Efficacy and Safety of Biologics for Chronic Rhinosinusitis With Nasal Polyps: A Meta‐Analysis of Real‐World Evidence

ABSTRACT

Dupilumab, omalizumab, mepolizumab, and benralizumab have demonstrated good efficacy and safety in the treatment of severe uncontrolled chronic rhinosinusitis with nasal polyps (CRSwNP) in phase 3 randomised controlled trials (RCTs). With recent regulatory approvals, there has been a surge in real‐world studies (RWSs). This systematic review and meta‐analysis aimed to summarise the efficacy and safety of these four biologics in real‐world settings. Primary outcomes were nasal polyp score and sino‐nasal outcome test‐22 score. Secondary outcomes included smell identification test score, loss of smell score, nasal congestion score, overall nasal symptom score, treatment response, and adverse events (AEs) prompting discontinuation. Efficacy outcomes at 4, 6, 12, and over 12 months were summarised, and meta‐analyses of mean changes from baseline were conducted. Sixty‐four RWSs involving 3921 patients were included. Significant improvements in clinical outcomes were observed at most follow‐up time points, with dupilumab showing particularly notable effects. The efficacy observed in these RWSs was superior to that demonstrated in phase 3 RCTs. All biologics exhibited low discontinuation rates due to AEs. Overall, biologic treatments for CRSwNP in real‐world settings demonstrate strong efficacy and good safety. However, the limitations in current RWSs highlight the need for long‐term, high‐quality multicentre prospective studies and comprehensive healthcare database analyses.

1. Introduction

Chronic rhinosinusitis with nasal polyps (CRSwNP) affects approximately 1%–2.6% of the general population [1], significantly impacting patients' health and quality of life. Traditional treatment options include topical and systemic corticosteroids, with endoscopic sinus surgery (ESS) reserved for severe uncontrolled cases [2]. Biologic therapies targeting type 2 immune pathways have recently provided new therapeutic options.

Dupilumab, omalizumab, and mepolizumab have been approved for the treatment of patients with CRSwNP. Additionally, benralizumab could be prescribed for CRSwNP in patients with severe eosinophilic asthma (SEA). Phase 3 randomised controlled trials (RCTs) have demonstrated the efficacy and safety of these four biologics in treating CRSwNP [3, 4, 5, 6, 7]. However, despite the high‐grade evidence provided by RCTs, their generalisability to real‐life outcomes is limited, as patient populations in clinical practice are more diverse and potentially less adherent to biological treatment.

With regulatory approvals, many real‐world studies (RWSs) are being conducted. Therefore, our study aimed to systematically review and perform meta‐analyses of real‐world effectiveness and safety data on dupilumab, omalizumab, mepolizumab, and benralizumab for CRSwNP treatment to provide comprehensive real‐world evidence for clinical practice.

2. Methods

2.1. Study Selection

We performed a literature search to identify RWSs for biologic therapy in CRSwNP. PubMed, EMBASE, Cochrane Library, and Web of Science were consulted using a combination of medical subject headings and free‐text terms (Table S1) until July 15, 2024.

Articles meeting the following criteria were selected: (1) population: adult patients (≥ 18 years old) diagnosed with CRSwNP; (2) treatment: the prescribed dupilumab, omalizumab, mepolizumab, or benralizumab; (3) follow‐up time: at least 16 weeks after the treatment initiation; (4) outcomes: at least one of the nasal polyp score (NPS), sino‐nasal outcome test (SNOT) score, overall nasal symptom score, nasal congestion score (NCS), self‐reported loss of smell score, and smell identification test (SIT) score were reported in 16 weeks/4 months, 24 weeks/6 months, 48 weeks/12 months, and/or more than 1 year after the treatment initiation without scale limitation; (5) study design: longitudinal observational studies, prospective, or retrospective; (6) language: English.

Exclusion criteria included the following: (1) patients receiving concurrent biological therapy with ESS or another biologic agent at baseline or during follow‐up; (2) unclear time points for follow‐up; (3) outcomes lacking uncertainty measures, such as confidence intervals (CIs), standard deviations (SDs), and quartiles; (4) treatment discontinuation for over 1 month during follow‐up; (5) pooled analysis of multiple biologics; and (6) report of conflicting outcome data.

2.2. Data Extraction

Data were collected by one author (Xu) and checked by another (Cai). Disagreements were resolved through consensus. The following data were collected: participants' inclusion criteria, baseline characteristics, treatment strategy, follow‐up period, outcome measurements of interest, results, potential sources of bias, and study design, which were obtained from texts, tables, figures.

