GP indicates general population; RCT, randomized clinical trial.
NA indicates not available; RR, relative risk. Square size indicates the weight of the study; blue diamonds, the pooled RR size.
BCC indicates basal cell carcinoma; CNS, central nervous system; FGO, female genital organs; NHL, non-Hodgkin lymphoma; RR, relative risk; and SCC, squamous cell carcinoma.
eTable 1. Database Search Strategy in MEDLINE and EMBASE
eTable 2. MOOSE Checklist for Meta-analyses of Observational Studies
eMethods. Method for Estimate Combination
eFigure. Risk of Overall Cancer Incidence and Mortality According to Severity of Psoriasis and Level of Adjustment for Confounders
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Trafford AM, Parisi R, Kontopantelis E, Griffiths CEM, Ashcroft DM. Association of Psoriasis With the Risk of Developing or Dying of Cancer: A Systematic Review and Meta-analysis. JAMA Dermatol. 2019;155(12):1390–1403. doi:10.1001/jamadermatol.2019.3056
What is the risk of cancer incidence and cancer mortality in people with psoriasis?
This systematic review and meta-analysis included 58 unique observational studies (50 reporting on cancer incidence and 15 on cancer mortality, including 7 reporting on both). The overall risk of developing cancer was significantly elevated in people with psoriasis and for a number of site-specific cancers; the risk of cancer mortality was found to be elevated in those with severe psoriasis.
These findings suggest that cancer is an important comorbidity in people with psoriasis, and dermatologists should be aware of this increased risk; further studies are needed to improve our knowledge about the underlying mechanisms of this increased risk.
The risk of cancer developing in people with psoriasis has raised some concern, with little clarity regarding differentiation in risk according to psoriasis severity.
To conduct a systematic review and meta-analysis of observational studies on the risk of cancer incidence and mortality in people with psoriasis.
Six electronic databases (MEDLINE, Embase, MEDLINE in Process, Cochrane Central Register, Web of Science, and LILACS [Literatura Latino-Americana e do Caribe em Ciências da Saúde]) were searched from inception to November 15, 2017, for eligible studies.
Cohort and case-control studies that provided estimates of the risk of cancer incidence or cancer mortality associated with psoriasis were included.
Data Extraction and Synthesis
Data were extracted relating to study design, study population, and risk estimates. Study-specific estimates of the relative risk (RR) were combined using a random-effects model. Heterogeneity was quantified using the I2 statistic. Data were analyzed from April 9, 2018, through February 22, 2019.
Main Outcomes and Measures
Pooled RR estimates for cancer incidence and cancer mortality for psoriasis cohorts compared with people without psoriasis.
A total of 58 unique studies were included, with quality varying for the incidence and the mortality studies. Severe psoriasis (RR, 1.22; 95% CI, 1.08-1.39 [9 studies]) and all severities of psoriasis (RR, 1.18; 95% CI, 1.06-1.31 [7 studies]) were associated with an increased risk of cancer (overall), and associations were found for a range of site-specific cancers, including colon (RR, 1.18 [95% CI, 1.03-1.35]), colorectal (RR, 1.34 [95% CI, 1.06-1.70]), kidney (RR, 1.58 [95% CI, 1.11-2.24]), laryngeal (RR, 1.79 [95% CI, 1.06-3.01]), liver (RR, 1.83 [95% CI, 1.28-2.61]), lymphoma (RR, 1.40 [95% CI, 1.24-1.57]), non-Hodgkin lymphoma (RR, 1.28 [95% CI, 1.15-1.43]), keratinocyte cancers (RR, 1.71 [95% CI, 1.08-2.71]), esophageal (RR, 2.05 [95% CI, 1.04-4.07]), oral cavity (RR, 2.80 [95% CI, 1.99-3.93]), and pancreatic (RR, 1.41 [95% CI, 1.16-1.73]). Overall cancer mortality risk was higher in patients with severe psoriasis (RR, 1.22; 95% CI, 1.08-1.38 [4 studies]). Specifically, liver (RR, 1.43 [95% CI, 1.09-1.88]), esophageal (RR, 2.53 [95% CI, 1.87-3.41]), and pancreatic (RR, 1.31 [95% CI, 1.02-1.69]) cancer mortality were found to be elevated in those with severe psoriasis. The heterogeneity of estimates was often very high despite stratification. Marked attenuation of risk was found in those studies that adjusted estimates for smoking, alcohol consumption, and obesity.
