Incidence of Cardiovascular Disease in Individuals with Psoriasis: A Systematic Review and Meta-Analysis

Transcription

1 ORIGINAL ARTICLE See related commentary on pg 2308 Incidence of Cardiovascular Disease in Individuals with Psoriasis: A Systematic Review and Meta-Analysis Eleanor J. Samarasekera 1, Julie M. Neilson 1, Richard B. Warren 2,JillParnham 1 and Catherine H. Smith 3 The relationship between psoriasis and increased risk of cardiovascular disease (CVD) is controversial. We critically evaluate 14 cohorts and meta-analyze the magnitude of CVD risk for the primary outcomes of CVD mortality, stroke, and myocardial infarction (MI), and establish subgroup risk for different psoriasis severities and age groups. Increased CVD risk was identified only in individuals with severe psoriasis (defined as requiring systemic therapy or hospital admission): the risk ratio relative to the general population was 1.37 (95% confidence interval (CI) ) for CVD mortality, 3.04 (95% CI ) for MI, and 1.59 (95% CI ) for stroke. The relative risks of CVD were highest in the younger, severe psoriasis population (e.g., 3.10 (95% CI ) for MI at 30 years), and absolute risks were greatest in older individuals with severe psoriasis (e.g., 23.2 excess MIs per 10,000 person-years at 60 years). Uncertainty remains about whether CVD risk is directly attributable to psoriasis, as the majority of studies failed to adequately adjust for key traditional risk factors. Journal of Investigative Dermatology (2013) 133, ; doi: /jid ; published online 16 May 2013 INTRODUCTION Psoriasis is one of the most common immune-mediated inflammatory disorders, with prevalence estimates worldwide ranging from 0 91% (USA) to 8 5% (Norway) (Parisi et al., 2012). Extent varies from a few scaly plaques at extensor sites to whole-body involvement, and for many individuals it results in profound functional, psychological, and social morbidity (Rapp et al., 1999). Some population-based cohort studies also suggest that individuals with psoriasis (Mallbris et al., 2004; Gelfand et al., 2006; Kurd et al., 2010), in common with other inflammatory autoimmune diseases such as rheumatoid arthritis (Wolfe et al., 2003; Maradit-Kremers et al., 2005) and inflammatory bowel disease (Yarur et al., 2011), are at an increased risk of cardiovascular disease (CVD). These findings have led some investigators to hypothesize, first, that the documented association between psoriasis and CVD reflects a causal relationship (i.e., systemic inflammation is driving CVD risk), and, second, that the use of systemic treatment for psoriasis with methotrexate and biological drugs will therefore reduce CVD risk (Reich, 2012). Such hypotheses remain to be proven, but there 1 National Clinical Guideline Centre, Royal College of Physicians of London, London, UK; 2 Dermatology Centre, The University of Manchester, Manchester Academic Health Sciences Centre, Salford Royal Foundation Trust, Manchester, UK and 3 St John s Institute of Dermatology, Division of Medicine and Molecular Genetics, Guy s Hospital, London, UK Correspondence: Catherine H. Smith, St John s Institute of Dermatology, Division of Medicine and Molecular Genetics, Guy s Hospital, London SE1 9RT, UK. catherine.smith@kcl.ac.uk Abbreviations: CI, confidence interval; CVD, cardiovascular disease; MI, myocardial infarction; SMR, standardized mortality ratio Received 15 October 2012; revised 15 February 2013; accepted 8 March 2013; accepted article preview online 21 February 2013; published online 16 May 2013 are at least five trials registered to investigate the impact of biological therapies on various surrogate markers of cardiovascular health US National Institute of Health ( clinicaltrials.gov/, accessed 9 January 2013). This reported association between psoriasis and CVD risk has been rapidly incorporated into clinical practice, with psoriasis cited as an independent risk factor for CVD in recent European guidelines on CVD prevention (Fifth Joint Task Force of the European Society of Cardiology, 2012). However, the prevalence of traditional risk factors for CVD, including smoking, excess alcohol intake, hypertension, hyperlipidemia, obesity, and insulin resistance (manifest clinically as the metabolic syndrome), are also reported to be higher in individuals with psoriasis (Naldi et al., 2005; Sommer et al., 2006; Gisondi et al., 2007). These comorbidities may account for some or all of the increased CVD risk, and may be more important indicators of risk than psoriasis. Furthermore, if future research does indicate improvement in CVD risk with biological therapies, understanding the size of CVD risk directly attributable to psoriasis becomes important. We aimed to summarize the evidence on the risk of CVD associated with psoriasis in different populations and across different disease severities. We pooled summary statistics of risk estimates, and identified specific subgroups at particular risk. We also reviewed the evidence for a causal relationship between psoriasis and CVD. These data are key to informing clinical decisionmaking and healthcare planning. RESULTS Study characteristics Of the 14 included studies, 10 were population-based cohorts, and sample sizes in the psoriasis group ranged from Journal of Investigative Dermatology (2013), Volume 133 & 2013 The Society for Investigative Dermatology

