The Association With Cardiovascular Disease and Type 2 Diabetes in Adults With Atopic Dermatitis

A Systematic Review and Meta-analysis

J.P. Thyssen; A.-S. Halling-Overgaard; Y.M.F. Andersen; G. Gislason; L. Skov; A. Egeberg

Disclosures

The British Journal of Dermatology. 2018;178(6):1272-1279. 

In This Article

Results

A total of 2855 citations were identified in the screening process, of which 53 were considered relevant based on the title and abstract review. The remaining 2802 articles did not include information on any of the study end points. Overall, 37 manuscripts were excluded owing to reasons listed in the Preferred Reporting Items for Systematic Reviews and Meta–Analyses flowchart (Figure 1). A total of 16 publications were included in the qualitative analysis and further details about the publications are presented online and below (Table S1 Table S2 Table S3 Table S4 Table S5 Table S6; see Supporting Information).[4–10,12–20]

Meta–analysis Based on Crude Data

A total of 13 studies were included in the quantitative meta–analysis on crude data. Studies originated from Taiwan (two),[15,18] Germany (one),[6] Denmark (five),[4,9,10,12,14] Canada (one)[20] and the U.S.A. (four).[5,7,8,13] Two of the publications contained data from multiple cohorts (n = 3),[5,8] which were treated separately in the analyses, i.e. all cohorts were included. All studies included in the quantitative analysis had a NOS score of 7 or higher, which indicates good study quality. The studies were grouped according to the end points, i.e. myocardial infarction, presumed type 2 diabetes, hypertension, angina pectoris and stroke. Each end point in the meta–analysis comprised studies from at least two continents. A test for publication bias yielded asymmetrical funnel plots except for the category 'diabetes'; however, the number of studies in each category was low, thereby limiting the interpretation. In particular, the funnel plot for hypertension was skewed to the far right, indicating considerable publication bias.

A total of seven studies including 100 382 patients with AD and 1 657 116 reference individuals were analysed in the category 'myocardial infarction' (Figure 2).[6–9,20] Five of these studies originated from North America and two from Europe. Overall, there was no significant association between AD and myocardial infarction (pooled OR 1·14, 95% CI 0·83–1·56; I2 95·5%). Most of the nine studies included in the diabetes category did not specify the type of diabetes (type 1 vs. type 2). The meta–analysis based on these nine studies,[4,5,7,10,12,13,18,20] which included a total of 80 357 patients with AD and 579 036 reference individuals, showed no association between AD and presumed type 2 diabetes (pooled OR 1·11, 95% CI 0·87–1·42; I2 97·1%) (Figure 3a). For stroke, a total of eight studies[6–9,15,20] including 120 706 patients with AD and 1 677 441 reference individuals were included in the meta–analysis, but again showed no association with AD (pooled OR 1·15, 95% CI 0·95–1·39; I2 89·8%) (Figure 3b). Similarly, no association between AD and hypertension was observed in a meta–analysis based on seven studies,[5–7,14,18,20] which included 93 837 patients with AD and 1 626 771 reference individuals (pooled OR 1·16, 95% CI 0·98–1·37; I2 98·9%) (Figure 3c). Pooled analysis for angina pectoris showed a significant association based on 41 660 patients with AD and 1 203 168 reference individuals from four cohorts (pooled OR 1·73, 95% CI 1·27–2·37; I2 89·3%) (Figure 3d).[6,8] No quantitative analyses were performed for the end points of 'coronary artery disease/ischaemic heart disease', 'cardiovascular death' and 'heart failure' because of the low number of published studies.

Figure 2.

Odds ratio meta–analysis of association between atopic dermatitis and myocardial infarction based on crude data.

Figure 3.

Odds ratio meta–analysis of association between atopic dermatitis and (a) type 2 diabetes, (b) stroke, (c) hypertension and (d) angina pectoris, respectively, based on crude data.

