August 31, 2004





Gaithersburg Hilton

Gaithersburg, MD

Meeting of the Orthopaedic and Rehabilitation Devices Advisory Panel




August 31, 2004






Michael J. Yaszemski, M.D., Ph.D.


Executive Secretary

Janet L. Scudiero, M.S.


Voting Members

Maureen A. Finnegan, M.D.

John S. Kirkpatrick, M.D.

Stephen Li, Ph.D.

Sanjiv Naidu, M.D., Ph.D.


Deputized Voting Members

Fernando G. Diaz, M.D., Ph.D.

Jonas Ellenberg, Ph.D.

Choll W. Kim, M.D., Ph.D.

Sally A. Rudicel, M.D.


Consumer Representative

LeeLee Doyle, Ph.D.


Industry Representative

Sally L. Maher, Esq.


FDA Representatives

Celia M. Witten, M.D., Ph.D.

Director, Division of General, Restorative, and Neurological Devices


John P. Holden, Ph.D., Biomedical Engineer, Lead Reviewer

Barbara D. Buch, M.D., Medical Officer

Richard M. Kotz, M.S., Mathematical Statistician





Executive Secretary Janet L. Scudiero, M.S., called the meeting to order at 8:02 a.m. She stated that panel consultants Fernando G. Diaz, M.D., Ph.D., Jonas Ellenberg, Ph.D., Choll W. Kim, M.D., Ph.D., and Sally A. Rudicel, M.D., had been appointed to temporary voting status for the duration of the meeting. Ms. Scudiero then read the conflict of interest statement. A full waiver had been granted to Stephen Li, Ph.D., for his interests in firms that could be affected by the Panel’s recommendations. The Agency took into consideration other matters concerning Drs. Finnegan, Kim, Kirkpatrick, and Li, all of whom reported current or past interests in firms at issue but in matters not related to the day’s agenda. They could participate fully in the panel’s deliberations. Finally, Ms. Scudiero noted that the next panel meeting is tentatively scheduled for December 2 and 3, 2004.

Celia M. Witten, M.D., Ph.D., Director, Division of General, Restorative, and Neurological Devices, thanked Drs. Finnegan and Li and Ms. Maher for their participation on the panel and welcomed Drs. Kim and Rudicel as new voting members.

            Panel Chair Michael J. Yaszemski, M.D., Ph.D., stated that the purpose of the meeting was to make recommendations concerning a PMA for the St. Francis Medical Technologies, Inc. intraspinous process distraction system, the X-Stop®. He asked the panel members to introduce themselves, after which he noted that the members present constituted a quorum.



Ms. Scudiero read the Agency’s statement on transparency of the device approval process.

William Christianson, President, Orthopedic Surgical Manufacturers Association and Vice President, Clinical and Regulatory Affairs, DePuy Spine, urged the panel to adhere to the regulations and focus on safety and effectiveness based on the data provided. He reviewed the definitions of reasonable assurance of safety and effectiveness and valid scientific evidence and noted that requirements for additional data collection can be burdensome to manufacturers.

            Merrie F. Miller, Ellicott City, MD, a patient of Dr. Charles Hartjen, received the X-Stop® in March 2001. Her severe back pain related to spinal stenosis began in 2000 and ultimately curtailed her normal daily activities, including long walks, gardening, and tennis. Now that she has received the X-Stop®, she is free of pain and has normal movement. The implantation surgery lasted only an hour and was minimally invasive, and the associated pain was short lived. The X-Stop® has given her a new lease on life.

            Allyson Washburn, San Francisco, CA, a former control patient of Dr. Zuckerman, began experiencing back pain and weakness in her left leg when she was 50 years old. Other treatments, such as injections, had side effects and were not reliable. By the fall of 2000, she was often couchbound. She received the X-Stop® in July 2003. Within 2 weeks, she was able to entertain house guests and do much walking; she even traveled to Europe shortly after the surgery. She still has pain when she lies on her right side, but it is tolerable.



             Yvonne Lysakowski, R.N., M.S., Vice President, Regulatory and Clinical Affairs, St. Francis Medical Technologies, Alameda, CA, introduced the sponsor presentation and speakers. She reviewed the product development history and the proposed indication for use and noted that FDA had granted the application expedited review status.