2 to 130,976. All but one of the studies reported European or North American populations, with one (Lin et al., 2011) being from Taiwan (Supplementary Table S1 online). The most common limitations were retrospective analysis and inadequate adjustment for confounders, which varied between the studies (Table 1). Assessment of publication bias using formal measures of funnel plot asymmetry was not possible, and the small number of studies per subgroup meant that no conclusions could be drawn from visual inspection. Some studies stratified results for disease severity, but the definitions of mild and severe psoriasis differed between studies. Requirement for systemic treatments or hospital admission were used as surrogate measures indicating severe disease, rather than formal severity assessment scores. The diagnoses, of both psoriasis and comorbidities, were assessed according to nationally or internationally recognized classifications (e.g., ICD, Oxford Medical Information Systems, and Read codes recorded in medical records). Exceptions to this were the definition of psoriasis being based on prescriptions for vitamin D analogs (Ahlehoff, 2011), and both psoriasis and the outcomes being defined according to self-report of a physician diagnosis, although data limited to confirmed psoriasis cases according to the self-administered Psoriasis Screening Tool were also reported (Li et al., 2012). One cohort included only patients treated with psoralen plus UVA for psoriasis in a clinical trial setting, with the method of psoriasis diagnosis being unclear (Stern et al., 1984; Stern and Huibregtse, 2011). CVD mortality Five studies estimated the incidence of CVD mortality in individuals with psoriasis compared with individuals without ((Figure 1a); (Boffetta et al., 2001; Mallbris et al., 2004; Mehta et al., 2010; Ahlehoff et al., 2011a; Stern and Huibregtse, 2011). For mild psoriasis, one study suggested that psoriasis had a small protective effect (standardized mortality ratio (SMR): 0.94 (95% confidence interval (CI) )) (Mallbris et al., 2004), and the other study suggested that psoriasis was a small risk factor ((incidence rate ratio: 1.14 (95% CI )); (Ahlehoff et al., 2011a)) with significant heterogeneity present. In severe psoriasis, four studies (including one reporting a hazard ratio from multivariable analysis) demonstrated a large statistically significant relative increase in the incidence of CVD mortality (Boffetta et al., 2001; Mallbris et al., 2004; Mehta et al., 2010; Ahlehoff et al., 2011a), whereas one study suggested that there may be no difference (Stern and Huibregtse, 2011). The pooled SMR from four studies was 1.37 (95% CI ), which was significantly different from the findings in the mild subgroup (P ¼ 0 02). The absolute increase in incidence in individuals with severe psoriasis was 1 in 283 patients per year (or a 3 5% excess 10-year risk); (Mehta et al., 2010). Consistent with the increased risk of all cause CVD mortality, there was an increased incidence of death due to specific cardiovascular events in individuals with severe psoriasis. Cerebrovascular disease mortality (SMRs: 1.33 (95% CI ) and 1.63 (95% CI )) and arterial disease mortality (SMR: 1.34 (95% CI )) demonstrated a similar magnitude of risk to overall CVD mortality (Boffetta et al., 2001; Mallbris et al., 2004). A greater risk was reported for ischemic heart disease mortality (SMR: 1.86 (95% CI )) (Mallbris et al., 2004). These outcomes were not reported for mild psoriasis. Table 1. Adequacy of controlling for potential confounders that may contribute to CVD risk Study Age Sex Smoking Alcohol excess BMI or obesity Hyperlipidemia Hypertension Diabetes Calendar time (Ahlehoff et al., 2011a)* B B (Ahlehoff et al., 2011b)* B B (Ahlehoff et al., 2011c)* B B (Boffetta et al., 2001) (Brauchli et al., 2009)* (Gelfand et al., 2006) / (Gelfand et al., 2009) / (Kaye et al., 2008) (Li et al., 2012)* (Lin et al., 2011)* (Mallbris et al., 2004)* (Mehta et al., 2010) / (Stern and Huibregtse, 2011) (Wakkee et al., 2010) Abbreviations: BMI, body mass index; CVD, cardiovascular disease. Controlled for. BAdjusted for surrogate markers. /Adjusted in sensitivity analyses. *Excluded those with outcome of interest at study entry