Meta–analysis Based on Adjusted Data

In total, 12 studies were included in the quantitative meta–analysis on adjusted data and these originated from Taiwan (one),[15] South Korea (two),[16] Germany (two),[6,19] Denmark (four),[4,9,12,14] Canada (one)[20] and the U.S.A. (three).[5,7,8] A total of nine studies including 121 077 patients with AD and 1 686 548 reference individuals were analysed in the category 'myocardial infarction' (Figure S1; see Supporting Information).[6–9,15,19,20] Five of these studies originated from North America, three from Europe and one from Asia. Overall, there was no significant association between AD and myocardial infarction (pooled OR 1·03, 95% CI 0·88–1·21; I 2 74·1%).

A meta–analysis of five studies,[4,5,12,20] which included 56 688 patients with AD and 443 542 reference individuals, showed no association between AD and presumed type 2 diabetes (pooled OR 0·97, 95% CI 0·80–1·18; I2 81·6%) (Figure S2a; see Supporting Information). For stroke, a total of nine studies[6–9,15,19,20] including 121 078 patients with AD and 1 686 550 reference individuals were included in the meta–analysis. No association was found between AD and stroke (pooled OR 1·12; 95% CI 0·95–1·32; I2 81·3%) (Figure 1b). No association was observed between AD and hypertension in a meta–analysis based on six studies[5,6,14,16,20] including 71 040 patients with AD and 1 520 340 reference individuals (pooled OR 1·10, 95% CI 0·97–1·24; I2 94·5%) (Figure 1c). A total of four studies,[5,6] which included 41 660 patients with AD and 1 203 168 reference individuals, were used in the meta–analysis on angina pectoris, which showed a positive association (pooled OR 1·48, 95% CI 1·23–1·79; I2 44·3%). Three of the studies originated from North America[8] and one study originated from Germany (Figure 1d).[6]

Qualitative Analysis

In addition to results from the quantitative analysis, a general population study from Germany found no association between AD and coronary artery disease.[19] In an experimental study, 17 of 31 adult patients with AD from Denmark had a coronary artery calcium score > 0, in addition to the presence of atherosclerotic plaques, when assessed using cardiac computed tomography angiography. In data analysis, AD was significantly associated with mild single–vessel disease.[10] A Danish cohort study analysing data from the same national register as the study by Andersen et al.,[9] showed that the risk of myocardial infarction was increased among patients with at least two hospital diagnoses of AD (either inpatient or outpatient) (adjusted hazard ratio (HR) 1·74, 95% CI 1·21–2·49). However, this study lacked adjustment for important risk factors, such as smoking, unlike the study by Andersen et al.[9,21] In addition to the studies in the meta–analysis, no association between elevated blood pressure and AD was found in a German cohort (n = 2990, high systolic blood pressure; P = 0·6513, high diastolic blood pressure; P = 0·7972 in linear regression models).[6] Notably, another German study found a decreased prevalence ratio of hypertension in patients with AD when compared with reference individuals without AD (n = 1 312 215, prevalence ratio = 0·83; 95% CI 0·81–0·85).[17]

AD was associated with coronary artery disease in three U.S. cohorts [OR 1·96 (95% CI 1·02–3·77), OR 1·38 (95% CI 1·12–1·70) and OR 1·32 (95% CI 1·04–1·66)][8] and in one Taiwanese cohort (P < 0·001).[18] However, a German study found a decreased prevalence of coronary artery disease in patients with AD when compared with reference individuals (n = 1 312 215, prevalence ratio = 0·83; 95% CI 0·80–0·86).[17] Thus far, peripheral artery disease has been associated with AD in one German study (n = 1 180 678; OR 1·32, 95% CI 1·04–1·66)[6] and one U.S. study (n = 34 552; OR 1·90, 95% CI 1·62–2·22).[8] An increased risk of heart failure in patients with AD was reported from a Taiwanese cohort study (n = 40 646; HR 1·46, 95% CI 1·10–1·93),[15] while a U.S. cross–sectional study reported no association between AD and congestive heart failure (n = 4971; OR 1·01, 95% CI 0·40–2·57).[8]

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....