Conclusions and Relevance
In this study, people with psoriasis appeared to have an increased risk of cancer incidence and cancer-related mortality involving a range of site-specific cancers. Future research examining specific lifestyle factors, treatments, and the inflammatory processes that contribute to psoriasis may help provide additional information on the underlying mechanisms for the apparent increased cancer risk.
Psoriasis is a common, chronic inflammatory skin disease with prevalence estimates varying globally in adults from 0.91% in the United States to 8.50% in Norway.1 The condition also carries significant physical and psychosocial detriment.2 The extent of the burden imposed by psoriasis is further exacerbated through numerous comorbidities that include depression, cardiovascular disease, and psoriatic arthritis.3 In contrast to the relatively well-established nature of the associations between these comorbidities and psoriasis, the association between cancer and psoriasis is much less clear, particularly for site-specific cancers. Despite this lack of clarity, the plausibility of the association is supported through the role of inflammation in the pathogenesis of psoriasis, with chronic inflammation previously linked to increased cancer risk.4 Treatments, including the use of immunomodulatory therapies, may also increase the risk of developing cancer. Furthermore, the prevalence of known cancer risk factors such as smoking, excessive alcohol consumption, and obesity has been reported to be increased in people with psoriasis.3
An earlier systematic review5 reported an increased risk of developing several cancers in people with psoriasis. Despite the usefulness of this review, it included some cancer risk estimates that only considered psoriatic arthritis, thus limiting the generalizability of results to those people with psoriasis only. A further limitation of this study was a search strategy that resulted in the omission of several studies on multiple autoimmune conditions, which also reported separate risk estimates for psoriasis. Given these limitations, the main aim of our study was to conduct a comprehensive critical review and meta-analysis of observational studies to ascertain the risk of cancer incidence in people with psoriasis. In addition to providing the first meta-analysis, to our knowledge, of cancer mortality risk in people with psoriasis, we also aimed to understand how the risk of cancer varied according to psoriasis severity and the level of adjustment for lifestyle factors.
Six electronic databases (MEDLINE, Embase, MEDLINE in Process and other nonindexed citations, Cochrane Central Register, Web of Science, and LILACS [Literatura Latino-Americana e do Caribe em Ciências da Saúde]) were searched from inception to November 15, 2017, for relevant studies. The details of the search strategy for MEDLINE and Embase are presented in eTable 1 in the Supplement. Reference lists of included studies were also hand searched. All included studies had to meet the following eligibility criteria: cohort or case-control study design; at least 1 study group of patients with psoriasis; and a comparison group involving patients without psoriasis or the general population. Included studies were also required to investigate cancer occurrence or cancer mortality. Case-control studies that examined cancer risk for a range of autoimmune disorders were also eligible for inclusion if data were provided separately for those individuals with psoriasis. No restrictions were applied to language, geographic region, or study period. The University of Manchester institutional review board determined that approval was not required for this study design. This study followed the Meta-analysis of Observational Studies in Epidemiology (MOOSE) reporting guidelines (eTable 2 in the Supplement).
One author (A.M.T.) screened all titles and abstracts. Articles were sorted as definitely include, possibly include, and exclude. Articles classed as possibly include were reviewed by 4 authors (A.M.T., R.P., E.K., and D.M.A.) to determine whether the study was eligible for inclusion. Data extraction for all included studies was performed by 1 author (A.M.T.), with any issues requiring further clarification rectified through consultation with the other 3 reviewing authors. When more than 1 risk estimate was presented in a study report, the one with the greatest level of adjustment for potential confounding factors was extracted.6
To assess the risk of bias in the included cohort and case-control studies, the Newcastle-Ottawa Scale tool was used. Using the tool, each study was judged on 8 items in 3 categories, including the selection of the study groups, the comparability of the groups, and the ascertainment of the exposure of interest for case-control studies or the outcome of interest for cohort studies. Studies that received 8 or 9 of a possible 9 points were regarded as high quality, whereas studies that received 6 or 7 were regarded as fair quality, and those that received 5 or less were regarded as low quality.7
Data were analyzed from April 9, 2018, through February 22, 2019. To account for between-study heterogeneity,8 a meta-analysis was performed using a Der Simonian–Laird random-effects model, with effect estimates from different studies combined through the generic inverse variance approach. For cohort studies, risk estimates were reported through risk ratios (RR), incidence rate ratios, incidence density ratios, hazard ratios, or standardized incidence/morbidity or mortality ratios. These estimates were considered relative measures of risk and were therefore pooled. Included case-control studies reported risk estimates through odds ratios. Given the differences in study design, case-control studies were examined separately from cohort studies.
