If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Correspondence to: Health Sciences Centre, GF 543- 820 Sherbrook Street, Winnipeg MB, Canada R3A 1R9.
Affiliations
Department of Internal Medicine and Community Health Sciences, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, CanadaDepartment of Community Health Sciences, Max Rady College of Medicine, Rady Faculty of Health Sciences, University of Manitoba, Winnipeg, Canada
Administrative data can accurately identify persons with psoriais.
•
Nearly 5% of persons with MS have psoriasis.
•
The incidence of psoriasis is 54% higher in the MS population than in a matched population.
Abstract
Background
Psoriasis and multiple sclerosis (MS) share some risk factors, and fumarates are effective disease-modifying therapies for both psoriasis and MS, suggesting a common pathogenesis. However, findings regarding the occurrence of psoriasis in the MS population are inconsistent.
Objectives
We aimed to estimate the incidence and prevalence of psoriasis in the MS population versus a matched cohort from the general population.
Methods
We used population-based administrative data from the Canadian province of Manitoba to identify 4911 persons with MS and 23,274 age-, sex- and geographically-matched controls aged 20 years and older. We developed case definitions for psoriasis using ICD-9/10 codes and prescription claims. These case definitions were compared to self-reported psoriasis diagnoses. The preferred definition was applied to estimate the incidence and prevalence of psoriasis over the period 1998–2008. We used multivariable Cox regression to estimate the risk of psoriasis in the MS population at the individual level, adjusting for sex, age at the index date, socioeconomic status and physician visits.
Results
In 2008, the crude incidence of psoriasis per 100,000 person-years was 466.7 (95%CI: 266.8–758.0) in the MS population, and 221.3 in the matched population (95%CI: 158.1–301.4). The crude prevalence of psoriasis per 100,000 persons was 4666.1 (95%CI: 3985.2–5429.9) in the MS population, and 3313.5 (95%CI: 3057.4–3585.3) in the matched population. The incidence and prevalence of psoriasis rose slightly over time. After adjusting for sex, age at the index date, socioeconomic status and physician visits, the risk of incident psoriasis was 54% higher in the MS population (HR 1.54; 95%CI: 1.07–2.24).
Conclusion
Psoriasis incidence and prevalence are higher in the MS population than in the matched population.
The etiology of multiple sclerosis (MS) remains unknown but insight may be gained by studying comorbidities that occur with different frequency than expected in MS as compared to the general population. In this regard, immune-mediated disorders have been of particular interest. Psoriasis is an immune-mediated skin disorder characterized by scaly erythematous plaques. It is common in the general population with prevalence estimates ranging from 0.9% to 8.5% in adults depending on the region, and like MS, the prevalence appears to increase with increasing distance from the equator.(
Findings regarding the occurrence of psoriasis in the MS population are inconsistent, with the estimated incidence of psoriasis ranging from 0.17% to 1.63% while the estimated prevalence ranges from 0.39% to 7.74%.(
) However, few prior studies were population-based, and most were conducted in Europe. Whether psoriasis occurs more often than expected in those with MS remains uncertain.(
) One English study also found the prevalence of psoriasis to be higher than expected in the MS population based on the literature for the general population, but five other studies have reported no difference in prevalence.(
Therefore, we developed and validated an administrative case definition for psoriasis and determined the incidence and prevalence of psoriasis in the MS population as compared to a matched cohort drawn from the general population.
2. Materials and methods
2.1 Setting
We conducted this population-based study in the Canadian province of Manitoba.
2.2 Administrative data
Manitoba Health delivers publicly funded health services to nearly all provincial residents, and maintains electronic records of those health services. Since 1984, each health service encounter includes a unique personal health identification number (PHIN) identifying who received the service. We linked the anonymized versions of the population registry, hospital Discharge Abstract Database (DAD), physician claims and Drug Program Information Network (DPIN) datasets via scrambled PHIN (to protect confidentiality). The population registry captures demographic information (dates of birth and death, sex, postal code) and dates of insurance coverage. Hospital abstracts include admission and discharge dates, and up to 25 discharge diagnoses recorded using the International Classification of Disease (ICD)−9-CM or ICD-10-CA codes, depending on the year. Physician claims include the date of service, and one diagnosis, recorded using a three-digit ICD-9-CM code. The DPIN has captured the date of dispensation, drug name and identification number (DIN) for all outpatient prescriptions dispensed to Manitoba residents since 1996. Except for prescription claims, administrative data were available from 1984 to 2011.