Baseline and post‐treatment mean values and SDs, mean changes (MCs) from baseline and SDs, and sample sizes were collected for quantitative outcomes. If only the whole or part of the five‐number summary (minimum, quartiles, and maximum) were reported, the estimated mean and SD were treated as true sample values, regardless of skewness [8, 9]. For binary data, the numbers of participants experiencing the event and those assessed were collected.

2.3. Outcome Measurements

We selected the NPS (scale 0–8) and SNOT‐22 score (scale 0–110) as primary outcomes. Minimal clinically important change scores (MCIDs) were defined as ≥ 1 point improvement for NPS and ≥ 12 points improvement for SNOT‐22 [10]. Secondary outcomes included the SIT score (scale 0–16), loss of smell score (scale 0–10), NCS (scale 0–10), overall nasal symptom score (scale 0–10), treatment response, and adverse events (AEs) prompting treatment discontinuation. The assessment of biological response followed the European Position Paper on Rhinosinusitis and Nasal Polyps (EPOS)/European Forum for Research and Education in Allergy and Airway Disease (EUFOREA) recommendations [2, 11, 12, 13, 14]. See Table S2 for more details about these outcomes. Efficacy data were collected at three time points within 1 year of follow‐up (16 weeks/4 months, 24 weeks/6 months, and 48 weeks/12 months) and time points exceeding 1 year. The safety outcome was analysed after the follow‐up.

To enhance clinical interpretability, outcomes measured with different scales were first standardised to the more commonly used instruments through proportional scaling before analysis. NPS (scale 0–6) was converted to NPS (scale 0–8), SNOT‐20 to SNOT‐22, and nasal congestion and loss of smell score (scale 0–3) to the visual analogue scale (VAS, 0–10). Both the University of Pennsylvania SIT (scale 0–40) and the Sniffin' Sticks‐12 identification test were converted to a 0–16 scale and referred to collectively, along with the Sniffin' Sticks‐16 identification test, as the SIT score.

2.4. Quality Assessment

Included records were assessed using the quality assessment tool developed by the National Heart, Lung, and Blood Institute [15]. Twelve criteria were evaluated, and studies were classified into three quality ratings (good, fair, or poor). Two reviewers (Cai and Xu) independently conducted the assessments, resolving disagreements through consensus.

2.5. Statistical Analysis

Continuous variables were expressed as mean (±SD), while the proportion of events was calculated for binary outcomes. Baseline and post‐treatment mean (±SD) were pooled to summarise variability across multiple studies. Random‐effect meta‐analyses were conducted to assess the pooled effect of MCs from baseline within individual studies and discontinuation rates using the “metafor” R package (version 4.6‐0). This model accounted for potential study heterogeneity arising from patient characteristics and study design, providing MCs from baseline and 95% CIs for true effect sizes [16]. A correlation coefficient of 0.5 between baseline and post‐treatment measurements was assumed for calculating MCs if they were not reported directly. For primary outcomes, the meta‐analysed MCs from RWSs were compared with the results from previous phase 3 RCT studies (LIBERTY NP SINUS‐24/52 [3], POLYP 1/2 [4], SYNAPSE [5], MERIT [6], and OSTRO [7]) for each biologic. Subgroup analyses were conducted on study design, inclusion criteria, sample size, the proportion of patients with asthma, baseline blood eosinophil count (BEC), NPS, and SNOT‐22, and data skewness, with interaction tests (z test) identifying the potential effect modification. Statistical significance was defined as p‐value < 0.05. All analyses were conducted using R statistical software (version 4.3.1; R Foundation; Vienna, Austria).

3. Results

3.1. Included Studies

The literature search revealed 4058 records, with 3335 remaining after removing duplicates. We excluded 3211 records that did not meet the eligibility criteria based on their titles and abstracts. Full texts were not obtained for the four studies. In addition, 56 publications were excluded after full‐text appraisal. Finally, we included 64 RWSs [17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80] involving 3921 participants (Figure 1).

FIGURE 1.

PRISMA flowchart for study identification. PRISMA, preferred reporting items for systematic reviews and meta‐analysis.

The study summary (Table 1) and the participants' baseline characteristics (Table S3) are presented. The included studies were observational, conducted in real‐world settings, and focused on adult patients with CRSwNP. Four biologics were evaluated: 35 studies on dupilumab with 2708 participants, 9 on omalizumab with 188 participants, 8 on mepolizumab with 317 participants, 9 on benralizumab with 542 participants, and 3 studies comparing these biologics with a total of 166 participants. Most of the included studies were conducted in Europe (N = 54), with a smaller number in Asia (N = 5) and North America (N = 5) (Table 1). Regarding study design, 34.4% (22 out of 64) were prospective, 57.8% (37 out of 64) were retrospective, and one was ambispective. Among these, eight studies (12.5%) had a sample size of more than 100. Most studies involved using the standard doses recommended in guidelines, and the follow‐up periods ranged from 16 weeks to 36 months.