3 Myocardial infarction Seven studies reported the incidence of myocardial infarction (MI) (Figure 1b), and no other coronary heart disease outcomes were reported. Five studies included all psoriasis severities (Kaye et al., 2008; Brauchli et al., 2009; Wakkee et al., 2010; Lin et al., 2011; Li et al., 2012). Four of these CVD mortality Study or subgroup Weight (%) SMR (or HR) IV, random, 95% CI Mild psoriasis* - SMR (1.06, 1.22) Mallbris et al (0.89, 0.99) (0.86, 1.25) Heterogeneity: τ 2 = 0.02; χ 2 = 18.39, d.f. = 1 (P < ); I 2 = 95% Test for overall effect: Z = 0.34 (P = 0.73) Severe psoriasis - SMR (1.27, 1.94) Boffetta et al (1.35, 1.56) Mallbris et al (1.44, 1.60) Stern and Huibregtse (0.90, 1.16) (1.17, 1.60) Heterogeneity: τ 2 = 0.02; χ 2 = 33.08, d.f. = 3 (P < ); I 2 = 91% Test for overall effect: Z = 3.95 (P < ) Severe psoriasis - HR Mehta et al Heterogeneity: not applicable Test for overall effect: Z = 4.00 (P < ) 1.57 (1.26, 1.96) 1.57 (1.26, 1.96) Test for subgroup differences: χ 2 = 8.86, d.f. = 2 (P = 0.01), I 2 = 77.4% MI Study or subgroup All psoriasis Weight (%) IV. random, 95% CI Kaye et al (1.10, 1.33) Li et al (1.51, 4.59) Lin et al (1.28, 3.45) Wakkee et al (0.80, 1.11) (1.03, 1.89) Heterogeneity: τ 2 = 0.07; χ 2 = 20.85, d.f. = 3 (P = ); I 2 =86% Test for overall effect: Z = 2.17 (P = 0.03) Mild psoriasis* (1.12, 1.33) Gelfand et al. 2006A (1.24, 1.91) (1.07, 1.68) Heterogeneity: τ 2 = 0.02; χ 2 = 3.86, d.f. = 1 (P = 0.05); I 2 = 74% Test for overall effect: Z = 2.56 (P = 0.01) Severe psoriasis (1.10, 1.91) Gelfand et al. 2006A (3.06, 16.37) (0.65, 14.35) Heterogeneity: τ 2 = 1.16; χ 2 = 12.43, d.f. = 1 (P = ); I 2 = 92% Test for overall effect: Z = 1.41 (P = 0.16) SMR (or HR) IV, random, 95% CI Protective factor Risk factor IV. random, 95% CI Protective factor Risk factor Test for subgroup differences: χ 2 = 1.07, d.f. = 2 (P = 0.59), I 2 =0% Stroke Study or subgroup Weight (%) IV, random, 95% CI All psoriasis Kaye et al (1.00, 1.25) Li et al (0.76, 3.42) (1.01, 1.26) Heterogeneity: τ 2 = 0.00; χ 2 = 0.88, d.f. = 1 (P = 0.35); I 2 = 0% Test for overall effect: Z = 2.15 (P = 0.03) Mild* Ahlehoff et al. 2011b (1.17, 1.34) Gelfand et al (1.01, 1.11) (0.98, 1.35) Heterogeneity: τ 2 = 0.01; χ 2 = 15.13, d.f. = 1 (P = ); I 2 = 93% Test for overall effect: Z = 1.68 (P = 0.09) Severe Ahlehoff et al. 2011b (1.39, 2.11) Gelfand et al (1.10, 1.86) (1.34, 1.89) Heterogeneity: τ 2 = 0.00; χ 2 = 1.08, d.f. = 1 (P = 0.30); I 2 = 7% Test for overall effect: Z = 5.35 (P < ) IV, random, 95% CI Protective factor Risk factor Test for subgroup differences: χ 2 = 11.85, d.f. = 2 (P = 0.003), I 2 = 83.1% studies reported an increased incidence of MI, whereas one study suggested no effect of psoriasis (hazard ratio: 0.94 (95% CI )) (Wakkee et al., 2010). The pooled estimate from four independent cohorts (excluding the Brauchli study, which reported an overlapping cohort with Kaye et al.) was 1.40 (95% CI ); this should be interpreted with caution, owing to the high heterogeneity between studies (I 2 ¼ 86%). Two other studies (Gelfand et al., 2006; Ahlehoff et al., 2011a) reported relative incidence rates for mild and severe psoriasis separately. Both studies reported a statistically significant increased incidence of MI for mild and severe psoriasis, but the magnitude of the increase was greater in severe psoriasis (pooled estimate 3.40 (95% CI ) based on individual study estimates of 1.45 ( ) and 7.08 ( )) than in mild psoriasis (pooled estimate 1.34 (95% CI ) based on individual study estimates of 1.22 ( ) and 1.54 ( )). However, the difference between severity subgroups based on the pooled data was not statistically significant (P ¼ 0.31). The absolute increased MI incidence in individuals with psoriasis was reported as 1 in 2,174 individuals per year (0 46% increase in 10-year risk) and 1 in 645 individuals per year (1 55% increase in 10-year risk) for mild and severe psoriasis, respectively (Gelfand et al., 2006). The risk of all-cause mortality was not significantly different in individuals with psoriasis compared with the general population following first-time MI (hazard ratio: 1.18 (95% CI )); (Ahlehoff et al., 2011c). Stroke Five studies reported the incidence of stroke (Figure 1c); (Kaye et al., 2008; Brauchli et al., 2009; Gelfand et al., 2009; Ahlehoff et al., 2011b; Li et al., 2012). Three studies included all psoriasis severities (Kaye et al., 2008; Brauchli et al., 2009; Li et al., 2012). Each study individually showed no statistically significant difference between those with and without psoriasis; however, the pooled estimate from two independent cohorts (Brauchli study excluded as above) showed psoriasis to be a weak risk factor ((1.13 (95% CI )); (Kaye et al., 2008;Liet al., 2012)). Figure 1. Incidence estimates for cardiovascular disease outcomes in individuals with psoriasis compared with those without. (a) Cardiovascular disease (CVD) mortality; (b) myocardial infarction; (c) stroke. *Defined as vitamin D analog prescription claims without hospitalizations (Ahlehoff et al., 2011a), outpatients treated at dermatological wards with psoriasis as the main diagnosis (Mallbris et al., 2004), and psoriasis diagnosis but no history of systemic therapy (Gelfand et al., 2006; Gelfand et al., 2009). w Defined as hospitalizations (including outpatient visits) for psoriasis or psoriatic arthritis (Ahlehoff et al., 2011a), inpatients treated at dermatological wards with psoriasis as the main diagnosis (Mallbris et al., 2004; Boffetta et al., 2001), systemic therapy consistent with severe disease (Mehta et al., 2010), and psoriasis diagnosis plus a history of systemic therapy (Gelfand et al., 2006; Gelfand et al., 2009); the definition was unclear in one study, although most had received previous systemic therapy and the mean BSA coverage was 33% (Stern and Huibregtse, 2011). CI, confidence interval; HR, hazard ratio; IRR, incidence rate ratio; IV, inverse variance; SMR, standardized mortality ratio Journal of Investigative Dermatology (2013), Volume 133

