Comparison of the demographic, laboratory and clinical characteristics between ECRSwNP and nonECRSwNP
Thirty CRSwNP patients (62.5%) were classified into the eosinophilic group, and eighteen CRSwNP patients (37.5%) were classified into non-eosinophilic group. Significant differences were noted in the levels of tissue and blood eosinophils (all P < 0.001), tissue neutrophils (P = 0.003), olfaction scores (P = 0.002), total VAS scores (P = 0.021), E/M ratio (P = 0.003), and endoscopy scores (P = 0.003) between the two groups (Table 1). Then, ECRSwNP patients were more likely to have asthma (P = 0.003), allergic rhinitis (P = 0.044) and atopy (P = 0.034) (Table 1), as well as higher ALOX15 mRNA levels in nasal polyp tissues (P < 0.001) (Fig. 1). However, the two groups were not significantly different with other characteristics (shown in Table 1).
Table 1 Baseline characteristics of patients
|
ECRSwNP
(n=30)
|
NonECRSwNP
(n=18)
|
P-value
|
Demographic characteristics
|
Sex, male/female
|
17/13
|
11/7
|
0.762
|
Age, y (mean ± SD)
|
45.03±11.86
|
46.00±15.77
|
0.810
|
Previous surgery, n (%)
|
10 (33.33)
|
7 (38.89)
|
0.697
|
Asthma, n (%)
|
16 (53.33)
|
2 (11.11)
|
0.003
|
Allergic rhinitis, n (%)
|
9 (30.00)
|
1 (5.56)
|
0.044
|
Atopy, n (%)
|
16 (53.33)
|
4 (22.22)
|
0.034
|
Histopathological findings (H&E staining)
|
Tissue eosinophils (%), median (IQR)
|
57.27 (41.86-68.43)
|
4.19 (1.97-6.68)
|
<0.001
|
Tissue neutrophils (%),
median (IQR)
|
0.00 (0.00-2.18)
|
6.84 (2.53-13.39)
|
0.003
|
Self-assessed symptom scores (VAS)
|
Nasal obstruction score, median (IQR)
|
7.00 (6.00-8.00)
|
6.00 (6.00-8.00)
|
0.572
|
Rhinorrhea score,
median (IQR)
|
6.00 (5.00-7.00)
|
5.50 (0.00-6.00)
|
0.078
|
Olfaction score,
median (IQR)
|
9.00 (6.00-10.00)
|
6.00 (4.00-8.00)
|
0.002
|
Facial pain or headache score, median (IQR)
|
0.00 (0.00-5.00)
|
0.00 (0.00-5.25)
|
0.581
|
Total VAS score,
median (IQR)
|
22.50 (19.00-25.50)
|
19.50 (16.75-22.25)
|
0.021
|
Imaging examination scores
|
CT score, median (IQR)
|
17.50 (14.00-22.00)
|
17.50 (11.00-20.25)
|
0.326
|
E/M ratio, median (IQR)
|
2.33 (2.00-3.50)
|
2.00 (2.00-2.00)
|
0.003
|
Endoscopy score, median (IQR)
|
5.50 (4.00-7.00)
|
4.00 (2.75-5.00)
|
0.003
|
Routine blood tests
|
Blood eosinophils (%), median (IQR)
|
6.25 (5.05-9.20)
|
1.60 (0.78-4.95)
|
<0.001
|
Blood neutrophils (%) , median (IQR)
|
53.45 (47.03-56.68)
|
58.50 (48.18-69.75)
|
0.058
|
ECRSwNP eosinophilic chronic rhinosinusitis with nasal polyps, nonECRSwNP non-eosinophilic chronic rhinosinusitis with nasal polyps, y years, SD standard deviation, n number, H&E hematoxylin and eosin, IQR interquartile range, VAS visual analogue scale, CT computed tomography, E/M ethmoid/maxillary sinus. Bold values are significant at P < 0.05.
Associations between ALOX15 level and laboratory and clinical characteristics
The level of ALOX15 mRNA was significantly correlated with tissue eosinophils (r = 0.565, P < 0.001), olfaction scores (r = 0.400, P = 0.005), total VAS scores (r = 0.383, P = 0.007), E/M ratio (r = 0.463, P = 0.001), endoscopy scores (r = 0.409, P = 0.004), and blood eosinophils (r = 0.395, P = 0.006); but not with age, tissue neutrophils, nasal obstruction scores, rhinorrhea scores, facial pain or headache scores, CT scores, or blood neutrophils (shown in Table 2).
Table 2 Spearman correlations between ALOX15 mRNA level and laboratory and clinical characteristics
Correlation
|
|
r-value
|
P-value
|
ALOX15 mRNA level
(-ΔCt value)
|
Age
|
-0.116
|
0.433
|
Tissue eosinophils (%)
|
0.565
|
<0.001
|
Tissue neutrophils (%)
|
-0.167
|
0.256
|
Nasal obstruction score
|
0.149
|
0.312
|
Rhinorrhea score
|
0.225
|
0.124
|
Olfaction score
|
0.400
|
0.005
|
Facial pain or headache score
|
0.056
|
0.706
|
Total VAS score
|
0.383
|
0.007
|
CT score
|
0.150
|
0.309
|
E/M ratio
|
0.463
|
0.001
|
Endoscopy score
|
0.409
|
0.004
|
Blood eosinophils (%)
|
0.395
|
0.006
|
|
Blood neutrophils (%)
|
-0.159
|
0.280
|
ALOX15 arachidonate 15-lipoxygenase, VAS visual analogue scale, CT computed tomography, E/M ethmoid/maxillary sinus. Bold values are significant at P < 0.05.