TABLE 1.

Study country, enrolment design, sample size, inclusion criteria for adult patients with CRSwNP, dosing strategy, and follow‐up period of included studies.

Study	Country	Design a	Sample size	Inclusion criteria	Dosing b	Follow‐up
Dupilumab
van der Lans 2023 [77]	Netherlands	P	228	EPOS 2020	300 mg q2w and tapering	96 weeks
Trimarchi 2022 [59]	Italy	P	21	EPOS 2020	Standard	6 months
Albrecht 2023 [68]	Germany	P	81	EPOS 2020	Standard	12 months
Riva 2024 [28]	Italy	R	45	EPOS 2020	NA	12 months
Pecorari 2023 [72]	Italy	R	52	EPOS 2020	NA	12 months
Jansen 2023 [25]	Germany	R	40	EPOS 2020	Standard	13 months
Tanzini 2024 [80]	Italy	P	61	EPOS 2020	Standard	20 months
Gal 2024 [45]	France	R	47	EPOS/EUFOREA 2023	Standard	12 months
Gelardi 2024 [34]	Italy	NA	23	AIFA	Standard	12 months
De Corso 2022 [38]	Italy	NA	57	AIFA	Standard	12 months
Ottaviano 2022 [39]	Italy	NA	47	AIFA	Standard	12 months
Piazzetta 2023 [76]	Italy	R	14	AIFA	Standard	24 weeks
De Corso 2023 [30]	Italy	R	648	AIFA	Standard	12 months
Giombi 2023 [60]	Italy	R	53	AIFA	Standard	12 months
Loperfido 2023 [33]	Italy	R	130	AIFA	NA	18 months
Mocellin 2023 [73]	Italy	R	23	AIFA	Standard	6 months
Suzaki 2024 [70]	Japan	P	63	with eosinophilic CRSwNP; ≥ 1 ESS	Standard	24 weeks
Galletti 2024 [29]	Italy	NA	170	NPS ≥ 4; receiving CS in the past 2 years	Standard	12 months
Mustafa 2020 [26]	USA	P	10	with AERD; SNOT‐22 ≥ 19	Standard	6 months
La Mantia 2024 [36]	Italy	P	60	NPS ≥ 5; uncontrolled by medication	Standard	6 months
Ryser 2023 [32]	Switzerland	P & R	68	EUFOREA 2019	Standard	6 months
Campion 2023 [61]	Austria	R	97	EUFOREA 2021	Standard	6 months
Cantone 2022 [53]	Italy	R	53	NPS ≥ 5; uncontrolled by medication; ≥ 1 ESS	Standard	6 months
Napolitano 2021 [27]	Italy	R	19	with severe AD	Standard after 600 mg loading	24 weeks
Grose 2023 [23]	Canada	R	27	uncontrolled by medication; ineligible for surgery or preferring medication	NA	12 months
Garvey 2024 [58]	USA	R	26	NA	Standard	12 months
Rosso 2024 [75]	Italy	R	20	with severe asthma; ≥ 1 ESS; switching from other biologics	NA	18 months
Tajiri 2024 [43]	Japan	P	16	medically uncontrolled CRSwNP; with moderate‐to‐severe asthma	Standard after 600 mg loading	48 weeks
Alicandri‐Ciufelli 2024 [48]	Italy	R	145	AIFA; ≥ 1 ESS	Standard	12 months
Danisman 2023 [24]	Germany	P	20	NA	NA	48 weeks
Ottaviano 2023 [50]	Italy	R	147	AIFA	Standard	12 months
Schneider 2023 [31]	Austria	P	31	with AERD	Standard after 600 mg loading	24 weeks
Kilty 2022 [21]	Canada	R	53	≥ 1 ESS (if no contraindication)	NA	28 weeks