Augustus A. White III, M.D., Ph.D., Orthopedic Surgeon-in-Chief Emeritus, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, MA described the pathoanatomy of lumbar spinal stenosis (LSS) and presented clinical background on the condition. The condition often presents clinically as intermittent neurogenic claudication and is caused by a narrowed spinal canal. Disc bulge or spur can reduce the space available, as can changes in facet joints and ligaments. The degenerative process is slow and results from the aging process. Pain is associated with ambulation. Symptoms remain unchanged or even improve in most patients, although patients with severe symptoms tend to deteriorate. The natural course of the condition can be relatively benign, and diagnosis does not necessarily mean that surgery will be required or that full deterioration will occur. About 700,000 people have clinical signs of stenosis, and about 60,000 surgical decompressions are performed annually. Nonsurgical treatment options include analgesics, nonsteroidal anti-inflammatory drugs (NSAIDs), exercise, weight reduction, bracing, and steroid injections. The X-Stop® device is designed to keep the spine out of full extension and thus limit impingement. It is a straightforward mechanical solution to the problem.

            Scott Yerby, Ph.D., Director, Research and Development, St. Francis Medical Technologies, Alameda, CA, described the X-Stop® design and mechanics. The device is a titanium alloy implant designed to limit extension. It is inserted laterally. The oval spacer matches the interspinous space, and the “wings” prevent lateral and anterior migration. The insertion process preserves the supraspinous ligament intact, and because the device is placed posterior to neural structures, risk of neural injury is minimized. The device is not fixed to bony structures, so revision surgery is straightforward.

Dr. Yerby reviewed the biomechanical testing, which included tests to quantify X-Stop® insertion loads, loads between the implant and spinous processes, spinal canal and neural foramen dimensions, intervertebral kinematics, implant stability, disc pressure, and facet load. In cadaver specimens, changes in canal area and diameter following implant insertion were statistically significant, as were the changes in foramen area and width. Changes at adjacent levels are minimal. Research with cadaver specimens also found that the X-Stop® significantly decreased the range of motion (ROM) in flexion and extension, but not axial rotation ROM or lateral bending ROM. The device prevents the symptomatic position and increases the dimensions of the structures that cause neural compression. The insertion process involves minimal tissue disruption.

            Gunnar Andersson, M.D., Ph.D., Senior Vice President, Medical Affairs, Rush University Medical Center, Chicago, IL, described the rationale for the pivotal trial. He reviewed nonoperative therapy and decompressive surgery outcomes, complications, and risks and benefits. Research indicates that nonoperative therapy is successful for about one-third of patients. Decompressive surgery is successful for 57 to 69 percent of patients (depending on the study), but many patients (about 20 percent) experience serious complications, such as hematoma, infection, and neural injury. The risks are not insignificant, but outcomes are good. In the current standard of care, treatment options for lumbar spinal stenosis can be stratified by disease severity. Nonoperative therapy is the standard of care for mild to moderate LSS symptoms, and decompressive laminectomy is indicated for severe symptoms. The current treatment algorithm does not provide clear options for mild to moderate LSS patients who do not respond to nonoperative treatment.

The pivotal study was a prospective, randomized, multicenter, controlled clinical trial that compared continued nonoperative care (epidural injections) with X-Stop® implantation. The primary outcome measure was the Zurich Claudication Questionnaire (ZCQ), a validated, patient-reported questionnaire designed to measure LSS treatment outcomes. The ZCQ measures three domains (symptom severity, physical function, and patient satisfaction) and is reproducible, internally consistent, and highly responsive. Secondary effectiveness measures consisted of the General Health Index of the SF-36, subjective patient assessment of back and leg pain, and independent radiographic review. The target population consisted of LSS patients with mild to moderate symptoms. Success criteria included clinically significant improvement in physical function and symptom severity, patient satisfaction, and no additional surgery. Additional criteria for X-Stop® patients included maintenance of distraction at 24-month follow-up, no dislodgment of implant, and no device-related complications. Finally, Dr. Andersson reviewed the inclusion and exclusion criteria.