4 As with MI, the increase was greater for the severe group: 1.59 (95% CI ) compared with 1.15 (95% CI ) in the mild group based on a pooled estimate from two studies (Gelfand et al., 2009; Ahlehoff et al., 2011b). The difference between severity subgroups was statistically significant (P ¼ 0.006). The absolute increased incidence of stroke in individuals with psoriasis was estimated as 1 in 4,115 individuals per year (0.24% increase in 10-year risk) and 1 in 530 individuals per year (1.89% increase in 10-year risk) for mild and severe psoriasis, respectively (Gelfand et al., 2009). Increased risk was also reported for atherosclerosis, angina, peripheral vascular disease, and atrial fibrillation across all severities but not for ischemic heart disease (Kaye et al., 2008; Wakkee et al., 2010; Ahlehoff et al., 2011b); (Supplementary Table S2 online). CVD risk modification by age Four studies reported the relative risk of incident CVD at different ages for individuals with psoriasis compared with those without (Mallbris et al., 2004; Gelfand et al., 2006; Ahlehoff et al., 2011a, 2011b). The studies reported different outcomes: atrial fibrillation, stroke, MI, ischemic heart disease mortality, cerebrovascular disease mortality, and overall CVD mortality. For severe psoriasis, the relative increase in CVD incidence was generally greater in younger groups for CVD mortality, MI, stroke, and atrial fibrillation outcomes (Supplementary Table S3 online). However, the large relative increase in incidence among younger individuals equates to a low absolute incidence rate; for MI, in the youngest age group, a relative increase of 3.1 refers to an increased risk of 1 in 1,389 individuals per year in the severe psoriasis group. This is a lower absolute incidence of MI than that among the older age group (1 in 429 individuals per year), although the relative incidence was only 1 36 times higher (Gelfand et al., 2006). This study also reported that there was a significant interaction between MI incidence and age variables in the analysis (Gelfand et al., 2006). No consistent pattern of age-related incidence was seen for the mild and mixed psoriasis severity subgroups (Supplementary Table S4 online). DISCUSSION Main findings Our study shows that individuals with psoriasis requiring systemic therapy or hospital admission (severe disease) have a marked increase in the incidence of MI, stroke, and all-cause CVD mortality, including cerebrovascular, ischemic, and arterial disease mortality. However, the risk of CVD in cohorts not stratified for disease severity is modest, and data in cohorts with mild disease are inconsistent, with no effect or very small relative risks across all outcomes. Taken together, these data indicate a clinically relevant increase in CVD risk only in individuals with severe disease. We show that the relative risks are greatest in younger populations, whereas in absolute terms the number of CVD events attributable to psoriasis per year is modest across all disease severities and particularly low for young individuals and those with mild psoriasis. We also address the question of whether the reported CVD risk is directly attributable to psoriasis by limiting our analysis to data reporting incident CVD. However, confidence in the accuracy of the effect estimates is limited by incomplete controlling for important confounding factors in most studies, particularly traditional risk factors for CVD, and the CVD risk that is directly attributable to psoriasis remains unclear. Nevertheless, when considering only those studies that performed multivariable analysis of time-to-event data, the most appropriate type of analysis, there was consistent evidence across all primary outcomes, indicating that the incidence is significantly higher in severe psoriasis patients than in individuals without psoriasis (Gelfand et al., 2006, 2009; Mehta et al., 2010). The within-study differences between mild and severe psoriasis in the Gelfand cohorts were also significant (P ¼ for MI and P ¼ 0.03 for stroke). Study heterogeneity Heterogeneity was noted for all outcomes. For CVD mortality in severe psoriasis, in contrast to consistent findings among four other studies, one cohort of psoralen plus UVA-treated