Sarnoch 2024 [79]	Germany	R	104	uncontrolled by medication and surgery; anosmia, SNOT‐22 score > 40, with asthma, or type 2 inflammation	Standard	22 months
Nettis 2021 [44]	Italy	P	9	with moderate‐to‐severe AD	Standard after 600 mg loading	16 weeks
Omalizumab
Lombardo 2023 [66]	Italy	P	13	NPS ≥ 5 or SNOT‐22 score ≥ 50	Standard	16 weeks
Ruiz‐Hornillos 2020 [54]	Spain	P	16	with moderate–severe allergic asthma	NA	12 months
Bidder 2018 [56]	UK	P	13	with SAA	Standard	16 weeks
Zheng 2022 [22]	China	P	22	with asthma; uncontrolled by medication and surgery in the past year	Standard	24 weeks
Maza‐Solano 2023 [55]	Spain	R	22	with moderate–severe asthma; SNOT‐22 score ≥ 20; symptom VAS score > 3; met surgery criteria	Standard	24 months
Armengot‐Carceller 2020 [41]	Spain	R	23	Uncontrolled CRSwNP; with mild asthma; ≥ 2 ESS; serum tIgE levels > 30 kU/L; NPS ≥ 1 bilaterally; ≥ 3 cycles of OCS per year	Standard	12 months
Tiotiu 2020 [40]	France	R	24	with SAA	Standard	6 months
Sima 2023 [71]	China	P	22	recurrent CRSwNP	Standard	24 weeks
Quint 2022 [57]	Austria	P	33	with AERD; ≥ 1 ESS	Standard	24 weeks
Mepolizumab
Galletti 2024 [37]	Italy	P	29	NPS ≥ 4; receiving CS in the past 2 years	Standard	12 months
Detoraki 2021 [52]	Italy	P	44	with SEA	Standard	12 months
Domínguez‐Sosa 2023 [64]	Spain	R	55	NPS ≥ 2 bilaterally; symptom VAS score > 5; with asthma or multiple prior ESS; ≥ 1 polypectomies in the last 10 years	Standard	6 months
Gallo 2022 [51]	Italy	R	43	with SEA	NA	12 months
Yilmaz 2020 [46]	Turkey	R	16	with OCS‐dependent SEA	Standard	6 months
Cavaliere 2024 [67]	Italy	R	20	severe uncontrolled CRSwNP; BEC > 150 cells/mL	Standard	12 months
Kurosawa 2019 [47]	Japan	P	11	with SEA	Standard	48 weeks
Bagnasco 2023 [49]	Italy	P	99	with SEA	Standard	36 months
Benralizumab
Lombardo 2020 [65]	Italy	R	10	with SEA; NPS ≥ 2 out of 6; ≥ 1 polypectomies	Standard	24 weeks
Bagnasco 2020 [42]	Italy	R	34	with SEA	NA	24 weeks
Pini 2024 [78]	Italy	R	60	with SEA	Standard	36 months
Cavaliere 2022 [20]	Italy	R	11	with SEA; NPS ≥ 3 out of 6	Standard	12 months
Le 2024 [19]	5 countries	R	233	with SEA	NA	12 months
Nolasco 2021 [18]	Italy	R	79	with SEA	Standard	24 weeks
Chitguppi 2021 [35]	USA	R	23	with SEA	NA	4 months
Pelaia 2024 [74]	Italy	R	82	with SEA	Standard	24 months
Buonamico 2020 [17]	Italy	R	10	with SEA	NA	6 months
Dupilumab and Omalizumab
Haxel 2022 [69]	Germany	P	70	uncontrolled CRSwNP; NPS ≥ 2 bilaterally	NA	6 months
Mümmler 2021 [63]	Germany	R	24	with severe asthma	NA	6 months
Omalizumab, Mepolizumab, and Benralizumab
Tiotiu 2023 [62]	France and Spain	R	72	with SAA	Standard	6 months

