This retrospective study was conducted between May 2014 and December 2019. One hundred and fifteen consecutive patients who were diagnosed with ECRS and underwent bilateral ESS at Faculty of Medicine, Kagawa University were enrolled in the retrospective study. Eighty adult patients who were followed for at least 1-year were included in the present study. Almost all excluded patients did not reach the 1-year observation period because they were followed-up at another hospital, and there was no patient who received revision surgery or biologics within 1-year after surgery. A definitive diagnosis of ECRS was reached according to the JESREC diagnostic criteria based on a JESREC score of higher than 11 and the presence of more than 70 eosinophils per high-power field in an average of three eosinophil-rich regions.1 Exclusion criteria were as follows: age younger than 20-years and patients who were using continuous systemic steroids or biologics at the time of surgery. This study design was approved by the Institutional Review Board, Faculty of Medicine, Kagawa University (approval nº 2020-201).

Bilateral ESS was performed on all patients under general anesthesia. ESS was performed based on the concept of complete resection of the sinus wall and complete opening of all sinuses to make a single sinus (full-house ESS).8 If polyps or edematous mucosa were present in the Olfactory Cleft (OC), they were shaved with a microdebrider. Septoplasty, submucosal inferior turbinectomy, and bilateral or unilateral submucosal middle turbinectomy were performed where indicated.9 Nasal packing was inserted into the middle meatus and an absorptive gelatin sponge was placed at the OC if the surgical treatment was performed there. All surgeries were performed by two otolaryngologists (the first and second authors). Oral Prednisolone (PSL) at a dose of 15–25 mg per day (0.3 mg/kg as a guide) with gradual tapering was administered from several days before to 2–3 weeks after surgery.10 Low-dose long-term clarithromycin macrolide therapy was continued for 3-months after surgery. All patients were followed up at 1- to 4-week intervals until 3-months after surgery to frequently remove secretions, blood crusts, and fibrin under endoscopy, and continuous long-term follow-up was performed at intervals of 1–3 months thereafter. A topical steroid spray was prescribed before and after surgery, and daily saline irrigation was performed by patients. Short bursts of oral PSL therapy (10–20 mg/day) for several days or nasal drop of betamethasone sodium for 1- to 2-weeks were added in the event of acute exacerbation. Antibiotics (amoxicillin or levofloxacin) were administered for a week when bacterial infection with purulent nasal discharge developed.

JESREC scores were assessed according to the JESREC scoring criteria based on individual CT, nasal findings, and blood sample tests (maximum score, 17).1 Patients underwent blood sampling tests, including allergen specific IgE (house dust mite, pollens, and fungi) at baseline, and patients who had specific symptoms and positive antigen tests (more than 0.7 UA/mL) were diagnosed with allergic rhinitis. The diagnosis of Bronchial Asthma (BA) was made by experienced pulmonologists based on bronchial symptoms and some lung function tests. Patients with a history of respiratory exacerbation after receiving Non-Steroidal Anti-Inflammatory Drugs (NSAIDs) were considered to have NSAID sensitivity. Computed Tomography (CT) was performed preoperatively and 3-months after ESS, and was scored according to the Lund-Mackay CT scoring system (score range between 0 and 24).11,12 The results of failure to complete removal of ethmoid cells were measured on 3-month CT by the Residual Ethmoid Cells (REC) score. The scoring was performed according to a previous report as follows: bilateral ethmoid sinuses were each divided into superior-anterior, inferior-anterior, and posterior areas, and the RECs in each area were assessed using a 3-graded scoring system (0, absence of any lamina or <4 mm in thickness; 1, presence of lamina ≥4 mm, but there is no continuity with other lamina; 2, at least 2 laminas exhibit continuity but they form incomplete cells; 3, complete cells remain), and the total bilateral possible score was 18.6 Endoscopic scoring was performed using a E scoring system with slight modification.7 Each sinus and OC were rated 0–2 (0, normal; 1, partially diseased; 2, completely closed or unobservable), and the total bilateral E score ranged from 0 to 2 4. The score was measured using a 4-mm, 30 ° rigid endoscope. Endoscopy procedures were repeated at every follow-up visit, and all exams were video recorded and saved on the server. The percentage of the total score to the maximum possible score for operated sinuses was rated as the Escore in the original method; however, we simply rated it as points in this study because all sinuses were bilaterally opened in all patients. The Escore at 3-months (3m-Escore), 6-months (6m-Escore), 1 year (1y-Escore), and subsequent years (±2-months) up to 5-years after ESS (2y- to 5y-Escore) were adopted in this study. When patients required continuous systemic steroids or biologics due to the exacerbation of upper/lower respiratory disease, revision surgery, or development of another new entry disease, their evaluations were stopped, and all subsequent observations were excluded from the study. Patients in whom the mucosal condition was poor, it was difficult to observe the middle meatus, and there was no improvement noted at the time of final observation were judged as failure to achieve endoscopic control of the disease (uncontrolled group).

