3.1 The key clinical outcomes presented in the decision problem were:
change in sinusitis symptoms
number of post‑procedure sinusitis episodes needing medication
number of postoperative debridements
change in ostial patency (assessed by endoscopy or CT scan)
number and types of sinus treated
length of hospital stay
procedure time and theatre/outpatient treatment room time
rate of revision surgery
number of sinus‑related follow‑up appointments
rate of readmission
rate and severity of nasal bleeding
device‑related adverse events.
3.2 The company conducted a literature search for evidence on XprESS and its predecessor device FinESS, which identified 13 papers describing 9 trials, 7 published and 2 unpublished. The retrieved papers included a meta‑analysis involving 6 of the 11 published studies.
3.3 The external assessment centre (EAC) judged the company's search terms to be appropriate, but could not fully reproduce them because the search strategies were not fully reported. The EAC re‑ran the company's searches and conducted its own search, which identified no further evidence.
3.4 The EAC considered that 1 included study, Eloy et al. (2012), should be excluded from further assessment because the population (patients who had previously had a failed frontal sinustomy) was not consistent with the scope. The EAC therefore assessed 12 publications, which reported on 1 randomised controlled trial and 7 observational studies, 2 of which were unpublished.
3.5 Three studies (Cutler et al. 2013, Bikhazi et al. 2014, Chandra at al. 2016) reported on the REMODEL trial, a prospective, multicentre, non‑inferiority, parallel, randomised clinical trial (the methodology is most comprehensively reported in Cutler et al. 2013). The REMODEL trial compared FESS with balloon dilation systems (FinESS and XprESS) in adult patients with uncomplicated chronic sinusitis or recurrent acute sinusitis associated with maxillary sinus disease with or without anterior ethmoid sinus disease. The split between XprESS and FinESS was not reported but the company has indicated it was approximately 50:50. Patients and clinicians were blinded to their allocation. Blinding could not be maintained after treatment allocation, but some post‑surgical assessments were done or audited by independent physicians. Following withdrawals after randomisation, there were 50 patients in the balloon arm and 42 in the FESS arm. A post hoc modified intention‑to‑treat analysis was done. The primary outcome measure was change in chronic sinusitis symptoms as measured by the Sino‑Nasal Outcome Test‑20 (SNOT‑20) scores at 6 months from baseline (pre‑procedure).
3.6 Cutler et al. (2013) reported outcomes up to 6 months after the procedure. At 1 week, the average change in SNOT‑20 scores in the balloon arm was −1.49 (standard deviation [SD]=0.87), compared with −0.96 (SD=1.12) in the FESS arm. At 1 month, the average change was −1.70 (SD=0.98) for the balloon arm and −1.62 (SD=0.95) for FESS. At 6 months, the change was −1.67 (SD=1.10) for the balloon arm and −1.60 (SD=0.96) for FESS. The changes from baseline were significant (p<0.001) in both groups at all time points, and because the change in score exceeded 0.8, the differences were judged to be clinically meaningful. With the exception of the results at 1 week (p=0.014), there was no statistically significant difference between the SNOT‑20 scores in the balloon dilation and FESS arms. This indicated non‑inferiority of the balloon procedures in terms of symptom improvement, with a potentially significant short‑term effect (at 1 week). The authors also reported significant (p<0.0001) and clinically meaningful improvements in each of the subscales of the SNOT‑20 at 6 months, with no statistically significant differences between the 2 arms. The same results were reported at 6 months for the following subgroups: maxillary only or maxillary and anterior ethmoid; presence or absence of accessory ostia; presence or absence of septal deviation and sinusitis diagnosis (chronic or recurrent acute). In the balloon arm, 92.0% (46/50) of patients did not need a postoperative debridement compared with 26.2% (11/42) of patients in the FESS arm. There was a mean of 0.1±0.6 postoperative debridements per patient in the balloon arm compared with 1.2±1.0 in the FESS arm (p<0.0001). No statistically significant differences were found between balloon dilation and FESS in terms of post‑discharge nausea or duration of over‑the‑counter pain medication. One patient in each arm had revision surgery.
