Materials and methods
This single-arm, longitudinal, single center cohort study assesses the impact of Dupilumab therapy on all patients with CRSwNP treated with Dupilumab at University Medical Center Hamburg-Eppendorf in Germany from November 2019 to May 2023. Informed consent was obtained from each participant. The study was approved by the Ethics Committee of the University of Hamburg (2020-10264-BO-ff).
All patients diagnosed with severe, uncontrolled CRSwNP according to EPOS criteria with subsequent guideline-based prescription of Dupilumab 300 mg s.c. every 2 weeks were included [
2]. The EPOS guidelines define CRSwNP as uncontrolled when typical therapy and surgical therapy do not provide sufficient symptom relief. Additionally, patients must have met the following criteria to receive Dupilumab: significant loss of olfaction, significant reduction in quality of life (SNOT-22 score > 40 points), presence of bronchial asthma, or evidence of type-2 inflammation (≥ 10 eosinophils per high-power field in tissue samples or ≥ 100 kU/L of total serum IgE or ≥ 250 eosinophils per μL in blood samples). The need for systemic glucocorticoids at least twice a year was considered but not mandatory. Patients with cystic fibrosis, immunosuppressive diseases, or those receiving treatment with other biologics were excluded from the study, as well as pregnant and breastfeeding participants.
The baseline assessment of outcome parameters was established before the first medically monitored administration of Dupilumab. Subsequentially, follow-up visits were held with a fixed schedule after 4 weeks and then every 12 weeks at month 4, 7, 10 and 13 after the baseline visit. Thereafter, depending on the success of the therapy, treatment was either continued in the outpatient setting or followed by semi-annually check-ups. All patients who had at least one follow-up visit at month 4 were included in the analysis. Data was collected up to a maximum of 22 months.
Study outcome parameters
Demographic data, body mass index (BMI), the Lund-Mackay computed tomography score (LMK), allergy status, presence of bronchial asthma, N-ERD status using the diagnostic criteria of the EACCI position paper [
8], and number of previous sinusitis surgeries were recorded during the baseline visit. The baseline and each subsequent visit included the collection of blood parameters such as ECP (µg/L), absolute BECs (billion/L) and total serum IgE (IU/mL). At each visit, pulmonary function diagnostics was performed for evaluating forced expiratory volume in 1 s value (FEV-1). General and subjective sinusitis symptom severity such as nasal obstruction, rhinorrhea, facial pain, and sleep disorder were determined using the Visual Analog Scale (VAS). Olfactory function was screened using Sniffin-Sticks 12-identification test (Burghart Messtechnik GmbH) with the best performing results included. The NPS of each nasal cavity was determined according to the classification system of Gaevert [
9,
10]. The Asthma Control Test (ACT) [
11] and Sino-Nasal-Outcome Test-22 (SNOT-22) [
12] were used.
NPS, VAS-scores and SNOT-22 were set as primary endpoints. Biomarkers (serum IgE, BECs, ECP), FEV-1, ACT and olfactometry performance were set as secondary endpoints.
Statistics
The Statistical analyses were conducted using the statistical software R version 4.2.2 [
23]. Missing data was not imputed. Descriptive statistics included mean and standard deviation (SD) for continuous variables, and frequencies with percentages for categorical variables.
To capture linear trends, mixed linear models were used, treating time as the fixed effect and patient ID as the random effect. The target variable was the respective change from baseline, with the baseline variable as a covariate. Results encompass adjusted mean changes from baseline with 95% confidence intervals (CI) for each time point.
Correlation analyses utilized Pearson correlation coefficient. Considering clustered data (multiple measurements per patient), correlations were calculated accounting for within-patient variability. First, random intercept models were employed to capture within cluster variability as random effects. Then, the conditional mean of these effects was subtracted from empirical absolute outcome values. Pearson correlation coefficient was computed for each comparison. All analyses are performed in an explorative manner. No adjustment for multiple testing was carried out.