patients (Stern and Huibregtse, 2011) showed no increase in CVD mortality compared with a matched general population cohort. The cohort represented a selected population comprising individuals who were recruited to a clinical trial, all having psoralen plus UVA exposure; thus, the findings may not be wholly generalizable. In addition, their definition of severe psoriasis differed from other studies; each center had to enter 50 or more patients with at least 30% BSA coverage, but subsequent patients could be enrolled with o30% coverage (the mean BSA was 33%). Of note, four out of five studies (including Stern and Huibregtse (2011)) only calculated a standard mortality ratio, rather than a multivariable analysis controlling for relevant confounders. All of these factors could be relevant when considering the difference in findings (see also Gelfand et al. (2013)). Interestingly, the subanalysis within the Stern cohort also supports the increased risk of CVD mortality in more severe disease, as multivariable analysis showed a trend toward increased risk in more extensive psoriasis even when the referent is moderate psoriasis (443% vs. o15% BSA; Supplementary Figure S1 online). For MI and stroke, among the studies reporting all psoriasis severities, one study (Li et al., 2012) reporting a female nurse population showed higher relative incidence rates of MI and stroke in the psoriasis group compared with other studies. The reason for this is unclear. Another study (Lin et al., 2011) also reported a higher relative incidence of MI, possibly because the study population consisted of those accessing ambulatory care, who are more likely to have severe disease. In studies reporting data for mild and severe psoriasis subgroups, heterogeneity may be driven by the effect of age on relative risk, as the effect estimates from the Gelfand study (Gelfand et al., 2006) differed by age. In support of this, risk estimates from age-stratified data showed consistent findings between two studies (Gelfand et al., 2006; Ahlehoff et al., 2011a) within the younger and older groups, with higher risk in the

5 younger group in both mild and severe psoriasis (Supplementary Figure S2 online). For each outcome, there was inconsistency in the mild psoriasis subgroup results, where the association is weak and thus susceptible to variations in baseline characteristics and study design. There was also variability in the definition of mild psoriasis, including the use of vitamin D analogs but not systemic psoriasis therapies in one study (Ahlehoff et al., 2011a), no use of systemic therapies (Gelfand et al., 2006), and psoriasis treated in an outpatient setting (Mallbris et al., 2004). Study strengths and limitations A key strength of our study is limiting the analysis to studies investigating incident (as opposed to prevalent) CVD. In addition, a comprehensive and transparent literature search identified all relevant reports, and the methodological quality of included studies was assessed according to standard and detailed criteria. However, we acknowledge important limitations. Our systematic review is limited by the quality of the included studies: first, the majority involved retrospective data analyses, which increase the risk of bias associated with the recording of baseline data, the need for imputation, and potential selection bias. Notably, very few of the studies reported how missing data were handled or if imputation was used. High risk of selection bias was noted in three studies (Boffetta et al., 2001; Mallbris et al., 2004; Stern and Huibregtse, 2011), as the exposed and unexposed group were sampled from different cohorts. Second, there was incomplete correction for relevant confounders in most of the studies. This is important when considering whether there is a causal relationship between CVD risk and psoriasis in the context of a population in which the incidence of traditional risk factors for CVD such as diabetes, hypertension (Qureshi et al., 2009), and smoking (Poikolainen et al., 1999) are increased. This problem was addressed partly by some investigators undertaking multiple sensitivity analyses to explore the robustness of results to changes in the analyses and assumptions. For example, the estimated magnitude of unmeasured confounders that could nullify the results for ischemic stroke, assuming a prevalence of 20%, would have to be greater than the effects and distribution of any of the measured confounders (valvular heart disease or prior MI); (Ahlehoff et al., 2011b). Two study groups (Gelfand et al., 2006, 2009; Mehta et al., 2010) also showed that the association between psoriasis and cardiovascular death held true when excluding those with psoriatic arthritis or those who had received methotrexate or ciclosporin. Third, psoriasis was ascribed as severe in most of the studies if systemic treatments were used, which may have resulted in misclassification bias. Therefore, an alternative explanation for findings in