3.7 Bikhazi et al. (2014) described 12‑month results for 89 of the 92 patients reported by Cutler (2013) who completed 1‑year follow‑up (48 balloon, 41 FESS). Changes in SNOT‑20 scores from baseline remained statistically significant and clinically meaningful in both groups, and confirmed non‑inferiority at 12 months between the 2 interventions on this measure (balloon arm: −1.64±1.06, FESS arm: −1.65±0.94; p<0.0001). In both arms patients reported significant reductions (p<0.0001) in sinusitis episodes at 12 months following surgery compared with the year before (4.2 in the balloon arm, 3.5 in the FESS arm), although the comparison between the 2 was not statistically significant. Overall patency (maxillary ostia) in those with an evaluable CT scan at 12 months was 96.7% in the balloon arm and 98.7% in the FESS arm but this was not statistically significant. Both treatments had positive effects in all the domains of the Work Productivity and Activity Impairment (WPAI) survey, except for FESS in the absenteeism domain (p=0.169).
3.8 All eligible patients in Chandra et al. (2016) reported longer‑term outcomes at 18 months (n=66) and 24 months (n=25), and included an additional cohort who had been subsequently randomised (a total of 135 patients, 133 patients at 6 months and 130 patients at 12 months). Mean changes in SNOT‑20 scores at 6 and 12 months were statistically significantly lower than baseline and clinically meaningful in both arms in this enlarged cohort (6 months, balloon arm −1.56, FESS arm −1.60; 12 months, balloon arm −1.59, FESS arm −1.60). Mean changes in SNOT‑20 scores were also statistically significantly lower than baseline and clinically meaningful in the patients from the original cohort followed up at 24 months (balloon arm −1.65, FESS arm −1.45). There were no statistically significant differences between the 2 arms. Overall revision rates at 18 months were 2.7% in the balloon arm and 6.9% in the FESS arm (not statistically significant).
3.9 The company and EAC identified a number of observational studies which compared balloon dilation (XprESS or FinESS) with baseline data. The EAC considered them to be lower quality evidence. Symptom improvement data from some of these studies were pooled in a meta‑analysis reported in Chandra et al. (2016).
3.10 The XprESS Multi‑Sinus Study (Gould et al. 2016) was a single‑arm, prospective observational study which enrolled 82 adults with chronic sinusitis or acute recurrent sinusitis; the method of recruitment was not reported. Patients had to have maxillary sinus disease as a minimum, although patients with additionally affected sinuses (frontal, sphenoid or ethmoid) were also included. The study found a significant and clinically meaningful improvement in the primary outcome, change in mean SNOT‑20 score at 12 months, compared with baseline (−1.57, p<0.0001). At 12 months there were also statistically significant reductions in Rhinosinusitis Symptoms Inventory (RSI) major symptoms score, medication use, absenteeism, and acute sinus infection and sinus‑related physician visits. The authors reported that the procedure was a technical success in 307 of 313 sinuses operated on (98.1%), with only 1 patient needing revision of the procedure at 12 months (1.3%), with no serious device or procedural adverse events. The procedure appeared to be well tolerated (mean pain VAS 2.8±2.2), with a high degree of patient satisfaction (87.8%).
3.11 The XprESS registry (Brodner et al. 2013) was the first full clinical study of XprESS. This was a prospective observational study that enrolled 175 patients needing treatment of the frontal recess and sphenoid sinus ostium, who had previously been scheduled for FESS. The primary outcome was safety, although effectiveness outcomes were also prespecified. Of the targeted sinuses, 96% (479/497) were successfully accessed and treated with the balloon, including 276 frontal recesses, 131 sphenoid ostia, and 72 maxillary ostia/ethmoid infundibula. In 4 the balloon did not inflate, and in 10 the ostia could not be accessed using XprESS so FESS was used instead. Over 90% (448 of 497) of sinuses were treated using a hybrid procedure of FESS and XprESS. Because these results were not disaggregated, they were not included in the Chandra (2016) meta‑analysis, and the EAC considered them to be of limited relevance. Results were similar to the other observational studies employing standalone balloon dilation only, and included statistically significant reductions at 3 and 12 months in SNOT‑20 score (−1.1), and in medication use, work or school days missed and sinus‑related physician visits in the year following surgery compared with the year before. There was no statistically significant reduction in acute sinus infections reported after the procedure, and no serious adverse events reported.
