The chronic inflammatory skin disease psoriasis has a prevalence around 2% in Western countries1 and is recognized to be the most common immunologic disorder. The prevalence is 2.5% in Peru.2 Severity ranges from mild forms affecting small areas of the body to serious disease causing lesions extending over large areas and greatly affecting patients’ quality of life.3

Various cardiovascular risk factors have been associated with psoriasis in recent years, attributable to chronic systemic inflammation that leads to atherosclerosis and finally cardiovascular disease.4 Metabolic syndrome, which doubles the risk of coronary disease, offers a way to assess a patient for cardiovascular disease.5 This syndrome involves a constellation of conditions (obesity, dyslipidemia, diabetes mellitus, and hypertension) that have a greater effect on cardiovascular disease when they occur together than when they develop in isolation. About 25% of Latin Americans (17% in Peru) have metabolic syndrome 17%.6 Insulin resistance and abnormal functioning of adipose tissue are thought to be the underlying mechanism.7

Recent meta-analyses have found that patients with psoriasis are at risk for metabolic syndrome, with odds ratios (ORs) between 1.84 and 2.268) and that severe psoriasis puts patients at greater risk (OR, 1.98; 95%CI, 1.62–2.43) than mild forms of the disease (OR, 1.22; 95%CI, 1.11–1.35).9 The analyzed studies have not included Latin American populations, however.

The prevalence of metabolic syndrome in patients with psoriasis in Latin America ranges from 40% in Mexico10 to 70% in Honduras.11 In Peru the prevalence is 46%.12 Few original studies have addressed this question in this part of the world, however.

Consequently, although current recommendations suggest we screen for metabolic syndrome in patients with psoriasis, the strength of association between the 2 conditions has not been studied systematically in Latin America.

We undertook a meta-analysis with the main objective of assessing our current understanding the association between metabolic syndrome and psoriasis in Latin American patients.

We followed the MOOSE reporting guidelines (for Meta-analysis of Observational Studies in Epidemiology)13 to design a systematic review of the literature. Included in the search were observational studies of various types: retrospective and prospective, cross-sectional, and case–control or cohort designs. We sought studies published between January 1980 and January 2016. Included studies had to assess the association between psoriasis and metabolic syndrome diagnosed by physical examination following predefined criteria or by a review of medical records. Studies also had to have 2 arms (psoriasis patients and controls) and report the incidence or prevalence of metabolic syndrome in each arm. Articles could be published in either English or Spanish.

Population characteristics were men or women aged 18 years or older who were native to any Latin American country and who had been diagnosed with psoriasis based on specific criteria. We excluded studies for which full datasets were unavailable even if we made the effort to contact authors.

The literature search was carried out in MEDLINE, Scopus, SciELO, Google Scholar, Science Direct, LILACS and the virtual libraries of universities. The following search terms were used: ([psoriasis] or [vulgar psoriasis]) and ([metabolic syndrome] or [metabolic syndrome X]) and ([insulin resistance syndrome X] or [syndromeX, metabolic] or [metabolic X syndrome]). The search string was adapted to the syntax of each database. The initial search yielded 1493 titles. After eliminating studies not done in Latin American populations, we were left with 75 articles. Eight additional studies were identified by checking the references lists of the 75 indexed studies and from other sources of information. Two authors (M.R. and F.C.) screened the titles and abstracts of the 83 candidate studies to determine whether they seemed to meet the inclusion criteria. Ten were initially chosen but after careful review, 5 of them were excluded because 2 had no control group, 2 studied individual components of metabolic syndrome, and 1 used a population already included in another study. Thus, 5 articles met the inclusion criteria for systematic review and meta-analysis (Fig. 1). If the screening authors did not agree on whether to include or exclude an article, the third author (E.Q.) was consulted to resolve the disagreement.

Figure 1.

Flow chart of the process of screening observational studies on the association between psoriasis and metabolic syndrome.

(0.15MB).

The following data were gathered into a model spreadsheet in MS Excel for analysis: year of publication, country where the patients lived, design, number of participants in each group, mean age, prevalence of metabolic syndrome in each group, and adjustment for covariates. Authors M.R. and F.C. checked data twice for accuracy; discrepancies were resolved in consultation with E.Q.

Metabolic syndrome was defined in accordance with specific criteria. According to the Third Report of the National Cholesterol Education Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (the Adult Treatment Panel III [ATP-III]),14 metabolic syndrome is defined by the presence of 3 out of 5 of the following factors: 1) high blood pressure of at least 130/85mmHg (or on antihypertensive medication), 2) high fasting blood sugar level of 100mg/dL or higher (or on medication to lower sugar levels), 3) triglyceride level of 150mg/dL or higher (or on medication to lower the level), 4) a high-density lipoprotein level of 40mg/dL or less in men or 50mg/dL or less in women (or on therapy to raise that level), and 5) waist circumference of 102cm or more in men or 88cm or more in women.

