Biologics for chronic rhinosinusitis

Summary of findings 1. Anti‐IL‐4Rα mAb (dupilumab) compared to placebo (on top of topical steroids) for chronic rhinosinusitis

Outcomes	Number of participants (studies)	Relative effect (95% CI)	*Anticipated absolute effects^ (95% CI)**			Certainty of the evidence (GRADE)	What happens
Anti‐IL‐4Rα mAb (dupilumab) compared to placebo (on top of topical steroids) for chronic rhinosinusitis
Patients or population: patients with severe chronic rhinosinusitis with nasal polyps Setting: tertiary care Intervention: anti‐IL‐4Rα mAb (dupilumab) Comparison: placebo (on top of topical steroids)
Outcomes	Number of participants (studies)	Relative effect (95% CI)	Without anti‐IL‐4Rα mAb (dupilumab)	With anti‐IL‐4Rα mAb (dupilumab)	Difference	Certainty of the evidence (GRADE)	What happens
Health‐related quality of life ‐ disease‐specific (SNOT‐22, range 0 to 110, lower = better) Follow‐up (range): 16 to 24 weeks	784 (3 RCTs)	—	The median disease‐specific health‐related quality of life score without anti‐IL‐4Rα mAb (dupilumab) was 40.5 points	—	MD 19.61 points lower (22.54 lower to 16.69 lower)	⊕⊕⊕⊕ HIGH	At up to 24 weeks, aspects of health‐related quality of life that are directly impacted by chronic rhinosinusitis were better in participants who received dupilumab. The size of the difference is clinically significant.
Disease severity ‐ VAS (range 0 to 10, lower = better) Follow‐up (range): 16 to 24 weeks	784 (3 RCTs)	—	The median disease severity score without anti‐IL‐4Rα mAb (dupilumab) was ‐1.3 points	—	MD 3 points lower (3.47 lower to 2.53 lower)	⊕⊕⊕⊝ MODERATE¹	Overall chronic rhinosinusitis symptoms were probably better in participants who received dupilumab.
Serious adverse events Follow‐up (range): 16 to 52 weeks	782 (3 RCTs)	RR 0.47 (0.29 to 0.76)	Study population			⊕⊕⊝⊝ LOW²	Participants who had dupilumab may have had fewer serious adverse events than participants who received placebo in 3 RCTs (28/470 with dupilumab versus 39/312 with placebo), but we have limited confidence in this estimate because the sample size may be too small to estimate this accurately, or capture the range of adverse events that could possibly occur in a larger population or with longer follow‐up.
Serious adverse events Follow‐up (range): 16 to 52 weeks	782 (3 RCTs)	RR 0.47 (0.29 to 0.76)	12.5%	5.9% (3.6 to 9.5)	6.6% fewer (8.9 fewer to 3 fewer)	⊕⊕⊝⊝ LOW²
Avoidance of surgery ‐ number of patients who had surgery as rescue treatment Follow‐up (range): 24 to 52 weeks	725 (2 RCTs)	RR 0.17 (0.05 to 0.52)	Study population			⊕⊕⊕⊝ MODERATE³	Patients who had dupilumab probably have lower risk of requiring surgery due to severe chronic rhinosinusitis symptoms after 24 to 52 weeks of treatment. We have moderate confidence in this estimate as we are not sure which criteria were used to determine the need for 'rescue surgery'.
	725 (2 RCTs)	RR 0.17 (0.05 to 0.52)	7.7%	1.3% (0.4 to 4)	6.4% fewer (7.3 fewer to 3.7 fewer)	⊕⊕⊕⊝ MODERATE³
Extent of disease: endoscopic nasal polyp score (range 0 to 8, lower = better) Follow‐up (range): 16 to 24 weeks	784 (3 RCTs)	—	The median nasal polyp score without dupilumab was 5.94 points.	—	MD 1.80 points lower (2.25 lower to 1.35 lower)	⊕⊕⊕⊝ MODERATE¹	Dupilumab probably results in a reduction in nasal polyp score by 24 weeks of follow‐up. This is likely to be a large effect, however we have moderate confidence in the estimate as it is unclear whether the scoring system used for nasal polyps is validated.
Extent of disease: CT scan score (Lund‐Mackay, range 0 to 24, lower = better) Follow‐up (range): 16 to 52 weeks	784 (3 RCTs)	—	The median CT scan score without anti‐IL‐4Rα mAb (dupilumab) was 17.9 points	—	MD 7 points lower (9.61 lower to 4.39 lower)	⊕⊕⊕⊕ HIGH	At up to 24 weeks, the extent of disease as assessed by CT scan was less severe in participants who received dupilumab ‐ the difference is likely to be a large effect.
Health‐related quality of life ‐ generic (EQ‐5D visual analogue scale, range 0 to 100, higher = better) Follow‐up (range): 16 to 24 weeks	766 (3 RCTs)	—	The median change in generic HRQOL for the placebo group was an increase of 3.01 points	—	MD 8.29 points higher (5.73 higher to 10.85 higher)	⊕⊕⊕⊝ MODERATE⁴	The overall quality of life or health status, as assessed by the EQ‐5D visual analogue scale was probably slightly higher in participants who received dupilumab. However, we are not sure if the size of this difference is noticeable or would be considered important enough by most patients.
Adverse events ‐ nasopharyngitis, including sore throat (longest available data) Follow‐up (range): 16 to 52 weeks	783 (3 RCTs)	RR 0.95 (0.72 to 1.25)	Study population			⊕⊕⊝⊝ LOW²	We are uncertain whether there is an important difference in the risk of nasopharyngitis. Adverse events were reported by 94/470 participants who took dupilumab versus 66/313 who took placebo.
	783 (3 RCTs)	RR 0.95 (0.72 to 1.25)	21.1%	20.0% (15.2 to 26.4)	1.1% fewer (5.9 fewer to 5.3 more)	⊕⊕⊝⊝ LOW²
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; CT: computerised tomography; MD: mean difference; RCT: randomised controlled trial; RR: risk ratio; SNOT‐22: Sino‐Nasal Outcome Test‐22; VAS: visual analogue scale
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
¹Downgraded by one level due to study limitations: methods or criteria used in the measurement of the outcome were not validated. ²Downgraded by two levels due to imprecision and indirectness: small sample size for the outcome estimated resulting in an imprecise estimation of effect size. Moreover, some serious adverse events are relatively rare; a larger and more heterogeneous population or longer periods of treatment and follow‐up may be needed. ³Downgraded by one level due to serious limitations: the criteria used for requiring/not requiring 'rescue surgery' were unclear. ⁴Downgraded by one level for imprecision: the confidence interval crosses the minimally important difference (8 points), therefore the difference may or may not be of importance to participants.

Summary of findings 2. Anti‐IL‐5 mAb (mepolizumab) compared to placebo (on top of topical steroids) for chronic rhinosinusitis

Outcomes	Number of participants (studies)	Relative effect (95% CI)	*Anticipated absolute effects^ (95% CI)**			Certainty of the evidence (GRADE)	What happens
Anti‐IL‐5 mAb (mepolizumab) compared to placebo (on top of topical steroids) for chronic rhinosinusitis
Patients or population: patients with severe chronic rhinosinusitis with nasal polyps Setting: tertiary care Intervention: anti‐IL‐5 mAb (mepolizumab) Comparison: placebo (on top of topical steroids)
Outcomes	Number of participants (studies)	Relative effect (95% CI)	Without anti‐IL‐5 mAb (mepolizumab)	With anti‐IL‐5 mAb (mepolizumab)	Difference	Certainty of the evidence (GRADE)	What happens
Health‐related quality of life ‐ disease‐specific (SNOT‐22, range 1 to 100, lower = better) Follow‐up: 25 weeks	105 (1 RCT)	—	The mean disease‐specific health‐related quality of life score without anti‐IL‐5 mAb (mepolizumab) was 40.36.	—	MD 13.26 lower (22.08 lower to 4.44 lower)	⊕⊕⊝⊝ LOW¹	Aspects of health‐related quality of life that are directly impacted by chronic rhinosinusitis may have been better in participants who received mepolizumab but we are uncertain about this estimate.
Disease severity ‐ VAS (range 0 to 10, lower = better) Follow‐up: 25 weeks	72 (1 RCT)	—	The mean disease severity score without anti‐IL‐5 mAb (mepolizumab) was 6.21.	—	MD 2.03 lower (3.65 lower to 0.41 lower)	⊕⊝⊝⊝ VERY LOW^1,2	We are very uncertain about the impact of mepolizumab on overall chronic rhinosinusitis symptom severity.
Serious adverse events Follow‐up (range): 25 to 40 weeks	135 (2 RCTs)	RR 1.57 (0.07 to 35.46)	Study event rates³			⊕⊝⊝⊝ VERY LOW^1,4	We are very uncertain about the number of serious adverse events for chronic rhinosinusitis patients who use mepolizumab. The number of serious adverse events was 0/62 for placebo and 1/73 for mepolizumab.
			0.0%	1.37%

Avoidance of surgery ‐ patients still meeting the criteria for surgery At end of follow‐up (range): 25 to 40 weeks	135 (2 RCTs)	RR 0.78 (0.64 to 0.94)	Study population			⊕⊝⊝⊝ VERY LOW^1,2,4	We are very uncertain whether mepolizumab can help participants reduce the need for surgery.
	135 (2 RCTs)	RR 0.78 (0.64 to 0.94)	80.3%	62.7% (51.4 to 75.5)	17.7% fewer (28.9 fewer to 4.8 fewer)	⊕⊝⊝⊝ VERY LOW^1,2,4
Extent of disease ‐ endoscopic score Follow‐up (range): 25 to 40 weeks	137 (2 RCTs)	—	The mean endoscopic score without anti‐IL‐5 mAb (mepolizumab) ranged from 0 to ‐0.7.	—	MD 1.23 lower (1.79 lower to 0.68 lower)	⊕⊝⊝⊝ VERY LOW^1,2	We are very uncertain whether mepolizumab can reduce the extent of disease as measured by an endoscopic score.
Extent of disease ‐ CT scan score (Lund‐Mackay, range 0 to 24, lower = better) Follow‐up: 8 weeks	27 (1 RCT)	—	One study reported that CT scan scores were "not significantly different between groups"	—	—	⊕⊝⊝⊝ VERY LOW^1,5	We are very uncertain whether mepolizumab can reduce the extent of disease as measured by a CT scan score.
Health‐related quality of life ‐ generic, measured using the EQ‐5D visual analogue scale (range 0 to 100; 0 = worst imaginable health state, 100 = best imaginable health state) Follow‐up: at week 25	105 (1 RCT)	—	The mean generic health‐related quality of life score without anti‐IL‐5 mAb (mepolizumab) was 75.45	—	MD 5.68 higher (1.18 lower to 12.54 higher)	⊕⊕⊝⊝ LOW¹	We are uncertain about the impact of mepolizumab on overall quality of life or health status, as assessed by the EQ‐5D visual analogue scale.
Adverse events ‐ nasopharyngitis, including sore throat Follow‐up (range): 25 to 40 weeks	135 (2 RCTs)	RR 0.73 (0.36 to 1.47)	Study population			⊕⊕⊝⊝ LOW¹	We are uncertain about the risk of nasopharyngitis in chronic rhinosinusitis patients who used mepolizumab.
	135 (2 RCTs)	RR 0.73 (0.36 to 1.47)	22.6%	16.5% (8.1 to 33.2)	6.1% fewer (14.5 fewer to 10.6 more)	⊕⊕⊝⊝ LOW¹
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; MD: mean difference; RCT: randomised controlled trial; RR: risk ratio; SNOT‐22: Sino‐Nasal Outcome Test‐22; VAS: visual analogue scale
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
¹Downgraded by two levels due to imprecision: very small sample size resulting in a very imprecise estimation of effect sizes. ²Downgraded by one level due to study limitations: methods or criteria used in the measurement of the outcome were not validated. ³No events were reported in the placebo arm of these trials. We have therefore presented the study event rates rather than anticipated absolute events. ⁴Downgraded by one level due to indirectness: one study only assessed patients for two doses (Gevaert 2011). The other study evaluated six doses (24 weeks), but had a more than 30% dropout rate (Bachert 2017). Therefore, the length of follow‐up is inadequate and it is unclear whether this evidence related to safety is generalisable. ⁵Downgraded by one level due to study limitations: high risk of attrition bias, insufficient information to judge other aspects of study design and no numerical data presented for this outcome.

Summary of findings 3. Anti‐IgE mAb (omalizumab) compared to placebo (on top of topical steroids) for chronic rhinosinusitis

Outcomes	Number of participants (studies)	Relative effect (95% CI)	*Anticipated absolute effects^ (95% CI)**			Certainty of the evidence (GRADE)	What happens
Anti‐IgE mAb (omalizumab) compared to placebo (on top of topical steroids) for chronic rhinosinusitis
Patients or population: patients with chronic rhinosinusitis with nasal polyps Setting: tertiary care Intervention: anti‐IgE mAb (omalizumab) Comparison: placebo (on top of topical steroids)
Outcomes	Number of participants (studies)	Relative effect (95% CI)	Without anti‐IgE mAb (omalizumab)	With anti‐IgE mAb (omalizumab)	Difference	Certainty of the evidence (GRADE)	What happens
Health‐related quality of life ‐ disease‐specific (SNOT‐22, range 0 to 110, lower = better) Follow‐up: 24 weeks	265 (2 RCTs)	—	The mean change in disease‐specific HRQOL for the placebo group was ‐7.57 points	—	MD 15.62 points lower (19.79 lower to 11.45 lower)	⊕⊕⊕⊝ MODERATE¹	At 24 weeks, omalizumab probably results in an improvement in disease‐specific health‐related quality of life (as measured with the SNOT‐22 questionnaire). The size of the difference was clinically significant. However, we have limited confidence in this estimate because the sample size may be too small to estimate this accurately.
Disease severity, as measured by validated, patient‐reported symptom score	—	—	—	—	—	—	None of the studies reported this outcome.
Serious adverse events Follow‐up (range): 20 weeks to 6 months	329 (5 RCTs)	RR 0.32 (0.05 to 2.00)	2.5%	0.8% (0.1 to 5.1)	1.7% fewer (2.4 fewer to 2.5 more)	⊕⊝⊝⊝ VERY LOW^2,3	There is too little information. We are very uncertain whether omalizumab changes the incidence of serious adverse events because the sample size may be too small to estimate this accurately, or capture the range of adverse events that might occur in a larger population or with longer follow‐up. Serious adverse events were reported by 1/171 participants who took omalizumab versus 4/158 who took placebo.
Avoidance of surgery Nasal polyp score ≤4 (≤ 2 on each side) and an improvement in SNOT‐22 score of ≥ 8.9 points Follow‐up: 24 weeks	265 (2 RCTs)	RR 5.60 (1.99 to 15.76)	3.1%	17.1% (6.1 to 48.1)	14.0% more (3 more to 45.1 more)	⊕⊕⊕⊝ LOW^1,4	At up to 24 weeks, the evidence suggests that the number of participants in whom surgery was not thought to be necessary was greater in those who received omalizumab. However, we have limited confidence in this estimate because the sample size may be too small to estimate this accurately, and there are no widely agreed criteria to determine which patients need surgery for nasal polyps. Avoidance of surgery was reported in 23/134 participants who took omalizumab versus 4/131 participants who took placebo.
Extent of disease: endoscopic nasal polyp score (range 0 to 8, lower = better) Follow‐up: up to 24 weeks	312 (4 RCTs)	—	The median change in endoscopic nasal polyp score for the placebo group was ‐0.05 points	—	MD 1.26 points lower (2.2 lower to 0.31 lower)	⊕⊕⊝⊝ LOW^4,5	At up to 24 weeks, the evidence suggests that omalizumab may result in a reduction in the nasal polyp score. However, there are inconsistencies in the size of effect between studies, and it is unclear whether the method used is validated.
Extent of disease: CT scan (lower score = better) Follow‐up: 20 weeks	47 (2 RCTs)	—	The mean CT scan score without anti‐IgE mAb (omalizumab) ranged from ‐8.9 to 18.3	—	SMD 0.2 lower (1.55 lower to 1.14 higher)	⊕⊝⊝⊝ VERY LOW^2,6	There is too little information ‐ we are very uncertain whether there is a difference in the extent of disease with omalizumab. There are inconsistencies in the size and direction of effect. In the NCT01066104 study, the results favoured the placebo group, while in Gevaert 2013 they favoured the omalizumab group.
Health‐related quality of life ‐ generic (SF‐36) Follow‐up (range): 20 weeks to 6 months	38 (2 RCTs)	One study found no significant differences (P > 0.05, all comparisons) except for one domain, 'vitality' (omalizumab 9.4, placebo 12.5, P < 0.05). A second study found that physical health was significantly improved in the omalizumab group (P = 0.02) but not in the placebo group (P = 0.75). Mental health did not significantly improve in either treatment group.				⊕⊝⊝⊝ VERY LOW^7,8	We are very uncertain about the impact of omalizumab on health‐related quality of life.
Adverse events ‐ nasopharyngitis, including sore throat Follow‐up (range): 20 weeks to 6 months	329 (5 RCTs)	RR 0.71 (0.29 to 1.73)	6.9%	4.9% (2 to 12)	2.0% fewer (4.9 fewer to 5.1 more)	⊕⊕⊝⊝ LOW²	The evidence suggests that omalizumab may result in little to no difference in the incidence of nasopharyngitis, including sore throat. However, we have limited confidence in this estimate because the sample size may be too small to estimate this accurately. Nasopharyngitis or sore throat was reported by 8/170 participants who took omalizumab versus 11/159 who took placebo.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; CT: computerised tomography; RCT: randomised controlled trial; SMD: standardised mean difference; SNOT‐22: Sino‐Nasal Outcome Test‐22
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
¹Downgraded by one level due to imprecision: small sample size resulting in an imprecise estimate of effect size. ²Downgraded by two levels due to imprecision: small sample size for the outcome estimated resulting in an imprecise estimation of effect size; confidence interval includes potential for considerable benefit or considerable harm. ³Downgraded by one level due to indirectness: some serious adverse effects are relatively rare ‐ a larger and more heterogeneous population or longer period of treatment and follow‐up may be needed. ⁴Downgraded by one level due to study limitations: method of assessment not validated. ⁵Downgraded by one level due to inconsistency: high and unexplained heterogeneity as the size of effect differed between the studies (I² = 90%). ⁶Downgraded by one level due to inconsistency: high and unexplained heterogeneity as the size and direction of effect differed between the studies (I² = 80%). ⁷Downgraded by two levels due to imprecision: very small sample size for the outcome measured. ⁸Downgraded by one level due to indirectness: a larger range of treatment doses and duration, and a more heterogeneous population, may be required to identify the effect of the intervention on quality of life.

Background

This review is one of a suite of Cochrane Reviews looking at common management options for patients with chronic rhinosinusitis (Chong 2016a; Chong 2016b; Chong 2016c; Head 2016a; Head 2016b; Head 2016c; Head 2018).

Description of the condition

Chronic rhinosinusitis represents a common source of ill health; 11% of UK adults reported chronic rhinosinusitis symptoms in a worldwide population study (Hastan 2011). Symptoms including nasal obstruction, nasal discharge, facial pain, anosmia (loss of sense of smell) and sleep disturbance have a major impact on quality of life, reportedly greater in several domains of the SF‐36 than angina or chronic respiratory disease (Gliklich 1995). Acute exacerbations (worsening), inadequate symptom control and respiratory disease exacerbation are common. Complications are rare, but may include visual impairment and intracranial infection.

Two major phenotypes of chronic rhinosinusitis have been described based on the presence or absence of nasal polyps on examination. Nasal polyps are tumour‐like hyperplastic swellings of the nasal mucosa, most commonly originating from within the ostiomeatal complex (Larsen 2004). Chronic rhinosinusitis with nasal polyps (CRSwNP) is diagnosed when polyps are seen (on direct or endoscopic examination) in the middle meatus or nasal cavity. Chronic rhinosinusitis without nasal polyps (CRSsNP) is diagnosed when no polyps are observed on examination.