The patients were stratified into two groups based on their baseline biomarker levels to assess potential differential progression with respect to the outcome parameters during the course of therapy (threshold values: total serum IgE 100 IU/mL, BECs 0.5 and 0.3 billion/L, ECP 30 µg/L).
Discussion
Since its release, Dupilumab has extended available treatment options for patients with severe uncontrolled CRSwNP. Although there are a few real-world observational studies with 12-month follow-up, there is a need to evaluate the efficacy and safety of Dupilumab beyond 1 year. Long-term observation will reveal whether the effects may diminish over time, whether permanent remission can be achieved or whether side effects may occur. As these reports are limited to date, this study reports long-term follow-up incorporating a larger population and focusing on the analysis of easily obtainable biomarkers in clinical practice settings [
4,
24,
25].
Our findings confirm that treatment with Dupilumab results in rapid and consistent improvement throughout the treatment period for up to 22 months. This improvement is evident in the results of the primary endpoint variables SNOT-22 and VAS-scores, as subjective patient outcome measures, as well as in objective endoscopic NPS and olfactometry performance. It is noteworthy that a significant part of therapeutic efficacy manifests within the first 4 months, after which there is only a minor but continuous improvement up to 22 months, suggesting that continuous application of Dupilumab is effective for a longer treatment period. Especially the NPS and olfactometry appear to benefit particularly from prolonged therapy, reaching physiological levels not before month 22 in our observation (0.86 SD ± 0.90 and 9.67 SD ± 2.07). Side effects are only minor in the second year of therapy, and there are no side effects that have not already occurred after initiation of Dupilumab.
These findings increase the long-term understanding of efficacy and safety of Dupilumab therapy, making it possible to provide patients with information about expected effects and when to expect remission, which is important for expectation management and compliance. To further investigate in long-term observation of Dupilumab treatment in patients with severe CRSwNP, a national registry should be established, so that data from the outpatient setting can be included, as more patients can now be started on Dupilumab on an outpatient basis and fewer patients are available for long-term observation in centers. This is of particular interest because treatment with Dupilumab is currently cost-intensive, also considering a possibility of lifelong therapy, and the long-term recurrence of nasal polyps after achieving complete remission and subsequent discontinuation of treatment is still uncertain. With a NPS of 0 in combination with good SNOT and VAS-scores in 19 of our patients at their last documented observation, this might be the appropriate time to consider such an approach. A Japanese study with a small number of patients treated with Dupilumab for atopic dermatitis was able to show that more than half of the observed population achieved long-term remission after an attempted discontinuation of Dupilumab treatment [
26]. However, despite both diseases sharing the same pathophysiological basis, drawing conclusions to patients affected by CRSwNP might be incorrect. In contrast, the pivotal study suggests a need for continued suppression of type 2 inflammation to achieve long-term disease control [
19]. In our study, however, there was one patient who experienced significant subjective and objective symptom progression 3 months after discontinuation of Dupilumab, supporting the suggestion of continuous suppression.
Impairment of sense of smell is one of the most unpleasant and persistent symptoms in patients with CRSwNP [
27]. Studies revealed that loss of olfaction is associated with disease severity and significantly impacts quality of life [
28]. Our data align with these findings: SNOT-22- and VAS-scores of nasal blockage and rhinorrhea, as an expression of quality of life, exhibit an inverse correlation with the progression of olfactometry results (Fig.
2). With consistent improvement of olfactometry and NPS, along with a positive correlation, these data suggest a relationship between their development, whereas other studies have linked the change in olfactometry more to tissue eosinophilia-related neurotoxic effects rather than mechanical obstruction [
25]. Unfortunately, we were unable to investigate a correlation between tissue eosinophilia and olfactometry as tissue samples have not been collected for each patient and especially not at each timepoint of subsequent follow-up.
Dupilumab has demonstrated its efficacy in addressing several type 2 inflammatory diseases affecting the upper and lower airways at the same time [
29]. In CRSwNP patients with comorbid asthma, Dupilumab has shown significant improvements in lung function and asthma control [
29]. Consistent with those results, our study demonstrated mild enhancement in FEV-1 and ACT-values, which were found to be independent from baseline BECs.