We also separately analyzed study estimates restricted to people with severe psoriasis and those including people with all severities of psoriasis. Studies were categorized as involving people with severe psoriasis if their study groups involved people who had received systemic treatment for psoriasis (including methotrexate sodium, cyclosporin, acitretin, etretinate, phototherapy, or biologicals) or had been hospitalized for psoriasis.
Studies were also grouped into 2 categories according to the level of adjustment for potential confounding. Level 1 studies adjusted for age and sex, whereas level 2 studies adjusted for age, sex, and at least 1 other confounding factor (such as smoking status, alcohol consumption, or body mass index). When more than 1 study provided an estimate using the same base population, the study reporting on the most complete data was included. Estimates calculated for separate subgroups, such as mild psoriasis and severe psoriasis, within the same study were combined before meta-analysis if the pooled estimate was more relevant to the meta-analysis (eMethods in the Supplement).
Heterogeneity in the pooled analyses was evaluated through the I2 statistic8 and corresponding 95% CIs. The calculation of CIs was not possible for pooled estimates with less than 3 studies. All analyses were performed using Stata, version 15 (StataCorp LLC). Two-sided P < .05 indicated significance.
Excluding duplicates, 2302 studies were identified through database searching. Through title and abstract screening, 2206 studies were excluded for reasons noted in Figure 1. Full-text screening excluded a further 40 studies. Hand searching identified 2 additional studies, leaving the total number of included studies at 58, with 50 of those investigating the risk of cancer9-58 and 15 investigating cancer mortality13,15,32,33,41,48,54,59-66 (7 studies13,15,32,33,41,48,54 reported on both). Of the 50 cancer incidence studies, 28 considered only site-specific malignant neoplasms, with the remaining 22 considering all and site-specific malignant neoplasms. In contrast, of the 15 cancer mortality studies, only 1 study considered all and site-specific malignant neoplasms, with 6 studies considering only all malignant neoplasms and 8 considering only site-specific malignant neoplasms. The characteristics of the cancer incidence and cancer mortality studies are summarized in Table 1 and Table 2, respectively.
For the 50 studies examining cancer incidence, quality assessment ratings were mostly fair (29 [58%]) and the remainder graded as low (8 [16%]) and high (13 [26%]). The 15 studies of cancer mortality generally had much better quality, with most of these studies being high (8 [54%]); the remaining studies were graded as fair (5 [33%]) and low (7 [13%]). For cancer incidence and cancer mortality studies, quality was mostly lowered as a result of study participant selection. The comparability of study and comparison cohorts through matching or later adjustment also presented as an issue for a considerable number of studies. Low-quality outcome ascertainment and follow-up was the least common reason for low quality (10 studies [17%]).
A total of 22 cohort studies11,13,14,17-19,24,27-31,34,36,39,40,42,44,46,47,54,55 reported estimates of risk of all cancer incidence in people with psoriasis. Nine severe psoriasis studies11,18,30,34,40,42,46,47,55 were suitable for inclusion in the meta-analysis (Figure 2A), producing a significantly elevated pooled RR of 1.22 (95% CI, 1.08-1.39) with high heterogeneity (I2 = 89.8% [95% CI, 82.8%-93.9%]). When taking into account psoriasis of all severities, 7 studies13,14,18,24,36,39,47 were included in the meta-analysis (Figure 2A), producing a significantly elevated pooled RR of 1.18 (95% CI, 1.06-1.31) with similarly high heterogeneity (I2 = 90.2% [95% CI, 82.5%-94.6%]). In contrast, 1 hospital-based case-control study57 also reported on the risk of all cancer in psoriasis of all severities, with a significantly reduced estimate; the corresponding odds ratio was 0.27 (95% CI, 0.16-0.44).