3.12 The XprESS Maxillary Pilot Study (Gould et al. 2012) was a single‑arm, prospective observational study involving 21 adults with uncomplicated refractory chronic sinusitis or recurrent acute sinusitis of the maxillary or anterior ethmoid sinuses. All patients had the XprESS procedure under local anaesthesia, and the main outcome was change in SNOT‑20 score from pre‑procedure to up to 6‑months post‑procedure. The study was not peer reviewed.
3.13 The RELIEF study (Levine et al. 2013) was a single‑arm, prospective observational study involving 74 adult patients with refractory chronic sinusitis or recurrent acute sinusitis of the maxillary and anterior ethmoid sinuses. The primary outcome was quality of life as measured by SNOT‑20; this and most other outcomes were reported at 12 months. All patients had the procedure with FinESS, the predecessor device to XprESS. There was a statistically significant and clinically meaningful reduction in SNOT‑20 score (−1.2) compared with baseline. Statistically significant reductions were also reported in RSI major symptoms, medication use (intranasal corticosteroids, antihistamines, antibiotics), absenteeism, sinus‑related physician visits, and acute sinus infections. The procedure was reported as a technical success in 91.9% of sinuses operated on (124 of 135) with a revision surgery rate of 5.8% (4 of 69 patients). No serious adverse events were reported
3.14 The BREATHE study was published in 3 papers: Stankiewicz (2011 and 2012) and Cutler (2011). This was the first published study of an Entellus balloon product (FinESS) involving 71 patients with chronic sinusitis of the maxillary or ethmoid sinuses. The study was a single‑arm, prospective study. Follow‑up was 2 years with the primary outcome of quality of life improvement measured using SNOT‑20. There was a statistically significant and clinically meaningful improvement compared with baseline in SNOT‑20 at 1 year (−1.80) and 2 year (−1.86) follow‑up. At 1 year there was also a statistically significant reduction in WPAI survey score and on the Work Limitation Questionnaire (WLQ) compared with baseline. The technical success rate was reported as 97.7% (129 of 132 sinuses). Procedures were well tolerated with a mean pain VAS of 2.7, and 88% of patients were reported to have recovered within 2 days. Patient satisfaction rates were 89% after 1 year and 91.5% after 2 years. After 2 years, 4 of 59 patients (6.8%) needed revision surgery. One patient was reported as having suffered a serious procedure‑related adverse event following balloon dilation (subcutaneous emphysema).
3.15 The protocol for the FinESS registry study was published on ClinicalTrials.gov but was only provided as an abstract, and has not been subsequently published or peer reviewed. Because the EAC could not appraise this study, and only limited outcomes were reported, it did not consider it further. Data from the FinESS registry did contribute to the meta‑analysis by Chandra et al. (2016).
3.16 Soler et al. (2016) is a single‑arm, prospective observational study (n=50) expected to be published in 2016. It was provided to the EAC as an abstract that did not allow for critical appraisal, and only limited results were reported as academic in confidence. This was the only study that was reported on children. Although children were included the scope of the decision problem as a subgroup, the EAC understands through discussion with clinical experts that sinus surgery is rarely done in children in England. Because of this, the EAC did not consider the study any further.
3.17 Chandra et al. (2016) undertook a meta‑analysis of the observational studies (excluding the XprESS registry) and the REMODEL trial to evaluate the clinical effectiveness of Entellus balloon dilation devices in a larger population. Results on SNOT‑20, RSI scores and short‑term outcomes were reported. The authors had access to individual patient data so the EAC could not replicate the meta‑analyses. The authors reported that there was no statistical difference in SNOT‑20 outcomes between studies (REMODEL FESS arm, REMODEL balloon dilation arm or pooled observational studies), measured at 6, 12 and 24 months. There were significant reductions (p<0.0001) from baseline to 12 months in the standalone balloon dilation studies in absenteeism (5.0 days±9.5), homebound because of nasal problems (6.3 days±11.3), number of physician/nurse visits because of nasal problems (4.5±11.5), number of infections of nose/sinuses (3.9±4.5), and number of antibiotic courses (2.9±3.1).