            Charles Hartjen, M.D., Medical Director, The Spine Center at GBMC, Baltimore, MD, presented the pivotal trial results. The randomization process used a block size of 2. Patient assignment was not revealed to him or any other study coordinators. A total of 100 patients received the X-Stop®; the 91 control patients received epidural injection, NSAIDs, analgesics, and physical therapy, as appropriate. Date of study entry was either the date of surgery or the date of injection. No deaths due to the study or related problems occurred.

The two groups were comparable for demographics and comorbidities. Few adverse events related to the device or procedure occurred; significantly more control patients than X-Stop® patients experienced pain, however, and significantly more X-Stop® patients than control patients experienced musculoskeletal adverse events unrelated to LSS treatment. The adverse events are expected in an elderly patient population, however, and the adverse event rate was surprisingly low in the control group. More control patients terminated the study early after having laminectomy. X-Stop® patients had a greater number of comorbidities at study entry. The investigators suggested that “return to activity” may have aggravated preexisting conditions and increased symptoms; neurogenic pain may have created a “masking effect” of preexisting conditions.

Clinically significant improvement was based on ZCQ scores and defined as reduction of symptom severity of 0.5 points or more on a scale of 1 to 5 and improvement in physical function of 0.5 points or more from baseline on a scale of 1 to 4. The differences in symptom severity between the X-Stop® and control group were statistically significant at all follow-up intervals (6 weeks, 6 months, 1 year, and 2 years). Likewise, X-Stop® patients had significantly greater improvement in physical function than control patients at all follow-up intervals. Patient satisfaction and the composite ZCQ success rate were significantly greater for X-Stop® patients.

The radiographic findings indicated no significant difference between control and X-Stop® patients at 12 and 24 months for disc height, curvature and angulation of the spine, and degree of spondylolisthesis. The X-Stop® maintained distraction in 96 percent of all levels measured.

Dr. Hartjen presented additional information on physical function and symptom severity ZCQ score changes along with subgroup analyses for 1- versus 2-level X-Stop® implantation and symptom duration. He noted that St. Mary’s had a success rate that was much higher than that of the other eight centers; it was noted St. Mary’s is also the site of the device’s inventer.

Variables correlated with a positive outcome included worse baseline scores in ZCQ physical function and SF-36 bodily pain and social functioning, absence of comorbid conditions, younger age, lower blood loss during surgery, positive femoral stretch test, and greater ROM in extension and rotation. SF-36 outcomes for the two groups were comparable at 2-year follow-up, but the proportion of participants who experienced improvement in frequency and severity of leg pain and back pain was greater in the X-Stop® group than in the control group at 24 months.

The X-Stop® is a safe product for many reasons. Implantation is usually a same-day procedure involving local anesthesia and negligible blood loss. Recovery is rapid, and the procedure carries a minimal risk of systemic and local complications. Implantation does not involve neural element manipulation. Participants experienced no neural injury, and musculoskeletal adverse events were related to deconditioning, preexisting degenerative joint disease, and age. The device does not compromise further surgical options and is suitable for patients who cannot tolerate major surgery or general anesthesia.

The X-Stop® produces immediate effectiveness, and its superiority over nonoperative therapy is consistent and sustained over time. The magnitude of improvement in X-Stop® patients in the pivotal trial exceeded the threshold definition of clinical significance. The ZCQ and SF-36 outcomes consistently demonstrated significant improvement, and patient back and leg pain significantly improved.

Dr. Andersson presented information on interpreting the pivotal study outcomes. He noted that the published literature focuses on outcomes of nonoperative therapy and laminectomy in patients with mild to moderate symptoms. Outcomes in patients who had laminectomy can be applied to matched patient populations. The criteria in the literature differ from those used in the pivotal trial; however, the pivotal trial criteria can be applied to studies using the ZCQ. The findings in the pivotal trial are consistent with outcomes from other studies. The X-Stop® provides an immediate, quantifiable benefit to patients suffering from stenosis and has a favorable risk–benefit profile. The data from the pivotal trial show reasonable assurance of safety and efficacy.



John P. Holden, Ph.D., Biomedical Engineer and lead reviewer, briefly described the X-Stop® and its development history and summarized the preclinical testing. He did not repeat the sponsor’s data, noting that the Agency summarized its testing in its review memorandum for the panel. The device material conforms to the ASTM F136 standard. Sizes are based on the minor diameter of the oval spacer component and range from 6 to 14 mm. The welded (second-generation) design was used in the pivotal study; the earlier version had problems that were resolved by design changes that included welding some components. Additional testing was performed to validate the new design.