Although the aetiology of chronic rhinosinusitis is not fully understood, it may involve abnormalities in the host response to irritants, commensal and pathogenic organisms and allergens, obstruction of sinus drainage pathways, abnormalities of normal mucociliary function, loss of the normal mucosal barrier or infection. Chronic rhinosinusitis is a heterogeneous group of diseases, but three main patterns of inflammation have been identified: type 1 driven, usually associated with chronic rhinosinusitis without nasal polyps; type 2 driven, usually associated with chronic rhinosinusitis with nasal polyps in Caucasian patients; and type 17 driven, associated typically with chronic rhinosinusitis with nasal polyps in Asian patients (Smith 2018). There is some overlap between phenotypes and inflammatory patterns and the current division of chronic rhinosinusitis into two main phenotypes, with and without polyps, is therefore likely to be inadequate for defining patient subgroups. Endotyping, using measurable biomarkers, is increasingly being performed but is not yet routinely incorporated into clinical practice.

Despite the differences in aetiology and phenotype, in clinical practice many treatments for chronic rhinosinusitis are initiated without knowledge of a patient's 'polyp status'. Even when it is known whether or not a patient with chronic rhinosinusitis has polyps, this knowledge does not always suggest adjustments to treatment. This review (and most of its companion reviews) considers patients with and without polyps together in the initial evaluation of treatment effects. However, as biologics are primarily used in hospital settings and in well‐defined patient populations, we planned subgroup analyses to explore potential differences between them (see below).

Description of the intervention

The term 'biologics' refers to medicinal products produced by a biological process. Monoclonal antibodies are one type of biologic. They target specific inflammatory mediators or immune cells in the pathophysiological pathways that produce chronic inflammatory diseases. Trials have evaluated these agents in conditions such as asthma and atopic dermatitis leading to growing interest in the possibility of using them to treat patients with chronic rhinosinusitis.

How the intervention might work

Monoclonal antibodies work on different target substances or receptors in the inflammatory pathway. The more we understand about the inflammatory pathways involved in chronic rhinosinusitis, the more we may be able to affect those pathways with biologics. Different biologics are likely to have very different efficacy in different patient populations depending on the pattern of inflammation in those patients. Recent trials in patients with chronic rhinosinusitis with nasal polyps have focused on biologics directed at the inflammatory mediators and receptors involved in type 2 pathways. As yet none have investigated the effectiveness of biologics in type 1 or type 17 driven inflammation.

Currently, biologics are mainly used in patients with severe chronic rhinosinusitis where pharmacological therapy does not provide adequate symptom control, with the aim of reducing those symptoms and leading to an improvement in their quality of life. Some patients with severe chronic rhinosinusitis undergo surgical treatment aimed at achieving these goals. If patients respond well to biologics, surgical intervention may be avoided. If biologics are successful in reducing inflammation and reducing the size of nasal polyps, this should also be visible using endoscopy and computerised tomography (CT) scans. These changes can be documented and quantified using the relevant scoring system.

Biologics are, however, associated with adverse reactions that may be immune‐related and can be serious ‐ such as anaphylaxis. Biologics are widely used in rheumatology and some of the serious adverse events documented in those patients include tuberculosis reactivation, lymphoma and severe infections (Singh 2011; Tarp 2017). Another adverse reaction is pharyngitis, which may be serious enough for patients to discontinue treatment.

The following are descriptions of a number of classes and mechanisms of actions of monoclonal antibodies (mAb) with some specific named biologics. This is not an exhaustive list. The field is growing and our understanding of the mechanisms of action may change over time. Biologics not listed here may be evaluated in future updates of this review.

Anti‐IL‐4Rα mAb and anti‐IL‐13 mAb

Dupilumab, delivered by subcutaneous injection, is a human monoclonal antibody of the IgG4 subclass that targets the IL‐4Rα subunit and disrupts IL‐4 and IL‐13 signalling. This is involved in the type 2 inflammatory pathway most typically seen in patients with chronic rhinosinusitis with nasal polyps. Trials of dupilumab in asthma have also shown improvement in the symptoms of coexisting chronic rhinosinusitis (Wenzel 2016). Lebrikizumab and tralokinumab are anti‐IL‐13 monoclonal antibodies.

Anti‐IL‐5 mAb

Mepolizumab, reslizumab and benralizumab are delivered subcutaneously or intravenously, and are human monoclonal (IgG₁) antibodies targeting interleukin 5 (IL‐5) or the IL‐5 receptor α subunit on the surface of eosinophil white blood cells. IL‐5 promotes eosinophil development survival, so targeting IL‐5 reduces blood and tissue eosinophil counts. Mepolizumab is currently approved by the UK's National Institute for Health and Care Excellence (NICE) for the treatment of severe eosinophilic asthma and, as IL‐5 has been suggested as a parallel marker for the severity of both asthma and chronic rhinosinusitis with nasal polyps, it has the potential to treat both simultaneously (Chupp 2017; Dasgupta 2017; Pavord 2012). Reslizumab and benralizumab have had early success in patients with poorly controlled asthma (DuBuske 2018; Máspero 2017).

Anti‐IgE mAb

Omalizumab, also delivered subcutaneously, is a recombinant DNA‐derived humanised (IgG_1k) monoclonal antibody that specifically binds to free human immunoglobulin E (IgE) in the blood and interstitial fluid, and to the membrane‐bound form of IgE (mIgE) on the surface of mIgE‐expressing B‐lymphocytes. It therefore has the effect of reducing the levels of IgE in the serum and tissues, with a subsequent blocking of the IgE‐mediated inflammatory cascade. This anti‐IgE treatment has to date been shown to be effective in allergic rhinitis and asthma (Casale 2001; Hanania 2011).

Further information about the mechanisms of action of biologics in this field can be found in Kariyawasam 2019.

Why it is important to do this review

To date much of the literature around the role of these new drugs has been focused on the allergy, asthma and immunology subspecialties. As the role for biologic therapies in chronic rhinosinusitis continues to be defined and pharmaceutical companies are now targeting this condition, it is increasingly important for practising otorhinolaryngologists, especially sub‐specialist rhinologists, to determine the place of biologics in the treatment cascade by keeping up‐to‐date on their progression. NICE is currently conducting a health technology appraisal of the clinical and cost‐effectiveness of mepolizumab for chronic rhinosinusitis with nasal polyps (NICE 2020). This Cochrane Review looks at the balance of benefits and harms for biologic drugs in the treatment of patients with chronic rhinosinusitis. It also serves to identify areas for future research, especially as the knowledge of specific chronic rhinosinusitis endotypes increases.

This review is a living systematic review, whereby we search key databases monthly and update the review as and when new important evidence is found. A living systematic review approach is appropriate for this review because: 1) the topic is important for health care decision‐making; 2) there is uncertainty about the existing evidence; and 3) this is a rapidly developing field where new trials are being actively planned and completed. We revisit the scope (population, intervention, comparison, outcomes) of the review yearly, or more frequently as appropriate, to ensure that new agents or uses are included as this field develops. In addition to having more data on safety and efficacy, our understanding of how biologics work, the best way to measure outcomes and how outcomes are interpreted will very likely change as more research is completed. Therefore, we will adapt our definition of what outcomes to measure and how outcomes should be measured and interpreted over time.

Objectives

Main objective

To assess the effects of biologics for the treatment of chronic rhinosinusitis.

Secondary objective

To maintain the currency of the evidence, using a living systematic review approach.

Methods

Criteria for considering studies for this review

Types of studies

Randomised controlled trials (RCTs) and quasi‐randomised trials, where trials were designed as RCTs but the sequence generation for allocation of treatment used methods such as alternate allocation, birth dates, alphabetical order etc.

We only considered cross‐over trials if there was sufficient evidence to suggest that the condition of patients was stable and the washout period was adequate. Otherwise, we only planned to use the first phase of cross‐over trials.

We only included studies where patients were followed up for at least three months, to reflect the importance of focusing on long‐term outcomes for a chronic condition.

Types of participants

Patients with chronic rhinosinusitis, whether with polyps (CRSwNP) or without polyps (CRSsNP).

We excluded studies that had included a majority of patients with:

cystic fibrosis;
allergic fungal sinusitis/eosinophilic fungal/mucinous rhinosinusitis;
antrochoanal polyps (benign polyps originating from the mucosa of the maxillary sinus);
malignant polyps;
primary ciliary dyskinesia;
a history of surgery for nasal polyps within three months of entry to the study.

Types of interventions

Intervention

All monoclonal antibodies used for the treatment of chronic rhinosinusitis. This included but was not limited to the following:

anti‐IL‐4Rα mAb (dupilumab);
anti‐IL‐13 (lebrikizumab, tralokinumab);
anti‐IL‐5 mAb (reslizumab, benralizumab, mepolizumab);
anti‐IgE mAb (omalizumab).

These are the biologics identified in November 2019 as most likely to be used in patients with chronic rhinosinusitis; they were identified through a scoping project for this suite of reviews on chronic rhinosinusitis (Scoping report - chronic rhinosinusitis). Additional monoclonal antibodies and other classes of biologics will also be included in this review when they are evaluated in patients with chronic rhinosinusitis.

All routes of administration, doses and duration of treatment were included. However, studies should have followed up participants for three months or more.

Comparison

Placebo or no treatment. Surgery will be an alternative treatment (comparison) when trials in the area become available.

Concurrent treatments

It was expected that most studies would have used intranasal steroids as a concurrent treatment. There was no limitation on the type of pharmacological concurrent treatments used.

Comparison pairs

The following main comparison pairs were proposed in the protocol (Chong 2019):

anti‐IL‐4Rα mAb plus intranasal steroids versus placebo/no treatment plus intranasal steroids;
anti‐IL‐13 plus intranasal steroids versus placebo/no treatment plus intranasal steroids;
anti‐IL‐5 mAb plus intranasal steroids versus placebo/no treatment plus intranasal steroids;
anti‐IgE mAb plus intranasal steroids versus placebo/no treatment plus intranasal steroids.

Types of outcome measures

We analysed the following outcomes in the review, but we did not use them as a basis for including or excluding studies.

Our primary intention was to assess the effects of assignment, rather than adherence, to treatment.

Primary outcomes

Health‐related quality of life, using validated disease‐specific health‐related quality of life scores, such as the Sino‐Nasal Outcome Test‐22 (SNOT‐22), Rhinosinusitis Outcome Measures‐31 (RSOM‐31) and SNOT‐20.
Disease severity, as measured by validated patient‐reported symptom score (such as the Chronic Sinusitis Survey (CSS) questionnaire and visual analogue scales). Where this was unavailable, we considered including data measuring the severity of individual symptoms (see below).
Serious adverse events (SAEs), measured by the number of participants affected. A serious adverse event is defined as "Death, a life‐threatening adverse event, inpatient hospitalisation or prolongation of existing hospitalisation, a persistent or significant incapacity or substantial disruption of the ability to conduct normal life functions, or a congenital anomaly/birth defect. Important medical events that may not result in death, be life‐threatening, or require hospitalisation may be considered serious when, based upon appropriate medical judgment, they may jeopardize the patient or subject and may require medical or surgical intervention to prevent one of the outcomes listed in this definition" (FDA 2018).

Many studies within this suite of reviews (Chong 2016a; Chong 2016b; Chong 2016c; Head 2016a; Head 2016b; Head 2016c; Head 2018) did not use/present data using instruments that were either validated or evaluated all four types of symptoms meeting the EPOS 2012 diagnostic criteria in a composite score (nasal blockage or congestion or obstruction, nasal discharge, facial pain or pressure and loss or reduction of the sense of smell). If data from a validated score were unavailable, we planned to analyse data related to each of these individual symptoms, if presented.

Secondary outcomes

Avoidance of surgery, measured by the number (proportion) of participants who had, or did not have, surgery for chronic rhinosinusitis symptoms, or who no longer fulfilled the eligibility criteria for surgery*. (See comments in Assessment of risk of bias in included studies).
Extent of disease as measured by either:
- endoscopic score (depending on population, either nasal polyps size score or other such as Lund‐Kennedy); and/or
- computerised tomography (CT) scan score (e.g. Lund‐Mackay with a range of 0 to 24, higher = worse).
Health‐related quality of life, using generic quality of life scores, such as the SF‐36, EQ‐5D and other well‐validated instruments.
Adverse effects: nasopharyngitis, including sore throat.

Outcomes were measured at 3 to 6 months, 6 to 12 months and more than 12 months. For adverse events, we analysed data from the longest time periods.

*We recorded and tabulated the eligibility criteria for surgery used in the included studies.

Search methods for identification of studies

The Cochrane ENT Information Specialist has conducted systematic searches for randomised controlled trials and controlled clinical trials. There were no language, publication year or publication status restrictions. The date of the latest search was 28 September 2020.

Electronic searches

As a living systematic review, the Information Specialist has conducted monthly searches of:

the Cochrane ENT Trials Register (search via the Cochrane Register of Studies to 28 September 2020);
the Cochrane Central Register of Controlled Trials (CENTRAL) (search via the Cochrane Register of Studies to 28 September 2020);
Ovid MEDLINE(R) Epub Ahead of Print, In‐Process & Other Non‐Indexed Citations, Ovid MEDLINE(R) Daily and Ovid MEDLINE(R) (1946 to 28 September 2020);
Ovid Embase (1974 to 28 September 2020);
Web of Knowledge, Web of Science (1945 to 28 September 2020);
ClinicalTrials.gov, www.clinicaltrials.gov (search via the Cochrane Register of Studies to 28 September 2020);
World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) (search via the Cochrane Register of Studies to 28 September 2020).

The Information Specialist conducts quarterly searches of the following sources, and prior to the publication of any update:

ClinicalTrials.gov (search via www.clinicaltrials.gov to 30 July 2020);
World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) (searched to 30 July 2020).

Details of when each of the databases was searched and the date restrictions used are available in Appendix 1.

The Information Specialist modelled subject strategies for databases on the search strategy designed for CENTRAL, Ovid MEDLINE and Ovid Embase. Where appropriate, they were combined with subject strategy adaptations of the highly sensitive search strategy designed by Cochrane for identifying randomised controlled trials and controlled clinical trials (as described in the Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0, Box 6.4.b. (Handbook 2011). Search strategies for major databases including CENTRAL are provided in Appendix 2.

Biologics are a new class of intervention. The search strategy developed is highly sensitive, in order to try to capture new interventions as they are introduced. The Information Specialist reviews the search methods (the sources and search frequency) and the search terms (index terms and free text terms) on an annual basis. The aim is to include new terms for new interventions as they are introduced, and to remove terms to increase precision as interventions are removed or withdrawn.

Searching other resources

We scanned the reference lists of identified publications for additional trials and contacted trial authors where necessary. In addition, the Information Specialist searched Ovid MEDLINE to retrieve existing systematic reviews relevant to this systematic review, so that we could scan their reference lists for additional trials. The Information Specialist also searched of the Web of Knowledge Science Citation Index for articles referencing the published review (Chong 2020) and the primary reference to the included studies (Bachert 2016; Bachert 2017; Gevaert 2011; Gevaert 2013; LIBERTY SINUS 24; LIBERTY SINUS 52; Pinto 2010), except for NCT01066104, POLYP 1 and POLYP 2 as these were not indexed on the Web of Science Citation Index at the time of searching.

These searches were last conducted on 25 August 2020.

We contacted the principal investigators of ongoing trials and asked them to advise when results would be available, or to share early or unpublished data. No results have been shared as of 16 September 2020.

We did not perform a separate search for adverse effects. We considered adverse effects described in included studies only.

Clinical study reports (CSRs) and other sources of evidence

This review meets many of the 18 criteria for considering clinical study reports as a source of evidence (Jefferson 2018). In particular, there is a concern about publication bias with a new class of drugs for this current condition.

There are no established search procedures to identify clinical study reports at the time of publication. We attempted to identify unpublished studies and clinical study reports. The Information Specialist searched:

Regulatory bodies: We searched the websites of the:
- US Food and Drug Administration (FDA): http://www.fda.gov and https://www.fda.gov/about-fda/about-website/fdagov-archive (searched 11 December 2019);
- European Medicines Agency (EMEA) (http://www.emea.europa.eu) (searched 18 November 2019);
- European Union Clinical Trials Register (EUCRT) (https://www.clinicaltrialsregister.eu/) (searched 15 November 2019).
Manufacturer‐specific clinical trial repositories and data sharing platforms:
- Novartis Clinical Trial Results Database (https://www.novctrd.com) (searched 18 November 2019);
- GSK Study Register (https://www.gsk-studyregister.com) (searched 18 November 2019).
Direct requests to manufacturers: We did not identify additional trials and therefore did not write to the manufacturer/sponsors. We plan to contact the principal investigators/manufacturers/sponsors of each of the known trials individually to ask for additional data as part of the planned update of this living systematic review. We did identify one clinical study report (Bachert 2017) and additional data from ClinicalTrials.gov and EUCTR for five included studies (Bachert 2016; Bachert 2017; LIBERTY SINUS 24; LIBERTY SINUS 52; NCT01066104), which were identified as part of the regular electronic searches.

Living systematic review considerations

We review on an ongoing basis (and at least every six months) the various sources to search for clinical study reports, updating the list of sources searched and when as required.

We have a number of plans to investigate further the identification of clinical study reports and other sources of evidence. These are ongoing and are detailed in Differences between protocol and review. We plan to incorporate the results of the these efforts at the next update of this living systematic review.

Data collection and analysis

Selection of studies

The Cochrane ENT Information Specialist used Cochrane's Screen4Me workflow to help assess the initial search results for the first iteration of this living systematic review because of the high number of results retrieved from the database searches. Screen4Me comprises three components:

Known assessments – a service that matches records in the search results to records that have already been screened in Cochrane Crowd and been labelled as 'a RCT' or as 'not a RCT'.
The machine learning classifier (RCT model) (Wallace 2017), available in the Cochrane Register of Studies (CRS‐Web), which assigns a probability of being a true RCT (from 0 to 100) to each citation. For citations that are assigned a probability score below the cut‐point at a recall of 99% we have assumed these to be non‐RCTs. For those that score on or above the cut‐point we either manually dual screened these results or sent them to Cochrane Crowd for screening.
Cochrane Crowd is Cochrane's citizen science platform where the Crowd help to identify and describe health evidence. For more information about Screen4Me and the evaluations that have been done, please go to the Screen4Me website on the Cochrane Information Specialist's portal and see Marshal 2018, McDonald 2017, Noel‐Storr 2018 and Thomas 2017.

At least two review authors (LYC/PP/KS/SS), the Cochrane ENT Information Specialist (SC, listed in the Acknowledgements) or one of two methodologists (AK/KW, listed in the Acknowledgements) acting as one screener, independently screened the remaining titles and abstracts to identify potentially relevant studies. At least two review authors (MB/PP/KS/SS), one of the two Cochrane ENT methodologists (AT/KW, listed in the Acknowledgements) or Information Specialist (SC), listed in the Acknowledgements) independently evaluated the full text of each potentially relevant study to determine whether it met the inclusion/exclusion criteria for this review.

We resolved any differences by discussion and consensus, with the involvement of a third author (KS) for clinical and/methodological input where necessary.

Living systematic review considerations

We immediately collate and screen any new citations retrieved by the monthly searches using the approach outlined above including, as a first step in monthly screening, applying the Screen4Me workflow starting with the RCT model.

Data extraction and management

At least two review authors (MB/KS/SS/KW) or one author and one Cochrane ENT methodologist (AT, listed in the Acknowledgements) independently extracted outcome data from each study using a standardised data collection form (see Appendix 3). Whenever a study had more than one publication, we retrieved all publications to ensure complete extraction of data. Where there were discrepancies in the data extracted by different review authors, we checked these against the original reports and resolved differences by discussion and consensus, with the involvement of a third author (MB) or a methodologist (LYC) where appropriate. We contacted the original study authors for clarification or for missing data whenever possible. If differences were found between publications of a study, we contacted the original authors for clarification. We used data from the main paper(s) if no further information was found.

In addition, we also compared trials identified through study registers with identified publications. If an unpublished trial was identified (registered in trial registry, but more than 12 months since completion of recruitment and no data/incomplete data published), we contacted the contact person listed in the trial registry websites for information. Whenever clinical study reports or data from regulatory bodies are available, we will compare these against the journal reports and use them as the primary source of data if there is a discrepancy in the information. However, current experience with the use of clinical study reports suggests that there is often a considerable time lag between requesting these data and obtaining them. Therefore, we will make use of data from journal reports as the main source of evidence as a starting point and then check the data against the clinical study reports and regulatory data as and when these are available.

