Incidence rate (IR) of psoriasis overall and by age group and sex.
Huerta C, Rivero E, Rodríguez LAG. Incidence and Risk Factors for Psoriasis in the General Population. Arch Dermatol. 2007;143(12):1559–1565. doi:10.1001/archderm.143.12.1559
MICHAELBIGBYMDDAMIANOABENIMD, MPHROSAMARIACORONADSc, MDURBÀGONZÁLEZMD, PhDABRAR A.QURESHIMD, MPHHYWELWILLIAMSMSc, PhD, FRCP
To study the clinical spectrum of psoriasis and the incidence in the general population and to identify risk factors associated with the occurrence of psoriasis.
Prospective cohort study with nested case-control analysis.
The data source was the United Kingdom General Practice Research Database containing computerized clinical information entered by general practitioners (GPs).
The study population comprised patients receiving a first-ever diagnosis of psoriasis between January 1, 1996, and December 31, 1997, and free of cancer.
Diagnosis of psoriasis was validated in a random sample of 14% of all ascertained cases requesting confirmation by the GPs. Nested case-control analysis included 3994 cases of psoriasis and a random sample of 10 000 controls frequency matched to cases by age, sex, and calendar year.
Main Outcome Measures
Incidence rate of psoriasis and estimates of the odds ratio (OR) and 95% confidence interval (CI) for psoriasis as associated with selected risk factors.
The incidence rate of psoriasis was 14 per 10 000 person-years. Patients with antecedents of skin disorders and skin infection within the last year carried the highest risk of developing psoriasis (OR, 3.6 [95% CI, 3.2-4.1], and OR, 2.1 [95% CI, 1.8-2.4], respectively). Also, smoking was found to be an independent risk factors for psoriasis (OR, 1.4 [95% CI, 1.3-1.6]). We did not find an association between risk of psoriasis and antecedents of stress, diabetes, hypertension, hyperlipidemia, cardiovascular disease, or rheumatoid arthritis.
The incidence rate in our study was higher than those published in other studies, probably owing to our case definition that considered cases recorded by the GPs independently of a specialist confirmation. Our results confirm the association between psoriasis, skin disorders, and smoking.
Psoriasis is a chronic, immunologically based, inflammatory disease of the skin and joints, which has been estimated to affect 1% to 3% of the population. The estimated prevalence rates are mostly derived from general population health surveys (rather than from disease-specific studies)1- 6 and from specific registries.7,8 Studies providing incidence rates of psoriasis are scarce. A study estimated an incidence rate of 60.4 per 100 000 person-years in a US general population.9
Psoriasis is subject to flares and remissions and comes in many different variations and degrees of severity. It may be symptomatic throughout life and may be progressive with age or wax and wane in its severity. The published studies on the natural history of psoriasis are scarce, have inherent design limitations, and address few aspects important to psoriasis. Frequently, they are cross-sectional surveys usually involving a large series of patients with psoriasis gathered from hospitals and from dermatology clinics10- 16 or patients identified from large-population, non–disease-specific surveys. Nevertheless, some studies provide data on specific aspects of the course of the disease, such as precipitating factors, comorbidity, incidence of cancer, and mortality.17- 19
We performed a prospective cohort study with a nested case-control analysis using data from the General Practice Research Database (GPRD) in the United Kingdom with the aims of estimating the incidence rate of psoriasis in the general population and identifying the risk factors associated with the occurrence of psoriasis.