3.18 Changes in WLQ score over 1 week, 1 month, 3 months, 6 months, 12 months, 18 months and 24 months compared with baseline were presented as a longitudinal graph. There were statistically significant and immediate reductions in several domains, which appeared maximal at 1 month before plateauing over 2 years. Revision rates at 12 months were 1.7% for the FESS arm of the REMODEL trial, 1.4% for the balloon dilation arm of the REMODEL trial and 3.2% for the pooled analysis (p=0.628). However, this analysis was based on a single patient in each of the REMODEL arms.
3.19 The company conducted a limited search for adverse events and identified 5 case reports of adverse events with a different balloon technology and 3 that did not specify which device was used). The EAC searched the FDA MAUDE database for Entellus and identified 12 reports, of which 8 involved XprESS. Of the reports, 6 described a cerebral spinal fluid leak in balloon‑only procedures (n=2), balloon with septoplasty (n=2), or hybrid endoscopic sinus surgery procedures (n=2). None noted any long‑term adverse health effects as a consequence. One report was a case of orbital wall damage identified by the company in its clinical evidence submission, which was reported to have had no long‑term adverse effect on the patient's vision. The eighth reported case was a death from massive intracranial bleed, shortly after successful completion of a bilateral maxillary balloon procedure. This was reported by the clinicians involved as unrelated to the device or procedure.
3.20 The EAC considered that the best evidence was from the REMODEL trial. This study design was assessed as being of high methodological quality, and internal validity was generally good. However, the EAC noted concerns about the high initial attrition rates in the FESS arm immediately following randomisation, which may have introduced differences between the characteristics of the 2 arms. The EAC was satisfied that the evidence showed balloon dilation to be non‑inferior to FESS in terms of the primary outcome (SNOT‑20) for up to 2 years post‑procedure. The EAC also judged that balloon dilation was equivalent to FESS in the secondary outcomes measured, such as maintaining ostia patency, reducing future episodes of sinusitis, and improving work and productivity. However, it noted that long‑term outcomes were assessed on small patient numbers. The EAC considered that balloon dilation with XprESS offers advantages over FESS by speeding recovery, reducing postoperative pain and reducing the need for nasal debridement.
3.21 The observational studies supplemented the evidence from REMODEL and were supportive of its results. However, the EAC noted a number of methodological weaknesses in all the observational studies which led it to conclude that the evidence from these studies was of limited quality to inform the decision problem. Although the studies matched the scope, the EAC was concerned about extrapolating the results from selected patient cohorts enrolled in trials in the US to the wider population of patients in the NHS. The EAC assumed equivalence between the FinESS and XprESS systems but considered there was only weak, indirect evidence to substantiate this assumption
3.22 The committee considered that the evidence from REMODEL demonstrated that balloon dilation (with either XprESS or FinESS) is clinically non‑inferior to FESS in terms of alleviating symptom in patients with uncomplicated chronic sinusitis.
3.23 The committee considered that although the single‑arm observational studies were of lower quality, the results were consistent with the findings of the REMODEL study. It considered that these studies provide evidence that balloon dilation is effective in improving other clinical outcomes including postoperative debridements, ostial patency, use of analgesic medication, time of recovery, and time taken to return to work.
3.24 The committee heard from the company that FinESS and XprESS function in the same way once inflated within the sinus ostia. However, it was informed that the trans‑nasal approach used for XprESS allows more sinuses to be treated than the trans‑antral approach used with FinESS.
3.25 The committee heard from experts that XprESS can be done using local anaesthetic and so allows patients to return to work on the same day. It further heard that balloon dilation reduces postoperative pain, preserves mucosa and bony structures, reduces scarring in the sinuses, and reduces nasal bleeding and the risk of damage to the ethmoidal artery.
3.26 The committee noted that the REMODEL study excluded patients with severe nasal polyposis, and it was advised by experts that balloon dilation is not suitable in these patients.