Biomarkers have a crucial role in guiding therapeutic decisions and monitoring treatment response. However, biomarkers that serve as predictors for the response of Dupilumab therapy in CRSwNP are not existing to date [
21]. In this study however, we found that patients with pretherapeutic increased BECs with threshold value at 0.5 billion/L, but not 0.3 billion/L, had a trend of better progression of SNOT-22 score, with a difference of − 13.55 points after 19 months of therapy, which surpasses the 12-point threshold for a minimally clinically important difference (MCID) that was found for the SNOT-22 test in medically managed patients with chronic rhinosinusitis [
30]. The potential role of blood eosinophils and their initial cell count as a predictor of therapy response has been discussed. Some studies found better outcomes in NPS and nasal congestion score with elevated baseline cell counts treated with Mepolizumab and Benralizumab, though without statistical significance [
31‐
33], while others attested no further capability in identifying responsiveness [
19,
22]. Bertlich et al. found the initial BECs to have at least some influence on the development of variables, although no underlying systemic effect was assumed [
34], whereas others found higher baseline BECs to predict better outcomes in SNOT-22 scores [
35,
36]. Along with these, our findings provide further evidence that baseline BECs > 0.5 billion/L may be a suitable biomarker for predicting, at least subjective, responsiveness to Dupilumab in patients with severe CRSwNP. Nonetheless, the result must be seen in the context that only SNOT-22 and none of the other variables showed this result (Fig.
5). Temporary elevations of BECs in patients undergoing Dupilumab treatment have been previously reported in the SINUS trials and were observed to resolve spontaneously within a few months without causing significant symptoms [
19,
37,
38]. However, there have been reported cases of hypereosinophilic syndrome (HES) or eosinophilic granulomatosis with polyangiitis (EGPA) in some patients [
38]. Consistent with those observations, we saw a transient increase in cell counts after initiation of therapy, which subsequently regressed back to baseline levels during the course. A transient hypereosinophilia with a cell count exceeding 3 billion/L was observed in two patients. However, this condition was self-limiting and resolved within a period of 3 months, without causing any noticeable symptoms. Regardless of hypereosinophilia, both subjective and objective outcomes improved consistently. Studies investigating the management of this condition suggest that hypereosinophilia can be considered benign when the BEC is less than 3 billion/L and no involvement of organs is seen. However, when the cell counts exceeded 3 billion/L, short-term corticosteroid therapy is recommended to reduce the number of eosinophils in the blood and prevent potential organ involvement [
39]. The transient increase of eosinophil cell count can be attributed to the hypothesis of Dupilumab blocking eotaxin-3, a chemokine which specifically attracts eosinophils to the site of inflammation [
40]. Consequently, more eosinophil cells remain in peripheral blood. Nevertheless, the lack of response in BECs after initiating therapy, while at the same time significant improvements in NPS, SNOT-22, and VAS-scores were seen, indicates a discrepancy between BECs and its influence on disease severity. These findings suggest one the one hand, that BECs are not reflecting those clinical improvements and effective suppression of type 2 inflammation processes, and on the other, that suppressing BECs may not be required to achieve therapeutic success. Either way, their ability to be used as a biomarker for monitoring therapeutic success can be denied.
Consistent with previous trials showing a reduction in serum IgE levels in patients treated with Dupilumab for CRSwNP, total serum IgE levels decreased continuously until the end of the observation period in our study [
11,
19,
25]. Thereby, the progression of total serum IgE levels showed moderate correlations with all clinical outcome parameters, with exception of FEV-1 and BECs (Fig.
2). While improvements in clinical symptoms are paramount, these findings suggest that insights into the improvement of the symptoms and inflammation severity could be inferred through the biomarker IgE, making it a good follow-up parameter for therapy response. On the other hand, the study showed that the pretherapeutic IgE levels do not allow for any conclusions regarding the long-term course of subsequent therapy, although a better NPS response in the first year was found in patients with elevated IgE levels. Taking into account the better SNOT-22 response in patients with elevated BECs, one may conclude that the more pronounced the type 2 endotype, the better the treatment response in the first year.