We included key characteristics of the studies, such as study design, setting, sample size, population and how outcomes were defined or collected in the studies. In addition, we also collected baseline information on prognostic factors or effect modifiers. For this review, this included:

presence or absence of nasal polyps;
polyp score (where applicable);
whether the patient has had previous sinus surgery.

The primary effect of interest is the effect of treatment assignment, which reflects the outcomes of treatment for people who were prescribed the intervention rather than per protocol analysis (the effect on people who completed the full course of treatment as planned). For the outcomes of interest to the review, we extracted the findings from the studies on an available case analysis basis, i.e. we included available data from all participants at the time points based on the treatment randomised whenever possible, irrespective of compliance or whether patients had received the treatment as planned.

In addition to extracting pre‐specified information about study characteristics and aspects of methodology relevant to risk of bias, we extracted the following summary statistics for each trial and each outcome:

For continuous data: the mean values, standard deviations and number of patients for each treatment group. Where endpoint data were not available, we extracted the values for change from baseline. We analysed data from measurement scales such as SNOT‐22 and EQ‐5D as continuous data.
For binary data: the number of participants experiencing an event and the number of patients assessed at the time point.
For ordinal scale data: if the data appeared to be approximately normally distributed or if the analysis that the investigators performed suggested parametric tests were appropriate, then we treated the outcome measures as continuous data. Alternatively, if data were available, we planned to convert into binary data.

We pre‐specified the time points of interest for the outcomes in this review. While studies may report data at multiple time points, we only extracted the longest available data within the time points of interest. For example, for 'short' follow‐up periods, our time point was defined as three to six months post‐randomisation. If a study reported data at three, four and six months, we only extracted and analysed the data for the six‐month follow‐up.

Assessment of risk of bias in included studies

Two review authors (KS/SS) or a Cochrane ENT methodologist (AT, listed in the Acknowledgements) independently assessed the risk of bias of each included study.

In the first and current version of the review, we have used the original version of the Cochrane 'Risk of bias' tool (ROB‐1) (Handbook 2011). For future versions of this living systematic review, we anticipate using the Cochrane 'Risk of bias 2.0' tool (ROB‐2) (Sterne 2019), according to the guidance in the latest version of the Cochrane Handbook for Systematic Reviews of Interventions (version 6; Handbook 2019).

When using the ROB‐1 tool, we followed the guidance in the Cochrane Handbook for Systematic Reviews of Interventions (version 5; Handbook 2011). We assessed the risk of bias as 'low', 'high' or 'unclear' for each of the following six domains:

sequence generation;
allocation concealment;
blinding of participants, personnel and outcome assessment;
incomplete outcome data;
selective reporting;
other sources of bias (if required).

In future iterations of this living systematic review, we plan to apply the ROB‐2 tool (rather than ROB‐1) according to the guidance in the latest version of the Cochrane Handbook for Systematic Reviews of Interventions (Handbook 2019). We will assess the risk of bias as 'low', 'high' or 'some concerns' for each of the following five domains:

risk of bias arising from the randomisation process;
risk of bias due to deviations from the intended interventions;
risk of bias due to missing outcome data;
risk of bias in measurement of outcome;
risk of bias in selection of the reported result.

For ROB‐2, we will only assess the outcomes included in the 'Summary of findings' table.

For the outcome 'disease severity, as measured by validated patient‐reported symptom score' we will only conduct a ROB‐2 assessment if this is reported. If only the results from individual symptoms, or non‐validated scores, are reported we will not individually assess these, as the risk of bias is likely to be present due to the choice of outcome measure and selective reporting of only certain aspects of the condition.

There is a particular risk of bias in assessing the outcome 'avoidance of surgery', as there are no widely accepted criteria to determine when patients should or should not have surgery. Unless studies explicitly specify what criteria are used for making judgements and both the investigator (offering/deciding on the surgery) and participants were blinded, there are potential biases in the decision‐making process of the study personnel in determining whether or not a participant fulfils the criteria for surgery and/or whether they should be offered the option of surgery. We assessed this in the 'Blinding, outcomes assessment' domain using the ROB‐1 tool and we will assess this in the 'Risk of bias in the measurement of outcome' domain when we are using the ROB‐2 tool.

Measures of treatment effect

We summarised the effects of dichotomous outcomes (e.g. proportion of patients with symptom resolution) as risk ratios (RR) with 95% confidence intervals (CIs). For the key outcomes that we presented in the 'Summary of findings' tables, we also expressed the results as absolute numbers based on the pooled results and compared to the assumed risk. If appropriate, we would also have considered calculating the number needed to treat to benefit (NNTB) using the pooled results. The assumed baseline risk is typically either (a) the median of the risks of the control groups in the included studies, this being used to represent a 'medium‐risk population' or, alternatively, (b) the average risk of the control groups in the included studies is used as the 'study population' (Handbook 2019). If a large number of studies are available, and where appropriate, we may also present additional data based on the assumed baseline risk in (c) a low‐risk population and (d) a high‐risk population.

For continuous outcomes, we expressed treatment effects as a mean difference (MD) with standard deviation (SD) or as a standardised mean difference (SMD) if different scales had been used to measure the same outcome. We provided a clinical interpretation of the SMD values using either Cohen's d or by conversion to a recognised scale if possible.

Unit of analysis issues

Cross‐over trials and cluster‐randomised trials are unlikely for this review topic. We did not plan to use data from phase II of cross‐over studies (unless there was sufficient evidence to suggest that the condition of patients was stable and the washout period was adequate). If these trial designs are found and deemed suitable to use in the future, we will seek advice from the Cochrane Bias Methods Group and use the latest version of the ROB‐2 tool for cross‐over and cluster‐randomised trials.

We expected that studies would take multiple measurements or observations of a single outcome in the same patients (repeated measurements). In these situations, we only extracted and analysed the data point for the longest available follow‐up specified in our protocol (Chong 2019).

Dealing with missing data

We tried to contact study authors via email whenever the outcome of interest was not reported, if the methods of the study suggest that the outcome had been measured. We did the same if not all data required for meta‐analysis had been reported, unless the missing data were standard deviations. If standard deviation data were not available, we approximated these using the standard estimation methods from P values, standard errors or 95% CIs where reported, as detailed in the Cochrane Handbook for Systematic Reviews of Interventions (Handbook 2019). If it was impossible to estimate these, we planned to contact the study authors.

Apart from imputations for missing standard deviations, we conducted no other imputations. We extracted and analysed all data using the available case analysis method.

Assessment of heterogeneity

We assessed clinical heterogeneity (which may be present even in the absence of statistical heterogeneity) by examining the included trials for potential differences between studies in the types of participants recruited, interventions or controls used and the outcomes measured.

We assessed statistical heterogeneity by visually inspecting the forest plots and by considering the Chi² test (with a significance level set at P value < 0.10) and the I² statistic, which calculates the percentage of variability that is due to heterogeneity rather than chance (Handbook 2019).

Assessment of reporting biases

We assessed reporting bias as between‐study publication bias and within‐study outcome reporting bias.

Outcome reporting bias (within‐study reporting bias)

We assessed within‐study reporting bias by comparing the outcomes reported in the published report against the study protocol, whenever this could be obtained. If the protocol or trial registry entry was not available, we compared the outcomes reported to those listed in the methods section. If results are mentioned but not reported adequately in a way that allows analysis (e.g. the report only mentions whether the results were statistically significant or not), bias in a meta‐analysis is likely to occur. We sought further information from the study authors. If no further information could be found, we planned to note this as being a 'high' risk of bias when the ROB‐1 tool was used. If there was insufficient information to judge the risk of bias we noted this as an 'unclear' risk of bias (Handbook 2011). When the ROB‐2 tool is used in the future, we will assess selective reporting bias in a similar way, according to the signalling questions in the 'risk of bias in selection of the reported result' domain (Handbook 2019). However, we will assess selective non‐reporting bias at the synthesis level, using the latest tools (e.g. ROB‐ME) if available.

Publication bias (between‐study reporting bias)

We planned to assess funnel plots if sufficient studies (more than 10) were available for an outcome. If we had observed asymmetry of the funnel plot, we would have conducted more formal investigation using the methods proposed by Egger 1997. We also report on whether there were any studies identified through trial registries and other sources (Searching other resources), with unpublished reports.

Data synthesis

We conducted all meta‐analyses using RevMan Web (RevMan Web 2019). For dichotomous data, we planned to analyse treatment differences as a risk ratio (RR) calculated using the Mantel‐Haenszel methods.

For continuous outcomes, if all the data were from the same scale, we pooled mean values obtained at follow‐up with change outcomes and reported this as a MD. However, if the SMD had to be used as an effect measure, we did not pool change and endpoint data.

We proposed using a random‐effects model since it was likely that there would be clinical heterogeneity in the response to different types of biologics or different types of monoclonal antibodies. However, we also planned to undertake a sensitivity analysis to examine the effects of using the alternative fixed‐effect model.

Living systematic review considerations

When new evidence will be incorporated into the living systematic review

Whenever new evidence (meaning studies, data or information) relevant to the review is identified, we extract the data and assess risk of bias, as appropriate. We immediately incorporate any important new evidence into the review.

We do not adjust the meta‐analyses to account for multiple testing, given that the methods related to frequent updating of meta‐analyses are under development (Simmonds 2017). We do not use sequential methods for updated meta‐analyses (Handbook 2019).

Subgroup analysis and investigation of heterogeneity

When studies had a mixed group of patients, we planned to analyse the study as one subgroup (rather than as a mixed group) if more than 80% of patients belonged to one category. For example, if 81% of patients had chronic rhinosinusitis without nasal polyps, we would analyse the study as that subgroup.

We planned to conduct subgroup analyses based on the phenotypes of patients (whether patients had chronic rhinosinusitis with or without nasal polyps, are a mixed group or the status of polyps is not known or not reported) regardless of whether statistical heterogeneity was observed, as these are widely suspected to be potential effect modifiers. Although there appears to be a considerable overlap between the two forms of chronic rhinosinusitis with regards to inflammatory profile, clinical presentation and effect of treatment (Cho 2012; DeMarcantonio 2011; Ebbens 2010; EPOS 2007; Ragab 2004; Ragab 2010; van Drunen 2009), there is some evidence pointing to differences in the respective inflammatory profiles (Kern 2008; Keswani 2012; Tan 2011; Tomassen 2011; Zhang 2008; Zhang 2009), and potentially even differences in treatment outcome (Ebbens 2011).

We planned to present this as the main subgroup analysis for effectiveness outcomes in this review. We planned to present all other subgroup analysis results in tables.

In addition to subgrouping by phenotype, we planned to conduct the following subgroup analyses in the presence of statistical heterogeneity:

Patients with asthma as a comorbidity. Patients with asthma may have different inflammatory markers and respond differently. In addition to chronic rhinosinusitis symptoms, they may also benefit from better control of asthma symptoms. However, there are no clear data to tell us which patients will benefit more or less from certain types of biologics, therefore the direction of effects is unclear.
Patients with non‐steroidal anti‐inflammatory drug (NSAID)‐exacerbated respiratory disease (N‐ERD). The rationale is similar to that for patients with asthma as a comorbidity.
Treatment regimens. For agents acting on the same target substance or receptor, treatment regimens such as dose and frequency of initial treatment and maintenance treatment are likely to be important. However, at the preparation of the protocol in 2019 there was not enough information to inform how these subgroups should be defined. We will revisit this question as part of our regular re‐evaluation of the review methods, as and when more data are available from trials.

As the vast majority of participants in the included studies were diagnosed with chronic rhinosinusitis with nasal polyps (1260 out of 1262), we were unable to conduct subgroup analysis according to the phenotype of patients, and the data reported relate to individuals who have both chronic rhinosinusitis and nasal polyps. Furthermore, because of the small number of included studies and sparse data for each comparison, we were unable to conduct meaningful subgroup analysis for the additional subgroup categories (asthma, N‐ERD and treatment regimens).

Sensitivity analysis

We planned to carry out sensitivity analyses to determine whether the findings are robust to the decisions made in the course of identifying, screening and analysing the trials. We planned to conduct sensitivity analysis for the following factors, if there were relevant data to do so:

risk of bias of included studies: excluding studies with high risk of overall bias for the results, as assessed using the Cochrane ROB‐1 and ROB‐2 tools;
impact of model chosen: fixed‐effect versus random‐effects model;
how outcomes were measured: we planned to investigate the impact of including data where the validity of the measurement was unclear.

If any of these investigations found a difference in the size of the effect or heterogeneity, we would mention this in the 'Effects of interventions' section. However, there were insufficient studies and data meeting these criteria and these analyses were not required.

Summary of findings and assessment of the certainty of the evidence

We used the GRADE approach to rate the overall certainty of evidence for each outcome using the GDT tool (https://gradepro.org/) for the main comparison pairs listed in the Types of interventions section. The certainty of evidence reflects the extent to which we are confident that an estimate of effect is correct and we applied this in the interpretation of results. There are four possible ratings: 'high', 'moderate', 'low' and 'very low'. A rating of 'high' certainty evidence implies that we are confident in our estimate of effect and that further research is very unlikely to change our confidence in the estimate of effect. A rating of 'very low' certainty implies that any estimate of effect obtained is very uncertain.

The GRADE approach rates evidence from RCTs that do not have serious limitations as high certainty. However, several factors can lead to the downgrading of the evidence to moderate, low or very low. The degree of downgrading is determined by the seriousness of these factors:

study limitations (risk of bias);
inconsistency;
indirectness of evidence;
imprecision;
publication bias.

The 'Summary of findings' tables present only the seven top priority outcomes (primary outcomes: disease‐specific health‐related quality of life, disease severity as measured by validated patient‐reported symptom score, serious adverse events (SAEs) and secondary outcomes: avoidance of surgery, extent of disease as measured by endoscopic score or CT scan score, generic health‐related quality of life and other adverse effects).

Methods for future updates

We will review the scope and methods of this review approximately yearly (or more frequently if appropriate) in the light of potential changes in the topic area, or the evidence being included in the review (for example, additional comparisons, interventions or outcomes, or new review methods available).

Conditions under which the review will no longer be maintained as a living systematic review

The review will no longer be maintained as a living systematic review once there is high‐certainty evidence obtained for the primary effectiveness outcomes of the review; new studies are not expected to be conducted regularly for the interventions included in this review; or the review topic is no longer a priority for health care decision‐making.

Results

Description of studies

Results of the search

Update searches (September 2019 to September 2020)

As of 28 September 2020 we have performed seven update searches (March, April, May, June, July, August and September 2020). These searches retrieved a total of 7065 records. This reduced to 4263 after removal of duplicates. The Cochrane ENT Information Specialist sent all 4263 records to the Screen4Me workflow. The Screen4Me workflow identified 210 records as having been previously assessed: 61 had been rejected as not RCTs and 149 had been assessed as possible RCTs. The RCT classifier rejected an additional 1460 references as not RCTs (with 99% sensitivity). We did not send any records to the Cochrane Crowd for assessment. Following this process, the Screen4Me workflow had therefore identified 2742 possible RCTs for title and abstract screening.

	Possible RCTs	Rejected
Known assessments	149	61
RCT classifier	2593	1460
Cochrane Crowd	n/a	n/a
Total	2742	1482

For further details of this process please see Selection of studies.

We subsequently identified 2235 additional duplicates, leaving 507 records to screen.

We screened the titles and abstracts of the remaining 507 references. We discarded 466 records and assessed 41 in full text. We linked 10 records to existing studies. Three additional duplicates were identified during screening.

Subsequently, we moved two completed studies from 'ongoing' to 'included' (POLYP 1; POLYP 2). We added five more ongoing studies (seven records) (EUCTR2020‐000421‐76; NAPPREB; NCT04362501; NCT04430179; ORCHID). We also identified additional data for the included study Bachert 2016. We added two more records to studies awaiting classification and excluded a further 19 records with reasons (see Excluded studies).

We also identified four studies from our supplementary searches that were subsequently excluded with reasons.

Original searches (September 2019)

The original searches (September 2019) retrieved a total of 4914 records. This reduced to 3341 after the removal of duplicates. The Cochrane ENT Information Specialist sent all 3341 records to the Screen4Me workflow. The Screen4Me workflow identified 399 references as having been previously assessed: 179 had been rejected as not RCTs and 220 had been assessed as possible RCTs. The RCT classifier rejected an additional 1253 records as not RCTs (with 99% sensitivity). The Cochrane Crowd assessed the remaining 1689 references, rejecting 1046 as not RCTs and identifying 643 as possible RCTs. Following this process, the Screen4Me workflow had therefore identified 863 possible RCTs for title and abstract screening.

	Possible RCTs	Rejected
Known assessments	220	179
RCT classifier	n/a	1253
Cochrane Crowd	643	1046
Total	863	2478

For further details of this process please see Selection of studies.

We subsequently identified six additional duplicates, leaving 857 references to screen.

We screened the titles and abstracts of the remaining 857 references. We discarded 778 references and assessed 79 full‐text articles. We discarded three additional references at the full‐text screening stage and identified one additional duplicate.

For all searches

We excluded 54 of these references (41 studies) with reasons recorded in the review (see Excluded studies).

We included 10 completed studies, where results were available (46 references) (Bachert 2016; Bachert 2017; Gevaert 2011; Gevaert 2013; LIBERTY SINUS 24; LIBERTY SINUS 52; NCT01066104; Pinto 2010; POLYP 1; POLYP 2). NCT01066104 is an unpublished study (no journal publications or abstracts found), but the results of the study were available on the clinicaltrials.gov website.

There is one reference to one study that completed in March 2017 where the results have not yet been published and no information on the findings is available on clinicaltrials.gov (NCT02772419). The study was conducted by Kyowa Kirin Co. Ltd. The company confirmed on 7 January 2019 that the study is complete and that they are considering publication of the results. We requested access to the study results or clinical study report on 7 January 2019. The response from Kyowa Kirin is shown in Appendix 4. This study is classified as ongoing.

We identified another 10 studies (14 references) that we classified as ongoing. Five studies are due to be completed during 2020 (NCT02799446; NCT03450083; NCT03614923; OSTRO; SYNAPSE). One study is due for completion in 2021 (NAPPREB), two studies are due for completion in 2022 (NCT04430179; ORCHID), one study is due to be completed in 2023 (NCT04362501) and one study registered in March 2020 does not state its completion date (EUCTR2020‐000421‐76).

See Characteristics of ongoing studies for further details of all 10 studies.

A flow chart of study retrieval and selection is provided in Figure 1.

Figure 1

PRISMA flow diagram

Included studies

We included a total of 10 completed RCTs (Bachert 2016; Bachert 2017; Gevaert 2011; Gevaert 2013; LIBERTY SINUS 24; LIBERTY SINUS 52; NCT01066104; Pinto 2010; POLYP 1; POLYP 2). All the studies were sponsored or supported by industry.

A summary of key participant characteristics, interventions, comparison pairs and outcomes measured and reported is provided in Table 1.