the severe cohort is that the treatments themselves drive the risk (for example, ciclosporin and acitretin cause hyperlipidemia). Fourth, there is a risk of detection bias, with a possible alternative explanation for the increased CVD incidence recorded in psoriasis being that individuals with psoriasis are more likely to seek medical attention and thus be screened for CVD; thus, the recorded increased incidence may be an artifact of more complete detection rather than psoriasis itself driving the effect. Fifth, in relation to our analysis, we note the inherent limitations of meta-analysis of observational data, particularly in light of the noted study variability and the small number of studies available for the severity subgroups, which poses a challenge for clear interpretation. In addition, pooling of the heterogenous results for MI in severe psoriasis in a randomeffects model produced a nonsignificant result, although both individual studies demonstrated severe psoriasis to be a risk factor. Finally, with regard to the age subgroups, we note the risk of chance effects owing to splitting the data set according to two different parameters (severity and age). Findings compared with other studies Our findings are consistent with those in three cross-sectional studies showing elevated prevalence of various CVD end points in psoriasis (Cohen et al., 2007; Prodanovich et al., 2009; Xiao et al., 2009), and with elevated estimated risks of coronary heart disease and stroke based on baseline demographic data from phase II and III psoriasis intervention studies (Kimball et al., 2010). One other systematic review has also reported an increased relative risk of the composite vascular end point of MI and stroke in psoriasis (Xu and Zhang, 2012). In rheumatoid arthritis (Avina-Zubieta et al., 2008; Meune et al., 2009), pooled estimates of relative risk are virtually identical to those reported in our study, and, notably, ascertainment of specific disease-associated risk is similarly complicated by confounders, where a similar spectrum of risk factors occurs (Boyer et al., 2011). In addition, one study (Ahlehoff et al., 2011a) showed that, based on the Danish patient register, the increase in CVD mortality risk associated with severe psoriasis was comparable to that associated with diabetes (incidence rate ratio 1.57 ( ) for severe psoriasis and 1.59 ( ) for diabetes; P ¼ 0.72). Study implications Our findings indicate that the risks of CVD are largely confined to severe disease and, specifically, to individuals requiring systemic therapy or hospital admission. Although there is a paucity of accurate epidemiological data about psoriasis, individuals with severe disease constitute a minority: in one US-based study, the proportion of individuals with body surface area involvement 410% (an accepted definition of severe) was 5 25%. In the Gelfand and Ahlehoff cohorts, the rates of severe psoriasis (variably defined) were % and 7.1%, respectively. Therefore, the diagnosis of psoriasis alone, without consideration of disease severity, is unlikely to be a clinically useful marker of CVD risk, and recommendations for formal cardiovascular risk assessment should be limited to individuals with severe disease, whereas proactive management of cardiovascular risk factors in individuals with psoriasis, as in the general population, is also important. There is some evidence to indicate that there may 2344 Journal of Investigative Dermatology (2013), Volume 133

6 be a causal link between severe psoriasis and CVD risk; however, given the serious methodological shortcomings in some of the studies in relation to correction for traditional risk factors, there is uncertainty about the size of CVD risk directly attributable to psoriasis. Choice of treatment strategies should therefore remain driven by skin disease severity rather than by any purported benefit to cardiovascular risk profile. CONCLUSIONS Severe psoriasis is associated with an increased risk of CVD. Research is required to better understand the complex relationship between psoriasis, traditional risk factors, and CVD. Long-term, large-scale cohort studies that adequately control for confounding factors and detection bias are required to address the question of whether aggressive treatment of severe psoriasis has an impact on clinically relevant CVD end points. MATERIALS AND METHODS Search strategy A systematic literature search of MEDLINE, Embase, Cinahl, and The Cochrane Library was conducted (Supplementary Table S5 online). A total of 4,646 abstracts were retrieved, 118 ordered, and 15 included in the review (Supplementary Figure S2 online). The definition of CVD was based on the World Health Organization classification (Mendis et al., 2011). Systematic reviews or cohort studies of the incidence of CVD reporting the presence of psoriasis as a prognostic factor were included. The primary outcomes were CVD mortality, MI, and stroke. Prognostic factors that allowed subgroups