The GPRD contains computerized information entered by general practitioners (GPs) in the United Kingdom. Most of the UK population is registered with a GP. Currently, approximately 1500 GPs participate in the GPRD, covering a population of approximately 3 million individuals, who are broadly representative of the population of the United Kingdom as a whole.20 The information recorded in the database includes demographic data, medical diagnoses from GP visits, specialist referrals and hospitalizations, results of laboratory tests as well as a free-text section, and all written prescriptions. Prescriptions are generated directly from the computer and recorded on the patient computerized file. An additional requirement is the recording of the indication for new courses of therapy. General practitioners send all this information anonymously to the Medicines and Healthcare Products Regulatory Agency (MHRA).20 The MHRA organizes this information to be used for research projects. Data from the GPRD on medical diagnoses and drug use are of satisfactory quality for epidemiologic research.20- 23 Previous studies have also confirmed the validity of using the GPRD for epidemiological research in the field of psoriasis.24,25
The study cohort comprised all persons who had been enrolled at least 2 years with a GP and who had at least 1 year since the first-recorded prescription between January 1, 1996, and December 31, 1997. All patients with cancer or who received a diagnosis of psoriasis before the start date of the study were excluded. All study members were then followed from the start date until the date of one of the following end points: a recorded diagnosis of psoriasis, cancer, death, or December 31, 1997, whichever came first. The study protocol was approved by the GPRD Scientific Ethical and Advisory Group.
For all patients identified with a first diagnostic code of psoriasis, we manually reviewed all their electronic medical records. We used the following algorithm to consider a patient as a possible case of psoriasis: referred to the specialist or hospitalized; receiving treatment recommended for psoriasis, including topical (corticosteroids, retinoids, vitamin D analogues, and anthralin) and/or systemic (methotrexate, cyclosporine, acitretin, and phototherapy including psoralen-UV-A and UV-B) therapy; or with more than 1 entry of psoriasis diagnosis in the computerized medical history. We considered a patient as a case of psoriasis even when a specialist did not make the diagnosis because, based on the existing literature, most cases in the United Kingdom are not referred for diagnosis to a specialist.26,27 Finally, after manual revision of the computer medical profiles, 4110 patients were considered as possible cases of psoriasis. We conducted a validation study to evaluate the accuracy of our psoriasis assessment and requested the confirmation of the diagnosis by the GPs in a random sample of 564 patients (approximately 14%). In the questionnaire, we asked the GP to provide the confirmation of the diagnosis, date of initial symptoms, whether the diagnosis had been confirmed by a specialist, type of psoriasis at onset, intensity of symptoms, whether the patient's psoriasis worsened over time, whether the patient had been hospitalized, and finally if there was a family history of psoriasis. We considered a patient as a confirmed case of psoriasis if the diagnosis was confirmed by the GP in the questionnaire and no other exclusion criteria was found such as history of psoriasis according to the date of the first diagnosis provided in the questionnaire by the GP. The response rate was high (96%). Our computer-based diagnosis of psoriasis was confirmed in 82% of patients. We excluded from the final list of psoriasis cases all individuals not confirmed by the GP in the validation study, resulting in 3994 cases of psoriasis that were included in our analysis.
The incidence rate (IR) within each sex and age stratum was calculated as the ratio of the estimated number of cases over the total number of person-years of follow-up in each respective stratum, weighted by the confirmation rate obtained in the validation study for each stratum. We used all 3994 incident cases of psoriasis as cases for the nested case-control study, and the date of the diagnosis was used as the index date. To select controls, a date during the study period was generated at random for every member of the study cohort. If the random date of a study member was included in his or her eligible person-time (follow-up period), we marked that patient as an eligible control and used this random date as the index date. This incidence density sampling method28 allows that the likelihood of being selected as a control is proportional to the person-time at risk. With this design, the odds ratio (OR) is an unbiased estimator of the IR ratio. Ten thousand controls free of psoriasis and cancer were frequency matched by sex, age (interval of 1 year), and calendar year from the list of all eligible controls. For each variable, we estimated the OR of psoriasis and the corresponding 95% confidence interval (CI). To account for the effects of potentially confounding factors, we calculated the OR using unconditional multiple logistic regression. To test whether a model including an independent variable provides more information about the dependent variable than a similar model without this independent variable, we used the likelihood ratio statistic. Frequency-matching variables were included in the model as well as number of visits to the GP in the previous year, smoking, and body mass index (BMI). All statistical analyses were conducted using STATA software (version 8.2; StataCorp, College Station, Texas).