Preclinical testing included mechanical and biomechanical tests; the presentation focused on expulsion testing (effects of placement on dislodgement), spinal canal and foramen dimensions, and spinal kinematics. Proper anterior placement of the X-Stop® is essential to preventing dislodgement of the device and deformation of the spinous processes. As a result of preclinical testing, the sponsor modified the surgical technique to emphasize correct placement within the concave space of the spinous processes. Surgeons should remove deformed facets.

The X-Stop® limits canal narrowing at the implanted level in extension. The device also tends to reduce dimensions in the flexed position. However, only in vitro measurements were taken, and they were not confirmed by in vivo measurements. The panel should consider the device’s effects on adjacent segments and spinal biomechanics, as reflected in the clinical data. No preclinical data were collected on the effects of two-level implantation. In addition, the clinical patients’ radiographs were not taken in flexed or extended positions.

Barbara D. Buch, M.D., Medical Officer, presented the Agency’s clinical review. She noted that the sponsor had already summarized the details of the pivotal trial. After reviewing the pilot studies and the first pivotal trial, Dr. Buch provided a brief overview of the second pivotal trial. The entry criteria included 2 objective measures of LSS, and success was defined as an 0.5 point improvement on the patient self assessment scale, the ZCQ. Maximum benefit in the questionnaire’s validation study was reached at 6 months. Dr. Buch noted that the ZCQ was validated at 6 months, but the sponsor used it in the X-Stop® study at longer follow-up time points. One of the Agency’s issues with the clinical study was that the follow-up provided no data beyond 2 years. She noted that it is important for orthopaedic device trials to look at pain relief and functional outcomes trends beyond 2 years.

The second issue was that the control subjects had already failed conservative treatment for periods ranging from 6 months to more than 2 years. They received additional conservative treatment, including various numbers of epidural injections. The control group experienced 95 percent failure of conservative treatment at the end of the study, and 26 percent had a laminectomy. However, one of the confounding factors in the trial was that the sponsor did not clearly pre-define the criteria for progression to laminectomy or additional epidural injections in the study protocol. Therefore, it appeared that decisions for secondary intervention were not applied across sites equally. Also, X-Stop® patients received epidural injections for pain. It is not clear whether the additional epidural delayed progression to laminectomy.

The safety data were unremarkable, except that musculoskeletal and accidental events occurred with greater frequency in X-Stop® than in control patients. The causes of this were not determined. The effectiveness data raise several questions. Were the enrollment criteria and patient demographics able to discern comparable patients? Did the entire study population require some sort of operative intervention at time of entry? The difference between achieved results (43 percent for X-Stop® patients and 4.4 percent for controls) and treatment goals (60 percent for X-Stop® and 37.5 percent for controls) raises questions. Although X-Stop® patients showed improvements in leg and back pain while standing and walking, no difference was found in mean back or leg pain scores for the two groups. Another important issue in the study was the fact that at the conclusion of the study at 24 months, patients who had improved pain and function scores at 6 and 12 months showed regression of pain and function toward baseline. This was shown by a chart comparing the symptom severity, functional and satisfaction scores of the ZCQ and the overall score at the different time points in the trial.  The chart shows that between 12 and 24 months the values decreased in each case toward the baseline values.

 Other issues discussed included that the preclinical cadaveric studies did not assess the effects of two-level X-Stop® implants on the biomechanics, the foramen volume, or the effect of flexed spine preinsertion. It was noted that patients treated at two levels had slightly better effectiveness outcomes and had fewer laminectomies and adverse events than those treated at one level. Not all patients with single-level implants had a single-level laminectomy; some had two level laminectomy.

In this study, no flexion or extension radiographs were taken. No significant radiographic differences between successful patients were found at 24 months. Of 113 levels treated with X-Stop®, 5 showed foramen decrease of >4 mm, and 50 showed no change. No summary radiographic evidence of limiting extension was provided.  Therefore, the question was posed to the panel—what is the best way to interpret the radiographic measurements as evidence of effectiveness? It is not clear that the measurements were able to predict success or failure when reviewing the plain radiographs provided. The fact that the rates of successful outcomes differed from site to site raises the issues of a possible learning curve or each center’s differential ability to select patients. The panel should consider these issues, including the adequacy of the control, the appropriateness of evaluations for pain and function, the utility of the radiographic evaluations for determining safety and effectiveness, and the device’s long-term biomechanical effects on the spine. The panel should also consider whether it is possible to know, from the data provided, in whom the device should be used.