The Escore system is a relatively newly proposed endoscopic scoring system, and its correlation with 2 established endoscopic scoring systems, Modified Lund-Kennedy endoscopic (MLK) score and the Perioperative Sinus Endoscopy (POSE) score, was confirmed. The MLK was measured for edema (0; absent, 1; mild, 2; severe), polyps (0; no polyps, 1; polyps in middle meatus, 2; beyond middle meatus), and discharge (0; no, 1; clear, thin discharge, 2; thick, purulent discharge).13 The scarring and crusting subscores were subtracted from the original LK score, and the total bilateral possible score was 12. The POSE score was rated based on the degree of 10 parameters (each 0–2 points: middle turbinate, middle meatus, maxillary sinus contents, ethmoid cavity; crusting, mucosal edema, polypoid change, polyposis, and secretions, frontal recess/sinus, and sphenoid sinus) for a bilateral total possible score of 40.14

All statistical analyses were performed with EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan). Differences between the 3m-Escore with the other observation points were assessed using the paired t-test. Correlations between the Escore at the final observation point (f-Escore) and 3m-Escore were evaluated by Spearman’s r-test. The significance of differences between the uncontrolled group with each dependent variable was screened by univariate analysis, and subsequently the potential predictors (p < 0.1 in the univariate analysis) were included in the multivariate regression test and narrowed down by backward stepwise selection to identify independent factors. To derive cut-off values of the 3m-Escore for the prediction of endoscopic long-term outcome, we constructed a ROC curve. Differences in patient characteristics between the groups were assessed using the Student’s t-test or Fisher’s exact test. Spearman’s r-test was performed to determine potential differences and correlation coefficients between the Escore and other endoscopic scoring systems. Significance was assumed when p < 0.05. All statistical studies were supervised by a statistician.

The baseline characteristics are described in Table 1. All patients were observed for longer than 1 year, and subsequent endoscopy evaluations were available for 63, 43, 26, and 13 patients at 2-, 3-, 4-, and 5-years, respectively. Twenty patients were excluded halfway through their observation periods: 7 received biologics, 4 required continued systemic steroid use (3 patients for EGPA and 1 patient for uncontrolled BA), 3 underwent revision surgery, 3 changed their follow-up hospital, and 3 withdrew. The number of patients who required short-burst steroids or antibiotics at least once during their follow-up was 28/80 (35%) and 29/80 (36.5%), respectively. Their average usage counts were 1.8 ± 1.9 times/year and 0.5 ± 0.3 times/years, respectively. All individual Escore changes from 3-months to 5-years postoperatively are described in Fig. 1A. Overall 3m, 6m, 1y, 2y, 3y, 4y, and 5y-Escores were 8.5 ± 4.1, 8.9 ± 5.0, 9.2 ± 5.5, 8.8 ± 5.1, 8.3 ± 5.8, 7.8 ± 3.3, and 7.4 ± 5.2, respectively (Fig. 1B). There was no significant difference between time periods. The f-Escore in 80 patients (median observation period was 3-years) slightly increased from 3-months (mean percent change of score 114 ± 74%), but no significant differences were observed compared with the 3m-Escore (9.2 ± 5.6 vs. 8.5 ± 4.1, p = 0.363). Typical endoscopic findings are shown in Fig. 2.

