Eosinophilic esophagitis assessment correlates with clinical, molecular features
Click Here to Manage Email Alerts
Key takeaways:
- Index of Severity for Eosinophilic Esophagitis scores clustered under 10 and at 15 and higher.
- Children had more severe symptoms and complications scores and adults had more severe fibrostenotic features.
Researchers found associations between Index of Severity for Eosinophilic Esophagitis scores and select clinical features across categories of disease severity, according to a study published in The Journal of Allergy and Clinical Immunology.
Scores also were associated with molecular features of non-severe categories of EoE, Evan S. Dellon, MD, MPH, director of the Center for Esophageal Diseases at the UNC School of Medicine, and colleagues wrote.
Proposed in 2022 by an international workgroup, the Index of Severity for Eosinophilic Esophagitis (I-SEE) comprises inactive (< 1 point), mild (1-6 points), moderate (7-14 points) and severe (15-78 points) categories.
“Many diseases, asthma being a prominent example, have linked clear ways to categorize severity with treatment and monitoring recommendations,” Dellon told Healio. “I-SEE is the first step towards this goal in EoE.”
However, Dellon continued, I-SEE still needs to be validated and correlated with other measures of severity in EoE. It also had yet to be assessed in the context of molecular severity, as measured by changes in esophageal tissue gene expression.
“This formed the basis for this study,” Dellon said.
Study design, results
The study included 209 adults (median age, 37.6 years; 37.3% women; 94.7 white) and 109 children (median age, 11.8 years; 28.4% girls; 89% white) with complete I-SEE and EoE diagnostic panel (EDP) data.
Median I-SEE scores included 7 for the total cohort (interquartile range [IQR], 4-18) and adult (IQR, 4.5-18) and child (IQR, 4-19) groups alike.
The researchers also noted a primary cluster with scores that ranged from 0 to 10 (median, 5) for patients in the inactive, mild and moderate categories and a secondary cluster with scores of 15 and higher (median, 20) for patients with severe disease.
Severity proportions included 4.4% of patients with inactive disease, 41.8% of patients with mild disease, 19.5% of patients with moderate disease and 34.3% of patients with severe disease.
The researchers additionally noted that the “symptoms and complications” domain in the I-SEE had a primary cluster between 0 and 5 with a secondary cluster at 15. Fibrostenotic features had a primary cluster between 0 and 4 with a secondary cluster at 15.
Median scores included 1 for symptoms and complications and 2 for inflammatory features and fibrostenotic features, which the researchers did not consider significantly different.
Proportions of patients with severe EoE included 33.5% of adults and 35.8% of children, which the researchers also called similar, although 33% of children and 5.3% of adults had severe symptom and complications scores (P < .001).
However, 29.2% of adults and 2.8% of children had severe fibrostenotic feature scores (P < .001). Adults had higher endoscopic and histologic inflammatory features as well, the researchers added (P < .001).
The severity of I-SEE categories also increased with duration of disease, including 8.2 years for patients with inactive disease and 12.1 years for patients with severe disease (P = .003).
The researchers further noted a mild inverse correlation between total I-SEE and EDP scores (r = –0.352; P < .001), adding that higher EDP scores indicate molecular features that are less severe.
Additionally, there were inverse correlations between EDP score and inflammatory (r = –0.665; P < .001) and fibrostenotic (r = –0.446; P < .001) features, the researchers said, but not between EDP score and symptoms and complications (r = –0.047).
Stratification into the adult and child age groups yielded similar results, the researchers said, although adults had significantly higher correlations between fibrostenotic features and EDP scores than children did.
There was no significant correlation between EDP score and the symptoms and complications domain, the researchers continued, even among patients who had never had an esophageal dilation.
Based on these findings, the researchers said that molecular profile may have a closer alignment with histologic and endoscopic features than clinical features, indicating that molecular biomarkers may play a role in disease pathogenesis beyond symptom reporting.
In all nonsevere categories, median EDP scores fell as total I-SEE scores increased in severity, including 363 (IQR, 322-379) in the inactive group, 312 (IQR, 208-377) in the mild group and 96 (IQR, 23-232) for the moderate group.
The higher EDP scores in the severe category (237; IQR, 109-326) compared with the moderate category were unexpected, the researchers said, so they compared patients in the first (n = 28) and fourth (n = 27) quartiles of EDP scores.