2.3 Study populations
First, we identified all persons with MS in Manitoba who met a validated administrative case definition for MS (
) between 1984 and 2011, and assigned the date of their first claim for demyelinating disease as the date of MS diagnosis (index date). Next, we identified a matched cohort from the general population by excluding individuals with any diagnostic codes for demyelinating disease and then matching up to 5 controls for each case on sex, exact year of birth and region of residence (postal code) to the MS cohort. Each control was assigned the same index date as their matched cases. MS cases and matched controls were included in the analyses from the index date until death, emigration, or end of the study, whichever came first.
2.4 Case definitions for Psoriasis
Based on previous studies that employed administrative data, or validated algorithms based on electronic medical records databases for the identification of psoriasis,(
Validity of diagnostic codes and prevalence of physician-diagnosed psoriasis and psoriatic arthritis in southern sweden – a population-based register study.
) we selected ICD-9/10 codes for psoriasis and psoriasis with psoriatic arthritis (696.0, 696.1, L40.xx, M07.0, M07.2, M07.3) for our candidate case definitions. We generated lists of prescription medications available for treatment of psoriasis in Canada using the Anatomic Therapeutic Chemical classification system. These included D05AC (antracen derivatives), D05AD (psoralens for topical use), D05AX (other antipsoriatics for topical use), D05BA (psoralens for systematic use), and D05BB (retinoids for treatment of psoriasis). Thereafter, we constructed candidate case definitions that varied the number of physician, hospital and prescription claims required and the number of years over which these were required, to classify a person as having psoriasis.
As described elsewhere, we employed a validation cohort that included 327 persons with MS who reported their MS history and comorbidities using a questionnaire and consented to linkage of their clinical information with their administrative data.(
) This cohort was constructed by asking Manitoba Health to identify all hospital, physician and prescription claims for demyelinating disease between April 1, 1984 and March 31, 2007. (
) Manitoba Health then mailed questionnaires to 2000 randomly selected individuals from this cohort who were: ≥ 18 years old as of January 1, 2007, alive and residing in Manitoba when the study started and who had ≥ 3 hospital, physician or prescription claims between 1984 and 2007; or ≥ 1 claim for persons resident in 2004 or later, where ≥ 1 claim was for MS or neuromyelitis optica. The validation cohort for the present study comprised the individuals with confirmed diagnoses of MS who responded. This questionnaire asked “Has a doctor ever told you that you have any of the following conditions?”, adding the condition of psoriasis to a validated questionnaire that listed multiple conditions.(
) The specificity of a self-reported diagnosis of psoriasis as compared to medical records was 96%. Within this validation cohort, we compared the performance of the candidate administrative case definitions for psoriasis with self-report, using sensitivity, specificity, positive predictive value and negative predictive value. Also, we compared the agreement between the two data sources using a kappa (κ) statistic where neither was considered the reference standard. We interpreted κ as: slight (0–0.20), fair (0.21–0.40), moderate (0.41–0.60), substantial (0.61–0.80), and almost perfect agreement (0.81–1.0).(
Based on findings in the validation cohort, we selected and applied a preferred case definition to both the MS and matched study populations. Because psoriasis is chronic, once an individual met the case definition, he/she was considered affected in all subsequent years while alive and resident in Manitoba. We estimated the prevalence on October 1 each year using the mid-year population figures as the denominators. To estimate incidence after MS diagnosis (index date) we used a 2-year run-in period preceding the first psoriasis claim to enhance the likelihood that cases were truly incident. However, as the preferred case definition included prescription claims which only became available in 1996, the first incident case could only be identified from 1998 onwards. Incidence may artificially drop at the end of a study period when there is insufficient time to meet the case definition. Thus, to reduce the likelihood of artefactual temporal trends we limited our reporting of incidence and prevalence to the 10-year period, 1998–2008.