Richard Kotz, M.S., Mathematical Statistician, presented the Agency’s statistical review, raising several issues about the pivotal trial design. Blinding was a problem because only the evaluating physician, not the patients, treating physician, or radiologist, was blinded. In addition, the randomization was stratified by site with a fixed block size of two, and patients were randomized upon determination of eligibility. With that block size, the investigator can predict which treatment the second patient of each pair in the block will receive. It would have been better to use a variable block size and to not randomize unblinded patients so far in advance of treatment. Patients who know the treatment can drop out of the study if they are unsatisfied with their treatment assignment. The primary endpoint was a composite based on subjective ZCQ scores, x-ray data, and complications.

            The patient accountability data raise several issues. Of 229 patients initially enrolled, 38 were not treated. Of the 191 treated patients, 82 X-Stop® and 54 control patients had completed follow-up at 24 months. The treatment and control groups had no significant differences with respect to demographics and most baseline characteristics, but 63 percent of X-Stop® patients versus 44 percent of control patients had had previous epidural injections, a statistically significant difference.

            Patients treated with X-Stop® had a statistically significantly greater proportion of successes at 2 years than the controls (43 percent vs. 4.7 percent, respectively). One site, however, had a significantly greater success rate among X-Stop® patients. Excluding that site, the X-Stop® success rate was 33 percent. Investigators at that site were inventors of the device and had considerable experience with it, raising issues of learning curve and patient selection.

Although the sponsor claims that X-Stop® success rates were comparable to laminectomy results, this comparison was not randomized, and most laminectomy patients were study failures. Laminectomy patients were pooled from dissimilar groups: 30 from the pivotal study, 7 from the unwelded device study (Pivotal Trial 1), and 7 untreated pivotal study patients.

In summary, the potential sources of bias in the study include the use of a fixed block size of two, a subjective primary endpoint, lack of blinding, and significant differences among sites in success rates. X-Stop® success was superior to controls, although less than expected. The adverse event profile was comparable for the two groups, except for musculoskeletal events, which the sponsor claims were unrelated to the device.



John S. Kirkpatrick, M.D., expressed his appreciation to the two patients who spoke during public hearing and noted that the sponsor “did a good job” of assembling data. He reviewed the types of lumbar stenosis and causes of neurogenic claudication. The X-Stop® attempts to take advantage of postural relief with flexion of the spine in some patients who have LSS and claudication. In extension, the X-Stop® appears to prevent natural canal narrowing and decrease in subarticular diameter and foramen area. It is unclear why flexion dimensions decrease as well.

Dr. Kirkpatrick reviewed the indications and contraindications for the device and reiterated many of the concerns expressed earlier about the preclinical testing. He noted that 38 of 229 enrolled patients were not treated.

Use of the ZCQ raises several concerns. Validation studies have found that although the questionnaire yields highly reproducible results, fairly large changes are required to determine that change has truly occurred. The questionnaire may not be reproducible enough to judge the outcome after surgery for an individual patient. The 0.5 point improvement that defines success represents a 10 percent change for symptoms and a 12.5 percent change for function. Does the minimal detectable difference equate to clinically meaningful improvement?

It is unclear whether the sponsor considered sciatic tension and radiographic signs for different types of LSS. Unilateral leg pain probably is not a stenosis issue. The radiographic data do not include a preoperative–postoperative comparison or flexion and extension views. It would have been useful for the sponsor to duplicate the MRI analysis that was done in the preclinical testing. Also, his random check of the data revealed several mathematical calculation errors.

The panel’s deliberations should focus on safety and effectiveness. The safety events were few or minor and are reasonable for a surgical procedure. Life-threatening complications were related more to the patient population than the intervention. Regarding efficacy, 45.7 percent is not a significant proportion for most surgical procedures. A better definition of the population might clarify the issue. The sponsor could stratify by type of stenosis, eliminate or limit comorbidity, and define a threshold ZCQ function score. The 0.5 level of improvement was selected by the sponsor and represents a 10 to 12.5 percent change in outcome score. It approximates a minimum clinically important difference described by developers of the measure. Is this amount of improvement significant to the individual patient?