For severe psoriasis studies,10,11,15,18,21,28,30,33,34,39-43,46,47,52,54,55 meta-analysis was possible for a total of 22 site-specific cancers. Of these, incident risk was found to be significantly elevated in 8 site-specific cancers (Figure 3): squamous cell carcinoma (3 studies9,34,46; RR, 11.74 [95% CI, 1.52-90.66]; I2 = 99.7% [95% CI, 99.5%-99.8%]), lymphoma (4 studies18,42,46,52; RR, 3.39 [95% CI, 1.34-8.62]; I2 = 85.6% [95% CI, 64.6%-94.2%]), basal cell carcinoma (3 studies9,28,46; RR, 3.17 [95% CI, 1.32-7.60]; I2 = 57.3% [95% CI, 0%-87.8%]), keratinocyte (7 studies18,21,30,40,42,43,46; RR, 2.44 [95% CI, 1.68-3.56]; I2 = 90.9% [95% CI, 83.8%-94.9%]), esophagus (3 studies30,34,46; RR, 2.34 [95% CI, 1.14-4.80]; I2 = 56.4 [95% CI, 0%-87.5%]), liver (2 studies30,34; RR, 1.94 [95% CI, 1.51-2.49]; I2 = 0 [95% CI, not applicable [NA]), pancreas (3 studies18,30,34; RR, 1.45 [95% CI, 1.15-1.83]; I2 = 0 [95% CI, 0%-89.6%]), and bladder (3 studies11,28,41; RR, 1.27 [95% CI, 1.06-1.52]; I2 = 0 [95% CI, 0%-89.6%]). Meta-analysis was possible for 27 site-specific cancers from studies of all severities of psoriasis.10,13,14,16-18,23,24,26,35,38,39,44,45,47 Risk was significantly elevated in 12 site-specific cancers (Figure 3), including the oral cavity (2 studies13,24; RR, 2.80 [95% CI, 1.99-3.93]; I2 = 0 [95% CI, NA]), squamous cell carcinoma (4 studies10,13,16,24; RR, 2.15 [95% CI, 1.32-3.50]; I2 = 91.4% [95% CI, 81.2%-96.1%]), esophagus (3 studies13,14,44; RR, 2.05 [95% CI, 1.04-4.07]; I2 = 49.1% [95% CI, 0%-85.2%]), liver (2 studies13,44; RR, 1.83 [95% CI, 1.28-2.61]; I2 = 0 [95% CI, NA]), larynx (2 studies13,24; RR, 1.79 [95% CI, 1.06-3.01]; I2 = 0 [95% CI, NA]), keratinocyte (4 studies18,24,38,45; RR, 1.71 [95% CI, 1.08-2.71]; I2 = 98.3% [95% CI, 97.3%-99.0%]), kidney (3 studies13,24,39; RR, 1.58 [95% CI, 1.11-2.24]; I2 = 31.6% [95% CI, 0%-92.9%]), pancreas (4 studies13,14,18,44; RR, 1.41 [95% CI, 1.16-1.73]; I2 = 18.3% [95% CI, 0%-87.5%]), lymphoma (4 studies14,18,26,38; RR, 1.40 [95% CI, 1.24-1.57]; I2 = 23.3% [95% CI, 0%-88.3%]), colorectum (3 studies14,17,39; RR, 1.34 [95% CI, 1.06-1.70]; I2 = 0 [95% CI, 0%-89.6%]), non-Hodgkin lymphoma (5 studies13,23,24,26,39; RR, 1.28 [95% CI, 1.15-1.43]; I2 = 4.4% [95% CI, 0%-80.1%]), and colon (4 studies13,18,24,47; RR, 1.18 [95% CI, 1.03-1.35]; I2 = 34.6% [95% CI, 0%-77.1%]). Meta-analysis of a further 7 site-specific cancers was also possible for 7 case-control studies of all severities of psoriasis.12,22,37,49,50,57,58 Cancer risk was significantly elevated in T-cell lymphoma (3 studies22,50,58; RR, 2.65 [95% CI, 1.30-5.41]; I2 = 80.2% [95% CI, 37.7%-93.7%]), marginal zone lymphoma (3 studies22,50,58; RR, 1.31 [95% CI, 1.06-1.62]; I2 = 0 [95% CI, 0%-89.6%]), and B-cell lymphoma (3 studies12,50,58; RR, 1.20 [95% CI, 1.03-1.39]; I2 = 0 [95% CI, 0%-89.6%]).