Figure 1.

(A) Changes in individual of Escore from 3-months to 5-years postoperatively in 80-patients. (B) Mean change in the Escore from 3-months to 5-years. There was no significant difference between time periods. Error bar shows standard deviation (m, month; y, year).

(0.26MB).

Figure 2.

Serial endoscopic records of the nasal cavity in typical cases. Well controlled case: (A) 3-months after surgery, right Escore is 4 (bilateral 7); (B) After 5-years, right Escore is 3 (bilateral 6). Uncontrolled case: (C) 3-months after surgery, right Escore is 8 (bilateral 16); (B) After 4-years, right Escore is 10 (bilateral 21). (◼; middle turbinate, *; nasal septum).

(0.41MB).

Twenty-one patients (21/80, 26.3%) were considered to have endoscopically uncontrolled ECRS at their final observation points. Subgroup analyses comparing the endoscopically uncontrolled and controlled groups are describe in Table 2. As baseline factors, the presence of NSAID sensitivity and CT score were significantly different, and some long-term evaluation factors, including postoperative CT score, prednisolone usage rate, 3m-Escore, and f-Escore, were significantly higher in the uncontrolled group. Gradual deterioration of the average Escore and a significant difference comparing the 3m-Escore with the f-Escore were observed only in the uncontrolled group (Fig. 3). Multivariate logistic regression analysis was performed to identify independent predictors of a poor endoscopic outcome in the long-term course. Some cofactors (Table 1, REC score, and 3m-Escore) were preliminarily screened for univariate significance. The potential predictors (p < 0.1 in the univariate analysis) were included in the multivariate logistic regression analysis and narrowed down by backward stepwise selection. Only the 3m-Escore was significant, with an adjusted odds ratios of 1.72 (95% CI: 1.30–2.27; p < 0.01). Therefore, the 3m-Escore was considered to be an independent predictive factor for endoscopic long-term outcome, and it had a strong correlation with the f-Escore (Spearman's rank correlation coefficient = 0.714, p < 0.01), (Fig. 4A). To determine the cut-off value for the prediction of endoscopic uncontrolled ECRS, the ROC curve was plotted (Fig. 4B). Lastly, the cut-off value of the 3m-Escore was defined as 12 (area under the curve = 0.894, 95% CI: 0.8–0.988), and the sensitivity and specificity were 75% and 96.7%, respectively. In the present study, 17 patients had a 3m-Escore of 12 or higher and the other 63 patient’s scores were less than 12. Sixteen/seventeen (94.1%) patients whose 3m-Escore was ≥12 had poor endoscopic control of the disease. In contrast, only 5/63 (7.9%) patients with a score <12 was endoscopically uncontrolled at the final observation point.

Figure 3.

(A) Mean change in the Escore by group. (B) Comparing the mean change in 3m-Escore at the final observation point. Error bar shows the standard deviation. E, Endoscopically; m, month; y, year. *Significantly different from 3m-Escore by paired t test (p < 0.05).

(0.16MB).

Figure 4.

(A) Correlation diagram between 3m-Escore and f-Escore (Spearman's rank correlation coefficient = 0.714, p < 0.01). (B) Receiver Operating Characteristics (ROC) curve of 3m-Escore and uncontrolled. The cut-off value of 3m-Escore was defined as 12 (area under the curve = 0.894, 95% CI: 0.8–0.988). 3m-Escore, Escore at 3-months; f-Escore, Escore at the final observation point.

(0.15MB).

In the 80 patients, the average MLK scores were 3.8 ± 1.9 and 4.7 ± 2.7 at 3-months and final observation point, respectively, and average POSE scores were 10.4 ± 6.2 and 12.3 ± 8.4 at 3-months and the final observation point, respectively. The Pearson correlation coefficients demonstrated a strong positive correlation between the Escore and 2 established endoscopic scoring systems (Table 3).