Incidence and prevalence estimates were age-standardized to the 2001 Canadian population, the census population closest to the study mid-point. Age-specific average annual incidence was reported using age groups 20–44, 45–59, and ≥ 60 years, to ensure adequate cell sizes to protect participant confidentiality and for consistency with prior work regarding the burden of comorbidity in MS.(
) Sex-specific estimates were also reported. We report 95% confidence intervals (CI) for each parameter based on the binomial distribution. We compared age-standardized incidence and prevalence estimates between groups using negative binomial regression, to account for overdispersion, adjusting for year. We report incidence rate ratios (IRR), prevalence ratios (PR) and the corresponding 95%CIs. Both the IR and PR have null values of 1, such that 95%Cis excluding this value can be considered statistically significant at the p=0.05 level.
2.6 Multivariable analysis
We used a multivariable Cox proportional hazards model to compare the risk of incident psoriasis in the MS and matched populations at the individual level. Zero time was the index date. Model covariates were index year (1998–2000, 2001–2003, 2004–2008), age at the index date (continuous), sex (male as reference), socioeconomic status (SES) in quintiles (lowest as reference group), comorbidity, and number of physician visits. SES was based on average household income in the postal code of residence by linkage to census data. Comorbidities of interest were those associated with psoriasis, (
) These included chronic lung disease, hypertension, inflammatory bowel disease, diabetes, hyperlipidemia, ischemic heart disease and any mood disorder, each of which was included in the model as a time-varying covariate as of the date of the first claim for the comorbidity. We included the annual number of physician visits other than those related to psoriasis as a time-varying covariate in the model, to account for possible surveillance bias due to increased contacts with the health system in the MS population. Model assumptions were tested using standard methods.(
We obtained ethics approval from the University of Manitoba Health Research Ethics Board, and approval to access administrative data from the Manitoba Health Information Privacy Committee. Statistical analyses were conducted using SAS 9.4 (SAS Institute Inc., Cary NC).
3. Results
3.1 Participants
Our study included 4911 persons identified with MS and 23,274 matched persons from the general population; over 70% were women (Table 1).(
) Most participants (n=430) were white (91.6%) and female (77.0%), with a mean age at MS symptom onset of 33.2 (11.1) years. The self-reported prevalence of psoriasis in this validation cohort was 4.2%.(
To allow 2 year run-in period, to ensure that all cases were identified using all 3 data sources (hospital, physician and prescription claims) and to avoid artifactual drops in incidence at the end of the study period due to inadequate follow-up time for new cases to quality, actual results are presented for 1998–2008, although all stated years of data were accessed;
MS cases, n
4911
Female, n (%)
3556 (72.4)
Age at index date, mean (SD)
43.6 (13.5)
Matched controls, n
23,274
Female, n (%)
16,903 (71.4)
Age at index date, mean (SD)
43.5 (13.5)
a To allow 2 year run-in period, to ensure that all cases were identified using all 3 data sources (hospital, physician and prescription claims) and to avoid artifactual drops in incidence at the end of the study period due to inadequate follow-up time for new cases to quality, actual results are presented for 1998–2008, although all stated years of data were accessed;
We tested multiple case definitions, labelled A to AT (Supplemental table 1). For any given number of years of data, requiring more physician visits reduced the sensitivity of the definition while increasing specificity. Adding prescription claims increased sensitivity without reducing specificity. Adding more years of data increased sensitivity for some definitions, and reduced specificity, although the declines in specificity were less than the gains in sensitivity. The highest sensitivity observed was 72.2%, for definitions requiring ≥ 1 hospital or ≥ 1 physician or ≥ 1 prescription claims, regardless of the number of years of data used (1–5 years). Specificities of all case definitions were high, ranging from 89.9% to 99%. The highest specificity observed (99%) was for the definition that required ≥ 3 hospital or physician claims in any combination over 1 year; specificity was similar (98.8%) for the definition that required ≥ 3 hospital or physician or prescription claims in any combination over 1 or 2 years. Agreement between medical records and the administrative case definitions was highest (κ=0.49) for the definitions that required ≥ 1 hospital or ≥ 2 physician or ≥ 2 prescription claims over two years or that required ≥ 2 hospital, physician or prescription claims in any combination over two years. Several other definitions reached similar levels of agreement.
