., Johnson & Johnson Consumer Products Company, LEO Pharma, US, Medicis Pharmaceutical Corporation, MELA Sciences, Novartis Pharmaceuticals Corp., Novo Nordisk A/S, Ortho Dermatologics, Peplin Inc., Pfizer Inc., Photocure ASA, Procter & Gamble Company, Stiefel, a GSK Company, and Valeant Pharmaceuticals International.