Jonas Ellenberg, Ph.D., panel statistical reviewer, raised several issues concerning the study methodology and statistical analysis. The two-person block size is problematic, in part because it enables an investigator to place patients on the arm they might do better with. In addition, the informed consent document has problems, so it is likely that patients in different arms had different expectations of outcome.

The association of baseline measures with later assessments of the ZCQ physical score component is troublesome. The relation between baseline physical score and change in score over time seems different in the X-Stop® and control arms. The more dramatic increase comes with X-Stop® implantation. Is there a group of patients in whom this procedure might be effectively used? It is clear that the patients are equivalent, but the outcomes could be treatment effects or patient related.

The centers’ differential success rates are a concern. The study had no protocol for use of laminectomy; consequently, use of laminectomy in the X-Stop® group could have inadvertently been forestalled. If laminectomy was delayed, that is not an objective outcome. Likewise, the study had no specific procedure for epidural use in the X-Stop® arm. It is not clear that the intervention was uniform. Other issues are that many outcome components are subjective, and it is not clear that there was equipoise in the informed consent process. The missing data and untreated patients have been taken into account in the sponsor’s analysis and are not a major issue.

It is not clear that the ZCQ cutpoint choice is clinically meaningful, although the shift in the X-Stop® group is clearly better than in the control group. Looking at actual values, 0.5 is about 1 standard deviation (SD) from baseline. Is 1 SD a clinically meaningful difference? The physical score is a complicated composite endpoint.

The Kaplan-Meier curve shows a significant difference between treatment and control groups in time to laminectomy. The results are surprising: One would have expected it to take longer for X-Stop® subjects to need laminectomy. Finally, the data indicate that the outcome measures regress toward baseline between 12 and 24 months; the sponsor should consider 36-month follow-up.



            Panel members asked whether the device is intended to be inserted by surgeons, as well as pain clinic doctors, and whether the x-rays were standing or supine and were taken under standardized conditions. Other questions were raised about flexibility and muscle testing, data to ascertain a potential psychological role in patients’ back pain, expulsion and other mechanical testing, the protocol for administering the ZCQ, the informed consent process, ability to track the patients who have completed the study, process by which patients were referred to the study, the sizing and insertion process and their correlation to clinical benefit, stratification on the basis of unilateral and bilateral leg pain or type of stenosis, and the clinical meaning of a 0.5 point improvement on the ZCQ. The sponsor provided a sample X-Stop® device that the panel members could examine.



1.      Please discuss the clinical significance of the musculoskeletal and other adverse events seen in the trial, and discuss whether the effects of the device on surrounding segments or on spinal biomechanics have been adequately addressed.

Panel members offered a spectrum of answers. They noted that it is unclear whether the device may lead to altered biomechanics, that the sponsor did not provide animal model data, that many patients will have back pain regardless of treatment, and that control patients were not followed as closely as treatment patients (although the panel believed that the discrepancy was not clinically significant). There appears to be a decline in results by 24 months, so the device may have limited usefulness.

2.      Based on your knowledge of the biomechanics of the spine and the nature of spinal stenosis, please discuss whether there is a clinical basis for pooling the outcomes of the one- and two-level patients.

The panel concurred that there is a clinical basis for pooling the data.

3.      Please discuss the interpretation of the measurements made on the clinical patients’ radiographs, as it relates to device effectiveness.

Panel members were not in consensus. Some members accepted the clinical data as evidence of effectiveness, whereas others believed that its effectiveness had not been demonstrated. Some panel members believed that a meaningful radiographic study would include CT or MRI data and wondered why the sponsor had not collected such data. At the very least, extension films would be needed to demonstrate effectiveness.

4a. Based on the data from this study, please discuss the appropriate population who might benefit from this device.

The panel believed that the sponsor had not provided sufficient information to answer this question. The pool of patients with clinical symptoms has enough variability that one cannot determine which patients will benefit. From a clinical symptom standpoint, the sponsor did not address the issue of unilateral versus bilateral leg symptoms. Intervention may be successful for only one type of stenosis. The sponsor should consider stratifying the data by type of stenosis or initial severity.