We applied the case definition (Supplemental table 1, definition ‘N′) that required ≥1 hospital or ≥ 2 physician or ≥ 2 prescription claims over two years to determine the incidence and prevalence of psoriasis as this definition had the highest agreement with self-report as well as high specificity (97.3) and adequate sensitivity (55.5). Of the definitions that did not use prescription claims ‘K′ had the highest agreement with medical records (κ=0.42) although sensitivity was only 44.4%.
3.3 Incidence
In 1998, the crude incidence of psoriasis in the MS population was 80.6 (95%CI: 9.8, 291.1) per 100,000 person-years, and it was 466.7 (95%CI: 266.8, 758.0) in 2008. In the matched population, the crude incidence of psoriasis per 100,000 person-years was 151.9 (95%CI: 90.0, 240.1) in 1998 and 221.3 (95%CI: 158.1, 301.4) in 2008. In the MS population crude incidence rates were more variable than in the matched population, with wider confidence intervals, reflecting the small numbers of individuals affected by psoriasis. After age-standardization, and adjustment for year the incidence of psoriasis remained higher in the MS population than in the matched population (IRR 1.48; 95%CI: 1.09, 2.00). Adjusting for MS status, the incidence of psoriasis rose slightly over time (IRR 1.05/year; 95%CI: 0.996, 1.10).
Over the 10-year study period, the average annual incidence in the MS population was 268.9 (95%CI: 216.0, 330.9) while it was 174.2 (95%CI: 154.8, 195.3) in the matched population. The peak incidence in both populations was at age ≥ 60 years (Table 2).
Table 2Age and sex-specific average annual incidence rates (95% confidence intervals) of psoriasis per 100,000 population, 1998–2008.
In 1998, the crude prevalence of psoriasis in the MS population was 2285.3 (95%CI: 1735.3, 2954.2) per 100,000 persons, and it was 4666.1 (95%CI: 3985.2, 5429.9) in 2008. The age-standardized prevalence of psoriasis in the MS population in 2008 was 3272.3 (95%CI: 2722.7, 3821.9). In the matched population, the crude prevalence of psoriasis per 100,000 persons was 1491.4 (95%CI: 1268.6, 1742.1) in 1998 and 3313.5 (95%CI: 3057.4, 3585.3) in 2008. The age-standardized prevalence of psoriasis in the matched population in 2008 was 2341.7 (95%CI: 2134.8, 2548.6). Over the study period, the age-standardized prevalence of psoriasis was higher in the MS population than the matched population (PR 1.31; 95%CI: 1.25, 1.37), and the prevalence increased over time (PR 1.06; 95%CI: 1.05–1.07; Fig. 1).
Fig. 1Crude prevalence of psoriasis in the multiple sclerosis and matched populations, 1998–2008.
The characteristics of the individuals with incident psoriasis in the two cohorts did not differ with respect to sex, SES, or region of residence, at the index date (Table 3). Although the age at diagnosis of psoriasis averaged 2.7 years older in the MS population than in the matched population, this difference was not statistically significant. The populations also did not differ with respect to the prevalence of specific individual comorbidities when psoriasis was diagnosed.
Table 3Characteristics of incident psoriasis cases in the multiple sclerosis (MS) and matched populations.