Information on comorbidities and drug use was obtained from the database. We ascertained information any time prior to the index date on skin disorders (infectious and noninfectious), cardiovascular disease (heart failure, ischemic heart disease, and hypertension), asthma, chronic obstructive pulmonary disease, gastrointestinal diseases (inflammatory bowel diseases and chronic liver disease), renal disease, diabetes, anemia, osteoarthritis, rheumatoid arthritis, systemic lupus erythematosus, and various surgical procedures. We also ascertained information, only in the last year, on pregnancy, stress, anxiety, or depression. Body mass index was calculated from recorded height and weight (weight in kilograms divided by height in meters squared). Alcohol intake and smoking status was used as directly recorded by the GP on computer files. We analyzed drug use before the index date: this included anti-infectious drugs, acetylsalicylic acid, nonsteroidal anti-inflammatory drugs, acetaminophen, antidepressants, antipsychotics, and antihypertensive agents. For each studied class of drugs, we defined 3 time windows of exposure: current use, past use, and nonuse. Current use was categorized as use that lasted until the index date or ended in the month prior to the index date based on the supply of drug therapy as prescribed by the GP. Past use was use that ended more than 1 month before the index date. Finally, the time window of nonuse was defined as nonuse of each respective drug group at any time before the index date. Duration of use was calculated as the period from the beginning of therapy with consecutive supplies (refill gap shorter than 2 months) among current users.
There were 3994 cases in 2 219 994 person-years. The overall incidence of psoriasis in our study after being weighted by the confirmation rate obtained in the validation study for each stratum was 14 per 10 000 person-years. The Figure shows the IR of psoriasis in male patients, female patients, and overall in different age groups. The IR did not increase markedly with age after the age of 20 years. The IR was slightly higher in male patients after the age of 30 years. Of the psoriasis cases, 40% were first diagnosed before reaching the age of 40 years.
Table 1 gives the estimates of developing psoriasis associated with selected risk factors. Visits to the GP in the last year were associated with an increased risk of psoriasis. Overweight individuals had only a slightly increased risk. Patients with psoriasis were more likely to be current smokers (OR, 1.5 [95% CI, 1.3-1.6]) compared with controls. History of any skin disorder diagnosis in the last year was associated with an OR of 3.6 (95% CI, 3.2-4.1). Having an episode of infectious disease in the last year increased the risk of psoriasis (OR, 1.6 [95% CI, 1.5-1.9]). Infectious skin disorders were associated with an OR of 2.1 (95% CI, 1.8-2.4). Of note, with regard to upper respiratory tract infections in the month prior to psoriasis occurrence, there was an increased risk among patients aged 21 to 40 years (OR, 2.4 [95% CI, 1.5-3.9]); in patients younger than 21 years or older than 40 years, we did not find a clear association (OR, 1.7 [95% CI, 0.7-1.9], and OR, 0.9 [95% CI, 0.6-1.4], respectively). Diabetes was not associated with an increased risk of psoriasis (OR, 0.7 [95% CI, 0.6-0.9]).
No clear association was found between risk of psoriasis and osteoarthrosis, rheumatoid arthritis, inflammatory bowel diseases (neither Crohn syndrome diseases nor ulcerative colitis), or chronic liver diseases. Pregnancy in the last year was associated with a decreased risk of psoriasis (OR, 0.7 [95% CI, 0.6-1.0]). No association was found with cardiovascular diseases (heart failure or ischemic heart disease), hypertension, hyperlipidemia, asthma, chronic obstructive airways disease, thyroid disease, or surgery in the last year (data not shown).