Patients with severe disease and low EDP scores had higher inflammatory features (P < .001) and fibrostenotic features (P < .001) scores than the patients with severe disease and high EDP scores, but there were no significant differences between symptoms and complications scores.
Correlations were moderate for total I-SEE score and EDP scores in the primary and secondary clusters of the non-severe (r = 0.572; P < .001) and severe (r = 0.596; P < .001) categories as well.
Considering these results, the researchers said there was a correlation between EDP scores and the sum score of I-SEE features that were assigned 1, 2 and 4 points regardless of whether I-SEE features assigned 15 points were present, further indicating a correlation between the EDP and severity categories between inactive and moderate.
The researchers additionally found positive correlations between 58 EDP genes and total I-SEE score (P < .001). Although it was mild, the researchers continued, PMCH had the highest positive correlation (r = 0.358; P < .001).
Among 20 genes with a negative correlation with total I-SEE score (P < .001), ENDOU’s was highest, even though it was slight (r –0.294; P < .001), the researchers said. There were no genes that correlated with symptoms and complications scores. Inflammatory features correlated with 87 genes, and fibrostenotic features correlated with 85 genes.
Directionality was the same among the genes that correlated with both inflammatory and fibrostenotic features. There were 63 genes that the researchers said correlated well with inflammatory features (absolute r > 0.4) and six genes that correlated well with fibrostenotic features (absolute r < 0.4).
CTSC, CA2, TNFAIP6, GLDC and PMCH all had positive correlations and overlapped between these groups, the researchers continued, adding that TSPAN12 was the only other gene that correlated well with fibrostenotic features. Its r value of –0.406 (P < .001) represented the highest negative correlation with fibrostenotic features.
The 138 patients with two visits for the Outcome Measures in Eosinophilic Gastrointestinal Disease Across the Ages, or OMEGA, study enabled the researchers to conduct a longitudinal analysis of the modified I-SEE at a median of 15.1 months, including 73 patients with active EoE and 65 with inactive EoE at baseline.
Modified I-SEE scores stratified by histologic status yielded 31 cases of response, 42 cases of nonresponse, 25 cases of relapse and 40 cases of sustained remission. The decrease in response from 6 to 3 (P < .001) and the increase in cases of relapse from 4 to 7 (P < .001) were both significant, the researchers said.
Cases of nonresponse and sustained remission did not see any significant changes, the researchers continued, nor were there any significant differences in symptom scores. Results for inflammatory and fibrostenotic features resembled the total modified I-SEE scores, the researchers added.
With 25 cases of modified I-SEE response and six cases of modified I-SEE nonresponse among paired pre-molecular and post-molecular profiles, there was a significant increase in median EDP score from 24 to 184 (P < .001) in modified I-SEE response. In the nonmodified I-SEE response, median EDP score fell from 85 to 79.
The increase in EDP scores was more pronounced in the modified I-SEE response than in the modified I-SEE nonresponse (185 vs. 57; P = .041), the researchers said, as well as in the inflammatory (212 vs. 76; P < .001) and fibrostenotic (190 vs. 134; P = .029) features but not in symptoms (185 vs. 162.3).
Conclusions, next steps
Overall, the researchers said these findings indicate the I-SEE’s inflammatory and fibrostenotic feature and total score components correlated with the molecular profiles of patients with inactive, mild and moderate EoE.
Similarly, the researchers concluded, molecular activity was reflected by the longitudinal changes of the modified I-SEE and the inflammatory and fibrostenotic feature scores, with further validation and study warranted.
“It was very interesting that the total I-SEE score, a clinical summary metric, correlated with molecular severity in the tissue, and particularly with the I-SEE inflammatory and fibrostenotic domains, but not with the symptoms/complications domain,” Dellon said.
Dellon added that this suggests that the molecular severity, which is not a component of the I-SEE metric, does relate to overall severity, though mostly with the biologic metrics rather than symptoms.
And while these results help to provide validity to the components that make up I-SEE, Dellon said that it is a bit premature for this paper to lead to improved care.
“The next steps in research related to I-SEE include continued validation and prospective study to link severity to treatment outcomes, so treatment patterns can be more standardized,” he said.
For more information:
Evan S. Dellon, MD, MPH, can be reached at evan_dellon@med.unc.edu.