Abbreviations: AD, atopic dermatitis; AERD, aspirin exacerbated respiratory disease; AIFA, Italian medicines agency; BEC, blood eosinophil counts; CRSwNP, chronic rhinosinusitis with nasal polyps; CS, corticosteroids; EMA, European medicines agency; EPOS, European position paper on rhinosinusitis and nasal polyps; ESS, endoscopic sinus surgery; EUFOREA, European forum for research and education in allergy and airway diseases; NA, not available; NPS, nasal polyps score; OCS, Oral corticosteroids; SAA, severe allergic asthma; SEA, severe eosinophilic asthma; SNOT‐22, 22‐item Sino‐Nasal Outcome Test; tIgE, total Immunoglobulin E; VAS, visual analogue scale.

^a Retrospective (R) or prospective (P).

^b Standard dosing strategies are as follows: 300 mg q2w for dupilumab; determined by serum total IgE level and body weight (q2w or q4w) for omalizumab; 100 mg q4w for mepolizumab; and 30 mg q4w for the first three doses, followed by q8w thereafter for benralizumab.

Studies on dupilumab primarily focused on patients with severe and uncontrolled CRSwNP. Among these, 10 followed the Italian Medicines Agency indication criteria [81] for biologic initiation, and seven followed the EPOS 2020 indication criteria [2]. 41.2% (14 out of 34) of the studies reported that the proportion of patients with asthma was ≥ 80%, and 76.7% (23 out of 30) indicated that the proportion of patients with previous nasal surgery was ≥ 80% (Table S3). In 53.8% (14 out of 26) of the studies, the baseline BEC was > 0.5 × 10⁹. Furthermore, 72.7% (24 out of 33) reported a baseline NPS ≥ 5 out of 8, and 79.4% (27 out of 34) recorded a baseline SNOT‐22 score ≥ 50. While for the other three biologics combined, the studies primarily involved patients with comorbid asthma. In 96.8% (30 out of 31) of the studies, the proportion of patients with asthma was ≥ 80%, and 52.2% (12 out of 23) reported that the proportion of patients with previous surgery was ≥ 80%. In 81.8% (18 out of 22) of the studies, the baseline BEC was > 0.5 × 10⁹. Four out of 16 (25.0%) studies indicated a baseline NPS ≥ 5 out of 8, and 57.1% (12 out of 21) reported a baseline SNOT‐22 score ≥ 50.

3.2. Risk of Bias

Among the 64 studies, overall quality was rated as good for 31, fair for 22, and poor for 11 (Figure S1). There were two primary concerns regarding unrepresentative participants (e.g., all patients with asthma) and high loss to follow‐up rates. Additional concerns included inadequate sample size, unclear patient selection criteria, undefined dosing strategy, and exclusion of patients with insufficient follow‐up and poor compliance.

3.3. Efficacy Outcomes

For primary outcomes (NPS and SNOT‐22 score), the aggregated data were summarised at three follow‐up time points (4, 6, and 12 months), along with the corresponding baseline data (Figure 2). Similarly, secondary outcome measures, including NCS and SIT, loss of smell, and overall nasal symptom scores, were detailed at different follow‐up intervals (Figure S2). Furthermore, meta‐analyses were performed on the MCs from baseline for each study. The summarised results for NPS and SNOT‐22 score are depicted in Figure 3, while the findings for the secondary outcomes are presented in Figure S3. The detailed meta‐analysis data for the four biologics are shown in Figures S4–S7. Additionally, results for primary outcomes beyond 1 year after treatment initiation are provided in Table S4. The proportion of responders meeting the MCID for NPS is presented in Table S5, while no data are available on the proportion of patients achieving the SNOT‐22 MCID of 12. The number of dupilumab responders according to EPOS/EUFOREA criteria is shown in Table S6; no such data were reported for omalizumab, mepolizumab, or benralizumab.

FIGURE 2.

Pooled NPS and SNOT‐22 score over different time points and the matched baseline data. NA, not available; NPS, nasal polyp score; SE, standard error; SNOT‐22, sino‐nasal outcome test‐22.

FIGURE 3.

Meta‐analysis forest plot of mean changes in NPS and SNOT‐22 score over different time points. CI, confidence interval; MC, mean change; NA, not available; NPS, nasal polyp score; SNOT‐22, sino‐nasal outcome test‐22.

3.4. Dupilumab

In the dupilumab studies, the mean NPS decreased from 5.5 at baseline to 1.8 at 6 months after treatment initiation (Figure 2), with a meta‐analysed MC of −3.5 [−3.7, −3.3] (Figure 3A). The NPS decreased from 5.6 at baseline to 1.4 at 12 months, with a pooled MC of −3.9 [−4.2, −3.5]. The SNOT‐22 score decreased from 56.8 at baseline to 18.5 at 6 months and from 58.1 to 16.3 at 12 months (Figure 2). The combined MCs were −36.3 [−39.1, −33.6] and −39.1 [−42.8, −35.3], respectively (Figure 3B). Five studies reported the sustained long‐term effectiveness of dupilumab, with follow‐up durations of 18–22 months, showing a decrease in NPS from an initial average of 4.6–5.7 to 0.5–1.3 and SNOT‐22 score from 53.6–62.6 to 8.3–28.7 (Table S4). Among these studies, three highlighted that 67%–99% of patients achieved the MCID for NPS over follow‐up periods of 16–96 weeks (Table S5).

Subgroup analysis of the 6‐ and 12‐month NPS and SNOT‐22 results was conducted in the dupilumab studies (Figure S8). The high baseline NPS group exhibited more pronounced reductions in NPS and SNOT‐22 score at 6 months. Differences in SNOT‐22 score reduction were also noted between baseline SNOT‐22 groups, with the high group showing a greater reduction at both time points. No significant differences were observed in other subgroup analyses.