Table 2 presents use of drugs and its association with newly diagnosed psoriasis. Users of antibiotics had an OR of 1.6 (95% CI, 1.3-1.8) compared with nonusers. We did not find any association with use of antihypertensive agents (OR, 0.9 [95% CI, 0.8-1.0]). We also examined the effect among various therapeutic groups of antihypertensive agents (β-blockers, angiotensin-converting enzyme inhibitors, and calcium channel blockers) and found that none was an independent risk factor for psoriasis (data not shown). Nonsteroidal anti-inflammatory drugs, acetaminophen, acetylsalicylic acid, or central nervous system drugs were not associated with developing psoriasis. Respiratory drugs were not associated with the onset of psoriasis (data not shown).
Results after stratification according to history of skin disorders were similar from the overall results (data not shown). We also analyzed the effect of history of skin disorders using a lag time of 1 year before the diagnosis of psoriasis, since skin disorders diagnosed in the year prior to the diagnosis of psoriasis could be related to that episode, and found no major difference with the main analysis (data not shown).
To our knowledge, there is only 1 study reporting incidence data of psoriasis, with an age- and sex-adjusted annual incidence of psoriasis of 6 per 10 000 individuals9; our IR was 14 per 10 000 person-years. Because of the intrinsic difficulty in studying this disease and in particular assessing the clinical onset, IRs would be expected to vary between different studies owing to different operational case definitions and methods of case detection. Our case definition did not require cases to be confirmed by a specialist. If we had imposed that criterion, our estimate would have been substantially lower. In our study, we manually reviewed all recorded clinical information using an algorithm that included treatment indicated for psoriasis, recording of diagnosis in 2 or more different dates, or confirmation by a specialist. Furthermore, in the validation study requesting confirmation of the diagnosis by the corresponding GP, we obtained an 82% confirmation rate. Most cases were patients with a diagnosis of psoriasis made directly by the GPs. Though it has been suggested that the ability of a primary care physician to diagnose skin complaints could be poor compared with a dermatologist,27 psoriasis is one of the skin disorders best diagnosed by the GP according to results from a study in the United Kingdom.27
In line with the study conducted in Rochester, Minnesota,9 we found the incidence to slightly increase gradually with age up to the seventh decade of life in men and women. Our data also show, as suggested in some case series of psoriasis, a bimodal curve in the age distribution of the incidence of psoriasis.13
Most previous reports on risk factors of psoriasis come from cross-sectional studies, or case-control studies using prevalent cases of psoriasis.29- 32 There are 3 case-control studies that used newly diagnosed cases of psoriasis to study risk factors for the onset of psoriasis.33- 35
Smoking has been more consistently associated in previous studies with psoriasis compared with alcohol.30,36 Moreover, whereas increased ingestion of alcohol has been described to be a factor triggering psoriasis, smoking increased the risk for the onset of the disease.33 Our study confirms smoking as a risk factor, whereas alcohol was not associated with the onset of psoriasis. Except for previous skin disorders, infectious disorders, and obesity, we did not confirm any of the other studied risk factors as risk factors for the onset of psoriasis.
With respect to infectious disease, clinical and immunological evidence has traditionally supported a link between an episode of streptococcal infection and the subsequent development of psoriasis, mainly in young people.35,37,38 We found an increased risk of psoriasis in patients with recent infectious disease, and the risk was greater within the first month after an upper respiratory tract infection, in particular among individuals aged 21 to 40 years. We did not have valid information to study the specific underlying infectious agent. As described by other authors,35Streptococcal pyogenes has been isolated in a proportion of patients with psoriasis ranging between 20% and 97%.
Obesity has also been associated with psoriasis in different studies.36 On the other hand, in recent years diabetes has been associated with psoriasis, either studied as an independent risk factor31,39 or as part of the metabolic syndrome.32 Results of studies on the association between diabetes and psoriasis indicate that risk for diabetes as well as for metabolic syndrome increases in patients according to duration and severity of the disease.31,32 It has also been described that the risk increased when patients with severe disease were compared with those with moderate disease.32
We did not find other cardiovascular diseases such as heart failure or ischemic heart disease, hypertension, or hyperlipidemia to be risk factors for a first episode of psoriasis. As for diabetes, studies have found that an association with cardiovascular diseases could be restricted to severe forms of psoriasis.31,32,40 A possible explanation could be that cardiovascular risk factors could be more important once the disease is clearly established than at the onset of the disease. It is likely that our newly diagnosed psoriasis cases probably involved patients with a more moderate or mild form of the disease than the prevalent cases included in prior studies.