A total of 7 cohort studies13,54,59,61,64-66 considered the risk of overall cancer mortality in people with psoriasis. Four studies of severe psoriasis59,64-66 produced a pooled RR of 1.22 (95% CI, 1.08-1.38) with considerable heterogeneity (I2 = 54.3% [95% CI, 0%-84.9%]) (Figure 2B). Only 2 studies61,66 reported mortality risk estimates from all cancers in people with psoriasis of all severities. Pooling of these 2 studies found no difference in risk of cancer mortality; the corresponding pooled RR was 1.05 (95% CI, 0.96-1.16) with slight heterogeneity (I2 = 30.5% [95% CI, NA]) (Figure 2B).
Meta-analysis of cancer mortality was possible for 7 site-specific cancers from studies involving people with severe psoriasis.15,32,33,60,63,64 Of these 7 individual cancers, mortality risk was found to be significantly elevated in the following 3 sites: esophagus (2 studies60,63; RR, 2.53 [95% CI, 1.87-3.41]; I2 = 0 [95% CI, NA]), liver (2 studies63,64; RR, 1.43 [95% CI, 1.09-1.88]; I2 = 0 [95% CI, NA]), and pancreas (2 studies63,64; RR, 1.31 [95% CI, 1.02-1.69]; I2 = 0.6% [95% CI, NA]).
Of the 9 studies included in the meta-analysis of all cancer risk for people with severe psoriasis, 7 studies11,30,34,40,42,46,55 conducted age and sex (level 1) adjustment, with the remaining 2 studies18,47 controlling for age, sex, and other factors (level 2) adjustment. Meta-analysis of studies with level 1 adjustment produced a pooled RR of 1.25 (95% CI, 1.08-1.45; I2 = 91.9% [95% CI, 85.9%-95.4%]) (eFigure in the Supplement). Conversely, meta-analysis of studies with level 2 adjustment produced attenuated results (RR, 1.09 [95% CI. 0.97-1.22]; I2 = 0 [95% CI, NA]). Marked attenuation of all cancer incidence risk for all severities of psoriasis was also observed when comparing level 1 adjustment studies (3 studies13,24,36; RR, 1.33 [95% CI, 1.23-1.44]; I2 = 49.9% [95% CI, 0%-85.5%]) with level 2 adjustment studies (4 studies14,18,39,47; RR, 1.07 [95% CI, 1.03-1.10]; I2 = 0 [95% CI, 0%-84.7%]). Stratification of studies by the level of adjustment for 6 site-specific cancers was also undertaken for studies involving people with all severities of psoriasis. For 5 of these cancers, the pooled estimates were lower in studies involving level 2 adjustment. The risk of all cancer mortality in severe psoriasis was higher in 4 studies with level 1 adjustment,13,54,59,66 with a pooled RR of 1.29 (95% CI, 1.17-1.42; I2 = 0 [95% CI, 0%-84.7%]) compared with RR of 1.15 (95% CI, 0.94-1.41; I2 = 81.2% [95% CI, NA]) for 2 studies with level 2 adjustment64,65 (eFigure in the Supplement). Low study numbers prevented further stratification for site-specific cancer mortality.
We found significant associations between psoriasis and the incidence of and mortality due to cancer involving a range of site-specific cancers. Cohort studies of all severities of psoriasis suggested a 1.18-fold increased risk of developing cancer compared with psoriasis-free populations. Evidence from cohort studies of severe psoriasis indicated a 1.22-fold increased risk of developing cancer when compared with populations without psoriasis, a result not significantly different from that of all psoriasis severities. With regard to cancer mortality, a 1.22-fold increased risk of dying due to cancer compared with psoriasis-free populations was observed in studies of severe psoriasis. In contrast, no significantly increased risk of cancer mortality was found in studies of all severities of psoriasis. Subgroup analysis by level of adjustment for confounders found marked attenuation of all cancer incidence and mortality risks in studies that additionally controlled for smoking, alcohol consumption, and obesity.