Table 1. Summary of characteristics of included studies

	SINUS 24 (n = 276)	SINUS 52 (n = 448)	Bachert 2016 (n = 60)	Bachert 2017 (n = 107)	Gevaert 2011 (n = 30)	Pinto 2010 (n = 14)	Gevaert 2013 (n = 24)	NCT01066104 (n = 27)	POLYP 1 (n = 138)	POLYP 2 (n = 127)
Population	Bilateral nasal polyps (mean 5.75 points) with symptoms of chronic rhinosinusitis despite intranasal steroids	Bilateral nasal polyps (mean 6.10 points) with symptoms of chronic rhinosinusitis despite intranasal steroids	Chronic sinusitis with nasal polyps (mean 5.8 points)	Severe, recurrent bilateral nasal polyposis requiring surgery (worst affected nostril ≥ 3 (on 4‐point scale), and symptoms score > 7 on 10 cm VAS despite intranasal steroids and/or previous oral corticosteroids Mean bilateral polyp score 6.29	Chronic rhinosinusitis with severe primary polyps (grade 3 to 4) or recurrent polyps (any grade) Failure of standard care for chronic rhinosinusitis	Chronic rhinosinusitis Polyps status: 7/7 in omalizumab and 5/7 in placebo had nasal polyposis	Chronic rhinosinusitis with nasal polyps Polyps status: TPS (total nasal endoscopic polyp score), median (IQR): 6 (4 to 6); 6 (6 to 8)	Chronic rhinosinusitis with nasal polyps Inclusion criteria state minimum polyp score of 4	Chronic rhinosinusitis with nasal polyps Inclusion criteria state minimum polyp score of 5	Chronic rhinosinusitis with nasal polyps Inclusion criteria state minimum polyp score of 5
Comorbidity	Asthma 58%	Asthma 60%	Asthma 58%	Asthma 78%	Asthma 43%	Inhaled asthma therapy taken by 72% (5/7) in omalizumab group and 43% (3/7) in placebo group	Asthma (100%)	No information	Asthma 54%	Asthma 60%
Eligible for surgery?	No information	No information	No information	Yes^a	No information	100% had undergone endoscopic sinus surgery, but no information on eligibility for more surgery	No information	No information	No information	No information
Intervention	Dupilumab 300 mg subcutaneously every 2 weeks	a) Dupilumab 300 mg subcutaneously every 2 weeks for 24 weeks, followed by every 4 weeks until 52 weeks b) Dupilumab 300 mg subcutaneously every 2 weeks for 52 weeks in total	Dupilumab 600 mg loading dose subcutaneously, followed by 300 mg every week	Mepolizumab 750 mg intravenously every 4 weeks	Mepolizumab 750 mg intravenously every 4 weeks	Omalizumab subcutaneously, once or twice monthly (dose dependent on participant weight and serum IgE level), for 6 months	Omalizumab subcutaneously every 2 weeks (8 injections in total) or every month (4 injections in total), based on total serum IgE levels and body weight, with a maximum dose of 375 mg	Omalizumab subcutaneously, every 2 to 4 weeks depending on baseline serum total IgE level and body weight	Omalizumab 75 mg to 600 mg subcutaneously, every 2 to 4 weeks depending on baseline serum total IgE level and body weight	Omalizumab 75 mg to 600 mg subcutaneously, every 2 to 4 weeks depending on baseline serum total IgE level and body weight
Comparison	Placebo subcutaneously every 2 weeks	Placebo subcutaneously every 2 weeks	Placebo subcutaneously every week	Intravenous placebo every 4 weeks	Intravenous placebo every 4 weeks	Placebo injection, same dose and frequency	Placebo injection, same dose and frequency	Stated as "Xolair placebo 150‐375 mg depending on baseline serum total IgE level and body weight"	Placebo injection at corresponding dose and frequency	Placebo injection at corresponding dose and frequency
Treatment length	24 weeks	52 weeks	15 weeks	24 weeks	8 weeks (2 doses)	26 weeks	16 weeks	22 weeks	24 weeks	24 weeks
Follow‐up length (total treatment and follow‐up period)	24 weeks	24 weeks and 52 weeks	16 weeks	25 weeks	48 weeks (most outcomes assessed after 8 weeks' treatment)	26 weeks	20 weeks (outcomes assessed after 16 weeks' treatment)	22 weeks	28 weeks (most outcomes assessed after 24 weeks' treatment)	28 weeks (most outcomes assessed after 24 weeks' treatment)
Specific HRQL	Measured and reported^b	Measured and reported^b	Measured and reported^b	Measured and reported^b	Not measured	Measured and reported^b	Measured and reported^c	Not measured	Measured and reported^b	Measured and reported^b
Disease severity (overall)	Measured and reported^d,e	Measured and reported^d,e	Measured and reported^d,j	Measured and reported^d	No global questionnaire reported Specific symptoms measured and reported^f	No global questionnaire reported Specific symptoms measured and reported^g,h	No global questionnaire reported Specific symptoms measured and reportedⁱ	No global questionnaire reported Measured but not reported^k	No global questionnaire reported Specific symptoms measured and reported^aa	No global questionnaire reported Specific symptoms measured and reported^aa
Severe adverse event	Measured and reported	Measured and reported	Measured and reported	Measured and reported	Measured and reported	Measured and reported	Not measured	Measured and reported	Measured and reported	Measured and reported
Avoidance of Surgery	Measured and reported^l,m	Measured and reported^l,n	Not measured	Measured and reported^o	Not measured	Not measured	Not measured	Not measured	Measured and reported^bb	Measured and reported^bb
CT scan	Measured and reported^p	Measured and reported^p	Measured and reported^p	Not measured	Measured and reported^q	Measured and reported^r	Measured and reported^p	Measured and reported^s	Not measured	Not measured
Polyps score	Measured and reported^t	Measured and reported^t	Measured and reported^t	Measured and reported^u	Measured and reported^t	Measured and reported^v	Measured and reported^t	Measured and reported^t	Measured and reported^t	Measured and reported^t
Generic HRQL	Measured and reported^w,m	Measured and reported^w,m	Measured and reported^w	Measured and reported^w,x	Not measured	Measured and reported^y	Measured and reported^y	Not measured	Measured, not reported^cc	Not measured
Nasopharyngitis	Measured and reported	Measured and reported	Measured and reported	Measured and reported	Measured and reported	Not measured^z	Not measured	Not measured	Measured and reported^dd	Measured and reported
Main data source	Publications; generic health‐related quality of life and avoidance of surgery data from trial registry only	Publications; generic health‐related quality of life and avoidance of surgery data from trial registry only	Publications	Publications	Publications	Publication	Publication	NCT record (no publications)	Publication	Publication Nasopharyngitis data for POLYP 2 alone (not pooled with POLYP 1) from NCT record

^aWorst affected nostril ≥ 3 (on a 4‐point scale), and symptoms score > 7 on 10 cm VAS despite intranasal steroids and/or previous oral corticosteroids.

^bSNOT‐22, scale 0 to 110, higher = worse, minimal clinically important difference (MID) ≥ 8.9 points.

^cRSOM‐31; AQLQ.

^dVisual analogue scale for rhinosinusitis: "how troublesome are your symptoms?", scale 0 to 10 cm, higher = worse.

^eTotal symptom severity score (including nasal congestion, rhinorrhoea and sense of smell, each rated between 0 and 3), total scale 0 to 9, higher = worse.

^fFour individual symptoms were measured (anterior rhinorrhoea, nasal obstruction, postnasal drip and loss of sense of smell); reported only as narrative summary.

^gTotal nasal symptom score (TNSS): nasal obstruction, nasal discharge, facial pain and hyposmia) each recorded on a 4‐point scale (0 = none, 1 = mild, 2 = moderate, 3 = severe); total scores summed.

^hOnly reported as 'no significant difference' ‐ no data presented.

ⁱDisease severity symptom score: nasal and asthma symptoms (patient‐reported, daily 'absent, mild, moderate or severe' (scores 0, 1, 2, 3).

^jSeverity scores for individual symptoms (nasal congestion, anterior and posterior rhinorrhoea, loss in sense of smell, nocturnal awakenings), range 0 to 3, higher = worse.

^kNCT record states that a total symptom score (TSS) and chronic rhinosinusitis facial pain/headache questionnaire were recorded daily; no outcome data presented in NCT record.

^lNumber of participants requiring rescue with nasal polyp surgery ‐ no definition for eligibility provided.

^mOutcome reported, but specific data only reported in trial registry (publication includes pooled data with SINUS 52 only).

ⁿOutcome measured but not reported (pooled data with SINUS 24 only, specific data for this trial not reported on trial registry or publication).

^oAt study endpoint, participants with a nasal polyp score of ≥ 3 were deemed as continuing to need surgery (regardless of VAS score). In addition, participants with a nasal polyp score of 2, who had a VAS score of > 7 were also viewed as requiring surgery.

^pLund‐Mackay CT score, range 0 to 24, higher = worse.

^qPublication reports proportion of participants who showed improvement in CT score during the study. Shown separately for three independent raters, with no summary measure reported.

^rMucosal thickness on CT scan.

^sCT scan scored using the Zinreich modification of the Lund‐Mackay scoring system.

^tBilateral "endoscopic nasal polyps score" (NPS) or total polyps score (TPS), range 0 to 8, higher = worse.

^uImprovement by at least one point in endoscopic nasal polyp score.

^vNasal endoscopy score (0 to 4). Unclear which scoring system used.

^wEQ‐5D visual analogue scale, range 0 to 100 (100 = best imaginable).

^xEQ‐5D index score, range 0 to 1, higher = better.

^ySF‐36.

^zOutcome not specifically mentioned, paper just states "No side effects or adverse events occurred during the study".

^aaTotal Nasal Symptom Score and individual components of this were reported, which included anterior rhinorrhoea, posterior rhinorrhoea, nasal congestion and loss of sense of smell. Each scored with a range of 0 to 3, higher = worse. Total score out of 12.

^bbAvoidance of surgery was defined as an improvement in SNOT‐22 score of at least 8.9 points and a nasal polyp score no greater than 4 points (with a unilateral score of no more than 2 on either side).

^ccProtocol states that EuroQol 5‐Dimension 5‐Level Questionnaire will be used, but results not reported.

^ddNasopharyngitis reported as pooled data with POLYP 2; however the data for POLYP 2 are also reported separately, therefore individual data for POLYP 1 can be calculated.

Study design

All studies were double‐blind RCTs and used a placebo. The shortest planned duration was eight weeks (Gevaert 2011), the longest was 52 weeks (LIBERTY SINUS 52). One study was stopped early and only had 14 participants (Pinto 2010). Some studies were phase II or proof of concept studies and had fewer than 30 patients in each treatment arm (Gevaert 2011; Gevaert 2013; NCT01066104; Pinto 2010).

Participants

A total of 1262 participants were included. With the exception of two participants in one study (Pinto 2010), all the participants were adults with chronic rhinosinusitiswith nasal polyps and a significant number of participants (43% to 100%) also had asthma as a co‐morbidity.

Interventions and comparisons

Studies were available to evaluate three of our four proposed comparison pairs. (No studies assessed the comparison anti‐IL‐13 plus intranasal steroids versus placebo/no treatment plus intranasal steroids). All studies compared a biologic against placebo and all participants received intranasal corticosteroids.

Comparison 1: Anti‐IL‐4Rα mAb versus placebo/no treatment (all receiving intranasal steroids)

Three RCTs (784 participants) investigated dupilumab 300 mg versus placebo.

LIBERTY SINUS 24 (276 participants) gave 300 mg (subcutaneous) dupilumab every two weeks and followed up patients for 24 weeks.
LIBERTY SINUS 52 (448 participants) randomised patients 1:1:1 into three arms (two dupilumab arms and one placebo arm): 300 mg subcutaneous dupilumab every two weeks for 52 weeks, or 300 mg subcutaneous dupilumab every two weeks for 24 weeks followed by 300 mg subcutaneous dupilumab every four weeks for another 28 weeks. The total period of follow‐up was 52 weeks and results were reported for both week 24 and 52. The study had prespecified that some of the data would be pooled across both studies and/or both treatment arms of dupilumab, and did not report the results of the individual trials separately. For the purpose of this review, we combined the results of the different dupilumab arms in the LIBERTY SINUS 52 study, but reported the results of SINUS‐52 and SINUS‐24 independently by using the data presented in trial registries whenever possible.
Bachert 2016 (60 participants) gave a 500 mg subcutaneous loading dose of dupilumab followed by 300 mg subcutaneous weekly for 15 weeks.

Comparison 2: Anti‐IL‐5 mAb versus placebo/no treatment (all receiving intranasal steroids)

Two RCTs were found for this comparison.

Bachert 2017 (107 participants).
Gevaert 2011 (30 participants).

Both studied mepolizumab 750 mg intravenously every four weeks for 24 weeks.

Comparison 3: Anti‐IgE mAb versus placebo/no treatment (all receiving intranasal steroids)

Five RCTs were found for this comparison.

POLYP 1 (138 participants).
POLYP 2 (138 participants).
Gevaert 2013 (24 participants).
NCT01066104 (27 participants).
Pinto 2010 (14 participants).

All studied subcutaneous omalizumab, at a dose dependent on the participants' weight and other characteristics, every two or four weeks for between 16 weeks and six months.

Outcomes

1. Health‐related quality of life (HRQL), using validated disease‐specific HRQL scores

Most studies measured and reported the SNOT‐22. Two did not: Gevaert 2011 and NCT01066104. SNOT‐22 has a range of 0 to 110 and the minimal clinically important difference (MCID) is 8.9 points (Hopkins 2009).

2. Disease severity, as measured by validated patient‐reported symptom score (such as the CSS questionnaire or visual analogue scales)

LIBERTY SINUS 24 used a 0 to 10 cm visual analogue scale (VAS) to measure overall (global) symptoms ("How troublesome are your symptoms?", 0 = "not troublesome", 10 = "worst thinkable troublesome"). Other studies either did not provide details or reported some variation in how the question was asked. Bachert 2017 reported using a VAS of 0 to 10 with the question, "How troublesome are your symptoms of nasal polyposis?", 0 = "not troublesome", 10 = "worst possible". These studies generally made reference to the recommendation in EPOS 2007 to use a VAS, but did not report whether or not the format or wording of the questions they used in the trials had been validated.

Other measures such as "total symptom score" (with a scale range of 0 to 9 points) or "total nasal symptoms score" (with a scale range of 0 to 12 points) were used by some studies. However, these scales only measured symptoms of rhinitis (posterior and anterior rhinorrhoea), loss of sense of smell and nasal blockage rather than the overall symptom score of chronic rhinosinusitis, and there was no evidence of validation. Data from these scales, and on those relating to specific, individual symptoms, are not considered in our meta‐analysis as they are not global symptom scores. For future updates of this review we intend to incorporate data from individual symptom scores in addition to the global symptom scores that are already included.

3. Serious adverse events

Most studies used the definition of treatment‐emergent serious adverse events, where the events and participants were accounted for according to the treatment actually received (rather than by randomised group) and at least one dose was taken.

4. Avoidance of surgery

A few studies attempted to measure the degree of improvement (or non‐improvement) experienced by participants, by identifying those participants who required some form of surgery to alleviate their symptoms. This took the form of determining the number of patients who required some form of 'rescue surgery', or the number of patients who met (or no longer met) the criteria for surgery. There are many issues and potential risks of bias associated with this measure. Table 2 summarises information for each included study about (a) whether or not the eligibility for surgery was defined at randomisation, and (b) in studies where the need for surgery was an 'outcome', what were the criteria for surgery in those circumstances?

See: Summary of findings 1 Anti‐IL‐4Rα mAb (dupilumab) compared to placebo (on top of topical steroids) for chronic rhinosinusitis; Summary of findings 2 Anti‐IL‐5 mAb (mepolizumab) compared to placebo (on top of topical steroids) for chronic rhinosinusitis; Summary of findings 3 Anti‐IgE mAb (omalizumab) compared to placebo (on top of topical steroids) for chronic rhinosinusitis

Table 2. Eligibility for surgery

Study name	Study	Eligibility for surgery: defined at randomisation?			Eligibility criteria for surgery: as recorded in results
Study name	Study	Yes	No	Description of how decisions were made to carry out/offer surgery	Yes	No	Remarks
Completed (included) studies
SINUS 52 (NCT02898454)	EUCTR2015‐001314‐10‐ES 2016		x	Not mentioned		x	Criteria not defined but one outcome was "Proportion of patients during study treatment receiving oral corticosteroid (OCS) for NP and/or planned to under surgery for nasal polyps"
SINUS 24 (NCT02898454)	Bachert 2019 NCT02898454		x	Not mentioned	x		Offered when there was worsening of signs and/or symptoms during the study Criteria not applied at baseline Who: not mentioned 28.3% nasal polyp surgery
	EUCTR2015‐003101‐42‐BG 2017 NCT02912468		x	Not mentioned		x	Criteria not defined but one outcome was "Proportion of patients during study treatment receiving oral corticosteroid (OCS) for NP and/or planned to under surgery for nasal polyps"
	Han 2019		x	Not mentioned		x	Full text not available but one outcome was "Reduction of surgery for nasal polyps"
NCT01066104	NCT01066104		x	Not mentioned		x
Pinto 2010 (NCT00117611)	NCT00117611 Pinto 2010 Mehta 2009		x	Not mentioned		x
Bachert 2017 (NCT01362244)	NCT01362244	x		Stated in the protocol Endoscopic nasal polyp score ≥ 3 and VAS > 7 Number of patients qualified at baseline: 105 Number of patients qualified at endpoint: 84 Number of patients who had surgery: not mentioned	x		Criteria for endoscopic nasal polyp score of ≥ 3, or nasal polyp score of 2 and a VAS symptom score of > 7 Criteria different from those applied at baseline Who: not mentioned 80% qualified for surgery
Bachert 2017 (NCT01362244)	EUCTR2008‐003772‐21‐NL 2009	x		Stated in the protocol refractory response to steroid therapy Number of patients qualified at baseline: 105 Number of patients qualified at endpoint: 79 Number of patients who had surgery: not mentioned	x		Criteria endoscopic nasal polyp score of ≥ 3, or nasal polyp score of 2 and a VAS symptom score of > 7 Criteria different from those applied at baseline Who: not mentioned 75% qualified for surgery
Gevaert 2013 (NCT01393340)	NCT01393340 Gevaert 2013 Gevaert 2012	x		Not mentioned
Bachert 2016 (NCT01920893)	NCT01920893 EUCTR2013‐001803‐35‐BE 2013 Bachert 2016 Other related publications: Bachert 2015 Schneider 2016 Willits 2016		x	Not mentioned		x
Gevaert 2011	Gevaert 2011		x	Not mentioned		x
POLYP 1 (NCT03280550)	NCT03280550		x	Stated in the protocol: Reduction in the need for surgery by week 24, as defined by a nasal polyps score of ≤ 4 (unilateral score of ≤ 2 on each side) and improvement in SNOT‐22 score of ≥ 8.9 Number of patients who qualified for surgery at baseline: not reported ‐ assumed all participants (inclusion criteria of nasal polyps score ≥ 5 with unilateral score of ≥ 2 for each nostril) Number of patients who qualified for surgery at endpoint: 123	x		No need for surgery when a nasal polyps score of ≤ 4 (unilateral score of ≤ 2 on each side) and improvement in SNOT‐22 score of ≥ 8.9 Criteria not reported at baseline Who: not mentioned
POLYP 2 (NCT03280537)	EUCTR2017‐001718‐28‐BE 2017 NCT03280537		x	Stated in the protocol: Reduction in the need for surgery by week 24, as defined by a nasal polyps score of ≤ 4 (unilateral score of ≤ 2 on each side) and improvement in SNOT‐22 score of ≥ 8.9 Number of patients who qualified for surgery at baseline: not reported ‐ assumed all participants (inclusion criteria of nasal polyps score ≥ 5 with unilateral score of ≥ 2 for each nostril) Number of patients who qualified for surgery at endpoint: 115	x		No need for surgery when a nasal polyps score ≤ 4 (unilateral score of ≤ 2 on each side) and improvement in SNOT‐22 score of ≥ 8.9 Criteria not applied at baseline Who: not mentioned
Included studies (not published)
NCT02772419	NCT02772419		x	Not mentioned		x
NCT02734849	NCT02734849		x	Not mentioned		x
Ongoing studies
NAPPREB (NCT04185012)	NCT04185012		x	Not mentioned on trial registry		x
ORCHID (NCT04157335)	NCT04157335	x		Stated on trial registry: Patients with bilateral sinonasal polyps that, despite treatment with standard of care including a history of treatment with systemic corticosteroids (oral, parenteral) or prior surgery for nasal polyps, have severity consistent with a need for surgery as described by: a minimum total nasal polyp score of 5 out of a maximum score of 8 (with a unilateral score of at least 2 for each nostril); ongoing symptoms for at least 12 weeks; patient‐reported moderate to severe nasal blockage (score 2 or 3 out of 3).			Ongoing study
OSTRO (NCT03401229)	NCT03401229	x		Stated in the protocol A minimum total nasal polyp score (NPS) of 5 out of a maximum score of 8 (with a unilateral score of at least 2 for each nostril) at V1 and continuously maintained at V2 to meet the randomisation criterion, as determined by the study Imaging Core Lab Ongoing symptoms for at least 12 weeks prior to V1 Patient‐reported moderate to severe nasal blockage score (NBS) 2 or 3 over the 2 weeks prior to V1 (2‐week recall assessment of symptoms, scores 0 (none) to 3 (severe)) Number of patients qualified at baseline: ongoing Number of patients qualified at endpoint: ongoing Number of patients who had surgery: ongoing			Ongoing study
SYNAPSE (NCT03085797)	NCT03085797	x		Stated in the protocol An overall VAS symptom score > 7, or an endoscopic bilateral nasal polyps score of at least 5 out of a maximum score of 8 (with a minimum score of 2 in each nasal cavity) Number of patients qualified at baseline: ongoing Number of patients qualified at endpoint: ongoing Number of patients had surgery: ongoing			Ongoing study
NCT02799446	NCT02799446		x	Not mentioned		x
NCT03614923	NCT03614923		x	Not mentioned		x
NCT03450083	NCT03450083		x	Not mentioned		x	Criteria not defined but one outcome was time to nasal polyp surgery
NCT04362501	NCT04362501		x	Not mentioned on trial registry		x
NCT044330179	NCT044330179		x	Not mentioned on trial registry		x

NP: nasal polyps
NPS: nasal polyps score
SNOT‐22: Sino‐Nasal Outcome Test‐22
VAS: visual analogue scale

In the two largest studies (724 participants), no specific criteria were given; it was stated that surgery was performed "when there was worsening of signs and/or symptoms during the study" (LIBERTY SINUS 24; LIBERTY SINUS 52).