A number of drugs have been reported to precipitate or exacerbate psoriasis in case reports and case series.41 Other drugs reported as a potential risk factor for psoriasis include antibiotics, lithium, antihypertensive agents (β-blocking agents, angiotensin-converting enzyme inhibitors, and calcium channel blockers), and nonsteroidal anti-inflammatory drugs.41- 43
Of all these drugs, we only found a small association between psoriasis and antibiotic use. This adverse effect has been described previously as inducing pustular psoriasis.44 Based on our data, it is more likely that there is an association between psoriasis and infections rather than with the drugs used to treat them. With respect to the other aforementioned drugs, it is possible that the effect of these drugs is only observed on exacerbation and not initiation of psoriasis. In addition, differences in the exposure ascertainment and definition between previous studies with prevalent cases and our study may account for the different associations found.
A number of considerations need to be addressed in the interpretation of our study. A major strength is that we are studying incident cases of psoriasis, and efforts have been made to assure this. To determine the date of onset of psoriasis is not a straightforward task. Given the lack of valid diagnostic criteria developed for clinical and population-based studies on psoriasis, one could surmise that clinical psoriasis begins with a series of symptoms that eventually lead to its diagnosis, but it is difficult to correctly identify the date the symptoms start. To determine whether this could have affected our results, first we stratified according to the presence of history of any skin disorders and observed similar results with this subanalysis. Second, assuming that history of skin disorders in the year prior to the index date could be related to the diagnosis of psoriasis, we did not count any skin disorder diagnosed in the year prior and also found no substantial change for any of the studied variables when stratifying according to this new definition of history of skin disorders. Furthermore, after the case validation process, we were not able to confirm 18% of patients as cases of psoriasis. To avoid the possible misclassification bias, data on incidence were corrected for by applying weights.
We studied only environmental risk factors associated with the onset of psoriasis. However, it is difficult to model causation in a disease such as psoriasis for which one would need to study interactions between genetic predisposition and environmental factors.30 We could not study the effect of an important risk factor such as family history of psoriasis because this information was missing for most of the study patients in the database.
In summary, our data confirmed that smoking, overweight, and infectious diseases are moderate risk factors for the onset of psoriasis, while no association was found with other characteristics previously described as risk factors in studies with prevalent cases. Psoriasis is a disease that may affect quality of life to a substantial degree. Population-based studies focused on the incidence of psoriasis are deemed necessary to know and confirm factors related to the onset of the disease.
Correspondence: Consuelo Huerta MD, MPH, Spanish Centre for Pharmacoepidemiologic Research (CEIFE), Almirante 28 (2°), Madrid 28004, Spain (email@example.com).
Financial Disclosure: Dr Rivero is employed by Novartis.
Accepted for Publication: April 20, 2007.
Author Contributions: Dr Huerta had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Huerta, Rivero, and García Rodríguez. Acquisition of data: García Rodríguez. Analysis and interpretation of data: Huerta, Rivero, and García Rodríguez. Drafting of the manuscript: Huerta and García Rodríguez. Critical revision of the manuscript for important intellectual content: Huerta, Rivero, and García Rodríguez. Statistical analysis: Huerta and García Rodríguez. Obtained funding: Rivero. Administrative, technical, and material support: Huerta. Study supervision: Huerta and García Rodríguez.
Funding/Support: This study was supported in part by an unrestricted grant from Novartis.
Additional Contributions: The general practitioners provided excellent collaboration, and the Boston Collaborative Drug Surveillance Program provided access to the General Practice Research Database. Miguel Gil, MD, PhD, assisted in defining an algorithm for the case ascertainment.