Assessment of the predictive value of the ALOX15 mRNA for ECRSwNP
To determine the predictive value of the ALOX15 mRNA for ECRSwNP, binary logistic regression analysis was conducted. Blood eosinophils, which has been shown to predict ECRSwNP previously [18], has also been introduced. The regression indicated that both ALOX15 mRNA (OR = 4.440, 95%CI = 1.614 - 12.213, P = 0.004) and blood eosinophils (OR = 1.457, 95%CI = 1.024 - 2.074, P = 0.036) showed the potential in predicting ECRSwNP (Table 3), and were thus, further subjected to ROC analysis.
Table 3 Binary logistic regression analysis for potential factors predicting ECRSwNP
Factors
|
OR
|
95% CI
|
P-value
|
ALOX15 (-ΔCt value)
|
4.440
|
1.614 - 12.213
|
0.004
|
Blood eosinophils (%)
|
1.457
|
1. 024 - 2.074
|
0.036
|
ECRSwNP eosinophilic chronic rhinosinusitis with nasal polyps, OR odds ratio, CI confidence interval, ALOX15 arachidonate 15-lipoxygenase.
ROC curve analysis showed that ALOX15 mRNA had a high accuracy, and blood eosinophils a moderate accuracy as predictors of ECRSwNP (AUC = 0.909, 95% CI = 0.828 - 0.990, P < 0.001; and AUC = 0.820, 95% CI = 0.687 - 0.954, P < 0.001, respectively) (Fig. 2). A maximal Youden index value of 0.700 demonstrated a cut-off value of -2.113 for ALOX15 to predict the diagnosis of ECRSwNP with a sensitivity of 83.3% and a specificity of 86.7% (Fig. 2). Similarly, a maximal Youden index value of 0.656 demonstrated a cut-off point of 3.45% for blood eosinophils in predicting the diagnosis of ECRSwNP with a sensitivity of 72.2% and a specificity of 93.3% (Fig. 2).
In view of the slightly lower accuracy and sensitivity of blood eosinophils in predicting ECRSwNP, we speculated that a combination of ALOX15 and blood eosinophils might improve the value of blood eosinophils in predicting ECRSwNP. A model comprising these combined predictors, derived from logistic regression analysis, was thus established as follows:
Model = ALOX15 mRNA level (-ΔCt value) + blood eosinophils (%) × 37.7/1.491
ROC curve analysis of the combined data for ALOX15 and blood eosinophils indicated an increased AUC of 0.933; with a cut-off value of -0.778 and sensitivity of 83.3% and specificity of 90.0%; which were improvements over the accuracy and sensitivity of blood eosinophils, as well as the specificity of ALOX15 mRNA. Although the AUC for the combination of ALOX15 and blood eosinophils was not significantly different from the AUC of ALOX15 mRNA (P = 0.363) alone, this was statistically different from the AUC of blood eosinophils (P = 0.028); indicating a higher value of the combination of ALOX15 mRNA and blood eosinophils in predicting ECRSwNP, than the value of blood eosinophils alone.
Comparisons of the laboratory and clinical characteristics between the high- and the low-combination level group
Based on the optimal cut-off value of -0.778 for the combination of ALOX15 mRNA and blood eosinophils, the CRSwNP patients were divided into a high combination level group (values ≥ -0.778, n = 30) and a low combination level group (values < -0.778, n = 18). The occurrence of ECRSwNP was significantly higher in the high combination level group (n = 27) than in the low combination level group (n = 3) (P < 0.001) (Fig. 3a).
Comparison of the demographic and clinical characteristics of the two groups demonstrated that the combined level was not significantly associated with the distribution of sex, age, previous surgery (Fig. 3b-d). However, patients in the high combination level group were more likely to suffer from asthma (P = 0.021), allergic rhinitis (P = 0.044), and atopy (P = 0.020) (Fig. 3e-g). Moreover, patients with high combination levels showed significantly higher ALOX15 mRNA levels and higher eosinophil, but not neutrophil infiltration, in both polyp tissue and peripheral blood than patients with low combination levels (all P < 0.001) (Fig. 4). Assessment of self-assessed symptom scores demonstrated that rhinorrhea scores (P = 0.042), olfaction scores (P = 0.007), and total VAS scores (P = 0.003) were significantly higher in patients with high combination levels (Fig. 5). Similarly, analysis of imaging examination scores indicated that E/M ratio (P = 0.032) and endoscopy scores (P = 0.001), but not CT scores, were significantly higher in patients with high combination levels (Fig. 6).