In Bachert 2017, a set of criteria was used at randomisation and a different set at the trial's endpoint, to determine "eligibility for surgery". The criteria used were hypothetical; it is unclear how many participants were offered or underwent surgery. Moreover, whether or not these criteria correlate with actual patients' decisions to accept (and undergo) surgery (if offered) is unclear. It is also uncertain whether patients fulfilling these criteria would actually benefit from surgery (i.e. whether surgery is appropriate in these cases).

In POLYP 1 and POLYP 2 this outcome was reported as the number of participants who had a nasal polyp score of ≤ 4 (with a unilateral score of ≤ 2 on each side) and a reduction in SNOT‐22 score of ≥ 8.9 points. As all participants had a nasal polyp score of ≥ 5 at baseline, we assumed that they met the criteria for surgery on entry to the trial.

Therefore, although we identified a number of attempts by trialists to provide an indicator of whether biologics could reduce the need for surgery in patients, none of the studies used a validated method that can provide conclusive answers.

5a. Extent of disease: endoscopic score

A number of studies reported using an "endoscopic nasal polyps score" (NPS) or total polyps score (TPS) and referenced Gevaert 2013, whereas the protocol for Bachert 2016 referenced a non‐related paper. These had the same scoring system, utilising the total scores from both sides (bilateral, range 0 to 8). Unlike the Lund‐Kennedy and other scales with reported validation, these scales focused on the size of polyps, and not other factors such as the presence of inflammation and secretions/mucus.

Table: Scoring system for endoscopic nasal polyps score (NPS), or total polyps score (TPS)

Polyp score	Polyp size
0	No polyps
1	Small polyps in the middle meatus not reaching below the inferior border of the middle turbinate
2	Polyps reaching below the lower border of the middle turbinate
3	Large polyps reaching the lower border of the inferior turbinate or polyps medial to the middle turbinate
4	Large polyps causing complete obstruction of the inferior nasal cavity

5b. Extent of disease: computerised tomography (CT) scan score

All studies (other than Bachert 2017) used the Lund‐Mackay score.

6. Health‐related quality of life (HRQL), using generic HRQL scores

Generic health‐related quality of life data were available from five studies. Data on the overall health status measured using the EQ‐5D visual analogue scale were commonly reported and were used in our meta‐analysis. A minimal clinically important difference (MCID) of 8 points has been reported by Hoehle 2019. Data from studies using the SF‐36 are reported narratively, as incompleteness of the information did not allow data analysis.

7. Adverse effects: nasopharyngitis, including sore throat

Most studies reported this outcome.

Excluded studies

We excluded 42 studies (54 references) after reviewing the full text. Further details of the reasons for exclusion can be found in the Characteristics of excluded studies table.

We excluded seven studies due to the population (ANDHI; Castro 2011; Hayashi 2020; Liberty Asthma Quest; MUSCA; NCT01285323; NCT02170337). NCT01285323 and MUSCA were in asthma patients. NCT02170337 was a safety study in healthy patients. Liberty Asthma Quest, Castro 2011 and ANDHI were studies in asthma patients with a subset of chronic rhinosinusitis patients. The chronic rhinosinusitis patients did not meet our inclusion criteria.

We excluded one study due to the intervention (Gevaert 2006). In this safety study a single dose of biologic was given, rather than a course of treatment, and the duration of follow‐up was insufficient (less than three months).

We excluded four studies identified via our supplementary searches (NCT03956862; NCT03688555; NCT03681093; NCT03028350), because we did not regard the interventions used to be 'biologics'.

We excluded one study due to the comparison (Wahba 2019). This study compared a biologic to 'standard care', which included antibiotics and steroids, rather than comparing to a placebo.

We excluded 27 studies that were not RCTs (Bachert 2020; Bagnasco 2020; Boguniewicz 2019; Corren 2020; De Schryver 2015; Desrosiers 2019; Dinakar 2018; Chan 2020; ChiCTR1900026575; EUCTR2017‐003450‐16; Gevaert 2008; Gonzalez‐Diaz 2014; Hellings 2017; Hoy 2020; Jain 2020; Katial 2019; Laidlaw 2019; Laidlaw 2019b; Laidlaw 2019c; Laidlaw 2020a; Mullol 2020; Mustafa 2020; Naclerio 2017; NCT02743871; Perez De Llano 2018; Tajiri 2013; Zangrilli 2019).

Two studies were withdrawn (NCT00603785; NCT02734849).

Risk of bias in included studies

We included 10 studies in this review. Overall the risk of bias was low or unclear for most domains.

See Figure 2 for the 'Risk of bias' graph (our judgements about each risk of bias item presented as percentages across all included studies) and Figure 3 for the 'Risk of bias' summary (our judgements about each risk of bias item for each included study).

Figure 2

'Risk of bias graph': review authors' judgements about each risk of bias item presented as percentages across all included studies.

Figure 3

'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.

Allocation

The risk of selection bias was low or unclear in the majority of studies. We considered the risk of bias to be low for both random sequence generation and allocation concealment in six studies (Bachert 2016; Bachert 2017; LIBERTY SINUS 24; LIBERTY SINUS 52; POLYP 1; POLYP 2), and the risk in both of these domains to be unclear for three studies (Gevaert 2011; NCT01066104; Pinto 2010). We considered the Gevaert 2013 study to be at low risk of bias for random sequence generation, but at high risk for allocation concealment, because a randomisation list was used.

Blinding

We considered nine of the 10 studies to be at low risk of performance bias, since all participants and personnel were blind to treatment allocation. Both the investigator and participants were blinded in the Gevaert 2013 study, but it is not clear whether or not the study personnel were also blind. We therefore marked this domain as being at unclear risk of bias.

In seven of the studies it was clear that people who were blind to treatment allocation assessed outcomes, so we considered these to be at low risk of detection bias (Bachert 2016; Bachert 2017; LIBERTY SINUS 24; LIBERTY SINUS 52; NCT01066104; POLYP 1; POLYP 2). We considered the remaining three studies to be at unclear risk of bias (Gevaert 2011; Gevaert 2013; Pinto 2010). Although Gevaert 2013 and Pinto 2010 mentioned that the CT scans were read by blinded assessors, it was not clear whether or not the nasal endoscopy outcome assessment was blind.

Incomplete outcome data

We assessed four of the studies to be at high risk of attrition bias (Bachert 2016; Bachert 2017; Gevaert 2011; LIBERTY SINUS 52), mostly due to high rates of discontinuation in these small studies. We assessed LIBERTY SINUS 52 to be at high risk because, although the investigators used a last observation carried forward (LOCF) imputation method, there were proportionally more discontinuations in the placebo arm. We assessed Gevaert 2013, NCT01066104, POLYP 1 and POLYP 2 to be at low risk of attrition bias, and we considered LIBERTY SINUS 24 and Pinto 2010 to be at unclear risk of bias for this domain.

Selective reporting

We only considered three of the studies to be at low risk of selective reporting (Bachert 2017; POLYP 1; POLYP 2). There were differences between the NCT trial registration and reported outcomes for Gevaert 2013 and NCT01066104, so we assessed these to be at high risk of reporting bias. We found the other trials to be at unclear risk of reporting bias.

Other potential sources of bias

There are concerns about whether or not appropriate and validated tools were used for some outcomes. None of the studies reported using validated methods for their endoscopic scoring systems. All of the studies either did not provide details of the method used or had reported using a scoring system that took into account only the size of the polyps and we did not find any references to the validation of this system. Similarly, whilst many studies reported using a VAS for overall symptom score, they made no reference to validation. Although a VAS is a well‐used type of scale, its validity needs to be confirmed in each specific population and for each outcome measured; factors such as the clarity of questions and the definition used for the 'best' and 'worst' points in the scale could affect a scale's validity.

The assessment of 'avoidance of surgery' (outcome 4 above) is fraught with difficulty; there is a high risk of bias in the included studies. Only a small number of studies defined eligibility for surgery at baseline. However, some of these studies did not use the same criteria for assessment of surgical eligibility at the trial's endpoint. Moreover, there is an absence of generally accepted or validated criteria as to what constitutes a situation that is 'severe' enough for patients to be willing to undergo surgery, or to benefit from it. Therefore, it is particularly unclear how these criteria were determined and/or the basis on which criteria were changed between entry and the endpoint of a study.

In those studies without any predefined or explicit criteria for surgery, it is even less clear how decisions were made to offer 'rescue surgery'. See Table 2 for further details.

Effects of interventions

Comparison 1: Anti‐IL‐4Rα mAb plus intranasal steroids versus placebo/no treatment plus intranasal steroids

Three studies (784 participants) investigated dupilumab (Bachert 2016; LIBERTY SINUS 24; LIBERTY SINUS 52). See summary of findings Table 1. Participants in these trials were relatively homogeneous, with a similar age profile and gender balance, nasal polyp scores, and similar proportions of participants who had previous surgery or co‐morbidities (such as asthma or N‐ERD).

1. Health‐related quality of life, using validated disease‐specific health‐related quality of life scores

Disease‐specific health‐related quality of life was measured with the Sino‐Nasal Outcome Test‐22 (SNOT‐22, range 0 to 110, minimal clinically important difference (MCID) 8.9 points).

At 24 weeks, the SNOT‐22 score was 19.61 points lower (better) in participants who received dupilumab (mean difference (MD) ‐19.61, 95% confidence interval (CI) ‐22.54 to ‐16.69; 3 studies; 784 participants; I² = 0%; high‐certainty evidence; Analysis 1.1). This is likely to be a large difference.

This effect was also seen at 52 weeks (MD ‐22.38, 95% CI ‐27.10 to ‐17.66; 1 study; 303 participants), but the certainty of evidence is moderate due to imprecision (Analysis 1.1).

2. Disease severity, as measured by validated patient‐reported symptom score

All of the studies used a 0 to 10 cm visual analogue scale (VAS) score to measure overall chronic rhinosinusitis symptoms. For the LIBERTY SINUS 24 and LIBERTY SINUS 52 studies (724 participants), the question asked was "How troublesome are your symptoms?". We found no evidence to indicate that this tool has been validated.

The pooled mean difference is ‐3.00 favouring the groups receiving dupilumab (95% CI ‐3.47 to ‐2.53; 3 studies; 784 participants; I² = 0%; moderate‐certainty evidence; Analysis 1.2). This is likely to be clinically significant.

3. Serious adverse events

The incidence of serious adverse events was measured over different periods: up to 16 weeks in Bachert 2016, 24 weeks in LIBERTY SINUS 24 and 52 weeks in LIBERTY SINUS 52. The number of serious adverse events seems to be lower in the treatment group (absolute risk 5.9% compared to 12.5%, risk ratio (RR) 0.47, 95% CI 0.29 to 0.76; 3 studies; 782 participants; I² = 0%; low‐certainty evidence; Analysis 1.3). This may be due, in part, to a reduction in the incidence of nasal polyps, chronic rhinosinusitis and asthma in those who received dupilumab. There were discrepancies in the numbers reported in the different publications reporting the results of LIBERTY SINUS 52 and LIBERTY SINUS 24. Therefore, we used the data that matched those reported in clinicaltrials.gov in this analysis.

4. Avoidance of surgery

Two studies reported the number of participants requiring "nasal polyps surgery (actual or planned) during the treatment period". Dupilumab may result in a large reduction in the number of patients who require nasal polyps surgery (RR 0.17, 95% CI 0.05 to 0.52; 2 studies; 725 participants; I²= 28%; moderate‐certainty evidence; Analysis 1.4). However, between baseline and endpoint there were changes in the criteria that determined whether or not a participant qualified for surgery. How many qualified for surgery compared with how many actually received surgery, and the specific factors that determined whether or not a patient received 'rescue' surgery during follow‐up were unclear. See Table 2 for more details on how this outcome was measured.

5a. Extent of disease: endoscopy score

All studies used a nasal polyps score, which summed the scores for both nostrils (0 to 8 points; 0 = no polyp, 4 = large polyps, for each nostril, with a lower score indicating smaller‐sized polyps). The differences between the intervention arms were large (Cohen's effect size > 0.7 = large effect), favouring the dupilumab group.

At 24 weeks follow‐up the mean difference was ‐1.80 (95% CI ‐2.25 to ‐1.35; 3 studies; 784 participants; I² = 65%; moderate‐certainty evidence; Analysis 1.5), with a corresponding effect size of standardised mean difference (SMD) ‐1.05 (95% CI ‐1.29 to ‐0.82). We found no evidence to indicate that this scoring system has been validated.

At 52 weeks, the mean difference was ‐2.34 (95% CI ‐2.77 to ‐1.91; 1 study; 303 participants; low‐certainty evidence; Analysis 1.5), and the corresponding effect size was SMD ‐1.24 (95% CI ‐1.48 to ‐0.99).

5b. Extent of disease: computerised tomography (CT) scan score

We pooled data from 16 weeks to 52 weeks as data were only available from one time point from each study.

The changes in the extent of disease were evaluated using a CT scan and scored using the Lund‐Mackay scale (0 to 24, higher = worse). The mean difference was ‐7.00 (95% CI ‐9.61 to ‐4.39; 3 studies; 784 participants; I² = 92%; high‐certainty evidence; Analysis 1.6), showing a large effect favouring the dupilumab group. The corresponding SMD was ‐1.50 (95% CI ‐1.84 to ‐1.15; Cohen's effect size > 0.7 = large effect). We considered the certainty of the evidence to be high despite the large I² value; there is no inconsistency in terms of direction or size of effects between the three studies.

6. Health‐related quality of life, using generic health‐related quality of life scores

All studies used the EQ‐5D visual analogue scale (0 to 100, higher = better) to measure the change in generic health‐related quality of life (overall health state). The pooled MD of three studies was 8.29 points (95% CI 5.73 to 10.85; 3 studies; 766 participants; I² = 0%; moderate‐certainty evidence; Analysis 1.7). This effect size is similar to the size of the MCID (8 points, as suggested by Hoehle 2019) and therefore there is probably a clinically important improvement in this outcome.

Bachert 2016 also reported change in scores on the SF‐36 questionnaire. The study authors reported a significant difference in the individual domains for 'vitality' and 'mental health', and an overall improvement in the mental component summary (least squares mean difference reported as 5.45 points higher in the dupilumab group, 95% CI 1.42 to 9.48 points; range 0 to 100, higher scores = better).

7. Adverse effects: nasopharyngitis, including sore throat

The pooled results indicate that there may be little or no difference in the risk of nasopharyngitis, but larger sample sizes are needed for a more precise estimate (RR 0.95, 95% CI 0.72 to 1.25; 3 studies; 783 participants; I² = 0%; low‐certainty evidence; Analysis 1.8).

Comparison 2: Anti‐IL‐5 mAb plus intranasal steroids versus placebo/no treatment plus intranasal steroids

Two studies evaluated mepolizumab (Bachert 2017; Gevaert 2011). See summary of findings Table 2. There was some clinical heterogeneity in the participants in these trials. Bachert 2017 included a majority of participants with asthma, whilst Gevaert 2011 had fewer than half of participants with asthma. Bachert also recruited participants with at least one previous operation for nasal polyps; this was not a requirement for Gevaert 2011, although the majority of participants had undergone previous surgery.

1. Health‐related quality of life, using validated disease‐specific health‐related quality of life scores

Data on disease‐specific health‐related quality of life as measured with the SNOT‐22 were only available from one study (Bachert 2017: data from the EudraCT website). Mepolizumab may result in a reduction (improvement) in SNOT‐22 score; the mean difference of ‐13.26 lower (better) with mepolizumab (95% CI ‐22.08 to ‐4.44; 1 study; 105 participants; low‐certainty evidence; Analysis 2.1) is greater than the MCID of 8.9 points.

2. Disease severity, as measured by validated patient‐reported symptom score

Bachert 2017 reported using a VAS of 0 to 10 with the question "How troublesome are your symptoms of nasal polyposis?" (0 = "not troublesome", 10 = "worst possible"). The MD was ‐2.03 (95% CI ‐3.65 to ‐0.41; 1 study; 72 participants; very low‐certainty evidence; Analysis 2.2). We are very uncertain about these data due to the very small sample size and the absence of evidence that a validated tool was used.

3. Serious adverse events (SAEs)

It is very uncertain whether or not there is a difference in the number of serious adverse events with mepolizumab (absolute risk 1.37% compared to 0%, RR 1.57, 95% CI 0.07 to 35.46; 2 studies; 135 participants; I² = 0%; very low‐certainty evidence; Analysis 2.3).

4. Avoidance of surgery

Each study applied different criteria for assessing the need for surgery (see Table 2). While Bachert 2017 reported the number of patients who still met the criteria for surgery at the end of trial, Gevaert 2011 reported the number that required surgery during the period of the trial. It is very uncertain whether or not the overall risk that patients still need surgery at the end of trial is lower in the mepolizumab group (RR 0.78, 95% CI 0.64 to 0.94; 2 studies; 135 participants; I² = 0%; very low‐certainty evidence; Analysis 2.4).

5a. Extent of disease: endoscopic score

The mean difference in the change of the nasal polyps score was 1.23 points lower in the mepolizumab group (MD ‐1.23, 95% ‐1.79 to ‐0.68; 2 studies; 137 participants; I² = 0%; very low‐certainty evidence; Analysis 2.5). This corresponds to a moderate effect size (SMD ‐0.69, 95% ‐1.04 to ‐0.34). We found no evidence to indicate that this scoring system has been validated.

5b. Extent of disease: computerised tomography (CT) scan score

Gevaert 2011 did not report the numerical values of the CT scan scores, but stated that at week eight the scores "were not significantly different between groups". Bachert 2017 did not measure CT scan scores. The evidence for this outcome was of very low certainty.

6. Health‐related quality of life, using generic quality of life scores

The mean difference on the EQ‐5D visual analogue scale was 5.68 in one study (95% CI ‐1.18 to 12.54; 1 study; 105 participants; low‐certainty evidence; Analysis 2.6), favouring the mepolizumab group (Bachert 2017). This difference is smaller than the MCID of 8 points.

7. Adverse effects: nasopharyngitis, including sore throat

There may be little or no difference in the risk of nasopharyngitis (RR 0.73, 95% 0.36 to 1.47; 2 studies; 135 participants; I² = 0%; low‐certainty evidence; Analysis 2.7).

