Discussion

Our Subset I Study indicated that a high TC level correlated with the severity of atopic dermatitis in school-age children, and our Subset II Study (a single center long-term study using OMOP-CDM) indicated that the high TC group had a slightly greater HR for the onset of atopic dermatitis than the normal TC group.
Several previous studies examined the relationship of hyperlipidemia with atopic dermatitis in adults.7,10-14 The results of the present study of children are consistent that reported a positive correlation of hyperlipidemia with atopic dermatitis in adults.7,10 However, very few studies have examined the association of atopic dermatitis with lipid parameters in children. A database study by Augustin et al.14 showed a significantly increased prevalence rate of hyperlipidemia in children with atopic dermatitis compared to those without atopic dermatitis (prevalence ratio: 1.29; 95% CI: 1.12, 1.49; P < 0.05). However, their study defined atopic dermatitis based only on ICD code, did not directly measure TC, LDL-C, HDL-C, or TG, did not report the severity of atopic dermatitis, and was a cross-sectional study.
Previous research indicated that dyslipidemia is a type of chronic inflammatory response.26 In particular, elevated TG and LDL-C lead to increased pro-inflammatory signaling and increased expression of TNF-α and interleukin (IL)-6.26,27 In contrast, HDL-C has an anti-inflammatory effect, in that it modulates T-cell activation.26,27 This supports the interpretation that the presence of a chronic inflammatory state (as indicated by dyslipidemia) is responsible for chronic skin inflammation, and provides a possible mechanism for the relationship between atopic dermatitis and hyperlipidemia.
We used the SCORAD score to grade the severity of atopic dermatitis, and found that this score had positive associations with TC and TG and a negative association with HDL-C. Thus, our results indicate an association of hyperlipidemia with the severity of atopic dermatitis. However, another previous case-control study of adults in Germany reported conflicting results in the relationship of dermatitis with HDL-C.7 This study found that the HDL-C level was significantly higher in subjects with atopic dermatitis (61.0 vs.54.9 mg/dL; P < 0.01) and that after stratification by gender, there was a significant positive association of HDL level and atopic dermatitis in men.7 Our results indicated that disease severity (SCORAD score), erythema, excoriation, lichenification, and pruritus scores in children had positive correlations with the levels of TC and TG, and negative correlations with HDL-C level. Thus, cholesterol might be related to the chronicity of atopic dermatitis in children.
The results of our Subset II Study indicated that the high TC group had a greater HR for the onset of atopic dermatitis during a median follow-up time of 1,750 days. The Observational Health Data Sciences and Informatics (OHDSI) initiative is a large-scale international collaborative that is investigating the value of analyzing health data.20
We used a retrospective long-term follow-up study (Subset II) to complement our cross-sectional study (Subset I) and examine the casual relationship of hyperlipidemia and atopic dermatitis. Nonetheless, there were several limitations to this study. We did not perform repeated objective clinical assessments and there are data quality issues inherent to working with health care databases, in that covariates, exposures, and outcomes may be inadequately or incorrectly measured. However, our use of data standardization into a common data model, propensity-score matching, and calibration via negative controls helped to prevent erroneous conclusions. Also, the case definition of atopic dermatitis was based on ICD-10 diagnostic codes, so there may be questions about the validity of our definition. To address this issue, we defined atopic dermatitis as at least 5 diagnoses of atopic dermatitis from ICD codes plus 2 treatment codes for atopic dermatitis drug therapy.22
Our study has many merits. To the best of our knowledge, this is the first large-sample study of urban children to comprehensively investigate the association between blood lipid level and the severity of atopic dermatitis. Moreover, trained physicians directly assessed the severity of atopic dermatitis. Most previous studies of the association between atopic dermatitis and hyperlipidemia were surveys, and none of these previous studies examined the disease severity. In our study, pediatricians specializing in allergies determined all SCORAD scores. Another strength is a longitudinal cohort analysis to supplement and valify the cross-sectional analysis.
In conclusion, our results indicated that hyperlipidemia in children has a positive association with the presence and severity of atopic dermatitis. These findings suggest that atopic dermatitis should be considered a systemic disorder that shares certain characteristics with other chronic inflammatory disorders.