Similarly, dupilumab significantly improved the secondary outcome measures. At 6 months, the summarised MC in the SIT score was 5.8 [5.1, 6.4], and it further increased to 7.2 [6.3, 8.1] at 12 months (Figure S3). This result was also corroborated by the loss of smell score, with summarised MCs of −5.5 [−6.4, −4.7] and −6.0 [−6.9, −5.1] at 6 and 12 months, respectively. In addition, dupilumab significantly improved NCS, with aggregated MCs of −5.6 [−6.5, −4.5] at 6 months and −6.4 [−6.9, −5.8] at 12 months. For the overall nasal symptom score, the pooled MCs were −4.7 at both 6 and 12 months. Responses were evaluated in twelve studies according to EPOS/EUFOREA criteria, with the majority indicating that over 90% of patients met at least three out of five criteria post‐treatment (Table S6).

3.5. Omalizumab

In patients treated with omalizumab, the NPS decreased from 3.7 at baseline to 2.3 at 6 months (Figure 2), with a pooled MC of −1.8 [−3.0, −0.7] (Figure 3A), while only one study reported a significant mean change of −3.8 in NPS at 12 months. The SNOT‐22 score decreased from 55.0 at baseline to 29.5 at 6 months, with a summarised MC of −22.4 [−30.7, −14.1] (Figure 3B). At 12 months, only two studies reported changes in SNOT‐22, with a pooled MC of −42.8 [−49.5, −36.2]. Regarding long‐term efficacy, one study reported a decreased SNOT‐22 score from 72.3 at baseline to 15.8 at 24 months (Table S4).

For the secondary outcome measures at 6 months, improvements in the loss of smell score of −3.0 [−3.8, −2.2], the NCS of −4.5 [−6.1, −3.0], and the overall nasal symptom score of −2.8 [−5.1, −0.5] were pooled from three studies, while only one study reported a 3.9 [1.8, 6.0] improvement in the SIT score (Figure S3). At 12 months, a cohort with 16 patients observed no significant results in the loss of smell and overall nasal symptom scores.

3.6. Mepolizumab

Mepolizumab reduced the NPS from 4.4 at baseline to 2.1 at 6 months (Figure 2), resulting in a pooled MC of −2.0 [−3.0, −0.9] (Figure 3A). At 12 months, the NPS decreased to 2.2 (from 4.1 at baseline), with a pooled MC of −2.2 [−3.4, −1.1]. Regarding SNOT‐22, the score dropped from 62.2 at baseline to 23.6 at 6 months, yielding an overall MC of −33.3 [−55.4, −11.2] (Figure 3B). At 12 months, the score decreased from 52.5 to 30.6, with an aggregated MC of −24.8 [−32.6, −17.0]. Concerning long‐term efficacy, one study recorded a reduced SNOT‐22 score from 51.0 at baseline to 34.0 at 24 months, with a further decrease to 26.0 at 36 months (Table S4). Only one study reported 63% of patients achieving MCID for NPS at 12 months (Table S5).

MCs in the loss of smell score of −3.2 [−6.0, −0.5] and the overall nasal symptom score of −3.5 [−5.0, −2.1] were pooled from two studies at 6 months (Figure S3). Only one study reported an improved SIT score, with 4.6 [4.1, 5.1] and 9.0 [8.2, 9.8] increases at 6 and 12 months, respectively, while one study reported a change in NCS of −4 [−5.2, −2.8] at 12 months.

3.7. Benralizumab

Benralizumab reduced the NPS from 5.4 at baseline to 3.1 (Figure 2), with a summarised MC of −2.0 [−2.3, −1.6] at 6 months and showed a similar reduction at 12 months with a pooled MC of −2.1 [−3.8, −0.3] (Figure 3A). For SNOT‐22, the score decreased from 53.1 at baseline to 36.0 at 6 months, resulting in a pooled MC of −19.4 [−24.6, −14.1] (Figure 3B). At 12 months, the score decreased from 53.1 at baseline to 36.1, with a comparable aggregated MC of −19.8 [−20.3, −19.2]. For long‐term efficacy, one study reported a decreased SNOT‐22 score from 47.2 at baseline to 28.6 at 24 months, with a further drop to 24.3 at 36 months (Table S4). One study reported that 49% of patients achieved MCID for NPS at 12 months (Table S5).

Improved NCS, loss of smell score, and overall nasal symptom score at 6 and 12 months were reported by only one study showing significant results, with no reports on the SIT score (Figure S3).

3.8. Comparisons With RCT Studies

The meta‐analysed MCs from RWSs were compared with the results from previous phase 3 RCT studies (Figure 4). For dupilumab, the reduction in NPS at 6 and 12 months in RWSs was nearly twice that observed in the SINUS‐24/52 trials (−3.5 vs. −1.8 and − 3.9 vs. −2.2, respectively). Similarly, the decrease in SNOT‐22 scores at 6 and 12 months in RWSs was greater than that in RCTs (−36.3 vs. −28.6 and − 39.1 vs. −30.1, respectively). For omalizumab, mepolizumab, and benralizumab, the changes in NPS were also better in RWSs than in RCTs; however, the results in RWSs showed greater variability. The changes in SNOT‐22 were comparable between the two.

