The best magazine
Methods
Data Sources
We conducted a population-based nationwide cohort study of Danish prospectively recorded administrative databases. All Danish citizens are allocated an individual and permanent personal civil registration number at birth, which allows linkage of data across the respective registers on an individual level.
Data on all dispensed drug prescriptions from Danish pharmacies were retrieved from National Prescription Registry, where all information is recorded according to the Anatomical Therapeutical Chemical (ATC) classification system since 1995. This register has a high accuracy due to the partial compensation of drug expenses by the government, which ensures complete registration of all dispensed prescriptions. Information on morbidity was obtained from the Danish National Patient Register, which holds data on all hospital admissions, out-patient consultations, diagnoses, and procedures, (recorded since 1978) listed according to the international classification of diseases (ICD). The Central Population Register and National Causes of Death Registry were utilized to identify all deaths and causes of deaths. An age-standardized index of socioeconomic status was defined as the individual average yearly gross income during a 5-year period before study start. The present study was conducted and reported in line with the guidelines for cohort study as defined in the Strengthening the Reporting of Observational Studies in Epidemiology recommendations.
Study Population
The entire Danish population aged ≥18 years, starting from 1 January 1997 was followed until diagnosis of AS, 31 December 2011, migration or death. Patients with psoriasis were identified by claimed prescriptions for topical vitamin D derivatives (ATC code D05AX). To ensure persistent medical treatment for psoriasis, patients were included when claiming their second prescription for these agents which are not accessible over the counter in Denmark and are used only for psoriasis as first-line treatment. Patients were classified as having severe psoriasis at the time of their third diagnosis, i.e. hospitalization or out-patient consultation for psoriasis (ICD-10 L40) or psoriatic arthritis (M070–M073). This identification and classification method for psoriasis and psoriasis severity has been used and validated in previous studies. Patients with dispensed prescriptions for vitamin D derivatives prior to study start and patients with a history of psoriasis and/or AS were excluded from the study at baseline.
Pharmacotherapy
Baseline treatment was defined by dispensed prescriptions up to 6 months prior to study inclusion date with the following medications: β-blockers (C07), loop diuretics (C03C), thiazides (C03AA), angiotensin-converting enzyme inhibitors/angiotensin 2 receptor blockers (ACEI/ARB) (C09), calcium-channel blockers (C08), vitamin K antagonists (B01AA), digoxin (C01AA), acetylsalicylic acid (B01AC06), anti-diabetic drugs (A10), platelet inhibitors (B01AC), cholesterol-lowering drugs (C10A), glucocorticoids (H02AB), and methotrexate (L01BA01).
Outcome
The main outcome for the study was new-onset AS, determined by first in- or out-patient diagnosis for non-rheumatic AS (ICD-10 codes: I350, 1352, and ICD-8 codes 42 410, 42 418, and 42 419) as recorded in the National Patient Registry.
Statistical Analysis
All statistical analyses were performed with the SAS statistical software version 9.2 (SAS Institute Inc. Cary, NC, USA) and STATA software version 11.0 (StataCorp, College Station, TX, USA).
Baseline characteristics are presented as frequencies and percentages for categorical variables and as means with standard deviations for continuous variables. Age and psoriasis were included as time-dependent variable, i.e. subjects were only considered at risk from the time they met the criteria for psoriasis. Calendar year was also included as time-dependent covariate, where bands were split in 1-year periods after 1 January 1997. Comorbidity was continuously updated throughout study follow-up and evaluated for diabetes (ICD-10 codes E10–E14 and ICD-8 codes 250), atrial fibrillation (ICD-10 code I48 and ICD-8 code 4279), hypertension (ICD-10 codes I10–I15), vascular disease (ICD-10 codes I21–I22, I70, and ICD-8 codes 410–440), thromboembolism (ICD-10 codes I26, I63, I64, I74, G458, G459, and ICD-8 codes 433–438, 444, and 450), and renal disease (N03, N04, N17–N19, I12, I13, R34 and ICD-8 codes 582–586, 588). Time-dependent multivariable Poisson regression models adjusted for confounding factors including age, gender, calendar year, comorbidity, socioeconomic status, and concomitant medications were fitted to estimate incidence rate ratios (IRRs) for AS. Incidence rates are presented as events per 10 000 person-years at risk.
Sensitivity Analyses
Patients with cardiovascular disease, e.g. atherosclerotic disease, are likely to undergo increased medical evaluation, including cardiac auscultation and medical tests such as echocardiography. Consequently, detection of AS may be favoured by more frequent in- or out-patient examination in this population group. To determine the potential detection bias, we therefore performed a sensitivity analysis where all subjects with atherosclerotic disease, i.e. atherosclerotic heart disease (ICD-10 code I251), myocardial infarction (ICD-10 codes I21, I22), stroke (ICD-10 codes I61, I63, and I64), and atherosclerotic vascular disease (ICD-10 code I70) were excluded at study baseline. We also continuously censored the subjects who developed atherosclerosis throughout the study period. Likewise, to address the impact of surveillance bias caused by increased healthcare consumption associated with our definition of psoriasis, we also performed analyses with alternative inclusion criteria for psoriasis, i.e. where patients with psoriasis were identified by the first claimed prescription for vitamin D derivatives and reclassified as having severe psoriasis at the time of their first in- or out-patient hospitalization with the psoriasis diagnosis. These alternative criteria for definition of psoriasis and psoriasis severity are associated with considerably less frequent physician and hospital visits compared with the primary psoriasis definition. Finally, we estimated age-stratified incidence rates of AS per 10 000 person-years in patients with psoriasis aged 18–40, 40–65, and 65–90 years, respectively.
For all analyses, a two-tailed P-value <0.05 was considered statistically significant and 95% confidence intervals (CIs) were also presented.
Source: ...