ECP is a cytotoxic secretory protein with bactericidal and antiviral properties, produced and released by activated degranulating eosinophils and considered as meaningful marker of eosinophilic inflammation [
41]. The connection between ECP and activated eosinophils in the blood is also evident in our study,
r = 0.64; 95% CI [0.58, 0.69] [
42]. Higher levels of ECP have been reported in recurrent CRSwNP, potentially contributing to epithelial damage in nasal mucosa [
42]. Pre-interventional ECP levels were found to be a reliable prognostic indicator of polyp recurrence after surgery [
43]. Whether it can also predict the response to Dupilumab therapy has, to the best of our knowledge, not yet been investigated. In our study, we observed that the progression of the clinical variables was irrespective of the initial ECP levels. There were no clinically relevant changes in serum ECP levels besides a small transient increase. Mean ECP levels remained elevated from baseline throughout the entire observation period. No correlations were found to the clinical parameters. In conclusion, these observations suggest that serum ECP cannot be used to predict or monitor treatment response. Instead, the ECP levels in nasal secretions may serve as a more suitable biomarker to monitor the success of Dupilumab therapy, since it was shown to decrease under therapy before [
11,
19]. For this purpose, further investigations with the collection of nasal secretion samples and statistical correlations are needed.
Since BECs and serum ECP did not provide any indication useful for monitoring the success of therapy, the question arises whether routine blood sampling at each follow-up could also be dispensed with. Especially in the field of outpatient care of CRSwNP patients treated with Dupilumab, the reduction of laboratory controls could bring significant relief and cost reduction. In general, Dupilumab had only very few side effects, and even in the very rare case of hypereosinophilia, patients did not only remain asymptomatic but consistently improved their clinical condition. Unfortunately, predicting possible hypereosinophilia is still challenging. Wechsler et al. and Ryser et al. proposed that only patients with higher baseline BECs may be at greater risk of developing transient hypereosinophilia [
35,
44]. Rampi et al. found the 2-months value of BECs as a possible predictor of long-lasting hypereosinophilia in patients treated with Dupilumab [
45]. Their results showed that hypereosinophilia did not develop within 2 years if the eosinophil count did not exceed 1.5 billion/L after 2 months. Incorporating our observation, one may suggest that laboratory controls of biomarkers may only be necessary for the first months of therapy, if normal values are observed during this period. Thereafter, a targeted anamnesis for potential eosinophil-related morbidity might provide similar safety as a laboratory check. If maintaining routine blood tests to monitor therapy success, serum IgE emerges as a more valuable parameter compared to the other biomarkers.
Several potential limitations apply to our study. The data for this study was collected retrospectively which resulted in missing data that could potentially impact the validity of our analyses. Furthermore, we were unable to control if all patients adhered to the prescribed administration of Dupilumab and intranasal corticosteroids during the treatment period. Finally, one major limitation should be pointed out: towards the end of observation period, only very few patients were available for observation.. This reduction in observations caused the data, like the subjective outcome variables SNOT-22-Score and VAS-scores, to inexplicably deviate towards poorer values at month 22. This deviation may potentially be explained by the fact that only the most severely affected patients are seen for follow-ups at our specialized center for an extended period, rather than receiving outpatient treatment. Their values may deviate towards the worse and may lead to a selection bias. The reduced sample size primarily results from the majority of patients being referred to outpatient care after only 1 year, as outpatient monitoring and initiation of Dupilumab is increasingly possible, often meaning shorter travel distances for patients. The strength of this study is the ability to demonstrate the effect of Dupilumab on CRSwNP patients with a larger and diverse population and a long-term observation period. This applies in particular for the analysis of the collected blood parameters in connection with clinical parameters, as they can be easily obtained in everyday settings.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.