FIGURE 4.

Forest plot comparing the NPS and SNOT‐22 scores for four biologics at 6 and 12 months between RWSs and RCTs. Data was presented as MC (95% CI). CI, confidence interval; MC, mean change; NA, not available; NPS, nasal polyp score; RCT, randomised controlled trial; RWS, real‐world study; SNOT‐22, sino‐nasal outcome test‐22.

3.9. Safety Outcome

In total, 27 dupilumab, seven omalizumab, seven mepolizumab, and four benralizumab studies reported AEs leading to treatment discontinuation, with discontinuation rates of 57/2164 (2.6%) for dupilumab, 1/144 (0.7%) for omalizumab, 4/265 (1.5%) for mepolizumab, and 0/205 (0%) for benralizumab (Table S7). Joint pain was the most common cause of treatment discontinuation, with eight cases in the dupilumab studies, one case in the omalizumab studies, and one case in the mepolizumab studies. Additionally, in the dupilumab studies, seven patients withdrew owing to elevated blood eosinophil levels. A meta‐analysis was conducted on the proportion of treatment discontinuations owing to AEs in 27 dupilumab studies, with a pooled rate of 1.3% (Figure S9).

4. Discussion

Biological therapy has revolutionised CRSwNP treatment. Between 2019 and 2024, seven phase 3 RCTs evaluating four biologics have been published. RWSs are crucial in supplementing efficacy and safety findings from RCTs. The accumulation of real‐world evidence now supports the first systematic review and meta‐analysis on this topic, offering a deeper understanding of the therapeutic landscape.

We primarily aimed to address a literature gap by providing the compiled evidence from RWSs on the efficacy and safety of biologics for CRSwNP treatment. We conducted meta‐analyses involving four biologics (dupilumab, omalizumab, mepolizumab, and benralizumab) across 64 RWSs with 3921 participants. Efficacy outcomes were primarily assessed at 6 and 12 months to evaluate durability, aligned with EPOS/EUFOREA expert recommendations. Four biologics, particularly dupilumab, demonstrated rapid and sustained effectiveness across primary and secondary outcomes, accompanied by favourable safety profiles. We also aimed to compare RWS findings with evidence from RCTs, a critical but rarely addressed aspect in the current literature. In our study, RWS and RCT outcomes demonstrated efficacy, with RWSs showing superior results to those of RCTs, suggesting promising implications for clinical practice.

The RWSs conducted so far have yielded encouraging results; however, the appropriate patient populations for biological therapy remain unknown. While guidelines and opinions vary, there is a consensus that biologics should be reserved for severe and uncontrolled cases. Among the included studies, substantial research focused on populations with concurrent asthma and CRSwNP. Since omalizumab and mepolizumab were initially approved for asthma treatment and benralizumab is currently only approved for SEA, patients receiving biological therapy for asthma may not meet the eligibility criteria for CRSwNP treatment. This discrepancy is reflected in the relatively lower baseline NPS and SNOT‐22 score and a lower proportion of prior nasal surgery, which raises concerns about the broader application of biologics in CRSwNP cases. It also indicates a need for comprehensive airway treatment strategies, as current guidelines do not fully consider the concurrent management of upper and lower airway diseases.

Additionally, most of the RWSs included a high percentage of patients with prior nasal surgery. Whether surgery is necessary before initiating biologic therapy remains inconclusive, although it is generally recommended. Some RWSs have compared the treatment outcomes between populations with and without a surgical history; however, the findings are inconclusive owing to the few patients in the no‐surgery subgroup [30, 61, 73]. Considering health economics, biologic agents are currently less cost‐effective than ESS and more recommended for patients requiring repeated surgeries or with contraindications to surgery [71, 82, 83]. However, there may be an increased use of biologics in non‐surgical populations in the future, considering expected cost reductions and patient preferences. Given the challenges in identifying suitable patients, factors such as treatment availability, economic considerations, and healthcare affordability should also be considered in different cases.

The comparison of RWSs and RCTs suggests that biologics exhibit greater efficacy in real‐world settings. This discrepancy may be partly attributed to the heterogeneity of study populations. Phase 3 clinical trials have indicated that certain patient subgroups, particularly those with more severe disease, type 2 inflammation, and related comorbidities, may derive greater benefits from biologic therapies, which has influenced the inclusion criteria in guidelines. Consequently, subsequent RWSs have tended to recruit patients who are more likely to benefit from these therapies.

