DISCUSSION
In this study, constipated patients had a 2.11-fold higher risk for AD than non-constipated patients, regardless of age, gender, or comorbidities. Best of our knowledge, this is the first and largest epidemiological research to utilize the nationwide population-based dataset to clarify the relationship between constipation and AD. This association could be of clinical and pathophysiological importance. Our findings highlight the considerably higher risk of AD in people with constipation. Constipation thus seems to be influential in the development of AD. Clinicians should be alert to the possibility of AD in patients with constipation. Similarly, constipated patients should be informed of the possible risk of AD and be provided with appropriate management for AD as required. Our findings further underline the importance of maintaining good bowel habits so as to avoid constipation, which would in turn mitigate risk of AD.
Our findings are in line with an epidemiological study conducted in Japan, which analyzed the risk factors for allergic disease in 21802 students, aged 15-18 years old, living in the same prefecture between 2012 and 2013.22 The results of the study indicated that constipated students had a 1.17-fold risk for developing AD, which was comparable to the 2.34-fold elevated risk found among teenagers in our study (95% C.I. 1.84-2.98 p<0.001). Moreover, we observed a significantly higher risk of AD not in only childhood, but also in adulthood.
We also noticed that the risk of AD disclosed significantly higher in patients with hypertension, chronic liver disease, COPD, or diabetes,. Most of these comorbidities, such as diabetes and COPD, were associated with chronic inflammatory status, which might lead to release of serum cytokines and T-cell activation. We also speculate that constipation might worsen pre-existing dysbiosis in vulnerable patients with diabetics who may therefore have an increased risk of AD. Intriguingly, a general population study conducted in Copenhagen, suggested that null mutations in the filaggrin gene in patients with diabetes were connected with a higher prevalence of atopic dermatitis. This result implies that poor protection of skin barrier in diabetes patients may lead greater to exposure allergens and increased epidermal water loss.23
The pathophysiological mechanisms underlying the relationship between constipation and atopic dermatitis remain unclear. In recent years, there has been considerable research conducted on the gut microbiota and a number of investigations have indicated that alterations in the gut microbiota might contribute to constipation and constipation-related symptoms.9,18 Compare to the healthy people, constipated individuals had relatively lower amount of obligate bacteria (e.g. Bifidobacterium, Lactobacillus, and Bacteroides spp.) and relatively higher amount of potentially pathogenic microbes, such as Pseudomonas aeruginosa. These changes in the intestinal environment could influence bowel motility by the active substances.24. Similar changes in gut microbiota were reported in patients with AD. For instance, a study in Japan indicated that lower Bifidobacterium counts were found in patients with AD compared with those of healthy people in the fecal microflora.25 In addition, other studies observed that there were low numbers of Bifidobacterium and Bacteroides spp. as well as a lack of intestinal microbiota variation in AD patients.26,27 Some studies demonstrated that microbial-derived metabolic products, specifically SCFAs, functioned as important drivers of T-cell subset proliferation and activity.28,29 Morevoer, it has been shown that production of intestinal microbial SCFAs might down-regulate proinflammatory reactions at the allergen-insult site.13,30,31 In contrast, immune homeostasis might be destroyed by typical ”Westernized” dietary intake, which can be characterized as being high proportion in fat and less proportion in fiber, altering the gut microflora, and resulting in reduced production of SCFAs.10,13,32–34 Furthermore, SCFAs might affect bowel motility via stimulating the contraction of colonic muscles, thereby helping to relieve or prevent constipation.35,36 Therefore, low fiber intake in constipated subjects might also have a key role in the development of atopy. Taken together the currently available evidence indicates that the gut microbiota might play a pivotal mechanistic role linking constipation and atopic dermatitis. It is not known how constipation changes the composition of the intestinal microbiome and how relevant this is to AD. Further comprehensive metagenomic and metabolomic analyses of the intestinal microbiota in constipated people are warranted to elucidate the mechanisms underlying these associations.
The major advantages of this study were the relatively long follow-up period and large sample size. A complete past history of utilized medical services was available for all cases. Thus, there was minimal selection, information, and recall bias. As such, it was possible to properly test our hypothesis. There were some potential limitations in our study. First, the diagnoses of constipation and AD were entirely dependent on the ICD-9 codes in the administrative dataset. We did not perform a review of the patients’ medical records so it was not possible to verify the accuracy of diagnoses, and therefore there may have been some misclassification. It is worth noting, however, that any misclassifications were more likely to be random, and associations are often underestimated rather than overestimated. In addition, clinical judgment might have varied among clinicians, and so diagnoses may not have been consistent, which might have influenced their validity. However, Taiwan’s NHI administration monitors the accuracy of the claims data to prevent violations. Second, the NHIRD does not contain data on covariates, such as social adversity, personal lifestyle, family history, laboratory data and environmental factors. Although we adjusted for various comorbidities and matched propensity scores, these unmeasured confounding factors could have affected our results. Third, it remains unclear as to whether the findings of our study may be extrapolated to other ethnic groups, as the majority of our patients were Taiwanese. Further clinical studies should include other ethnicities and nationalities to determine the generalizability of the associations observed herein.