Comparison 3: Anti‐IgE mAb plus intranasal steroids versus placebo/no treatment plus intranasal steroids

We identified five studies evaluating omalizumab (Gevaert 2013; NCT01066104; Pinto 2010; POLYP 1; POLYP 2). See summary of findings Table 3. Some clinical heterogeneity was present between the participants recruited to these studies. Sparse data were available for NCT01066104, therefore we were unable to identify the number of participants with asthma or N‐ERD. For the remaining trials in this comparison, Gevaert 2013 exclusively included participants with asthma, whilst around half of the participants in Pinto 2010, POLYP 1 and POLYP 2 had asthma.

1. Health‐related quality of life, using validated disease‐specific health‐related quality of life scores

Two studies reported the SNOT‐22 scores for participants (POLYP 1; POLYP 2). The mean difference in SNOT‐22 scores was ‐15.62 points lower in those who received omalizumab compared to those who received placebo (95% CI ‐19.79 to ‐11.45; 2 studies; 265 participants; I² = 0%; moderate‐certainty evidence; Analysis 3.1). This difference is greater than the MCID of 8.9 points.

Two studies reported alternative measures for disease‐specific health‐related quality of life. A narrative summary was reported in Gevaert 2013 (24 participants): "On the basis of the 31‐item Rhinosinusitis Outcome Measuring Instrument (RSOM‐31), sleep (P =0.03) and general symptoms (P = 0.01) showed a significant improvement in the omalizumab group, whereas in the placebo group no significant changes were seen".

Pinto 2010 reported that the median change in SNOT‐20 score was ‐1.05 for the omalizumab group and ‐0.20 for the placebo group (P < 0.78 for the difference between groups).

2. Disease severity, as measured by validated patient‐reported symptom score

No study used a global score of symptom severity, or a visual analogue scale, therefore no meta‐analysis has been conducted for this outcome.

Three studies assessed disease severity using "total nasal symptom" scores. However, these assessed nasal aspects of chronic rhinosinusitis only (such as anterior rhinorrhoea, posterior rhinorrhoea, nasal congestion) but not facial pain (Pinto 2010; POLYP 1; POLYP 2). For future iterations of this review we intend to incorporate these data, but as no study assessed global scores for chronic rhinosinusitis symptoms they are not included at present.

3. Serious adverse events (SAEs)

It is very uncertain whether omalizumab affects the occurrence of serious adverse events (RR 0.32, 95% CI 0.05 to 2.00; 5 studies; 329 participants; I² = 0%; very low‐certainty evidence; Analysis 3.2). The number of serious adverse events reported by the trials was small (five events in total), and the maximum duration of follow‐up was 24 weeks ‐ this may be inadequate to capture the full range of adverse events associated with treatment, and may not reflect the risks with longer treatment duration.

4. Avoidance of surgery

Two studies reported this outcome (POLYP 1; POLYP 2). Both studies used the same assessment method ‐ a reduction in the need for surgery was defined as a total endoscopic nasal polyp score of ≤ 4 (with a unilateral score of ≤ 2 on each side) and a reduction in SNOT‐22 score of ≥ 8.9 points. Omalizumab may result in a large reduction in the need for surgery with a RR of 5.60 (95% CI 1.99 to 15.76; 2 studies; 265 participants; I² = 0%; low‐certainty evidence; Analysis 3.3). However, we consider the evidence to be of low certainty as the sample size for this estimate may be too small to estimate this accurately. In addition, it is not clear whether the criteria used to determine "avoidance of surgery" are appropriate, as there are no widely agreed standards to establish this.

5a. Extent of disease: endoscopic score

Four studies evaluated and reported nasal polyp scores (range 0 to 8 points, higher = worse). The evidence suggests that omalizumab improves the endoscopic score, with a pooled mean difference of 1.26 points lower in the omalizumab group (95% CI 0.31 points lower to 2.2 points lower; 4 studies; 312 participants; I² = 90%, low‐certainty evidence; Analysis 3.4). However, there was high statistical heterogeneity in this result and the effect size for the individual trials varied considerably, therefore we considered this result to be of low certainty. One study showed no effect; the other three trials showed improvement with omalizumab, but to a varying degree (between ‐0.59 to ‐2.55 points).

Pinto 2010 reported that "There were no significant changes within in endoscopy scores for either group (data not shown). Net change across treatments were not significantly different (omalizumab 0, placebo –0.5, P < 0.58)". There was no information about what scoring system was used or whether one or both sides of the nose were assessed and scored. The paper reported using a 0‐ to 4‐point score, but referenced a paper using a 0‐ to 3‐point scale.

5b. Extent of disease: computerised tomography (CT) scan score

Gevaert 2013 reported the Lund‐Mackay scores at the endpoint whereas NCT01066104 reported the percentage change compared to baseline using a modification of the Lund‐Mackay score (no reports of validation). In both studies, lower scores mean a better outcome for the patients. The observed pooled results correspond to a small effect size (SMD ‐0.20, 95% CI ‐1.55 to 1.14; 2 studies; 47 participants; I² = 80%; Analysis 3.5).

Statistical heterogeneity is high and there are inconsistencies in the size and direction of effect. In the NCT01066104 study, the results favoured the placebo group, while in Gevaert 2013 they favoured the intervention group. The evidence for this outcome was of very low certainty.

6. Health‐related quality of life, using generic quality of life scores

Two studies used the SF‐36 to measure health‐related quality of life. Pinto 2010 reported that "Across treatments, there were also no significant differences (P > 0.05, all comparisons) except for one domain, Vitality (omalizumab 9.4, placebo 12.5, P < 0.05)." Gevaert 2013 reported, "After 16 weeks, the Short‐Form Health Questionnaire (SF‐36) for physical health was significantly improved in the omalizumab group (P = 0.02) but not in the placebo group (P = 0.75). Unlike physical health, mental health did not significantly improve in either treatment group." The evidence for this outcome was of very low certainty.

7. Adverse effects: nasopharyngitis, including sore throat

All five studies reported on the occurrence of nasopharyngitis. The evidence suggests that omalizumab may result in little to no difference in the occurrence of nasopharyngitis with a RR of 0.71 (95% CI 0.29 to 1.73; 5 studies; 329 participants; I² = 0%, low‐certainty evidence; Analysis 3.6). However, the total sample size may be too small to accurately estimate this effect, therefore our confidence in the estimate is low.

Discussion

Summary of main results

We identified randomised controlled trials (RCTs) evaluating the effectiveness of three different drugs, representing three different types of monoclonal antibodies. These were dupilumab (an anti‐IL‐4Rα mAb), mepolizumab (an anti‐IL‐5 mAb) and omalizumab (an anti‐IgE mAb). For this update of the review we identified two additional trials that provide evidence on mepolizumab.

All of the drugs were evaluated in adults with chronic rhinosinusitis and nasal polyps who were also using regular topical nasal steroids. In these patients, we found high‐certainty evidence from three studies (with nearly 800 participants) that dupilumab results in a large improvement in disease‐specific health‐related quality of life (HRQL) compared to placebo, and a large reduction in the extent of the disease as measured on a computerised tomography (CT) scan. Moderate‐certainty evidence shows that it probably also results in a large improvement in symptoms, increases generic HRQL (as measured by overall health status) and results in a large reduction in the size of polyps (as measured by nasal polyp scores). It probably results in a large reduction in the need for further surgery but it is difficult to interpret the clinical implications of this finding due to methodological limitations. There may be little or no difference in the risk of nasopharyngitis.

Mepolizumab has been evaluated in similar patients but the certainty of evidence is either low or very low. It may improve both disease‐specific and generic HRQL. It may also improve nasal polyp scores, but the evidence is very uncertain. We are very uncertain whether it reduces the need for surgery, as there are important limitations of the methodology that limit the clinical interpretation of the data. There may be little or no difference in the risk of nasopharyngitis. It is very uncertain if there is a difference in the risk of serious adverse events.

We identified moderate‐certainty evidence from two studies that omalizumab probably results in a large improvement in disease‐specific HRQL compared to placebo. It may also result in a large reduction in the need for surgery, but the evidence for this was of low certainty. Omalizumab may also result in a reduction in the extent of disease, as assessed with an endoscopic nasal polyps score, although there were differences in the extent of this change between the four studies that reported this measure. Similarly, when the extent of disease was assessed with CT scores, there were differences in the size and direction of effect in the two studies, and the evidence was of very low certainty. Omalizumab may result in little to no difference in nasopharyngitis when compared to placebo, although the risk of serious adverse events is very uncertain.

Overall completeness and applicability of evidence

There are four major limitations pertaining to the completeness and applicability of the evidence:

All but one study (Pinto 2010) recruited patients with moderate to severe chronic rhinosinusitis with nasal polyps, as defined by polyp size and need for systemic steroids and/or surgery, and at least half of the participants also had asthma as a comorbidity. Therefore, there is no evidence on whether or not patients with less severe disease (with or without nasal polyposis or asthma) would benefit as much or at all.
All studies were in adults. There are no data for children.
There is a lack of long‐term evidence. Whilst treatment with biologics is arguably a lifetime commitment, there was only one study with a 52‐week follow‐up. It was not always possible to compare the mid‐term (24‐week) data with the longer‐term data in this study. However, where data were published (SNOT‐22 and endoscopy score) the effect size was maintained (LIBERTY SINUS 52).
The sample sizes were insufficient and the length of follow‐up too short to comprehensively and adequately assess the risks of side effects.

Whilst the data on adverse effects included in this review are sparse, we acknowledge that some biologics have now been used for several years in other conditions without serious safety concerns. Data from asthma trials with larger study sizes and a longer duration of follow‐up indicate that the rate of serious adverse effects is small (Farne 2017). It is likely that there is considerable overlap in the patient population between these studies, as asthma and chronic rhinosinusitis frequently occur together, which may give further reassurance as to the safety profile in those with chronic rhinosinusitis.

For this review we have focused on global ratings of chronic rhinosinusitis symptoms as a primary outcome measure, rather than assessing the individual, specific symptoms. Global symptom scores are most important when considering the effectiveness of each individual biologic, but as more biologics become available, individual symptom scores may help to facilitate comparison between different drugs. It is possible that the efficacy of biologic agents may vary for different underlying symptoms. In particular, patients with recalcitrant chronic rhinosinusitis and nasal polyps (who may be candidates for biologic therapy) are likely to have considerable problems with olfaction, and it would be useful to ascertain whether biologics improve this symptom. For future iterations of this review we hope to be able to include more details on the individual symptom scores, to identify which of the four specific chronic rhinosinusitis symptoms improve with biologics.

Quality of the evidence

The primary reason for downgrading the quality of the available evidence was imprecision, where sample sizes were too small to provide a precise estimate.

In addition, the lack of evidence that validated scales or scoring systems were used was also a concern, especially for symptom scores and endoscopy scores. As in other studies found in this series of Cochrane Reviews, the lack of use of a globally validated symptom score scale, which focuses on overall disease severity, continues to be a problem. It is difficult to compare 'the overall improvement' of symptoms across trials or reviews if studies use different scales, with different weightings given to different types of symptoms. Although there have been improvements in methodology compared to previous studies, in the sense that studies attempted to use visual analogue scales, there was no evidence that these scales had been validated and that they are comparable across studies. In addition, many studies also used a scoring system for nasal endoscopy that only takes into account the size of polyps. There is no reference to how this scale has been validated against patient outcomes.

All but one study (Pinto 2010) focused (sometimes solely) on recruiting patients who had comorbid asthma and more severe nasal polyposis. However, notwithstanding this we did not further downgrade studies based on applicability.

It should also be noted that the evidence available is relatively short‐term; only one study was conducted for more than six months. We did not downgrade the evidence for indirectness due to the relatively short follow‐up.

Potential biases in the review process

None of the studies reported using a validated overall symptom score measure to assess changes in patients' symptom severity. Some studies reported specific types of chronic rhinosinusitis symptoms using different tools, for many of which there was no evidence of validation.

To provide the best possible picture of overall symptoms, we examined each reported tool carefully and used data from questions/questionnaires that asked about overall symptoms. We avoided using data from tools that only measured one or two specific symptoms of chronic rhinosinusitis. For example, we did not use data from the 'total symptom score' (TSS); this only measured symptoms of anterior and posterior rhinorrhoea and nasal blockage. The symptoms of loss of sense of smell and facial pain were not measured.

Whenever an overall symptom assessment was reported using a visual analogue scale, we recorded and used those data even though there were slight variations between studies in how the questions were worded.

Agreements and disagreements with other studies or reviews

Results from two of the larger trials that assessed omalizumab have not been included in any previous systematic reviews (POLYP 1; POLYP 2).

Two systematic reviews include a number of trials that are included in this review, either as included or ongoing studies (Laidlaw 2020b; Tsetsos 2020). Laidlaw 2020b does not include any meta‐analysis, but provides an overview of ongoing and completed trials. Tsetsos 2020 considered a change in University of Pennsylvania Smell Identification Test (UPSIT) score as their primary outcome measure. They included a small amount of meta‐analysis for dupilumab and found an improvement in UPSIT score for those receiving dupilumab as compared to placebo.

Three systematic reviews (Codispoti 2019; Iqbal 2020; Tsetsos 2018) reported five trials that we also included in this Cochrane Review (Bachert 2016; Bachert 2017; Gevaert 2011; Gevaert 2013; Pinto 2010) and one that we excluded (Gevaert 2006). The primary outcome for Tsetsos 2018 was total nasal endoscopic polyp score, and these data were also reported by Iqbal 2020. No study performed a meta‐analysis.

Rivero 2017 included randomised and non‐randomised studies in their systematic review and meta‐analysis. Three of our included studies were also included in their review (Gevaert 2011; Gevaert 2013; Pinto 2010). Nasal polyp score was their primary outcome of interest. The differences in the study types means that is not appropriate to compare the results of their meta‐analyses with those in this review.

An earlier systematic review, Hong 2015, only identified two RCTs (Gevaert 2013; Pinto 2010).

One further review considers the use of biologics in airway disease, but with a focus on asthma (Walter 2020). Only one trial that relates to individuals with chronic rhinosinusitis is included (Bachert 2016).

In summary, there are no systematic reviews or meta‐analyses with which it is appropriate to compare the results of the present review.

Figure 1

PRISMA flow diagram

Figure 2

'Risk of bias graph': review authors' judgements about each risk of bias item presented as percentages across all included studies.

Figure 3

'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.

Analysis 1.1

Comparison 1: Anti‐IL‐4Rα mAb (dupilumab) versus placebo (on top of topical steroids), Outcome 1: HRQL ‐ disease‐specific (SNOT‐22, 0 to 110, lower = better)

Analysis 1.2

Comparison 1: Anti‐IL‐4Rα mAb (dupilumab) versus placebo (on top of topical steroids), Outcome 2: Disease severity ‐ VAS (0 to 10, lower = better)

Analysis 1.3

Comparison 1: Anti‐IL‐4Rα mAb (dupilumab) versus placebo (on top of topical steroids), Outcome 3: Serious adverse events

Analysis 1.4

Comparison 1: Anti‐IL‐4Rα mAb (dupilumab) versus placebo (on top of topical steroids), Outcome 4: Avoidance of surgery ‐ number of patients who had surgery as rescue treatment

Analysis 1.5

Comparison 1: Anti‐IL‐4Rα mAb (dupilumab) versus placebo (on top of topical steroids), Outcome 5: Extent of disease ‐ endoscopy ('nasal polyps score', 0 to 8, higher = worse)

Analysis 1.6

Comparison 1: Anti‐IL‐4Rα mAb (dupilumab) versus placebo (on top of topical steroids), Outcome 6: Extent of disease ‐ CT scan (Lund Mackay, 0 to 24, higher = worse)

Analysis 1.7

Comparison 1: Anti‐IL‐4Rα mAb (dupilumab) versus placebo (on top of topical steroids), Outcome 7: HRQL ‐ generic (EQ‐5D VAS, 0 to 100, higher = better)

Analysis 1.8

Comparison 1: Anti‐IL‐4Rα mAb (dupilumab) versus placebo (on top of topical steroids), Outcome 8: Adverse events ‐ nasopharyngitis, including sore throat (longest available data)

Analysis 2.1

Comparison 2: Anti‐IL‐5 mAb (mepolizumab) versus placebo (on top of topical steroids), Outcome 1: HRQL ‐ SNOT‐22 (1 to 100, lower = better) up to 25 weeks

Analysis 2.2

Comparison 2: Anti‐IL‐5 mAb (mepolizumab) versus placebo (on top of topical steroids), Outcome 2: Disease severity ‐ VAS (0 to 10, lower = better)

Analysis 2.3

Comparison 2: Anti‐IL‐5 mAb (mepolizumab) versus placebo (on top of topical steroids), Outcome 3: Serious adverse events

Analysis 2.4

Comparison 2: Anti‐IL‐5 mAb (mepolizumab) versus placebo (on top of topical steroids), Outcome 4: Avoidance of surgery ‐ patients still meeting criteria for surgery at end of follow‐up

Analysis 2.5

Comparison 2: Anti‐IL‐5 mAb (mepolizumab) versus placebo (on top of topical steroids), Outcome 5: Extent of disease ‐ endoscopic score

Analysis 2.6

Comparison 2: Anti‐IL‐5 mAb (mepolizumab) versus placebo (on top of topical steroids), Outcome 6: HRQL ‐ generic measured using EQ‐5D VAS (range 0 to 100; 0 = worst, 100 = best imaginable health state) at week 25

Analysis 2.7

Comparison 2: Anti‐IL‐5 mAb (mepolizumab) versus placebo (on top of topical steroids), Outcome 7: Adverse events ‐ nasopharyngitis, including sore throat

Analysis 3.1

Comparison 3: Anti‐IgE mAb (omalizumab) versus placebo (on top of topical steroids), Outcome 1: HRQL disease‐specific ‐ SNOT‐22 (0 to 110, lower = better)

Analysis 3.2

Comparison 3: Anti‐IgE mAb (omalizumab) versus placebo (on top of topical steroids), Outcome 2: Serious adverse events

Analysis 3.3

Comparison 3: Anti‐IgE mAb (omalizumab) versus placebo (on top of topical steroids), Outcome 3: Avoidance of surgery

Analysis 3.4

Comparison 3: Anti‐IgE mAb (omalizumab) versus placebo (on top of topical steroids), Outcome 4: Extent of disease ‐ endoscopic score (nasal polyps score, range 0 to 8, lower = better)

Analysis 3.5

Comparison 3: Anti‐IgE mAb (omalizumab) versus placebo (on top of topical steroids), Outcome 5: Extent of disease ‐ CT scan (lower score = better)

Analysis 3.6

Comparison 3: Anti‐IgE mAb (omalizumab) versus placebo (on top of topical steroids), Outcome 6: Adverse events ‐ nasopharyngitis, including sore throat

Summary of findings 1. Anti‐IL‐4Rα mAb (dupilumab) compared to placebo (on top of topical steroids) for chronic rhinosinusitis

