Abstract
Objectives
The objectives of the study were to determine the prevalence of chronic rhinosinusitis (CRS) overall and its 2 phenotypic variants, CRS with and without polyposis (NP), in patients with chronic inflammatory comorbidities including autoimmune disorders, inflammatory bowel disease, and atopic dermatitis. These findings were compared with data in patients with asthma. Patients with hypertension were also used as a reference group to estimate the incidence of CRS in a group with regular medical follow-up.
Study design
A retrospective, cross-sectional query of a large tertiary care electronic medical record database was performed.
Results
Electronic medical record database prevalence of CRS in patients with hypertension was 4.4%. The prevalence of CRS was 18% in asthma ( P < .0001), 7% in atopic dermatitis, 3.5% in inflammatory bowel disease, and ranged from 1.4% to 5.9% in autoimmune disorders. The frequency of CRS patients exhibiting the NP phenotype was similarly low in patients with autoimmune disease and hypertension, but was significantly greater in patients with asthma ( P < .0001), inflammatory bowel disease ( P = .033), and atopic dermatitis ( P = .049),
Conclusions
These findings suggest similar prevalence of overall CRS in patients with autoimmune disease and inflammatory bowel disease, and background rates as estimated by observations in hypertension patients. Inflammatory bowel disease and atopic dermatitis patients with CRS exhibit some skewing toward the NP phenotype, as do asthmatics, where this association is well known.
1
Introduction
Chronic rhinosinusitis (CRS) is a heterogeneous disorder with multiple phenotypic presentations. Broadly, the disease is often subcategorized as either CRS with nasal polyps (CRSwNP) or CRS without nasal polyps (CRSsNP). Although this may reflect an oversimplification of a heterogeneous disorder, the distinction is convenient for clinical purposes; and pathophysiologic differences between these entities have been described. Specifically, CRSwNP, at least in Western populations, is characterized by a skewed T-helper 2 (Th2) responses and tissue eosinophilia; CRSsNP is more likely to exhibit T-helper 1 (Th1) responses and a predominantly neutrophilic inflammation .
The significant concordance between CRS and asthma is well known, and studies have demonstrated as high as 70% of asthmatics to have symptoms of CRS . Chronic rhinosinusitis also appears to be more frequent, and severe, in those with severe asthma . Furthermore, this association may be of even greater significance for patients with the CRSwNP phenotype ; and previous studies from our institution revealed that the presence of comorbid asthma was the most significant factor associated with the presence of NP in CRS patients . Taken together, these observations suggest that Th2-mediated disorders affecting the airway tend to coexist, in support of a unified airway concept .
Other investigators have examined the hypothesis that the presence of comorbid autoimmune mediated conditions may also impact the prevalence of CRS subtypes . Specifically, disorders associated with a bias toward Th1 immunity (eg, multiple sclerosis, psoriasis, rheumatoid arthritis) have been proposed to skew comorbidities away from manifestations conventionally attributed to Th2-mediated pathways . Given the presence of a Th1 autoimmune disorder, this hypothesis would predict a higher prevalence of CRSsNP and a lower prevalence of CRSwNP than observed in the general population. We sought to test this hypothesis by investigation of a large tertiary care electronic medical record (EMR) database to determine the frequency CRS overall and its 2 phenotypic variants (CRSsNP and CRSwNP) in patients with various autoimmune disorders.
Unlike rheumatologic conditions, CRS affects a barrier membrane that separates host tissues from the external environment. This has led to the hypothesis that defects in this intrinsic epithelial barrier may predispose patients to CRS . Similar mechanisms have been proposed for inflammatory bowel disease and atopic dermatitis , as these disorders also affect tissues at the interface between the host and the external environment. Thus, we also sought to determine among patients in the Northwestern EMR the frequency of CRS overall and its 2 phenotypic variants (CRSsNP and CRSwNP) in patients with inflammatory bowel disease and atopic dermatitis.
The prevalence of CRS and its phenotypic variants was determined and compared among patients with autoimmune conditions, inflammatory bowel disease, atopic dermatitis, and asthma, and was compared with a reference group of patients with hypertension.
2
Methods
The EPIC EMR database of the Northwestern Medical Faculty Foundation, Chicago, IL, has enrolled a total of 1 970 695 patients between December 1, 1998, and December 1, 2008. Fields in the database are established to capture comorbid conditions in an ongoing problem list. These data are typically entered by the primary care physician before the referral but can also be entered at the time of the otolaryngology visit (eg, self-referred patient). This database was queried by the International Classification of Diseases, Ninth Revision ( ICD-9 ) code. For patients within each disease category, a query was run to determine the frequency of patients also diagnosed with CRSsNP (ICD-9 code in the 473 series, but no 471 code) or with CRSwNP (ICD-9 code 471 with or without a 473 series code). Because all analyses were conducted on aggregated deidentified data, this study was approved by the institutional review board at the Northwestern University Feinberg School of Medicine and was deemed exempt from patient consent and periodic review. The relative proportion of CRS patients manifesting the NP phenotype was also determined for each of the selected diseases. These findings were compared statistically by χ 2 analysis, and P < .05 was considered significant.
2
Methods
The EPIC EMR database of the Northwestern Medical Faculty Foundation, Chicago, IL, has enrolled a total of 1 970 695 patients between December 1, 1998, and December 1, 2008. Fields in the database are established to capture comorbid conditions in an ongoing problem list. These data are typically entered by the primary care physician before the referral but can also be entered at the time of the otolaryngology visit (eg, self-referred patient). This database was queried by the International Classification of Diseases, Ninth Revision ( ICD-9 ) code. For patients within each disease category, a query was run to determine the frequency of patients also diagnosed with CRSsNP (ICD-9 code in the 473 series, but no 471 code) or with CRSwNP (ICD-9 code 471 with or without a 473 series code). Because all analyses were conducted on aggregated deidentified data, this study was approved by the institutional review board at the Northwestern University Feinberg School of Medicine and was deemed exempt from patient consent and periodic review. The relative proportion of CRS patients manifesting the NP phenotype was also determined for each of the selected diseases. These findings were compared statistically by χ 2 analysis, and P < .05 was considered significant.
3
Results
Patients with hypertension were used as a control group, to which patients with chronic inflammatory conditions were compared. The prevalence of any form of CRS in patients with autoimmune disorders (multiple sclerosis, psoriasis, lupus, rheumatoid arthritis, anyklosing spondylitis) ranged from 1.4% to 5.9% ( Fig. 1 ). Among the groups of patients with autoimmune conditions, the frequency of CRS was significantly greater in those with either rheumatoid arthritis or ankylosing spondylitis than in patients with lupus, psoriasis, or multiple sclerosis ( P < .0001). The prevalence of CRS in autoimmune patients overall (3.9%) was similar to that observed in hypertension patients (4.4%). Patients with inflammatory bowel disease also exhibited CRS at a similar frequency (3.5%). In contrast, the prevalence of CRS in patients with either atopic dermatitis (7.0%) or asthma (18%) was significantly greater than that observed in patients with hypertension ( P < .0001 for both) ( Table 1 ).
