Cohort studies on 71 outcomes among people with atopic eczema in UK primary care data

Atopic eczema (AE), also known as atopic dermatitis, is a prevalent chronic condition globally, imposing significant morbidity and healthcare costs. Beyond its established links with atopic diseases like allergies and asthma, AE's association with non-atopic conditions remains an area of active research. The underlying mechanisms potentially involve chronic inflammation (explaining cardiovascular outcomes), psychological stress, low self-esteem, and sleep deprivation (potentially linking to anxiety and depression). Existing guidelines, such as those from the American Academy of Dermatology (AAD) in 2022, have reviewed evidence for associations between AE and 32 adverse health outcomes. While the link between AE and other atopic conditions is well-established, evidence for many other potential associations (mental illness, cardiovascular disease, metabolic disease, osteoporosis, fractures) remains less clear. This study aimed to provide a comprehensive assessment of the associations between AE and a wide range of health outcomes using a large, population-based dataset to clarify these relationships and inform clinical practice and future research directions. The study's large sample size and broad range of outcomes allowed for a powerful comparison of the strength of different associations, moving beyond simple statistical significance and incorporating consideration of absolute risks to identify clinically relevant findings. The comprehensive approach aimed to minimize researcher bias in outcome selection and ensure a robust evaluation of the AE-outcome associations.

Previous research has highlighted the association between atopic eczema and a range of other health conditions. Studies using the Clinical Practice Research Datalink (CPRD) database have examined links with conditions such as anxiety, cardiovascular diseases (myocardial infarction, renal failure, stroke), fractures, and several cancers. These studies have provided valuable insights, but often focused on a limited set of outcomes. The 2022 AAD guidelines included studies on 32 different adverse health outcomes, providing a more comprehensive, but still limited, overview of the evidence. There is a need for large-scale studies examining a wider range of outcomes to provide a more complete understanding of the comorbidities associated with eczema and the relative strength of these associations. This study builds upon previous work by utilizing a significantly larger dataset and investigating a more extensive spectrum of health outcomes, creating a more comprehensive picture of the comorbidity landscape associated with eczema.

This study employed a matched cohort study design using de-identified data from the Clinical Practice Research Datalink (CPRD) Aurum, encompassing UK primary care electronic health records from April 1, 1997, to March 31, 2023. The CPRD Aurum database includes over 6 million individuals and is considered representative of the English population. Three cohorts were created based on minimum age at inclusion: any age, 18+, and 40+. Individuals with eczema were identified using a validated algorithm based on diagnostic codes and treatment records (emollients, topical glucocorticoids, topical calcineurin inhibitors, systemic immunosuppressants). Eczema-exposed individuals were matched to unexposed individuals based on age, sex, and general practice. Cox proportional hazards regression, stratified on matched sets, was used to estimate hazard ratios (HRs) for the effect of eczema on each outcome, considering minimally adjusted and comorbidity-adjusted models. Sensitivity analyses were conducted using different cohorts (considering various age cut-offs, those with more severe eczema), and models excluding non-consulters to assess the robustness of the findings. The study included 71 adverse health outcomes, encompassing a wide range of conditions. To address multiple testing, 99% confidence intervals were used, and Bonferroni correction was applied. A severity analysis was performed to explore potential associations between severity of eczema and various outcomes. Results were benchmarked against previous CPRD studies using similar designs.

The study confirmed strong associations between eczema and atopic/allergic conditions (food allergy, allergic conjunctivitis, allergic rhinitis, asthma, eosinophilic esophagitis), consistent with established knowledge. Strong associations were also found with immune-mediated conditions like alopecia areata and urticaria. Importantly, the study revealed strong associations, although with relatively small absolute rate differences, with Hodgkin's lymphoma, Crohn's disease, coeliac disease, ulcerative colitis, and autoimmune liver disease. More common outcomes with less strong but still significant associations included irritable bowel syndrome, oesophagitis, gastroesophageal reflux disease, thromboembolic disease, obesity, chronic obstructive pulmonary disease (COPD), gastritis and duodenitis, and peripheral neuropathies. Severity analyses indicated that some outcomes, such as cardiovascular conditions, osteoporosis, and fractures, were primarily associated with severe eczema. Most cancers and neurological conditions showed no significant association with eczema. Sensitivity analyses generally supported the main findings, although some variations were observed depending on the cohort and model used, highlighting the importance of considering age and severity in interpreting the results. The study also provides direct comparison of the magnitude of association between eczema and a wide variety of outcomes.

This study's extensive analysis of 71 outcomes provides a comprehensive picture of the comorbidity profile associated with atopic eczema, expanding upon previous work and offering a valuable comparison between the strength of associations with a diverse range of diseases. While the findings confirm existing knowledge regarding links between eczema and atopic diseases, the identification of strong but smaller-effect associations with specific immunological and gastrointestinal conditions necessitates further mechanistic investigation. The findings related to severity suggest that the impact of eczema on health outcomes may vary depending on the severity of the disease. This supports a more nuanced approach to management of eczema, emphasizing more aggressive treatment for those with severe symptoms. The study’s strength lies in the cross-outcome approach, which mitigates investigator bias in outcome selection and provides a direct comparison of effect sizes. Although causality cannot be definitively established, the significant associations identified highlight the importance of a multidisciplinary approach to eczema care, ensuring early detection and management of associated conditions.

This large-scale cohort study provides a comprehensive overview of the associations between atopic eczema and a wide range of adverse health outcomes. The findings confirm known links with atopic conditions and highlight significant associations with several immunological, gastrointestinal, and metabolic conditions. The severity analyses suggest a stronger association for some outcomes with severe eczema. This study’s novel cross-outcome approach facilitates direct comparison of effect sizes, minimizing investigator bias. Future research should focus on investigating the underlying mechanisms linking eczema to these conditions and exploring the potential for interventions to prevent or mitigate the development of comorbidities.

This study is based on primary care data and may miss diagnoses made in secondary care settings. There are limitations in assessing ethnic diversity due to missing data. Residual confounding due to lifestyle factors not fully captured in the data might influence the results. The definition of eczema severity may be limited by the reliance on prescriptions and hospital admissions and hence some caution is advised in interpreting the magnitude of effects in different severity groups. The broad scope of the study may limit the in-depth investigation of the mechanisms underlying specific associations. Despite utilizing a large dataset, there is still a potential for residual confounding and challenges with capturing the true onset of conditions due to the retrospective nature of the analysis.