Population-based studies on the natural history of psoriasis. and their role in the drug development process and in clinical practice

Pharmacoepidemiology has been defined as the study of the uses and the effects of drugs in large numbers of people and is important for the surveillance of drugs after marketing. With the recent movement from a reactive to a rather proactive pharmacovigilance, (pharmaco)epidemiological research plays an increasing role in basically all stages of the drug development process. Data on the disease planned to be treated with a new drug have to be gained which can be useful e.g. for the riskbenefit analysis of that compound (e.g. for the comparison of rates of adverse events in the treated population with the disease with rates of such events in the untreated population with the disease). Additionally, good knowledge of diseases is valuable for daily clinical practice. Hence, apart from the classical drug safety studies, pharmacoepidemiology groups conduct more and more disease epidemiology or drug utilisation studies in order to learn more about the natural history of diseases. The aim of this thesis was to increase the knowledge of psoriasis by providing new information and complementing existing data. Psoriasis is a chronic inflammatory skin disease which is common in certain parts of the world. The gain of new insights into the pathogenesis of this disease has prompted the recent development of new therapeutic drugs, primarily biologicals, and vice versa. The studies of this thesis were conducted with data from the General Practice Research Database, which contains longitudinal primary care clinical records from several million patients representative of the United Kingdom population. The general practitioners have been trained to record information on patient demographics and characteristics, lifestyle factors, symptoms, medical diagnoses, referrals to hospitals or specialists, and therapies in a standard and anonymous way. Several hundred studies have been conducted using this extensively validated database. In the first three case-control studies, the influence of beta-blockers and other antihypertensives (Study 3.1), of lithium and antipsychotics (Study 3.2), and of thiazolidinediones and other antidiabetics (Study 3.3) on the risk of developing psoriasis were investigated. The study population consisted of 36,702 patients with a first-time psoriasis diagnosis between 1994 and 2005 and the same number of patients without psoriasis, matched on age, sex, index date, general practitioner, and history in the database. Exposure to the drug classes was evaluated taking duration and timing of use and potential confounding into account. In contrast to the notion in the literature (including standard dermatology textbooks), which was mainly based on data from case reports and case series, use of beta-blockers and other antihypertensives did not materially alter the risk of incident psoriasis. On the contrary, the second study confirmed the suggestion that long-term exposure to lithium can induce psoriasis. Furthermore, for atypical antipsychotics, primarily olanzapine, a statistically significantly decreased psoriasis risk was found for current exposure of longer duration. This observation needs further confirmation. Small clinical trials had shown potential clinical benefits of thiazolidinediones on psoriasis symptoms. Study 3.3 additionally suggested that longer-term exposure to thiazolidinediones reduces the risk of developing psoriasis. The risk also tended to be decreased after use of metformin, however, this needs further investigation. Studies 3.4 to 3.6 were cohort studies with a nested case-control analysis in which the study population defined for Studies 3.1 to 3.3 (= cohort population) was followed for identification of incident diabetes mellitus (Study 3.4), myocardial infarction (MI) or stroke / transient ischaemic attack (TIA) (Study 3.5), and cancer (Study 3.6) in patients with or without psoriasis. Incidence rates (IRs) and unadjusted incidence rate ratios (IRRs) were calculated. In the nested case-control analysis, patients with the outcome of interest were matched on age, sex, and index date to four control patients from the cohort population, and the psoriasis history stratified by duration and severity (using treatment as proxy) was compared by calculating adjusted odds ratios (ORs). The overall diabetes IR in psoriatic patients was about 35% higher than in psoriasis-free patients. Psoriasis patients with intensive systemic treatment for their skin disease and a disease history of longer duration showed an about 2.5 times increased risk of diabetes compared to psoriasis-free patients. For MI and stroke / TIA the overall risk was not increased, but further analyses showed increased risks in subpopulations (e.g. severe psoriasis patients or patients <60 years of age [for MI]). The risk of lymphohaematopoietic or certain types of solid cancers was statistically significantly increased in patients with psoriasis, for solid cancers primarily in patients with a longer-term disease history. These large population-based studies further analysed existing hypotheses and raised new ones. The results may be valuable for healthcare professionals in their daily clinical practice and for pharmaceutical companies in the risk-benefit analysis of their drugs. Additionally, the example of the association between use of betablockers and psoriasis showed that there should be no place for dogmas in medicine and that conclusions can be challenged.