4b. Given the historical success rates for laminectomy, please discuss what impact the effectiveness results of this study have in relation to our interpretation of the risks and benefits of treatment with the X-Stop® device.

Panel members provided a spectrum of responses. The population that will receive the device is nonhomogeneous, making the decision to proceed with surgery difficult. Many study patients had severe, not mild to moderate, symptoms. The device could be appropriate for people who have problems with general anesthesia or are otherwise not good candidates for laminectomy. Some panel members suggested that the appropriate comparison in the study should have been to laminectomy, rather than to a nonoperative control group. Panel members again suggested stratifying the study patients by initial severity.

 5. Please describe the potential impact on the interpretation of the study result of these confounding factors.

The panel believed that a host of factors confounded interpretation of the study data, particularly the lack of a protocol for epidural injection or laminectomy, the possibly biased informed consent process, and the subjective nature of other outcomes. Epidural after X-Stop® implantation should have been considered failure. The results are nevertheless impressive, and a real difference between the X-Stop® and control patients exists.

6. Do the clinical data in the PMA provide reasonable assurance that the device is safe?

The panel concurred that the device is safe.

7. Do the clinical data in the PMA provide reasonable assurance that the device is effective?

The panel did not reach consensus on whether the data demonstrate effectiveness. Many panel members believed that the success rate was rather low and that it is unclear what patient population would benefit from the device. X-Stop® is a safe device that might prove to be a stopgap on the way to laminectomy. The study patients appear to have mixed etiology for their back pain.


No comments were made.



No comments were made.



Dr. Andersson stated that the sponsor met the primary endpoint, despite a high bar, and would choose nonoperative treatment as the control group again. The patient population was not homogenous, but it could not be because spinal stenosis is such a varied disease entity with a wide variety of etiology and presentations. The X-Stop® is an innovative device that addresses a problem in the spinal canal without entering it. It leaves the door open for additional procedures should they become necessary. He stated that a recommendation of approval would be a step in the right direction.  



Ms. Scudiero read the voting options. The panel voted five to three that the device was not approvable.



Panel members voting for the not approvable recommendation believed that the clinical study did not verify the theory of device, did not involve an adequately defined clinical population, and used overly subjective endpoints. The study involved too many potential sources of bias; for example, many patients were not mentally capable of answering questions themselves. Posterior dislodgment was a concern. The study had too many inconsistencies, and the discrepancy in results among the centers raised questions. The ZCQ was validated at 6 months, but the pivotal study used it as an outcome measure at 2 years, and it was not clear that that use of the ZCQ was validated at 24 months. Ultimately, the sponsor did not demonstrate efficacy.

Panel members voting against disapproval believed that the sponsor had demonstrated safety and efficacy of the device. The X-Stop® is minimally invasive and safer than open laminectomy.

The panel encouraged the sponsor to more precisely define which patients will benefit from the device by stratifying the study population by etiology, pain level, and bilateral versus unilateral leg symptoms. The device biomechanics should be studied, and radiographic confirmation of results—CT or MRI—would help demonstrate enlargement of the foramen or canal. Three-year follow-up and further analysis of the data to address potential biases also were recommended. The X-Stop® is an “elegantly simple” device with good potential, but the sponsor needs to demonstrate the mechanism of action.

            LeeLee Doyle, Ph.D., Consumer Representative, stated that she supported disapproval of the PMA in part because at two years, the outcomes seem poorer than at one year. However, the device seems safe, and it is unfortunate that patients will not have it as an option.

Sally L. Maher, Esq., Industry Representative, stated that the study did not have significant bias and that the device had a satisfactory risk–benefit profile. She did not support disapproval of the PMA.



Dr. Witten thanked the participants. Dr. Yaszemski thanked the panel and extended special thanks to the three outgoing panel members. The meeting was adjourned at 3:05 p.m.


I certify that I attended this meeting of the Orthopaedic and Rehabilitation Devices Advisory Panel on August 31, 2004, and that these minutes accurately reflect what transpired.



Janet L. Scudiero, M.S.

Executive Secretary





I approve the minutes of the August 31, 2004, meeting

as recorded in this summary.




Michael J. Yaszemski, M.D., Ph.D.

Acting Chairperson