In a univariate model, the risk of incident psoriasis was higher in the MS population than the matched population (HR 1.64; 95%CI: 1.14, 2.37). In a multivariable model, adjusting for index year, sex, age at the index date, and SES, this increased risk of incident psoriasis remained unchanged (Table 4). Women and those of higher SES were at greater risk for psoriasis. After adjustment for physician visits, the risk of incident psoriasis in the MS was attenuated slightly (HR 1.54; 95%CI: 1.07, 2.24; Table 4). Additional adjustment for the presence of specific comorbidities did not affect the association between MS and psoriasis, and these comorbidities were not associated with the risk of psoriasis (data not shown). We did not identify any interactions between sex and population nor between sex and age.
Table 4Multivariable association of MS with incident psoriasis.
We evaluated the incidence and prevalence of psoriasis in the MS population and a matched cohort from the general population using population-based administrative health data over a 10-year period. In 2008, the prevalence of psoriasis in the matched population was 3.3%, similar to the prevalence reported in the United States during a similar time period (3.1%).(
) The incidence and prevalence of psoriasis increased over time in the MS and matched populations, also consistent with recent findings in the general population globally.(
All case definitions for psoriasis that we tested had high specificity, limiting the probability of false positives. Although sensitivity of the definitions with the highest specificity was low, we achieved a sensitivity of 72% while maintaining a relatively high specificity of 90%. Our literature review did not identify any previously validated administrative case definitions for psoriasis in an MS cohort and studies that have previously validated case definitions in the general population have generally used different data sources, specifically electronic medical records.(
) One study in Sweden that used information from computerized medical records and administrative data sources and identified psoriasis using ICD-10 codes L40.1, L40.2, L40.4, L40.5, L40.8, L40.9, found that the presence of one of these codes had a positive predictive value of 81%.(
Validity of diagnostic codes and prevalence of physician-diagnosed psoriasis and psoriatic arthritis in southern sweden – a population-based register study.
Two European studies estimated the incidence of psoriasis in the MS population to range from 0.17% to 1.6%, lower than our estimate. The latter study identified hospitalized persons with MS living in Northern Greece, and captured psoriasis using a questionnaire followed by medical records review.(
) They reported an incidence of 1.6% over a mean follow-up of 7.1 years though the annual incidence rate was not stated. The former study was population-based and linked the Danish MS Register to a hospital register but also did not report annual incidence rate.(
) This study required ≥ 2 hospital claims for psoriasis, and found newly diagnosed psoriasis in 0.17% of their MS cases. However, by virtue of being limited to hospital records, milder cases of psoriasis diagnosed and treated by primary care providers or by specialists practicing outside hospitals would not have been captured, potentially leading to the lower estimate of incidence obtained.
A prior systematic review identified six studies in which the prevalence of psoriasis in the MS population was reported, with estimates ranging from 0.39% to 7.74%.(
) The highest of these was from the lone population-based study, although the sample was relatively small, including only 155 definite, probable and possible MS cases.(
) At 4.7%, our estimate of the prevalence of psoriasis falls within the range of these prior estimates.
We found that the incidence and prevalence of psoriasis were higher in the MS population than in the matched population. The increased risk of incident psoriasis was nearly identical to that reported in Denmark previously (HR 1.5; 95%CI: 0.95–2.4).(
) In another Danish study, the risk of incident MS was increased among individuals with mild (IRR 1.84; 95%CI: 1.46, 2.30) and severe psoriasis (IRR 2.61; 95%CI: 1.44, 4.74).(
) After we adjusted for health care utilization the association was attenuated slightly, but remained statistically significant, suggesting that our observations do not simply reflect improved ascertainment in the MS population due to increased contacts with the health system. The later mean age at diagnosis of psoriasis in the MS cohort also suggests that our findings are not due to ascertainment biases. The increased prevalence of psoriasis was consistent with our findings of increased incidence, and with the findings of one English study which reported a two-fold increase in prevalence.(
) The differences in our findings and those earlier studies may reflect prior limitations which included small samples with few individuals affected by psoriasis in either population, the use of hospital-based controls, or the use of spousal controls which may lead to overmatching and attenuation of associations.