With regard to site-specific cancers, we observed elevated incident cancer risks for lymphoma, keratinocyte, esophageal, liver, and pancreatic cancers in studies of severe psoriasis and in those involving people with all severities of psoriasis. Similarly, esophageal, liver, and pancreatic cancers also demonstrated an increased risk when considering cancer mortality in studies of severe psoriasis.
A number of hypothesized explanations exist for the increased risk of cancer in people with psoriasis. First, psoriasis is a chronic inflammatory disease, and the link between chronic inflammation and cancer has been well reported in other conditions, such as Crohn disease and Barrett esophagus.4 This explanation may hold particular sway when considering the suggested increased risk of pancreatic cancer incidence in people with psoriasis. Indeed, previous studies have noted an association between psoriasis and pancreatitis,67 with acknowledged links between pancreatitis and pancreatic cancer. Beyond the underlying inflammatory mechanisms of psoriasis, the use of immunomodulatory agents and potentially carcinogenic therapies in psoriasis treatment has also been explored as a mechanism for increased cancer risk, particularly for cancers such as lymphoma. The use of immunomodulatory agents and increased cancer incidence has previously been suggested in other conditions, such as rheumatoid arthritis. However, a meta-analysis of the association between rheumatoid arthritis and cancer suggested a lower risk of all cancer (standardized incidence ratio, 1.05; 95% CI, 1.01-1.09) compared with that found in psoriasis.68 Of particular relevance in this regard are biological therapies, which are being increasingly used for the management of severe psoriasis. Although preliminary studies have suggested little to no increased risk of cancer incidence in patients with psoriasis receiving these therapies, further study allowing greater follow-up and increased power is required to properly examine the potential cancer risk, particularly for site-specific cancers.69 In contrast to this, the association between phototherapy in psoriasis and cancer incidence has been extensively explored, with an increased risk of squamous cell carcinoma and basal cell carcinoma being well established.9 Furthermore, the increased prevalence of known cancer risk factors in people with psoriasis, such as smoking, excessive alcohol consumption, and obesity, has also been posited as a plausible explanation for an association with cancer. The potential role of these lifestyle factors is strengthened by the attenuation of risk in those studies14,18,39,47,54,64 that adjusted estimates for lifestyle factors and the increased risk of site-specific cancers, such as esophageal cancer and liver cancer, which have been reported to be independently associated with obesity, smoking, and higher alcohol consumption.70
Despite these findings, cancer currently receives relatively little focus in guidelines for the management of psoriasis, compared with other comorbidities, such as cardiovascular disease. The National Institute for Health and Care Excellence guidelines for the assessment and management of psoriasis71 and the Canadian guidelines for psoriasis management72 do not give any consideration to cancer as a comorbidity of psoriasis. In addition, guidelines produced by the Scottish Intercollegiate Guidelines Network and the American Academy of Dermatology73 give consideration only to lymphoma. The evidence from this meta-analysis not only suggests that cancer should be given more consideration as an important comorbidity of psoriasis but also begins to present evidence that this risk could be alleviated to some extent through lifestyle behavior change. Although potential drivers of lifestyle behavior change, such as structured weight reduction and psychological counseling for alcohol use disorders, have been considered as a means to reduce the severity of psoriasis and alcohol-related mortality,74 their implementation may have the concurrent effect of attenuating some of the cancer risk. Although it has been noted that lifestyle behavior change is challenging for health care professionals to implement,75 the importance of a more holistic approach to psoriasis care involving lifestyle behavior change is reinforced through the results of this meta-analysis.
To our knowledge, this is the only meta-analysis of cancer risk that has stratified by psoriasis severity and examined the risk of cancer mortality in people with psoriasis. The stratification provided greater insight into the factors that may be implicit in the association between psoriasis and cancer, because certain treatments are more prevalent in those with severe disease. Further value is provided by this study in the consideration of site-specific cancers.
Potential limitations also need to be acknowledged in interpreting the findings, particularly the high levels of heterogeneity between some studies. Given the observed heterogeneity, we used a random-effects model to meta-analyze studies and conducted subgroup analyses to try to explain heterogeneity. Of interest, heterogeneity was largely reduced for overall cancer risk for severe and all severities of psoriasis through the grouping of studies that adjusted risk estimates for established cancer risk factors.
