Efficacy and Mechanism Evaluation (Aug 2023)
Subacromial spacers for adults with symptomatic, irreparable rotator cuff tears: the START:REACTS novel group sequential adaptive RCT
Abstract
Background A balloon spacer is a relatively simple addition to an arthroscopic debridement procedure for irreparable rotator cuff tears. Objective To evaluate the clinical and cost-effectiveness of a subacromial balloon spacer for individuals undergoing arthroscopic debridement for irreparable rotator cuff tears. Design A multicentre participant-and assessor-blinded randomised controlled trial comparing arthroscopic debridement with the InSpace® (Stryker, Kalamazoo, MI, USA) balloon to arthroscopic debridement alone, using a novel adaptive design. Pretrial simulations informed stopping boundaries for two interim analyses, using outcome data from early and late time points. Setting A total of 24 NHS centres. Participants Adults with a symptomatic, irreparable rotator cuff tear for whom conservative management had been unsuccessful. Interventions Arthroscopic debridement of the subacromial space plus insertion of the InSpace balloon compared with arthroscopic debridement alone. Main outcome measures Oxford Shoulder Score at 12 months. Results A predefined stopping boundary was met at the first interim analysis. Recruitment stopped with 117 participants randomised. We obtained primary outcome data on 114 participants (97%). The mean Oxford Shoulder Score at 12 months was 34.3 in the debridement-only group (59 participants of 61 randomised) and 30.3 in the debridement with balloon group (55 participants of 56 randomised; mean difference: −4.2; 95% confidence interval −8.2 to −0.26; p = 0.037). There was no difference in safety events. In the cost-effectiveness analysis, debridement-only dominated with a probability of <1% that the device is cost-effective. Magnetic resonance imaging substudy To evaluate the function of the balloon, we developed a dynamic magnetic resonance imaging protocol to induce humeral movement by activating the deltoid. The pandemic restricted recruitment, so the sample size was small (n = 17). Statistical methodology study We applied the novel adaptive design approach to data from seven previous randomised controlled trials. The method would have been applicable to five of these trials and would have made substantial savings in time to recruitment, without compromising the main findings of the included trials. Interim analysis interpretation study We asked potential data monitoring committee members to review interim analysis reports presented using Bayesian and frequentist frameworks. They did not always follow the stopping rules and would benefit from additional information to support decision-making. Limitations The InSpace balloon could be beneficial in a different population although we are not aware of it being widely used for other purposes. As a result of the pandemic, we were not able to complete data collection for objective measures. Conclusions In this efficient adaptive trial, clinical and cost-effectiveness favoured the control treatment without the InSpace balloon. Therefore, we do not recommend this device for the treatment of irreparable rotator cuff tears. Future work There is an urgent need for high-quality research into interventions for people with irreparable rotator cuff tears as there is a lack of good evidence for all available treatment options at present. Trial registration This trial is registered as ISRCTN17825590. Funding This project (project reference 16/61/18) was funded by the Efficacy and Mechanism Evaluation (EME) Programme, a Medical Research Council and National Institute for Health and Care Research (NIHR) partnership. The trial is co-sponsored by the University of Warwick and University Hospitals Coventry and Warwickshire NHS Trust. This study will be published in full in Efficacy and Mechanism Evaluation; Vol. 10, No 3. See the NIHR Journals Library website for further project information. Plain language summary Tears of the rotator cuff tendons of the shoulder are very common. Some people with a rotator cuff tear have pain and loss of function that is not improved by simple treatments, and they may undergo surgery. Many people have the tear repaired. if this cannot be done, keyhole surgery can be used to clear space around the tendons and allow the shoulder to move better. This procedure is called an arthroscopic debridement. The InSpace® (Stryker, Kalamazoo, MI, USA) balloon is a dissolvable device that is placed by the surgeon above the shoulder joint, after an arthroscopic debridement. We wanted to know if it improved results for patients and was good value. We compared the standard operation, arthroscopic debridement, with the same procedure with the InSpace balloon inserted. We collected data from 117 people with rotator cuff tears from 24 NHS hospitals. Because of the COVID-19 pandemic, we could not directly measure strength or movement and used well-established questionnaires instead. Twelve months after their operation, most people had improved. People who had the standard operation alone, without the balloon, had less pain and could use their shoulder more. We calculated the costs of each treatment, including lost earnings. The InSpace balloon was more expensive and was poor value for money for the NHS. We developed a new method for doing an early statistical analysis to decide whether we could stop the study early. Because of this, around half the number of people were needed in the study, compared with the number we would normally need to do this work. We did additional research and found that this new method would work for some other studies and would give much quicker results about which treatments work best. We developed a new method for doing magnetic resonance imaging scans of the shoulder, although data collection was limited by the COVID-19 pandemic. Scientific summary Background Tears of the rotator cuff tendons of the shoulder are a common cause of shoulder pain and disability. Individuals often present with pain and restricted movement, as well as loss of strength and function affecting even simple activities of daily living, such as hair brushing. Of those presenting with a rotator cuff tear, around half are treated surgically but roughly one-third of tears cannot be repaired. Compared with those who have a repair, people with irreparable tears have more severe pain, worse outcomes from treatment and more limited management options. New surgical techniques have been introduced to improve care, including the InSpace® (Stryker, Kalamazoo, MI, USA) subacromial balloon spacer. The InSpace balloon is a saline-filled biodegradable balloon that is inserted surgically in the space between the humerus and acromion at the end of an arthroscopic debridement for people with an irreparable rotator cuff tear. By acting as a cushion between the acromion and the humerus, and potentially reducing friction, the device aims to improve the mechanics of the affected shoulder and aid rehabilitation. It deflates between 3 and 6 months after insertion, by which time it is hoped that the mechanics of the shoulder have improved. The National Institute for Health and Care Excellence called for a randomised trial, recommending that the device should be used in research only. It received Food and Drug Administration (FDA) clearance in the United States in July 2021. Approximately 29,000 devices had been implanted outside the United States prior to FDA approval. Encouraging clinical results were observed from small early case series but some studies have reported poor results or cases of inflammation and pain. Novel surgical procedures can expose patients to harm and should be carefully evaluated before widespread use, ideally with clinical trials. Surgical trials can provide high levels of evidence but can take a long time to complete. Adaptive designs can reduce the time needed to perform trials, potentially exposing fewer people to risk than with traditional fixed sample designs. Surgical trials typically use outcomes of 12 months or more, but by using the correlation between early and late outcomes, the advantages of adaptive designs could be extended to trials of new surgical techniques. Objective The primary objective was to assess the clinical effectiveness of arthroscopic debridement with the InSpace subacromial spacer balloon compared with arthroscopic debridement alone for people with symptomatic irreparable tears of the rotator cuff, based on the Oxford Shoulder Score (OSS) at 12 months. Our secondary clinical objectives were: to compare the two interventions using both patient-reported and objective outcome scores at 3, 6 and 12 months to perform an economic analysis to assess the comparative cost-effectiveness of the two interventions to perform a magnetic resonance imaging (MRI) substudy to evaluate the proposed mechanism of the balloon, with scans taken at 8 weeks and at least 6 months after treatment. The methodological objectives of the study were: to implement an efficient adaptive clinical trial design with the potential to stop early either for futility or efficacy, using outcome data available at 3, 6 and 12 months to evaluate the novel adaptive trial design using real trial data from a number of previous high-impact orthopaedic trials to explore the challenges of supporting adaptive design decision-making health economic analyses to compare the use of frequentist and Bayesian design and analysis on the conduct and interpretation of an adaptive surgical clinical trial with reference to decision-making by data monitoring committees (DMCs) during the study. Trial methods Design We performed a participant- and assessor-blinded multicentre superiority randomised controlled trial (RCT; IDEAL stage 3) across 24 centres in the UK using a group sequential adaptive design with two preplanned interim analyses. Participants Adults with a rotator cuff tear with intrusive symptoms (pain and loss of function) deemed by the treating clinician to be technically irreparable, for whom conservative management had been unsuccessful. People were ineligible if any of the following conditions were met: advanced shoulder osteoarthritis on preoperative imaging; subscapularis deficiency or pseudoparalysis (these three criteria are contraindications for the InSpace balloon) the clinician had determined that interposition grafting or tendon transfers were indicated an unrelated ipsilateral shoulder disorder neurological or muscular conditions that would interfere with strength measurement or rehabilitation previous proximal humeral fracture previous entry into the trial (i.e. for the other shoulder) those unable to complete trial procedures and those unfit for surgery. All participants gave prior written consent. Eligibility was assessed prior to consent on the morning of surgery and intraoperatively (after assessment of the tear and surrounding structures in the shoulder) immediately prior to randomisation. Intervention The control group (debridement only) underwent arthroscopic debridement of the subacromial space and biceps tenotomy (if not already torn). Surgery was performed by subspecialty trained shoulder surgeons, using their normal surgical technique, within the confines described in a trial-specific surgical manual and surgical video. The intervention group (debridement with InSpace balloon) underwent the same arthroscopic debridement procedure, followed by insertion of the InSpace balloon. The manufacturer’s recommended technique was followed as documented in the surgical manual and was confirmed with them before distributing to surgeons. Surgical training was offered and a training course was run at the start of the trial. A company representative was invited to attend cases for technical support. Participants were offered the same rehabilitation, including a home exercise programme and at least three face-to-face physiotherapy sessions. For both groups, fidelity was assessed with an operative record form, and arthroscopic photographs. The number of physiotherapy visits for each participant in both groups was also documented. Outcome measures The primary outcome was the OSS 12 months after randomisation. The OSS is a 12-item participant-reported measure of shoulder-related pain and function. A higher score (0–48) corresponds with a better outcome. The study was originally designed with the Constant score as the primary outcome. However, during the COVID-19 outbreak in March 2020 (in the recruitment phase of the trial), it was decided to change this to the OSS. This was because the Constant score requires face-to-face contact to measure and is usually assessed in hospital clinics; this would have exposed participants to unnecessary risk during the height of the pandemic. The OSS correlates well with the Constant score; both outcomes assess pain and shoulder function, are similarly responsive and have comparable worthwhile effect sizes in rotator cuff pathology. We collected secondary outcomes at baseline and at 3, 6 and 12 months post-randomisation, including (where possible) the Constant score, range of pain-free movement and strength of shoulder abduction and flexion the shoulder, the Western Ontario Rotator Cuff (WORC) index (scored 0–100), health utility assessed using the EuroQol-5 Dimensions, five-level version (EQ-5D-5L), Participant Global Impression of Change scale, health-care resource use, analgesia use and adverse events. We defined adverse events as any condition of the affected shoulder or any event related to the anaesthetic or rehabilitation. Trial results Of 317 eligible people, 249 (79%) agreed to join the trial. A predefined futility stopping boundary was met at the first interim analysis and recruitment stopped with 117 participants randomised. A total of 61 (52%) participants were randomised to receive debridement surgery alone and 56 (48%) were randomised to receive debridement with the InSpace balloon. Twelve-month primary outcome data were obtained from 114 of the 117 participants (97%). Of the three items of missing primary outcome data, two participants had died (neither death was trial related) and one participant could not be contacted. Scores improved compared with baseline in both groups. In the primary intention-to-treat analysis, the mean OSS at 12 months was 34.3 in the debridement-only group (n = 59) and 30.3 in the debridement with InSpace balloon group [n = 55; mean difference −4.2, favouring control; 95% confidence interval (CI) −8.2 to −0.26; p = 0.037]. Using a prespecified secondary adjusted model to account for the baseline OSS, sex, tear size and age group, a similar treatment effect was observed (effect −4.2, 95% CI −7.8 to −0.6; p = 0.026). There was no difference in safety events. The Constant score, range of flexion and abduction and WORC index results were consistent with the primary analysis (the Constant score and range of motion measures had a high amount of missing data due to COVID-19 restrictions). The differences in WORC index and EQ-5D-5L at 12 months were not statistically significant (WORC index: −8.4, 95% CI −16.8 to −0.1; p = 0.055; EQ-5D-5L: −0.056, 95% CI −0.15 to 0.03; p = 0.24). In both cases the direction of change favoured debridement-only. In cost-effectiveness analyses, quality-adjusted life-years were higher in the debridement-only group and costs were lower compared with the debridement with the InSpace device, in terms of both direct health-care costs and wider societal costs. As a result, debridement-only dominated with a probability of <1% that the InSpace device is cost-effective. Magnetic resonance imaging substudy We developed and refined a technique for dynamic MRI of a shoulder under deltoid load. We undertook a developmental work package in which we applied the technique to participants awaiting rotator cuff repair surgery who underwent electromyography evaluation of deltoid muscle contraction, which allowed us to determine the most appropriate technique to consistently achieve muscle activation in the narrow confines of a MRI scanner. We then piloted the MRI technique using this muscle activation protocol and a fast acquisition sequence for collecting both resting and active images. The main outcome of interest was the acromiohumeral distance, which was used as a marker of humeral migration under load. We applied this technique in a mechanistic substudy, with scans taken at 8 weeks and at least 6 months after randomisation. Recruitment was severely hampered both by the early adaptive stop and the effects of the pandemic, which prevented continuing data collection. Despite the small sample size, we were able to observe narrowing of the acromiohumeral distance under load, demonstrating that the MRI technique was effective. We did not observe any between-group differences, although numbers were very low for this analysis. Adaptive designs for surgical trials We applied the novel adaptive design methodology that was developed for the main study to a number of previous high-impact randomised trials in trauma and orthopaedic surgery. The study assessed whether each of a selected number of RCTs, originally implemented using conventional sample size designs, would have stopped early if a group sequential trial design had been used, and what the final outcome would have been had they done so. We received anonymised data from seven large multicentre trials: Wound Management of Open Lower Limb Fractures (WOLLF); Distal Radius Acute Fracture Fixation Trial (DRAFFT); UK Fixation of Distal Tibia Fractures (UK FixDT); UK Full Randomised Controlled trial of Arthroscopic Surgery for Hip Impingement versus best Conventional (FASHIoN); the Warwick Arthroplasty Trial (WAT); Can Shoulder Arthroscopy Work? (CSAW); the Total or Partial Knee Arthroplasty Trial (TOPKAT). The temporal sequence of data accumulation was replicated in exactly the manner it was in the original study, using the dates when each outcome measure was taken. We selected planned interim analyses and stopping boundaries and simulated how each study might have progressed using the methodological approach described in this monograph. The results for five of the studies (WOLLF, FixDT, DRAFFT, FASHIoN and WAT) showed that adaptive design using early outcome data would have been feasible and likely to provide designs that were at least as efficient, and possibly more efficient, than the original fixed sample size designs. For WOLLF and FixDT the simulations showed that it was highly likely that these studies would have (correctly) stopped early for futility, both over one year early, saving potential considerable effort and resources. The two studies that showed modest effect estimates at interim analyses in favour of the test treatment (WAT and DRAFFT) did not stop early, which was consistent with the final results of these studies. The FASHIoN trial showed good evidence in favour of the test surgical intervention in the final analysis but fell short of stopping at the interim analyses. For this study, it would have been possible to select different but sensible boundaries that would have resulted in early stopping for efficacy. TOPKAT and CSAW would not have been suitable for the adaptive design methods in their current form, as they had either longer primary outcomes times or less early time point data, although it is reasonable to think that these issues could have been resolved if the method was applied prospectively. For all the studies, it was clear that the feasibility and practicality of using the proposed adaptive design methods was determined in large part by: (1) whether the timing of recruitment allows for interim analyses; (2) the availability of early outcome data and correlations with final outcomes; (3) recruitment and outcome data accrual profiles; and (4) the estimates of correlation and covariance parameters at the design planning stage. Interim analysis report study To understand the influence of different approaches to adaptive trial design on the conduct of DMC decisions, we performed an exploratory study of the interpretation of Bayesian and frequentist interim analysis reports by potential committee members. We found that potential DMC members do not always choose to follow the stopping rules that are presented and would benefit from more ancillary information to support their decision-making and understanding of the analysis, regardless of the statistical framework used. Conclusions We used a blinded RCT with predefined stopping boundaries to test whether the InSpace balloon was of benefit for people with irreparable rotator cuff tears. The study stopped at just over half the maximum sample size, allowing us to report the findings early. In the primary analysis, arthroscopic debridement was found to be superior to arthroscopic debridement with the InSpace balloon for people with an irreparable cuff tear of the shoulder, based on the OSS 12 months after surgery. Secondary outcomes and cost-effectiveness analysis agreed, and effectively exclude the possibility of any meaningful benefit for the InSpace balloon. This trial has delivered evidence that the InSpace balloon is not an effective treatment, could be harmful, and is very unlikely to be cost-effective. Randomised trials are needed early in the evaluation of new technologies to prevent harm to patients and cost to society, but also to allow effective treatments to be offered widely. We have demonstrated that using adaptive designs in surgical trials are possible and practical. By delivering efficient, effective trial designs early in the introduction of new procedures and technologies, we will make major cost savings for the health service and deliver better patient outcomes both now and into the future. Trial registration This trial is registered as ISRCTN17825590. Funding This project (project reference 16/61/18) was funded by the Efficacy and Mechanism Evaluation (EME) programme, a Medical Research Council and National Institute for Health and Care Research (NIHR) partnership. The trial is co-sponsored by the University of Warwick and University Hospitals Coventry and Warwickshire NHS Trust. This study will be published in full in Efficacy and Mechanism Evaluation; Vol. 10, No. 3. See the NIHR Journals Library website for further project information.
Keywords