Female sex and higher SES were associated with increased risk of psoriasis in the MS and matched populations. A prior systematic review found the association between sex and incidence of psoriasis in the general population to be inconsistent, with some studies reporting the incidence to be higher in women and others reporting the opposite.(
) but we did not observe these associations in our study populations. This may reflect several factors. First, we had lower power to detect such associations due to the relatively small number of individuals affected by both psoriasis and these comorbidities. Second, while several of these comorbidities have been reported to be more common among individuals with a current diagnosis of psoriasis,(
) we evaluated these associations before the onset of psoriasis.
The population-based design of this study and 10-year study period are strengths. Limitations of the study should be noted as well. The validation cohort may not fully represent the MS population. As we lacked access to the complete medical record for the validation cohort, we compared the administrative case definitions to self-reported diagnoses of psoriasis. Although a recent study suggests that self-report is valid for assessing psoriasis, (
) it may not capture undiagnosed scalp psoriasis. While specificity of the definition we used for psoriasis was high, sensitivity was modest, therefore we likely underestimated the true burden of psoriasis in both populations.
5. Conclusions
The incidence and prevalence of psoriasis are higher in MS than in a matched cohort from the general population. This association may reflect shared genetic factors between psoriasis and MS and shared environmental factors such as smoking that can be targets for intervention.(
) Since fumarates are effective in both conditions, further investigation of this association may yield etiologic insights into MS. Regardless, given that psoriasis affects nearly 5% of the MS population, and is itself associated with increased comorbidity and lowered quality of life,(
) this condition warrants attention in the context of comprehensive MS care.
Acknowledgement/Funding source
This study was supported (in part) by operating and Don Paty Career Development grants from the Multiple Sclerosis Society of Canada, CIHR (CBG 101829), and the Waugh Family Chair in Multiple Sclerosis. The funding source(s) had no role in the study design, collection, analysis or interpretation of the data, nor in the decision to submit the article for publication. The results and conclusions presented are those of the authors. No official endorsement by Manitoba Health is intended or should be inferred.
Disclosures
Ruth Ann Marrie receives research funding from: Canadian Institutes of Health Research, Public Health Agency of Canada, Multiple Sclerosis Society of Canada, Multiple Sclerosis Scientific Foundation, Rx & D Health Research Foundation, National Multiple Sclerosis Society, Crohn's and Colitis Canada, the Consortium of MS Centers, and has conducted clinical trials funded by Sanofi-Aventis.
Christina Wolfson receives research funding from the Multiple Sclerosis Society of Canada, Canadian Institutes of Health Research, Canada Foundation for Innovation, the National MS Society and has received one speaking honorarium from Novartis.
Helen Tremlett is the Canada Research Chair for Neuroepidemiology and Multiple Sclerosis. She currently receives research support from the National Multiple Sclerosis Society, the Canadian Institutes of Health Research, the Multiple Sclerosis Society of Canada and the Multiple Sclerosis Scientific Research Foundation. In addition, in the last five years she has received research support from the Multiple Sclerosis Society of Canada (Don Paty Career Development Award); the Michael Smith Foundation for Health Research (Scholar Award) and the UK MS Trust; speaker honoraria and/or travel expenses to attend conferences from the Consortium of MS Centers (2013), the National MS Society (2012, 2014, 2016), Teva Pharmaceuticals (2011), ECTRIMS (2011, 2012, 2013, 2014, 2015, 2016), UK MS Trust (2011), the Chesapeake Health Education Program, US Veterans Affairs (2012), Novartis Canada (2012), Biogen Idec (2014), American Academy of Neurology (2013, 2014, 2015, 2016). All speaker honoraria are either declined or donated to an MS charity or to an unrestricted grant for use by her research group.
John Fisk receives research funding from the Canadian Institutes of Health Research, Rx & D Health Research Foundation MS Society of Canada, National Multiple Sclerosis Society, and the Dalhousie Medical Research Foundation.
Validity of diagnostic codes and prevalence of physician-diagnosed psoriasis and psoriatic arthritis in southern sweden – a population-based register study.
30038-X&id=gr1.jpg){kind=link}