of the psoriasis population to be investigated were prespecified as disease severity and age. Case control or cross-sectional studies were excluded, as were papers lacking full reports. Data extraction and quality assessment The studies were all large cohorts (Supplementary Table S1 online). Methodological quality was assessed according to the risk of bias across five domains based on established criteria (Hayden et al., 2006); (Supplementary Table S6 online). This included adjustment for key confounding factors (age, sex, smoking status, alcohol intake, body mass index or obesity, hyperlipidemia, hypertension, and diabetes) (Table 1). The other domains were attrition, prognostic factor, and outcome biases. Where possible, data for both minimally and maximally adjusted risk estimates were extracted, although maximally adjusted risk estimates are reported. Where more than one study reported on the same or largely overlapping cohorts for the same outcome, the study with the greatest number of person-years of follow-up was used. If this was not distinguishable, the study using the most appropriate analysis was used. Statistical analysis The relative risk estimates were combined using a random-effects (DerSimonian and Laird) meta-analysis to provide a conservative estimate of the effect owing to potential population differences between the studies. The log of the risk estimate and its standard error were calculated from the estimate, and its 95% CI was reported in the studies, and these data were used to input into Review Manager 5.1 (Cochrane Collaboration, Copenhagen, Denmark). A value of Pr0.05 was considered to be statistically significant. Data were stratified a priori according to psoriasis disease severity, with the risk for mild and severe psoriasis being presented separately when possible. Heterogeneity was assessed by the w 2 test for significance at Po0.1 or an I 2 inconsistency statistic of 450% to indicate significant heterogeneity. Assessments of potential differences in effect between subgroups were based on the w 2 test for heterogeneity statistics between subgroups. We planned to assess publication bias using formal tests for funnel plot asymmetry. CONFLICT OF INTEREST The authors state no conflict of interest. ACKNOWLEDGMENTS This work was undertaken by the National Clinical Guideline Centre, which received funding from the National Institute for Health and Clinical Excellence. The views expressed in this publication are those of the authors and not the Institute. No funding was received that directly influenced the preparation or decision to publish this review. CHS acknowledges financial support from the Department of Health via the National Institute for Health Research (NIHR) comprehensive Biomedical Research Centre award to Guy s & St Thomas NHS Foundation Trust in partnership with King s College London and King s College Hospital NHS Foundation Trust. RBW is an NIHR Senior Clinical Lecturer. We thank Maggie Westby for methodological advice and Jill Cobb and Paul Miller for literature searching and referencing. Author contributions EJS determined the structure of the review, prepared the first draft, and performed the quality rating. EJS and JMN selected the references and extracted the data. CHS helped determine the concept and structure of the review, and contributed to the writing and interpretation of the findings. RBW and JP contributed to the writing and interpretation of the findings. SUPPLEMENTARY MATERIAL Supplementary material is linked to the online version of the paper at REFERENCES Ahlehoff O, Gislason GH, Charlot M et al. (2011a) Psoriasis is associated with clinically significant cardiovascular risk: a Danish nationwide cohort study. J Intern Med 270: Ahlehoff O, Gislason GH, Jorgensen CH et al. (2011b) Psoriasis and risk of atrial fibrillation and ischaemic stroke: a Danish nationwide cohort study. Eur Heart J 33: Ahlehoff O, Gislason GH, Lindhardsen J et al. (2011c) Prognosis following firsttime myocardial infarction in patients with psoriasis: a Danish nationwide cohort study. J Intern Med 270: Ahlehoff O (2011) Psoriasis and cardiovascular disease. Dan Med Bull 58:B4347 Avina-Zubieta JA, Choi HK, Sadatsafavi M et al. (2008) Risk of cardiovascular mortality in patients with rheumatoid arthritis: a meta-analysis of observational studies. Arthritis Rheum 59: Boffetta P, Gridley G, Lindelof B (2001) Cancer risk in a population-based cohort of patients hospitalized for psoriasis in Sweden. J Invest Dermatol 117: Boyer JF, Gourraud PA, Cantagrel A et al. (2011) Traditional cardiovascular risk factors in rheumatoid arthritis: a meta-analysis. Joint Bone Spine 78: Brauchli YB, Jick SS, Miret M et al. (2009) Psoriasis and risk of incident myocardial infarction, stroke or transient ischaemic attack: an inception cohort study with a nested case-control analysis. Br J Dermatol 160:

7 Cohen AD, Shapiro Y, Davidovici B et al. (2007) Psoriasis and ischemic heart disease: a case-control study. G Ital Dermatol Venereol 142: Gelfand J, Mehta NN, Langan SM (2013) Psoriasis and cardiovascular risk: strength in numbers, part II. J Invest Dermatol 131: Gelfand JM, Dommasch ED, Shin DB et al. (2009) The risk of stroke in patients with psoriasis. J Invest Dermatol 129: Gelfand JM, Neimann AL, Shin DB et al. (2006) Risk of myocardial infarction in patients with psoriasis. JAMA 296: Gisondi P, Tessari G, Conti A et al. (2007) Prevalence of metabolic syndrome in patients with psoriasis: a hospital-based case-control study. Br J Dermatol 157:68 73 Hayden JA, Cote P, Bombardier C (2006) Evaluation of the quality of prognosis studies in systematic reviews. Ann Intern Med 144: Kaye JA, Li L, Jick SS (2008) Incidence of risk factors for myocardial infarction and other vascular diseases in patients with psoriasis. Br J Dermatol 159: Kimball AB, Guerin A, Latremouille-Viau D et al. (2010) Coronary heart disease and stroke risk in patients with psoriasis: retrospective analysis. Am J Med 123:350 7 Kurd SK, Troxel AB, Crits-Christoph P et al. (2010) The risk of depression, anxiety, and suicidality in patients with psoriasis: a population-based cohort study. Arch Dermatol 146:891 5 Li W-Q, Han J-L, Manson JE et al. (2012) Psoriasis and risk of nonfatal cardiovascular disease in U.S. women: a cohort study. Br J Dermatol 166:811 8 Lin HW, Wang KH, Lin HC et al. (2011) Increased risk of acute myocardial infarction in patients with psoriasis: a 5-year population-based study in Taiwan. J Am Acad Dermatol 64: Mallbris L, Akre O, Granath F et al. (2004) Increased risk for cardiovascular mortality in psoriasis inpatients but not in outpatients. Eur J Epidemiol 19: Maradit-Kremers H, Crowson CS, Nicola PJ et al. (2005) Increased unrecognized coronary heart disease and sudden deaths in rheumatoid arthritis: a population-based cohort study. Arthritis Rheum 52: Mehta NN, Azfar RS, Shin DB et al. (2010) Patients with severe psoriasis are at increased risk of cardiovascular mortality: cohort study using the General Practice Research Database. Eur Heart J 31: Mendis S, Puska P, Norrving B (2011) Global Atlas on Cardiovascular Disease Prevention and Control. World Health Organization: Geneva Meune C, Touze E, Trinquart L et al. (2009) Trends in cardiovascular mortality in patients with rheumatoid arthritis over 50 years: a systematic review and meta-analysis of cohort studies. Rheumatology (Oxford) 48: Naldi L, Chatenoud L, Linder D et al. (2005) Cigarette smoking, body mass index, and stressful life events as risk factors for psoriasis: results from an Italian case-control study. J Invest Dermatol 125:61 7 Parisi R, Symmons D, Griffiths CEM et al. (2012) Global epidemiology of psoriasis: a systematic review of incidence and prevalence. J Invest Dermatol 133: Poikolainen K, Karvonen J, Pukkala E (1999) Excess mortality related to alcohol and smoking among hospital-treated patients with psoriasis. Arch Dermatol 135: Prodanovich S, Kirsner RS, Kravetz JD et al. (2009) Association of psoriasis with coronary artery, cerebrovascular, and peripheral vascular diseases and mortality. Arch Dermatol 145:700 3 Qureshi AA, Choi HK, Setty AR et al. (2009) Psoriasis and the risk of diabetes and hypertension: a prospective study of US female nurses. Arch Dermatol 145: Rapp SR, Feldman SR, Exum ML et al. (1999) Psoriasis causes as much disability as other major medical diseases. J Am Acad Dermatol 41:401 7 Reich K (2012) The concept of psoriasis as a systemic inflammation: implications for disease management. J Eur Acad Dermatol Venereol 26(Suppl 2):3 11 Sommer DM, Jenisch S, Suchan M et al. (2006) Increased prevalence of the metabolic syndrome in patients with moderate to severe psoriasis. Arch Dermatol Res 298:321 8 Stern RS, Huibregtse A (2011) Very severe psoriasis is associated with increased noncardiovascular mortality but not with increased cardiovascular risk. J Invest Dermatol 131: Stern RS, Laird N, Melski J (1984) Cutaneous squamous-cell carcinoma in patients treated with PUVA. New Engl J Med 310: US National Institutes of Health (accessed 9 January 2013) Wakkee M, Herings RM, Nijsten T (2010) Psoriasis may not be an independent risk factor for acute ischemic heart disease hospitalizations: results of a large population-based Dutch cohort. J Invest Dermatol 130:962 7 Wolfe F, Freundlich B, Straus WL (2003) Increase in cardiovascular and cerebrovascular disease prevalence in rheumatoid arthritis. J Rheumatol 30:36 40 Xiao J, Chen LH, Tu YT et al. (2009) Prevalence of myocardial infarction in patients with psoriasis in central China. J Eur Acad Dermatol Venereol 23: Xu T, Zhang Y-H (2012) Association of psoriasis with stroke and myocardial infarction: meta-analysis of cohort studies. Br J Dermatol 167: Yarur AJ, Deshpande AR, Pechman DM et al. (2011) Inflammatory bowel disease is associated with an increased incidence of cardiovascular events. Am J Gastroenterol 106: Journal of Investigative Dermatology (2013), Volume 133