Similar criteria were proposed by the International Diabetes Federation (IDF),15 but they named obesity as a necessary condition, defined by a body mass index greater than 30kg/m2 and specifying waist measurements for different ethnic groups. Although measurements for a Latin American population were not given, the IDF advised using those for a South Asian population (waist circumference of 90cm in men and 80cm in women.

The main outcome measure was the association between psoriasis and metabolic syndrome, reflected by an OR and 95% CI. We also calculated pooled ORs and 95% CIs by means of a random effects model. Significance was set at P<.05. Results for subgroups formed according to patient characteristics and criteria used to define metabolic syndrome were also analyzed. Heterogeneity between studies was measured with the I2 statistic. A value for I2 greater than 50% indicated heterogeneity.16 Descriptive statistics (percentages and means) were tabulated for the included studies; they were analyzed and figures created with the Stata13 program and Review Manager Version 5.1

The quality of studies was evaluated with a modified version of the Cochrane risk of bias tool for observational studies. The tool lists 6 traits to score as 0 (absent) or 1 (present). A total score of 3 or less indicated low quality and a score of 4 or 5 indicated high quality, as described elsewhere.17 These scores were used in the subgroup analysis.

Publication bias was measured by constructing funnel plots. A symmetric, inverted funnel indicated a high probability that no publication bias was present.

Finally, sensitivity was assessed by removing a study to test its effect on the pooled OR. Similarity between pooled ORs with and without the article removed would indicate a high degree of certainty of results.

We included 5 trials from 3 Latin American countries (Table 1). The trials analyzed 545 patients (241 with psoriasis and 304 controls). The meta-analyzed sample included 159 patients who had a recent diagnosis of psoriasis or who had not received systemic treatment (vs 218 controls); 82 patients with chronic psoriasis or who were on a systemic treatment were also included (vs 86 controls). All the trials were prospective case–control studies The ATP-III criteria for metabolic syndrome were applied in 2 of the studies,10,12 and the IDF diagnostic criteria were applied in the others18–20 (Table 1).

Metabolic syndrome was diagnosed in 46.5% in the psoriasis arm and in 24.7% of the controls (pooled OR, 2.63; 95%CI, 1.11–6.23; P=.03; I2=78%) (Fig. 2).

Figure 2.

Forest plot of the prevalence of metabolic syndrome in psoriasis patients in the meta-analyzed observational studies. A Mantel-Haenszel (M-H) fixed-effect method was used to calculate the ORs and 95%CIs. Pso refers to psoriasis at any level of severity or point in diagnosis; OR, odds ratio.

(0.21MB).

The funnel plots revealed asymmetry along the mean, suggesting high likelihood of publication bias (Fig. 3).

Figure 3.

Funnel plot of the prevalence of metabolic syndrome in patients with psoriasis according to quality of study design used in the meta-analyzed studies. OR refers to odds ratio; SE, standard error.

(0.07MB).

The subgroup analysis showed that the association between metabolic syndrome and psoriasis was stronger in the studies that used the ATP-III criteria (pooled OR, 3.97; 95%CI, 1.27–12.42) than in those that used the IDF criteria (pooled OR, 2.75; 95%CI, 0.50–15.27). Likewise, studies that included patients recently diagnosed with psoriasis, or not yet prescribed systemic treatment (psoriasis de novo subgroup), detected a significantly lower risk of metabolic syndrome (pooled OR, 1.50; 95%CI, 0.52–4.30) than those with only patients with chronic disease or treated with systemic drugs (psoriasis-in-treatment subgroup) (pooled OR, 6.65; 95%CI: 3.32–13.31) (Fig. 4).

Figure 4.

Forest plots of the prevalence of metabolic syndrome in subgroups of patients with psoriasis formed according to patient characteristics. A Mantel-Haenszel (M-H) fixed-effect method was used to calculate the ORs and 95%CIs. Pso refers to psoriasis at any level of severity or point in diagnosis; OR, odds ratio.

(0.3MB).

Finally, analysis by quality of research design showed that risk of metabolic syndrome in psoriasis patients was higher in high-quality studies (pooled OR, 3.48; 95%CI, 2.22–5.45) than in low-quality studies (pooled OR, 1.44; 95%CI, 0.75–2.77). Fig. 5 shows the results of an analysis of risk of bias in these studies.

Figure 5.

Assessment of risk of bias. A, Within studies. B, Between studies.

(0.22MB).

Analysis of sensitivity (Table 2) showed that the results for effect direction and size were not significantly affected by removing any study's data from the meta-analysis, suggesting the overall findings are robust.

This systematic review and meta-analysis shows that risk for metabolic syndrome is nearly 3-fold higher in Latin American patients with psoriasis than in the general population. Furthermore, the risk is higher for patients with severe or chronic psoriatic disease. Finally, we advise using the ATP-III criteria in future studies on this subject.

The authors declare that they have no conflicts of interest.