While primary and secondary outcomes were significantly improved at 6 and 12 months after treatment initiation, few RWSs comprehensively monitored response rates. Furthermore, RWSs lasting beyond 1 year are lacking, and follow‐up data after treatment cessation are scarce, leaving the duration of treatment efficacy beyond 12 months unclear. Among the limited studies with follow‐up periods exceeding two years [49, 55, 74, 77, 78], findings suggest that biologics provide sustained improvements in NPS, quality of life, sinonasal symptoms, and asthma control. Therapeutic efficacy is typically achieved within the first 6 months and maintained thereafter. However, further long‐term research is needed to confirm these outcomes.

Another challenge is biologic selection. While head‐to‐head RCTs remain lacking, several previous meta‐analyses of RCTs have indirectly compared the biologics, consistently concluding that dupilumab has superior efficacy [84, 85, 86, 87]. Comparisons are more feasible in RWSs, where dupilumab often outperforms other biologics [63, 88], and switching to dupilumab may significantly improve outcomes when other biologics fail [75, 89]. However, due to significant baseline differences and unequal sample sizes across biologics, we were unable to directly compare the four biologics in this meta‐analysis.

Ongoing safety monitoring is also essential, given the increasing number of patients on these treatments. Previous systematic reviews based on RCTs, summarising AEs and incidence, have generally concluded that biologics are safe [90, 91, 92]; however, their follow‐up times were relatively short for assessing long‐term safety. An FDA AE database study identified 911 dupilumab‐associated adverse reactions for CRSwNP treatment, with 121 (13.3%) classified as serious and three resulting in death, showing lower rates compared to asthma and atopic dermatitis treatment [93]. Another comparative database analysis indicated possible links between dupilumab and ophthalmologic/injection‐site AEs, with omalizumab as the only biologic showing a positive correlation with anaphylaxis [94]. Due to inconsistencies in AE reporting and definitions of serious AEs across studies, this study focused on AEs leading to treatment discontinuation, aiming to identify both non‐serious AEs causing concerns and serious AEs. Most patients completed treatment, and discontinuation rates due to AEs were similar to those reported in RCTs: 11/440 (2.5%) in SINUS‐24/52 [3] for dupilumab; 0/135 (0%) in POLYP 1/2 [4] for omalizumab; 4/206 (1.9%) in SYNAPSE [5] and 2/85 (2.4%) in MERIT [6] for mepolizumab; 6/203 (3.0%) in OSTRO [7] for benralizumab; and 26/900 (2.9%) across seven RCTs for placebo. These findings suggest that biologics in RWSs are well tolerated with acceptable safety profiles. Given the small sample sizes and limited number of studies, we conducted a meta‐analysis only for dupilumab and were unable to directly compare AE incidence across biologics.

Lastly, it is important to highlight that the results observed in real‐world settings require careful interpretation owing to RWS limitations. Over half of the studies were retrospective, potentially introducing selection and recall biases. The absence of control groups in RWSs limits assessment of the placebo effect, and the lack of blinding in patients and researchers may lead to overreporting of treatment effects. Additionally, most studies were conducted in Europe, limiting the generalisability of findings to non‐European populations. One possible reason is that many countries have not yet approved biologics for the indication of CRSwNP treatment, such as China. Moreover, ten of the 64 studies had a high loss to follow‐up, although overall follow‐up rates for each outcome were considered acceptable. Longer treatment durations led to a higher loss to follow‐up, possibly owing to significant efficacy leading to treatment discontinuation, lack of response, side effects, or non‐adherence to treatment protocols, which can bias the results. Furthermore, converting between various scoring systems across studies, although minimal in our analysis, may not fully capture original values. Despite the meaningfulness of the combined results, substantial heterogeneity was observed across nearly all pooled outcomes. Sources of bias may include study design, inclusion criteria, and medication regimens. Variability in patients' baseline characteristics, such as age, sex, race, and underlying disease status, also contributes to this heterogeneity, although these characteristics are essential for reflecting real‐world clinical practice. Given these considerations, we analysed results for each biologic individually rather than making direct comparisons.

Overall, our current findings indicate the significant and sustained effectiveness and good safety of dupilumab, omalizumab, mepolizumab, and benralizumab for CRSwNP treatment in real‐world settings, demonstrating superior efficacy than that reported in RCTs. However, the current limitations in RWSs indicate a need for long‐term, high‐quality multicentre prospective studies and healthcare database analyses.

Author Contributions

Zhang, Cai, and Xu contributed to the design of the research. Cai, Xu, and Zhao performed the data collection and analysis as well as wrote the manuscript. Zhang contributed amendments to the manuscript and revised it critically. All authors approved the final version to be published.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.