Outcomes	Number of participants (studies)	Relative effect (95% CI)	*Anticipated absolute effects^ (95% CI)**			Certainty of the evidence (GRADE)	What happens
Anti‐IL‐4Rα mAb (dupilumab) compared to placebo (on top of topical steroids) for chronic rhinosinusitis
Patients or population: patients with severe chronic rhinosinusitis with nasal polyps Setting: tertiary care Intervention: anti‐IL‐4Rα mAb (dupilumab) Comparison: placebo (on top of topical steroids)
Outcomes	Number of participants (studies)	Relative effect (95% CI)	Without anti‐IL‐4Rα mAb (dupilumab)	With anti‐IL‐4Rα mAb (dupilumab)	Difference	Certainty of the evidence (GRADE)	What happens
Health‐related quality of life ‐ disease‐specific (SNOT‐22, range 0 to 110, lower = better) Follow‐up (range): 16 to 24 weeks	784 (3 RCTs)	—	The median disease‐specific health‐related quality of life score without anti‐IL‐4Rα mAb (dupilumab) was 40.5 points	—	MD 19.61 points lower (22.54 lower to 16.69 lower)	⊕⊕⊕⊕ HIGH	At up to 24 weeks, aspects of health‐related quality of life that are directly impacted by chronic rhinosinusitis were better in participants who received dupilumab. The size of the difference is clinically significant.
Disease severity ‐ VAS (range 0 to 10, lower = better) Follow‐up (range): 16 to 24 weeks	784 (3 RCTs)	—	The median disease severity score without anti‐IL‐4Rα mAb (dupilumab) was ‐1.3 points	—	MD 3 points lower (3.47 lower to 2.53 lower)	⊕⊕⊕⊝ MODERATE¹	Overall chronic rhinosinusitis symptoms were probably better in participants who received dupilumab.
Serious adverse events Follow‐up (range): 16 to 52 weeks	782 (3 RCTs)	RR 0.47 (0.29 to 0.76)	Study population			⊕⊕⊝⊝ LOW²	Participants who had dupilumab may have had fewer serious adverse events than participants who received placebo in 3 RCTs (28/470 with dupilumab versus 39/312 with placebo), but we have limited confidence in this estimate because the sample size may be too small to estimate this accurately, or capture the range of adverse events that could possibly occur in a larger population or with longer follow‐up.
Serious adverse events Follow‐up (range): 16 to 52 weeks	782 (3 RCTs)	RR 0.47 (0.29 to 0.76)	12.5%	5.9% (3.6 to 9.5)	6.6% fewer (8.9 fewer to 3 fewer)	⊕⊕⊝⊝ LOW²
Avoidance of surgery ‐ number of patients who had surgery as rescue treatment Follow‐up (range): 24 to 52 weeks	725 (2 RCTs)	RR 0.17 (0.05 to 0.52)	Study population			⊕⊕⊕⊝ MODERATE³	Patients who had dupilumab probably have lower risk of requiring surgery due to severe chronic rhinosinusitis symptoms after 24 to 52 weeks of treatment. We have moderate confidence in this estimate as we are not sure which criteria were used to determine the need for 'rescue surgery'.
	725 (2 RCTs)	RR 0.17 (0.05 to 0.52)	7.7%	1.3% (0.4 to 4)	6.4% fewer (7.3 fewer to 3.7 fewer)	⊕⊕⊕⊝ MODERATE³
Extent of disease: endoscopic nasal polyp score (range 0 to 8, lower = better) Follow‐up (range): 16 to 24 weeks	784 (3 RCTs)	—	The median nasal polyp score without dupilumab was 5.94 points.	—	MD 1.80 points lower (2.25 lower to 1.35 lower)	⊕⊕⊕⊝ MODERATE¹	Dupilumab probably results in a reduction in nasal polyp score by 24 weeks of follow‐up. This is likely to be a large effect, however we have moderate confidence in the estimate as it is unclear whether the scoring system used for nasal polyps is validated.
Extent of disease: CT scan score (Lund‐Mackay, range 0 to 24, lower = better) Follow‐up (range): 16 to 52 weeks	784 (3 RCTs)	—	The median CT scan score without anti‐IL‐4Rα mAb (dupilumab) was 17.9 points	—	MD 7 points lower (9.61 lower to 4.39 lower)	⊕⊕⊕⊕ HIGH	At up to 24 weeks, the extent of disease as assessed by CT scan was less severe in participants who received dupilumab ‐ the difference is likely to be a large effect.
Health‐related quality of life ‐ generic (EQ‐5D visual analogue scale, range 0 to 100, higher = better) Follow‐up (range): 16 to 24 weeks	766 (3 RCTs)	—	The median change in generic HRQOL for the placebo group was an increase of 3.01 points	—	MD 8.29 points higher (5.73 higher to 10.85 higher)	⊕⊕⊕⊝ MODERATE⁴	The overall quality of life or health status, as assessed by the EQ‐5D visual analogue scale was probably slightly higher in participants who received dupilumab. However, we are not sure if the size of this difference is noticeable or would be considered important enough by most patients.
Adverse events ‐ nasopharyngitis, including sore throat (longest available data) Follow‐up (range): 16 to 52 weeks	783 (3 RCTs)	RR 0.95 (0.72 to 1.25)	Study population			⊕⊕⊝⊝ LOW²	We are uncertain whether there is an important difference in the risk of nasopharyngitis. Adverse events were reported by 94/470 participants who took dupilumab versus 66/313 who took placebo.
	783 (3 RCTs)	RR 0.95 (0.72 to 1.25)	21.1%	20.0% (15.2 to 26.4)	1.1% fewer (5.9 fewer to 5.3 more)	⊕⊕⊝⊝ LOW²
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; CT: computerised tomography; MD: mean difference; RCT: randomised controlled trial; RR: risk ratio; SNOT‐22: Sino‐Nasal Outcome Test‐22; VAS: visual analogue scale
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
¹Downgraded by one level due to study limitations: methods or criteria used in the measurement of the outcome were not validated. ²Downgraded by two levels due to imprecision and indirectness: small sample size for the outcome estimated resulting in an imprecise estimation of effect size. Moreover, some serious adverse events are relatively rare; a larger and more heterogeneous population or longer periods of treatment and follow‐up may be needed. ³Downgraded by one level due to serious limitations: the criteria used for requiring/not requiring 'rescue surgery' were unclear. ⁴Downgraded by one level for imprecision: the confidence interval crosses the minimally important difference (8 points), therefore the difference may or may not be of importance to participants.

Summary of findings 1. Anti‐IL‐4Rα mAb (dupilumab) compared to placebo (on top of topical steroids) for chronic rhinosinusitis

Summary of findings 2. Anti‐IL‐5 mAb (mepolizumab) compared to placebo (on top of topical steroids) for chronic rhinosinusitis

Outcomes	Number of participants (studies)	Relative effect (95% CI)	*Anticipated absolute effects^ (95% CI)**			Certainty of the evidence (GRADE)	What happens
Anti‐IL‐5 mAb (mepolizumab) compared to placebo (on top of topical steroids) for chronic rhinosinusitis
Patients or population: patients with severe chronic rhinosinusitis with nasal polyps Setting: tertiary care Intervention: anti‐IL‐5 mAb (mepolizumab) Comparison: placebo (on top of topical steroids)
Outcomes	Number of participants (studies)	Relative effect (95% CI)	Without anti‐IL‐5 mAb (mepolizumab)	With anti‐IL‐5 mAb (mepolizumab)	Difference	Certainty of the evidence (GRADE)	What happens
Health‐related quality of life ‐ disease‐specific (SNOT‐22, range 1 to 100, lower = better) Follow‐up: 25 weeks	105 (1 RCT)	—	The mean disease‐specific health‐related quality of life score without anti‐IL‐5 mAb (mepolizumab) was 40.36.	—	MD 13.26 lower (22.08 lower to 4.44 lower)	⊕⊕⊝⊝ LOW¹	Aspects of health‐related quality of life that are directly impacted by chronic rhinosinusitis may have been better in participants who received mepolizumab but we are uncertain about this estimate.
Disease severity ‐ VAS (range 0 to 10, lower = better) Follow‐up: 25 weeks	72 (1 RCT)	—	The mean disease severity score without anti‐IL‐5 mAb (mepolizumab) was 6.21.	—	MD 2.03 lower (3.65 lower to 0.41 lower)	⊕⊝⊝⊝ VERY LOW^1,2	We are very uncertain about the impact of mepolizumab on overall chronic rhinosinusitis symptom severity.
Serious adverse events Follow‐up (range): 25 to 40 weeks	135 (2 RCTs)	RR 1.57 (0.07 to 35.46)	Study event rates³			⊕⊝⊝⊝ VERY LOW^1,4	We are very uncertain about the number of serious adverse events for chronic rhinosinusitis patients who use mepolizumab. The number of serious adverse events was 0/62 for placebo and 1/73 for mepolizumab.
			0.0%	1.37%

Avoidance of surgery ‐ patients still meeting the criteria for surgery At end of follow‐up (range): 25 to 40 weeks	135 (2 RCTs)	RR 0.78 (0.64 to 0.94)	Study population			⊕⊝⊝⊝ VERY LOW^1,2,4	We are very uncertain whether mepolizumab can help participants reduce the need for surgery.
	135 (2 RCTs)	RR 0.78 (0.64 to 0.94)	80.3%	62.7% (51.4 to 75.5)	17.7% fewer (28.9 fewer to 4.8 fewer)	⊕⊝⊝⊝ VERY LOW^1,2,4
Extent of disease ‐ endoscopic score Follow‐up (range): 25 to 40 weeks	137 (2 RCTs)	—	The mean endoscopic score without anti‐IL‐5 mAb (mepolizumab) ranged from 0 to ‐0.7.	—	MD 1.23 lower (1.79 lower to 0.68 lower)	⊕⊝⊝⊝ VERY LOW^1,2	We are very uncertain whether mepolizumab can reduce the extent of disease as measured by an endoscopic score.
Extent of disease ‐ CT scan score (Lund‐Mackay, range 0 to 24, lower = better) Follow‐up: 8 weeks	27 (1 RCT)	—	One study reported that CT scan scores were "not significantly different between groups"	—	—	⊕⊝⊝⊝ VERY LOW^1,5	We are very uncertain whether mepolizumab can reduce the extent of disease as measured by a CT scan score.
Health‐related quality of life ‐ generic, measured using the EQ‐5D visual analogue scale (range 0 to 100; 0 = worst imaginable health state, 100 = best imaginable health state) Follow‐up: at week 25	105 (1 RCT)	—	The mean generic health‐related quality of life score without anti‐IL‐5 mAb (mepolizumab) was 75.45	—	MD 5.68 higher (1.18 lower to 12.54 higher)	⊕⊕⊝⊝ LOW¹	We are uncertain about the impact of mepolizumab on overall quality of life or health status, as assessed by the EQ‐5D visual analogue scale.
Adverse events ‐ nasopharyngitis, including sore throat Follow‐up (range): 25 to 40 weeks	135 (2 RCTs)	RR 0.73 (0.36 to 1.47)	Study population			⊕⊕⊝⊝ LOW¹	We are uncertain about the risk of nasopharyngitis in chronic rhinosinusitis patients who used mepolizumab.
	135 (2 RCTs)	RR 0.73 (0.36 to 1.47)	22.6%	16.5% (8.1 to 33.2)	6.1% fewer (14.5 fewer to 10.6 more)	⊕⊕⊝⊝ LOW¹
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; MD: mean difference; RCT: randomised controlled trial; RR: risk ratio; SNOT‐22: Sino‐Nasal Outcome Test‐22; VAS: visual analogue scale
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
¹Downgraded by two levels due to imprecision: very small sample size resulting in a very imprecise estimation of effect sizes. ²Downgraded by one level due to study limitations: methods or criteria used in the measurement of the outcome were not validated. ³No events were reported in the placebo arm of these trials. We have therefore presented the study event rates rather than anticipated absolute events. ⁴Downgraded by one level due to indirectness: one study only assessed patients for two doses (Gevaert 2011). The other study evaluated six doses (24 weeks), but had a more than 30% dropout rate (Bachert 2017). Therefore, the length of follow‐up is inadequate and it is unclear whether this evidence related to safety is generalisable. ⁵Downgraded by one level due to study limitations: high risk of attrition bias, insufficient information to judge other aspects of study design and no numerical data presented for this outcome.

Summary of findings 2. Anti‐IL‐5 mAb (mepolizumab) compared to placebo (on top of topical steroids) for chronic rhinosinusitis

Summary of findings 3. Anti‐IgE mAb (omalizumab) compared to placebo (on top of topical steroids) for chronic rhinosinusitis

Outcomes	Number of participants (studies)	Relative effect (95% CI)	*Anticipated absolute effects^ (95% CI)**			Certainty of the evidence (GRADE)	What happens
Anti‐IgE mAb (omalizumab) compared to placebo (on top of topical steroids) for chronic rhinosinusitis
Patients or population: patients with chronic rhinosinusitis with nasal polyps Setting: tertiary care Intervention: anti‐IgE mAb (omalizumab) Comparison: placebo (on top of topical steroids)
Outcomes	Number of participants (studies)	Relative effect (95% CI)	Without anti‐IgE mAb (omalizumab)	With anti‐IgE mAb (omalizumab)	Difference	Certainty of the evidence (GRADE)	What happens
Health‐related quality of life ‐ disease‐specific (SNOT‐22, range 0 to 110, lower = better) Follow‐up: 24 weeks	265 (2 RCTs)	—	The mean change in disease‐specific HRQOL for the placebo group was ‐7.57 points	—	MD 15.62 points lower (19.79 lower to 11.45 lower)	⊕⊕⊕⊝ MODERATE¹	At 24 weeks, omalizumab probably results in an improvement in disease‐specific health‐related quality of life (as measured with the SNOT‐22 questionnaire). The size of the difference was clinically significant. However, we have limited confidence in this estimate because the sample size may be too small to estimate this accurately.
Disease severity, as measured by validated, patient‐reported symptom score	—	—	—	—	—	—	None of the studies reported this outcome.
Serious adverse events Follow‐up (range): 20 weeks to 6 months	329 (5 RCTs)	RR 0.32 (0.05 to 2.00)	2.5%	0.8% (0.1 to 5.1)	1.7% fewer (2.4 fewer to 2.5 more)	⊕⊝⊝⊝ VERY LOW^2,3	There is too little information. We are very uncertain whether omalizumab changes the incidence of serious adverse events because the sample size may be too small to estimate this accurately, or capture the range of adverse events that might occur in a larger population or with longer follow‐up. Serious adverse events were reported by 1/171 participants who took omalizumab versus 4/158 who took placebo.
Avoidance of surgery Nasal polyp score ≤4 (≤ 2 on each side) and an improvement in SNOT‐22 score of ≥ 8.9 points Follow‐up: 24 weeks	265 (2 RCTs)	RR 5.60 (1.99 to 15.76)	3.1%	17.1% (6.1 to 48.1)	14.0% more (3 more to 45.1 more)	⊕⊕⊕⊝ LOW^1,4	At up to 24 weeks, the evidence suggests that the number of participants in whom surgery was not thought to be necessary was greater in those who received omalizumab. However, we have limited confidence in this estimate because the sample size may be too small to estimate this accurately, and there are no widely agreed criteria to determine which patients need surgery for nasal polyps. Avoidance of surgery was reported in 23/134 participants who took omalizumab versus 4/131 participants who took placebo.
Extent of disease: endoscopic nasal polyp score (range 0 to 8, lower = better) Follow‐up: up to 24 weeks	312 (4 RCTs)	—	The median change in endoscopic nasal polyp score for the placebo group was ‐0.05 points	—	MD 1.26 points lower (2.2 lower to 0.31 lower)	⊕⊕⊝⊝ LOW^4,5	At up to 24 weeks, the evidence suggests that omalizumab may result in a reduction in the nasal polyp score. However, there are inconsistencies in the size of effect between studies, and it is unclear whether the method used is validated.
Extent of disease: CT scan (lower score = better) Follow‐up: 20 weeks	47 (2 RCTs)	—	The mean CT scan score without anti‐IgE mAb (omalizumab) ranged from ‐8.9 to 18.3	—	SMD 0.2 lower (1.55 lower to 1.14 higher)	⊕⊝⊝⊝ VERY LOW^2,6	There is too little information ‐ we are very uncertain whether there is a difference in the extent of disease with omalizumab. There are inconsistencies in the size and direction of effect. In the NCT01066104 study, the results favoured the placebo group, while in Gevaert 2013 they favoured the omalizumab group.
Health‐related quality of life ‐ generic (SF‐36) Follow‐up (range): 20 weeks to 6 months	38 (2 RCTs)	One study found no significant differences (P > 0.05, all comparisons) except for one domain, 'vitality' (omalizumab 9.4, placebo 12.5, P < 0.05). A second study found that physical health was significantly improved in the omalizumab group (P = 0.02) but not in the placebo group (P = 0.75). Mental health did not significantly improve in either treatment group.				⊕⊝⊝⊝ VERY LOW^7,8	We are very uncertain about the impact of omalizumab on health‐related quality of life.
Adverse events ‐ nasopharyngitis, including sore throat Follow‐up (range): 20 weeks to 6 months	329 (5 RCTs)	RR 0.71 (0.29 to 1.73)	6.9%	4.9% (2 to 12)	2.0% fewer (4.9 fewer to 5.1 more)	⊕⊕⊝⊝ LOW²	The evidence suggests that omalizumab may result in little to no difference in the incidence of nasopharyngitis, including sore throat. However, we have limited confidence in this estimate because the sample size may be too small to estimate this accurately. Nasopharyngitis or sore throat was reported by 8/170 participants who took omalizumab versus 11/159 who took placebo.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; CT: computerised tomography; RCT: randomised controlled trial; SMD: standardised mean difference; SNOT‐22: Sino‐Nasal Outcome Test‐22
GRADE Working Group grades of evidence High certainty: We are very confident that the true effect lies close to that of the estimate of the effect Moderate certainty: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different Low certainty: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect Very low certainty: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect
¹Downgraded by one level due to imprecision: small sample size resulting in an imprecise estimate of effect size. ²Downgraded by two levels due to imprecision: small sample size for the outcome estimated resulting in an imprecise estimation of effect size; confidence interval includes potential for considerable benefit or considerable harm. ³Downgraded by one level due to indirectness: some serious adverse effects are relatively rare ‐ a larger and more heterogeneous population or longer period of treatment and follow‐up may be needed. ⁴Downgraded by one level due to study limitations: method of assessment not validated. ⁵Downgraded by one level due to inconsistency: high and unexplained heterogeneity as the size of effect differed between the studies (I² = 90%). ⁶Downgraded by one level due to inconsistency: high and unexplained heterogeneity as the size and direction of effect differed between the studies (I² = 80%). ⁷Downgraded by two levels due to imprecision: very small sample size for the outcome measured. ⁸Downgraded by one level due to indirectness: a larger range of treatment doses and duration, and a more heterogeneous population, may be required to identify the effect of the intervention on quality of life.