With regard to adjustment for potential confounders, a possible limitation to the wider consideration of the association between psoriasis and cancer is the lens through which smoking, alcohol consumption, and obesity are considered, as confounding factors or ones that lie on the causal pathway. Because it is possible that these factors, particularly obesity, lie on the causal pathway, it may not be appropriate to consider them simply by adjustment.
Two other limitations, similarly to this study and the wider examination of the association between psoriasis and cancer, are the classification of psoriasis severity and the reporting of duration of treatment exposure. Because a large number of contemporary studies, including a considerable number in this meta-analysis, use electronic health record databases, which do not include the psoriasis area and severity index (PASI), little information is available regarding the risk of cancer according to PASI-classified psoriasis severity. Indeed, only Dai et al (2016)10 presented the risk of cancer according to PASI. Due to the lack of PASI data, studies were unable to account for psoriasis severity or, as in this study, were required to use potentially suboptimal proxies to determine severe psoriasis, such as the receiving of systemic therapies or hospitalization for psoriasis. Although this study did not focus on cancer risk according to specific treatments for psoriasis, it was notable that very few included studies reported on the length of treatment exposures, and in those that did, duration of exposure was often only provided for phototherapy. Future studies providing detailed information regarding the length of exposure to treatments, using registries such as the British Association of Dermatologists Biologic and Immunomodulators Register, are essential to understand the role of psoriasis treatments in the association between the condition and cancer.
People with psoriasis have an elevated risk of developing or dying of cancer, particularly for a number of site-specific cancers. To improve the understanding of the link between psoriasis and cancer, further population-based studies are needed. Where possible, considering the association according to different severities of psoriasis would be beneficial. Understanding the role of lifestyle factors in any increased cancer risk remains challenging, but studies giving greater consideration of these factors would be of benefit.
Accepted for Publication: August 9, 2019.
Corresponding Author: Alex Trafford, MSc, Division of Pharmacy and Optometry, School of Health Sciences, Faculty of Biology, Medicine, and Health, University of Manchester, Oxford Road, Stopford Bldg, Room 1.136, Manchester M13 9PT, United Kingdom (firstname.lastname@example.org).
Published Online: October 16, 2019. doi:10.1001/jamadermatol.2019.3056
Author Contributions: Mr Trafford had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: All authors.
Acquisition, analysis, or interpretation of data: Trafford, Ashcroft, Parisi, Griffiths.
Drafting of the manuscript: Trafford, Ashcroft.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Trafford, Parisi, Kontopantelis.
Obtained funding: Ashcroft, Griffiths.
Supervision: Ashcroft, Parisi, Griffiths, Kontopantelis.
Conflict of Interest Disclosures: Dr Griffiths reported receiving honoraria and/or research grants from AbbVie, Almirall, Bristol-Myers Squibb, Celgene Corporation, Eli Lilly and Company, Galderma, Janssen Pharmaceutica, Leo Pharma, Novartis International, Sandoz, and UCB. Dr Ashcroft reported receiving research grants from AbbVie, Almirall, Celgene Corporation, Eli Lilly and Company, Novartis International, UCB, and the Leo Foundation.
Funding/Support: We acknowledge the key role played by the GPA Collaborating Organisations in the establishment and organisation of the Global Psoriasis Atlas (GPA): International Federation of Psoriasis Associations; International League of Dermatological Societies; and International Psoriasis Council. The GPA is funded through contributions from the GPA Collaborating Organisations and through supporters who are listed on the GPA website (https://globalpsoriasisatlas.com/supporters/).
Role of the Funder/Sponsor: The sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclaimer: The views and opinions expressed herein are those of the authors and do not necessarily reflect those of the GPA collaborating organizations.
Additional Contributions: The GPA Project teams at the University of Manchester and University of Hamburg made substantial contributions to the administration of the project. Christopher Griffiths, MD (GPA project director), Darren Ashcroft, PhD (workstream 1 lead), Matthias Augustin, PhD (workstream 2 lead), and Rebekah Swan, PgCert (programme manager) served as the GPA executive committee. We acknowledge the enthusiastic support of all the members of the GPA board of governors, steering committee, and regional coordinators. In addition, we recognize the support of the NIHR and the Biomedical Research Centre. There was no collaboration with anyone other than the named authors and no one else provided any of the data included in this systematic review.
Create a personal account or sign in to: