Inspace Balloon: Real Solution or Hot Air?
In March 2024 The Journal of Orthopedic Experience and Innovation (JOEI) ran an article on Dr. Abboud’s experience with the Inspace bioabsorbable balloon spacer (Stryker, Kalamazoo, MI). The Inspace balloon was approved by the FDA in 2021 as a treatment for irrepairable rotator cuff tears (Abboud 2024). Surgeons implant the balloon in the subacromial space in a relatively quick arthroscopic procedure and it dissolves in about 6 months’ time.
I like and respect Dr. Abboud tremendously. We both have strong Philadelphia connections. And we are both share passions for the American Shoulder and Elbow Surgeons Foundation (ASESF). As far as the Inspace balloon goes however, we have different perspectives. In this article you will hear the story of why I do not recommend the Inspace balloon spacer for my patients.
_dr._dipaola_(right).jpeg)
My Experience
Like Dr. Abboud, I participated in the FDA IDE for the Inspace balloon. While I was not part of the initial study design team, I did contribute a significant number of patients and followed the study through to completion. In deference to Dr. Abboud’s story of how he came to lead the Inspace trial I will digress a bit to highlight my circuitous route to becoming a site coordinator for the Inspace balloon trial (Abboud 2024).
In September of 2016 we hosted Dr. Michael McKee at the University at Buffalo Orthopedic Department for grand rounds. The night before Dr. McKee’s lectures, we were out at dinner and I asked Dr. McKee a question about a challenging patient that I had seen with an irrepairable rotator cuff.
“What would you do?” I asked
“Why don’t you try the balloon?” He said.
Balloon? I remember reading about the balloon a year or so prior but had not seen it in practice here in the United States. After looking it up again I found out that it was not available in the U.S. but, there was an FDA trial ongoing.
Interestingly, my mentor Laith Jazrawi MD at The Hospital for Joint Diseases was part of the FDA IDE trial. I thought of sending my patient down to him and called him. He said: “why don’t you guys get involved with the study?” As it turns out, they were still looking for trial sites. So, we did the requisite paperwork and showed the study group that we had the infrastructure to participate, and we were off to the races.
Our site enrolled six patients who had surgery. Half received the balloon and half did not. Both groups of patients did well overall and improved. Although I noticed that at least one of my patients who got the balloon felt that the benefit mostly wore off by about a year after surgery. One of the others who did not get the balloon told me that surgery was the best thing he had ever done and went back to work as a tree cutter.
Since the trial ended, I have only put in 2 balloons. Both patients sought me out for the treatment. Both ultimately did well with typical improvement as Dr. Abboud describes in his article but one had an issue that I will mention in more depth later.
I currently do not actively recommend the balloon for my patients and only consider placing them if patients seek me out specifically for it. The real reason I do not use the balloon comes down to two things: 1) the data and 2) the lack of value.
Disclosure: my institution did receive research support to help perform the FDA IDE trial for the Inspace balloon. But I do not currently receive any monetary compensation from Stryker the company that now distributes the Inspace balloon.
What Does the Data Show?
Since my experience alone is small enough to be considered anecdotal, it is important to consider it in light of the available literature. It is rare in orthopedics to have level I data to guide our decisions. So, in some ways we are enormously blessed to have 2 randomized controlled trials (RCT) from which to draw conclusions about the Inspace Balloon. We will examine in more depth the best literature available for the balloon device.
At the same time that the FDA IDE study was published in JBJS, another independent RCT on the Inspace balloon was published in The Lancet (Verma, Srikumaran, Roden, et al. 2022; Metcalfe, Parsons, Parsons, et al. 2022). The Lancet study was performed in England and was called START: REACTS (Metcalfe, Parsons, Parsons, et al. 2022). In short, neither the FDA IDE trial nor the START:REACTS trial showed any clinically significant differences in patient reported outcomes for patients that received the Inspace balloon compared to patients who had debridement (START: REACTS) or debridement with partial rotator cuff repair after 1 year. In fact, the START:REACTS trial suggested that the balloon may impart a statistically significant net negative effect, although it did not reach the minimal clinically important difference (MCID).
The START:REACTS trial has gotten a lot less play in the United States than it has in Great Britain and the rest of Europe. Below we will compare the two RCTs and review some of the critiques of each study starting with study design and rounding out the review with an analysis of the results. In the end, I will explain why I do not feel that there is a convincing evidence to use balloon spacers for patients with irrepairable rotator cuff tears.
Study Design
There were some inherent differences in study design between the FDA IDE study and START: REACTS trial. We will discuss them below and how they may impact your interpretation of the data.
The Subscapularis
The authors of the FDA IDE study clash with the authors of the START:REACTS trial about the subscapularis. The inclusion criteria for the FDA trial defined an eligible patient as one who has a large irrepairable posterior/ superior rotator cuff tear (5 cm), who had failed non operative treatment and had preserved passive range of motion of the shoulder (see page 1251 of the referenced article for full criteria) (Abboud 2024). Some believe that the balloon device is only appropriate for patients who have an intact subscapularis which is why the FDA IDE was designed this way.
In contrast the designers of the START: REACTS trial allowed patients with partial, upper border subscapularis tears to participate (Metcalfe, Parsons, Parsons, et al. 2022). Authors of the FDA IDE study criticize the architects of the START: REACTS trial for including patients with any level of subscapularis tear stating that these patients are “contraindicated” for the balloon device (Verma, et al. 2022). Dr. Metcalfe, lead author of the START: REACTS trial disagrees.
Dr. Metcalfe has stated, it is not known whether the Inspace balloon device is only appropriate for patients with an intact subscapularis since this population had not yet been studied rigorously at the time of the study design. (Metcalfe 2022) Additionally, when you review the START:REACTS trial, you note that not only are the subscapularis tears solely upper border partial tears (< 1 cm), the tears are also fairly evenly distributed through the two groups of patients: 12/61 (20%) in the debridement group and 14/56 (25%) in the debridement plus balloon group (START REACTS) (Metcalfe, Parsons, Parsons, et al. 2022). Thus, one would expect that any effect produced by the presence of a subscapularis tear should likely wash out between groups.
To summarize, not only do we not truly know whether the device can work in patients with subscapularis tears, the effect of partial subscapularis tears in the START:REACTS study design might be negligible, because they were distributed evenly through both the control group and the intervention group.
A Skewed Patient Population?
Did the START:REACTS group study a skewed patient population, a patient population over weighted with patients that had worse active shoulder range of motion (ROM) compared to the patients in the FDA IDE trial? After some exposition below you will see that the answer to that question is a resounding “no.” In fact, the overweighting of patients with “far worse” active forward elevation in the START: REACTS trial turned out to be a misunderstanding of the reported data by the critics.
The two RCTs had different inclusion criteria as it pertained to ROM of the affected shoulder. The FDA IDE trial required eligible patients to have preserved passive range of motion of the affected shoulder (Verma, Srikumaran, Roden, et al. 2022). Only the START:REACTS trial used active range of motion of the shoulder in the inclusion/exclusion criteria for their study, excluding patients with pseudoparalysis defined as 20 degrees or less of active forward elevation (note that there is no true consensus definition for pseudoparalysis in the literature) (Metcalfe, Parsons, Parsons, et al. 2022). The FDA IDE study design did not officially exclude patients with “pseudo paralysis” although in my experience the study coordinators verbally discouraged us from enrolling them.
Both studies did record AROM of the shoulder as part of the initial assessment with the FDA IDE trial recording the AROM throughout the study (Verma, Srikumaran, Roden, et al. 2022; Metcalfe, Parsons, Parsons, et al. 2022). The START:REACTS trial later shifted away from measuring AROM as they ran into the COVID restrictions and could not see patients in person for a large portion of study enrollment (Metcalfe, Parsons, Parsons, et al. 2022).
When the authors of the START: REACTS trial reported their results they received criticism from the lead authors of the FDA IDE trial authors in a November 5, 2022, letter to the editor in Lancet. In the letter the FDA study’s authors stated:
“Patients [in the START REACTS trial] had preoperative elevation far worse than expected in tears indicated for this device, which is concerning for involvement of the subscapularis, a contraindication to the device.” (Verma, et al. 2022)
The critique levied by the FDA trial authors in November of 2022 echoed what I was hearing in October 2022 at the ASES annual closed meeting. At that meeting some of the lead authors of the FDA IDE Inspace trial gave a discussion on the Inspace balloon study and results. They argued then, that the authors of the START:REACTS trial had included far too many patients with poor baseline active forward elevation and that this skewed the START:REACTS population to have a starting average forward elevation of 71 degrees.
Concerned that the authors of the FDA IDE trial might have a valid point, I poured over both papers again including inclusion and exclusion criteria. If you look at the JBJS FDA IDE article you do not see baseline AROM of motion reported (Verma, Srikumaran, Roden, et al. 2022). I requested this information from Stryker, but did not receive a response. I then contacted the lead author of the START:REACTS trial, Andy Metcalfe. What he said to me over email in October 2022 echoed his response in the Lancet from November 2022 (shared with permission from Dr. Metcalfe):
“The ROM point is interesting, we used pain-free active range of motion, not full active range of motion, which meant that the data we presented looked like the ROM was lower than it would be, had we collected the full active ROM…We collected pain-free active range of motion using the Constant score method – that is, we asked people to stop the measurement when the pain started (or, if they had pain at rest, when the pain worsened). I’ve attached the Constant score manual that we gave to sites – pages 3, 4 and 5 are the relevant parts. The result of that is that the reported ranges are much lower than the full active ranges.”
Returning back to the original papers on the reporting of the Constant score reveals as Dr. Metcalfe pointed out, the Constant score performed correctly, records “pain free” range of motion. Pain free range of motion is typically much lower than plain active “range of motion.” (Ban et al. 2013)
Most patients that have pain from rotator cuff tears begin to have pain in the region that the START:REACTS trial did: at or just below shoulder height. So, it’s likely not that the START:REACTS trial patients were materially different in their starting ROM, but that the recording method of the authors between studies was different.
Dr. Metcalfe, replied to the authors of the FDA trial’s criticisms that they had studied a skewed patient population in a follow up letter in the Lancet explaining:
“They misinterpret our study; further details to address these queries are available in the supplementary data and published protocol (Abboud 2024). We collected data on pain-free range of movement with the method set out by the Constant Score; the full range of movement would inevitably be greater than the pain-free range (Verma, Srikumaran, Roden, et al. 2022; Metcalfe, Parsons, Parsons, et al. 2022; Ban et al. 2013). We used a published, widely used, and reproducible method of assessment that does not induce pain. The papers quoted by Verma and colleagues are about a different measurement.” (Metcalfe 2022)
In conclusion, the FDA IDE authors likely misinterpreted the low AROM numbers garnered for the Constant score in the START: REACTS trial as absolute AROM not pain free ROM. This misinterpretation weakens the critique and should draw us to take deeper look at the START:REACTS conclusions.
Did the FDA IDE Trial have a Critical Design Flaw?
Another concern with the FDA IDE trial and why you should consider the START:REACTS trial when making clinical decisions about the Inspace balloon, is that the FDA IDE trial suffered from a design flaw that START:REACT did not: The FDA IDE trial did not have a pure control group.
In a well-designed RCT if you are going to draw a causal inference about an intervention you want to isolate just one variable. The FDA IDE trial changed 2 variables: the presence of the ballon and presence of partial rotator cuff repair. The Inspace FDA IDE study compared: 1) a group of patients that received a shoulder arthroscopy with debridement and the balloon and 2) a “control” group that received shoulder arthroscopy, debridement, partial rotator cuff repair and no balloon. Stated differently if the study was done with a pure control, the only change between control group and intervention group would have been the addition of the balloon.
The FDA required the authors of the IDE design team to include partial rotator cuff repair in the control group as part of the study design. Surgeons had to place at least one anchor to partially repair a large superior cuff tear. This critique is not to cast blame on anyone but simply to point out that the addition of the partial rotator cuff repair in the control group adds a layer of uncertainty to the conclusions that one can draw about the role the balloon is playing in the outcome of these procedures.
For example, we notice that in the FDA trial the patients that received the balloon had somewhat higher forward elevation after 2 years of follow up. But we cannot definitively say that the increased forward elevation was caused by the balloon alone. Since the control group had a partial rotator cuff repair, we must leave open the possibility that the partial repair may have impaired patients’ motion in the control group. Perhaps the partial repair overtightened the control group shoulder instead of the balloon causing the relative improvement in motion.
This is where the START: REACTS trial gets it right. The authors of the START: REACTS trial designed their trial in such a way that they only changed one variable between the 2 study groups. The only difference between their test group and their control group was the balloon: the test group received the balloon, and the control group did not. Statistically speaking this is more rigorous design and can lead one to draw better causal inferences.
Results of the START: REACTS Trial Raise Concerns
Based on their evidence the authors of the START: REACTS trial recommended against the balloon as they found no benefit in the intervention group and in some subgroups found net harm compared to the control group. They noticed worse clinical outcome results especially in women, with some showing an increase in pain in the test group that received the balloon. The START: REACTS trial authors did not have a definitive explanation for this potential increase in pain in the female population and recommended further study.
To be fair the FDA IDE authors rebutted this claim in their letter to the editor saying, “statement of device harm is not supported by data: the difference in Oxford Shoulder Score did not exceed the minimal clinically important difference, indicating that the difference is not clinically important, there was similar satisfaction between groups, and there were no device-related complications.”
While anecdotal, I did observe an increased pain response in one of the female patients in whom I placed the balloon after the FDA trial. The patient was a woman in her 60’s who developed what seemed to be an inflammatory reaction with increased subacromial pain about 2 months after the balloon was placed. Eventually the pain did subside, and she improved compared to baseline. However, it took a corticosteroid injection and a disconcerting few months to resolve.
What Else Can We Say About These Two Studies?
In my opinion it does appear that he authors of the START: REACTS trial did a thorough amount of preparation for the study design. The supplemental pre study design paper was 90 pages in length and outlined the design rationale and parameters in detail. The START:REACTS authors took some flak for changing their primary outcome measure from the Constant score to the OSS during enrollment. One could argue that changing the primary outcome metric was unusual. However, the study was powered in such a way that this did not require any adjustments and the authors were adapting to the medical climate influenced by the worldwide pandemic. Since OSS and ASES are both well accepted and validated outcome measures that do not require in person measurements, it seems reasonable to assume that both primary outcome measures have the potential to provide sound feedback.
The FDA Inspace trial also was a rigorously performed trial that was executed professionally and with great attention to detail in all phases. The published data indicate that patients see statistically equivalent functional improvement with either the balloon or the debridement/ partial rotator cuff repair intervention. Although patients who received the balloon did get better more quickly. However as mentioned earlier, patients who received the implant did not undergo as long of a surgery and had no surgery done on the bone. Did it take longer for these patients to recover because they had more extensive surgery? Or was this a result of the balloon? We cannot know for certain due to the study design.
Value
The Inspace balloon is now distributed by Stryker and is marketed as a potential expeditious and minimally invasive solution for some patients with large irreparable rotator cuff tears. In the U.S. market the balloon costs in the range of $5000-7000 per implant. As a comparison, many reverse total shoulder replacement implants cost close to the same.
If we are to critically assess the value of the Inspace balloon intervention, we must compare it to the alternative of arthroscopic debridement alone which is much less costly without the implant. If we define value as clinical outcomes divided by cost, one could argue that the Inspace balloon does not provide good value since neither RCT showed materially better long-term functional outcomes compared to debridement (or debridement with partial cuff repair). And we cannot know for certain whether any advantage in early improvement with the balloon as seen in the FDA IDE trial was the result of the balloon or just a smaller surgery since we did not see this quicker early recovery in the START:REACTS trial.
Most medical care in the United States is covered under some form of insurance where the costs are often borne by the insurer. Although with high deductible insurance plans we now often see a larger chunk of that cost borne by the patient. One wonders with insurers less willing to pay for even an acromioplasty in the setting of a rotator cuff repair these days, how they can justify such costs for an implant with equivalent clinical benefit to a debridement alone. Patients, surgeons, insurers and government entities who ultimately approve and pay for innovations should critically examine and be aware of the value proposition of implants like the Inspace balloon when making clinical decisions about a costly device such as this.
We know that many patients with large/massive rotator cuff tears have excellent function and minimal, if any pain. What separates those patients from the ones with pain? We don’t fully know the answer to this question? Our healthcare landscape is built in a way that favors the study of new and costly technology. Perhaps more resources need to be allocated to underfunded treatment options such as physical therapy.
Conclusions
The Inspace balloon is a safe option in patients with irreparable rotator cuff tears of the superior cuff with or without a tear of the upper border of the subscapularis. It does not appear to be significantly better in the long term than debridement alone or with partial cuff repair. In my opinion and experience it is not a significantly better value than arthroscopy with debridement alone.