Summary of findings 3. Anti‐IgE mAb (omalizumab) compared to placebo (on top of topical steroids) for chronic rhinosinusitis

Table 1. Summary of characteristics of included studies

	SINUS 24 (n = 276)	SINUS 52 (n = 448)	Bachert 2016 (n = 60)	Bachert 2017 (n = 107)	Gevaert 2011 (n = 30)	Pinto 2010 (n = 14)	Gevaert 2013 (n = 24)	NCT01066104 (n = 27)	POLYP 1 (n = 138)	POLYP 2 (n = 127)
Population	Bilateral nasal polyps (mean 5.75 points) with symptoms of chronic rhinosinusitis despite intranasal steroids	Bilateral nasal polyps (mean 6.10 points) with symptoms of chronic rhinosinusitis despite intranasal steroids	Chronic sinusitis with nasal polyps (mean 5.8 points)	Severe, recurrent bilateral nasal polyposis requiring surgery (worst affected nostril ≥ 3 (on 4‐point scale), and symptoms score > 7 on 10 cm VAS despite intranasal steroids and/or previous oral corticosteroids Mean bilateral polyp score 6.29	Chronic rhinosinusitis with severe primary polyps (grade 3 to 4) or recurrent polyps (any grade) Failure of standard care for chronic rhinosinusitis	Chronic rhinosinusitis Polyps status: 7/7 in omalizumab and 5/7 in placebo had nasal polyposis	Chronic rhinosinusitis with nasal polyps Polyps status: TPS (total nasal endoscopic polyp score), median (IQR): 6 (4 to 6); 6 (6 to 8)	Chronic rhinosinusitis with nasal polyps Inclusion criteria state minimum polyp score of 4	Chronic rhinosinusitis with nasal polyps Inclusion criteria state minimum polyp score of 5	Chronic rhinosinusitis with nasal polyps Inclusion criteria state minimum polyp score of 5
Comorbidity	Asthma 58%	Asthma 60%	Asthma 58%	Asthma 78%	Asthma 43%	Inhaled asthma therapy taken by 72% (5/7) in omalizumab group and 43% (3/7) in placebo group	Asthma (100%)	No information	Asthma 54%	Asthma 60%
Eligible for surgery?	No information	No information	No information	Yes^a	No information	100% had undergone endoscopic sinus surgery, but no information on eligibility for more surgery	No information	No information	No information	No information
Intervention	Dupilumab 300 mg subcutaneously every 2 weeks	a) Dupilumab 300 mg subcutaneously every 2 weeks for 24 weeks, followed by every 4 weeks until 52 weeks b) Dupilumab 300 mg subcutaneously every 2 weeks for 52 weeks in total	Dupilumab 600 mg loading dose subcutaneously, followed by 300 mg every week	Mepolizumab 750 mg intravenously every 4 weeks	Mepolizumab 750 mg intravenously every 4 weeks	Omalizumab subcutaneously, once or twice monthly (dose dependent on participant weight and serum IgE level), for 6 months	Omalizumab subcutaneously every 2 weeks (8 injections in total) or every month (4 injections in total), based on total serum IgE levels and body weight, with a maximum dose of 375 mg	Omalizumab subcutaneously, every 2 to 4 weeks depending on baseline serum total IgE level and body weight	Omalizumab 75 mg to 600 mg subcutaneously, every 2 to 4 weeks depending on baseline serum total IgE level and body weight	Omalizumab 75 mg to 600 mg subcutaneously, every 2 to 4 weeks depending on baseline serum total IgE level and body weight
Comparison	Placebo subcutaneously every 2 weeks	Placebo subcutaneously every 2 weeks	Placebo subcutaneously every week	Intravenous placebo every 4 weeks	Intravenous placebo every 4 weeks	Placebo injection, same dose and frequency	Placebo injection, same dose and frequency	Stated as "Xolair placebo 150‐375 mg depending on baseline serum total IgE level and body weight"	Placebo injection at corresponding dose and frequency	Placebo injection at corresponding dose and frequency
Treatment length	24 weeks	52 weeks	15 weeks	24 weeks	8 weeks (2 doses)	26 weeks	16 weeks	22 weeks	24 weeks	24 weeks
Follow‐up length (total treatment and follow‐up period)	24 weeks	24 weeks and 52 weeks	16 weeks	25 weeks	48 weeks (most outcomes assessed after 8 weeks' treatment)	26 weeks	20 weeks (outcomes assessed after 16 weeks' treatment)	22 weeks	28 weeks (most outcomes assessed after 24 weeks' treatment)	28 weeks (most outcomes assessed after 24 weeks' treatment)
Specific HRQL	Measured and reported^b	Measured and reported^b	Measured and reported^b	Measured and reported^b	Not measured	Measured and reported^b	Measured and reported^c	Not measured	Measured and reported^b	Measured and reported^b
Disease severity (overall)	Measured and reported^d,e	Measured and reported^d,e	Measured and reported^d,j	Measured and reported^d	No global questionnaire reported Specific symptoms measured and reported^f	No global questionnaire reported Specific symptoms measured and reported^g,h	No global questionnaire reported Specific symptoms measured and reportedⁱ	No global questionnaire reported Measured but not reported^k	No global questionnaire reported Specific symptoms measured and reported^aa	No global questionnaire reported Specific symptoms measured and reported^aa
Severe adverse event	Measured and reported	Measured and reported	Measured and reported	Measured and reported	Measured and reported	Measured and reported	Not measured	Measured and reported	Measured and reported	Measured and reported
Avoidance of Surgery	Measured and reported^l,m	Measured and reported^l,n	Not measured	Measured and reported^o	Not measured	Not measured	Not measured	Not measured	Measured and reported^bb	Measured and reported^bb
CT scan	Measured and reported^p	Measured and reported^p	Measured and reported^p	Not measured	Measured and reported^q	Measured and reported^r	Measured and reported^p	Measured and reported^s	Not measured	Not measured
Polyps score	Measured and reported^t	Measured and reported^t	Measured and reported^t	Measured and reported^u	Measured and reported^t	Measured and reported^v	Measured and reported^t	Measured and reported^t	Measured and reported^t	Measured and reported^t
Generic HRQL	Measured and reported^w,m	Measured and reported^w,m	Measured and reported^w	Measured and reported^w,x	Not measured	Measured and reported^y	Measured and reported^y	Not measured	Measured, not reported^cc	Not measured
Nasopharyngitis	Measured and reported	Measured and reported	Measured and reported	Measured and reported	Measured and reported	Not measured^z	Not measured	Not measured	Measured and reported^dd	Measured and reported
Main data source	Publications; generic health‐related quality of life and avoidance of surgery data from trial registry only	Publications; generic health‐related quality of life and avoidance of surgery data from trial registry only	Publications	Publications	Publications	Publication	Publication	NCT record (no publications)	Publication	Publication Nasopharyngitis data for POLYP 2 alone (not pooled with POLYP 1) from NCT record
^aWorst affected nostril ≥ 3 (on a 4‐point scale), and symptoms score > 7 on 10 cm VAS despite intranasal steroids and/or previous oral corticosteroids. ^bSNOT‐22, scale 0 to 110, higher = worse, minimal clinically important difference (MID) ≥ 8.9 points. ^cRSOM‐31; AQLQ. ^dVisual analogue scale for rhinosinusitis: "how troublesome are your symptoms?", scale 0 to 10 cm, higher = worse. ^eTotal symptom severity score (including nasal congestion, rhinorrhoea and sense of smell, each rated between 0 and 3), total scale 0 to 9, higher = worse. ^fFour individual symptoms were measured (anterior rhinorrhoea, nasal obstruction, postnasal drip and loss of sense of smell); reported only as narrative summary. ^gTotal nasal symptom score (TNSS): nasal obstruction, nasal discharge, facial pain and hyposmia) each recorded on a 4‐point scale (0 = none, 1 = mild, 2 = moderate, 3 = severe); total scores summed. ^hOnly reported as 'no significant difference' ‐ no data presented. ⁱDisease severity symptom score: nasal and asthma symptoms (patient‐reported, daily 'absent, mild, moderate or severe' (scores 0, 1, 2, 3). ^jSeverity scores for individual symptoms (nasal congestion, anterior and posterior rhinorrhoea, loss in sense of smell, nocturnal awakenings), range 0 to 3, higher = worse. ^kNCT record states that a total symptom score (TSS) and chronic rhinosinusitis facial pain/headache questionnaire were recorded daily; no outcome data presented in NCT record. ^lNumber of participants requiring rescue with nasal polyp surgery ‐ no definition for eligibility provided. ^mOutcome reported, but specific data only reported in trial registry (publication includes pooled data with SINUS 52 only). ⁿOutcome measured but not reported (pooled data with SINUS 24 only, specific data for this trial not reported on trial registry or publication). ^oAt study endpoint, participants with a nasal polyp score of ≥ 3 were deemed as continuing to need surgery (regardless of VAS score). In addition, participants with a nasal polyp score of 2, who had a VAS score of > 7 were also viewed as requiring surgery. ^pLund‐Mackay CT score, range 0 to 24, higher = worse. ^qPublication reports proportion of participants who showed improvement in CT score during the study. Shown separately for three independent raters, with no summary measure reported. ^rMucosal thickness on CT scan. ^sCT scan scored using the Zinreich modification of the Lund‐Mackay scoring system. ^tBilateral "endoscopic nasal polyps score" (NPS) or total polyps score (TPS), range 0 to 8, higher = worse. ^uImprovement by at least one point in endoscopic nasal polyp score. ^vNasal endoscopy score (0 to 4). Unclear which scoring system used. ^wEQ‐5D visual analogue scale, range 0 to 100 (100 = best imaginable). ^xEQ‐5D index score, range 0 to 1, higher = better. ^ySF‐36. ^zOutcome not specifically mentioned, paper just states "No side effects or adverse events occurred during the study". ^aaTotal Nasal Symptom Score and individual components of this were reported, which included anterior rhinorrhoea, posterior rhinorrhoea, nasal congestion and loss of sense of smell. Each scored with a range of 0 to 3, higher = worse. Total score out of 12. ^bbAvoidance of surgery was defined as an improvement in SNOT‐22 score of at least 8.9 points and a nasal polyp score no greater than 4 points (with a unilateral score of no more than 2 on either side). ^ccProtocol states that EuroQol 5‐Dimension 5‐Level Questionnaire will be used, but results not reported. ^ddNasopharyngitis reported as pooled data with POLYP 2; however the data for POLYP 2 are also reported separately, therefore individual data for POLYP 1 can be calculated.

Table 1. Summary of characteristics of included studies

Table 2. Eligibility for surgery

Study name	Study	Eligibility for surgery: defined at randomisation?			Eligibility criteria for surgery: as recorded in results
Study name	Study	Yes	No	Description of how decisions were made to carry out/offer surgery	Yes	No	Remarks
Completed (included) studies
SINUS 52 (NCT02898454)	EUCTR2015‐001314‐10‐ES 2016		x	Not mentioned		x	Criteria not defined but one outcome was "Proportion of patients during study treatment receiving oral corticosteroid (OCS) for NP and/or planned to under surgery for nasal polyps"
SINUS 24 (NCT02898454)	Bachert 2019 NCT02898454		x	Not mentioned	x		Offered when there was worsening of signs and/or symptoms during the study Criteria not applied at baseline Who: not mentioned 28.3% nasal polyp surgery
	EUCTR2015‐003101‐42‐BG 2017 NCT02912468		x	Not mentioned		x	Criteria not defined but one outcome was "Proportion of patients during study treatment receiving oral corticosteroid (OCS) for NP and/or planned to under surgery for nasal polyps"
	Han 2019		x	Not mentioned		x	Full text not available but one outcome was "Reduction of surgery for nasal polyps"
NCT01066104	NCT01066104		x	Not mentioned		x
Pinto 2010 (NCT00117611)	NCT00117611 Pinto 2010 Mehta 2009		x	Not mentioned		x
Bachert 2017 (NCT01362244)	NCT01362244	x		Stated in the protocol Endoscopic nasal polyp score ≥ 3 and VAS > 7 Number of patients qualified at baseline: 105 Number of patients qualified at endpoint: 84 Number of patients who had surgery: not mentioned	x		Criteria for endoscopic nasal polyp score of ≥ 3, or nasal polyp score of 2 and a VAS symptom score of > 7 Criteria different from those applied at baseline Who: not mentioned 80% qualified for surgery
Bachert 2017 (NCT01362244)	EUCTR2008‐003772‐21‐NL 2009	x		Stated in the protocol refractory response to steroid therapy Number of patients qualified at baseline: 105 Number of patients qualified at endpoint: 79 Number of patients who had surgery: not mentioned	x		Criteria endoscopic nasal polyp score of ≥ 3, or nasal polyp score of 2 and a VAS symptom score of > 7 Criteria different from those applied at baseline Who: not mentioned 75% qualified for surgery
Gevaert 2013 (NCT01393340)	NCT01393340 Gevaert 2013 Gevaert 2012	x		Not mentioned
Bachert 2016 (NCT01920893)	NCT01920893 EUCTR2013‐001803‐35‐BE 2013 Bachert 2016 Other related publications: Bachert 2015 Schneider 2016 Willits 2016		x	Not mentioned		x
Gevaert 2011	Gevaert 2011		x	Not mentioned		x
POLYP 1 (NCT03280550)	NCT03280550		x	Stated in the protocol: Reduction in the need for surgery by week 24, as defined by a nasal polyps score of ≤ 4 (unilateral score of ≤ 2 on each side) and improvement in SNOT‐22 score of ≥ 8.9 Number of patients who qualified for surgery at baseline: not reported ‐ assumed all participants (inclusion criteria of nasal polyps score ≥ 5 with unilateral score of ≥ 2 for each nostril) Number of patients who qualified for surgery at endpoint: 123	x		No need for surgery when a nasal polyps score of ≤ 4 (unilateral score of ≤ 2 on each side) and improvement in SNOT‐22 score of ≥ 8.9 Criteria not reported at baseline Who: not mentioned
POLYP 2 (NCT03280537)	EUCTR2017‐001718‐28‐BE 2017 NCT03280537		x	Stated in the protocol: Reduction in the need for surgery by week 24, as defined by a nasal polyps score of ≤ 4 (unilateral score of ≤ 2 on each side) and improvement in SNOT‐22 score of ≥ 8.9 Number of patients who qualified for surgery at baseline: not reported ‐ assumed all participants (inclusion criteria of nasal polyps score ≥ 5 with unilateral score of ≥ 2 for each nostril) Number of patients who qualified for surgery at endpoint: 115	x		No need for surgery when a nasal polyps score ≤ 4 (unilateral score of ≤ 2 on each side) and improvement in SNOT‐22 score of ≥ 8.9 Criteria not applied at baseline Who: not mentioned
Included studies (not published)
NCT02772419	NCT02772419		x	Not mentioned		x
NCT02734849	NCT02734849		x	Not mentioned		x
Ongoing studies
NAPPREB (NCT04185012)	NCT04185012		x	Not mentioned on trial registry		x
ORCHID (NCT04157335)	NCT04157335	x		Stated on trial registry: Patients with bilateral sinonasal polyps that, despite treatment with standard of care including a history of treatment with systemic corticosteroids (oral, parenteral) or prior surgery for nasal polyps, have severity consistent with a need for surgery as described by: a minimum total nasal polyp score of 5 out of a maximum score of 8 (with a unilateral score of at least 2 for each nostril); ongoing symptoms for at least 12 weeks; patient‐reported moderate to severe nasal blockage (score 2 or 3 out of 3).			Ongoing study
OSTRO (NCT03401229)	NCT03401229	x		Stated in the protocol A minimum total nasal polyp score (NPS) of 5 out of a maximum score of 8 (with a unilateral score of at least 2 for each nostril) at V1 and continuously maintained at V2 to meet the randomisation criterion, as determined by the study Imaging Core Lab Ongoing symptoms for at least 12 weeks prior to V1 Patient‐reported moderate to severe nasal blockage score (NBS) 2 or 3 over the 2 weeks prior to V1 (2‐week recall assessment of symptoms, scores 0 (none) to 3 (severe)) Number of patients qualified at baseline: ongoing Number of patients qualified at endpoint: ongoing Number of patients who had surgery: ongoing			Ongoing study
SYNAPSE (NCT03085797)	NCT03085797	x		Stated in the protocol An overall VAS symptom score > 7, or an endoscopic bilateral nasal polyps score of at least 5 out of a maximum score of 8 (with a minimum score of 2 in each nasal cavity) Number of patients qualified at baseline: ongoing Number of patients qualified at endpoint: ongoing Number of patients had surgery: ongoing			Ongoing study
NCT02799446	NCT02799446		x	Not mentioned		x
NCT03614923	NCT03614923		x	Not mentioned		x
NCT03450083	NCT03450083		x	Not mentioned		x	Criteria not defined but one outcome was time to nasal polyp surgery
NCT04362501	NCT04362501		x	Not mentioned on trial registry		x
NCT044330179	NCT044330179		x	Not mentioned on trial registry		x
NP: nasal polyps NPS: nasal polyps score SNOT‐22: Sino‐Nasal Outcome Test‐22 VAS: visual analogue scale

Table 2. Eligibility for surgery

Comparison 1. Anti‐IL‐4Rα mAb (dupilumab) versus placebo (on top of topical steroids)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 HRQL ‐ disease‐specific (SNOT‐22, 0 to 110, lower = better) Show forest plot	3		Mean Difference (IV, Random, 95% CI)	Subtotals only

1.1.1 Up to 24 weeks	3	784	Mean Difference (IV, Random, 95% CI)	‐19.61 [‐22.54, ‐16.69]
1.1.2 At 52 weeks	1	303	Mean Difference (IV, Random, 95% CI)	‐22.38 [‐27.10, ‐17.66]
1.2 Disease severity ‐ VAS (0 to 10, lower = better) Show forest plot	3	784	Mean Difference (IV, Random, 95% CI)	‐3.00 [‐3.47, ‐2.53]

1.3 Serious adverse events Show forest plot	3	782	Risk Ratio (M‐H, Fixed, 95% CI)	0.47 [0.29, 0.76]

1.4 Avoidance of surgery ‐ number of patients who had surgery as rescue treatment Show forest plot	2	725	Risk Ratio (M‐H, Random, 95% CI)	0.17 [0.05, 0.52]

1.5 Extent of disease ‐ endoscopy ('nasal polyps score', 0 to 8, higher = worse) Show forest plot	3		Mean Difference (IV, Random, 95% CI)	Subtotals only

1.5.1 Up to 24 weeks	3	784	Mean Difference (IV, Random, 95% CI)	‐1.80 [‐2.25, ‐1.35]
1.5.2 Up to 52 weeks	1	303	Mean Difference (IV, Random, 95% CI)	‐2.34 [‐2.77, ‐1.91]
1.6 Extent of disease ‐ CT scan (Lund Mackay, 0 to 24, higher = worse) Show forest plot	3	784	Mean Difference (IV, Random, 95% CI)	‐7.00 [‐9.61, ‐4.39]

1.7 HRQL ‐ generic (EQ‐5D VAS, 0 to 100, higher = better) Show forest plot	3	766	Mean Difference (IV, Random, 95% CI)	8.29 [5.73, 10.85]

1.8 Adverse events ‐ nasopharyngitis, including sore throat (longest available data) Show forest plot	3	783	Risk Ratio (M‐H, Random, 95% CI)	0.95 [0.72, 1.25]

Comparison 1. Anti‐IL‐4Rα mAb (dupilumab) versus placebo (on top of topical steroids)

Comparison 2. Anti‐IL‐5 mAb (mepolizumab) versus placebo (on top of topical steroids)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 HRQL ‐ SNOT‐22 (1 to 100, lower = better) up to 25 weeks Show forest plot	1	105	Mean Difference (IV, Random, 95% CI)	‐13.26 [‐22.08, ‐4.44]

2.2 Disease severity ‐ VAS (0 to 10, lower = better) Show forest plot	1	72	Mean Difference (IV, Random, 95% CI)	‐2.03 [‐3.65, ‐0.41]

2.3 Serious adverse events Show forest plot	2	135	Risk Ratio (M‐H, Fixed, 95% CI)	1.57 [0.07, 35.46]

2.4 Avoidance of surgery ‐ patients still meeting criteria for surgery at end of follow‐up Show forest plot	2	135	Risk Ratio (M‐H, Random, 95% CI)	0.78 [0.64, 0.94]

2.4.1 Patients still meeting criteria for surgery at 24 weeks	1	105	Risk Ratio (M‐H, Random, 95% CI)	0.78 [0.64, 0.95]
2.4.2 Patients requiring 'rescue' surgery during trial	1	30	Risk Ratio (M‐H, Random, 95% CI)	0.67 [0.18, 2.42]
2.5 Extent of disease ‐ endoscopic score Show forest plot	2	137	Mean Difference (IV, Random, 95% CI)	‐1.23 [‐1.79, ‐0.68]

2.6 HRQL ‐ generic measured using EQ‐5D VAS (range 0 to 100; 0 = worst, 100 = best imaginable health state) at week 25 Show forest plot	1	105	Mean Difference (IV, Random, 95% CI)	5.68 [‐1.18, 12.54]

2.7 Adverse events ‐ nasopharyngitis, including sore throat Show forest plot	2	135	Risk Ratio (M‐H, Random, 95% CI)	0.73 [0.36, 1.47]

Comparison 2. Anti‐IL‐5 mAb (mepolizumab) versus placebo (on top of topical steroids)

Comparison 3. Anti‐IgE mAb (omalizumab) versus placebo (on top of topical steroids)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 HRQL disease‐specific ‐ SNOT‐22 (0 to 110, lower = better) Show forest plot	2	265	Mean Difference (IV, Random, 95% CI)	‐15.62 [‐19.79, ‐11.45]

3.2 Serious adverse events Show forest plot	5	329	Risk Ratio (M‐H, Fixed, 95% CI)	0.32 [0.05, 2.00]

3.3 Avoidance of surgery Show forest plot	2	265	Risk Ratio (M‐H, Random, 95% CI)	5.60 [1.99, 15.76]

3.4 Extent of disease ‐ endoscopic score (nasal polyps score, range 0 to 8, lower = better) Show forest plot	4	312	Mean Difference (IV, Random, 95% CI)	‐1.26 [‐2.20, ‐0.31]

3.5 Extent of disease ‐ CT scan (lower score = better) Show forest plot	2	47	Std. Mean Difference (IV, Random, 95% CI)	‐0.20 [‐1.55, 1.14]

3.6 Adverse events ‐ nasopharyngitis, including sore throat Show forest plot	5	329	Risk Ratio (M‐H, Random, 95% CI)	0.71 [0.29, 1.73]

Comparison 3. Anti‐IgE mAb (omalizumab) versus placebo (on top of topical steroids)