Retraction Note: Comparison of two FDA-approved interspinous spacers for treatment of lumbar spinal stenosis: Superion versus X-STOP—a meta-analysis from five randomized controlled trial studies

The Editor-in-Chief has retracted this article [1] because of an error in the meta analysis. Re-examination of the data has showed that there is only one published randomized controlled trial comparing Superion with XStop. Due to a misunderstanding of the published clinical data, the conclusions drawn in the article are incorrect. Author Xing Yu approved this retraction, none of the other authors replied to correspondence from the publisher about this retraction.


Background
Lumbar spinal stenosis (LSS) is characterized as narrowing of the spinal canal, lateral recesses, or neuroforamina, resulting in entrapment of the cauda equine and exiting nerve roots due to hypertrophy of the osseous and soft tissue structures surrounding the lumbar spinal canal [1]. The classic manifestation is intermittent neurogenic symptom including low back and leg pain exacerbated by standing, ambulatory, or trunk extension. With the progression of compression, neurogenic claudication secondary to LSS may occur. The annual incidence of LSS is 5 per 100,000 people, most commonly affects the middle-aged and elderly population [2,3]. As society grows older, LSS becomes one of the main causes of medical visits and responsible for a tremendous impact in developed countries, from a social, economic, and medical point of view. Many studies indicate that conservative care (CC) for treatment of LSS cannot hinder the progression of the disease. Thus, decompressive laminectomy (DI) is a standard operation for LSS patient with severe claudication symptoms who fail to CC. However, patients whose symptom severity does not meet undergoing invasive surgery make therapeutic options into dilemma [4][5][6]. Interspinous spacers (ISP) bridge the gap between surgical interventions and CC in the management of LSS representing a promising minimally invasive procedure. Those devices are implanted between contiguous spinous processes via minimal surgical incision and alleviate neurogenic claudication symptoms by limiting spinal extension. Nowadays, various interspinous spacers such as Coflex, Wallis, DIAM, X-STOP, and Extersure Felix trial are still practiced in clinical [7]. It is noteworthy that, in 2005, the X-STOP became the first Food and Drug Administration (FDA)-approved interspinous process spacer for treatment of neurogenic claudication secondary to LSS. Since then, no other interspinous process spacers have received FDA approval. However, 10 years later, Superionas the second "stand-alone" ISP approved by the FDA on May 20, 2015, got the relay baton after X-STOP was ceased to sale and distribution by its manufacturer (Medtronic, Inc., Minneapolis, MN, USA) in 2015.
Although two stand-alone IPSs possessed similar mechanism in treatment for lumbar spinal stenosis, there were inherent differences in surgical placement technique and device design between two spacers. Different from X-STOP, the distinct merits of Superion procedure is minimally invasive approach with smaller blood loss and shorter hospital stay [8], and it does not cause substantial alterations or disruptions to the spinal anatomy, which likely reduces the complexity of future surgical options in the event that revision becomes necessary to Fig. 1 Flow chart of study selection process address progressive degenerative changes and/or reemergence of symptoms. Besides the operative benefits, there are also biomechanical properties compared with X-STOP on the basis of the radiographical data demonstrating Superion group generated significant lower rate of dislodgements and migrations than in the X-STOP group. In terms of device design, the wings of the X-STOP is more slender than Superion result in providing less stability between the spinous processes and exhibit greater pain, loss of function, and a higher rate of second operation [9]. Although, many previous published articles are mainly focused on comparing traditional decompressive surgery with interspinous spacer, we are aiming to compare those two types of interspinous spacer in treatment of lumbar spinal stenosis [10][11][12]. The purpose of this meta-analysis is to compare Note: 1 ZCQ patient satisfaction score, 2 axial pain severity, 3 extremity pain severity, 4 back-specific functional impairment, 5 reoperation, 6 complication the clinical outcomes in patients treated with two FDA-approved interspinous spacers.

Search strategy
To search all of the relevant literature, we systematically searched for literature published in the database (PubMed, Embase, MEDLINE, Cochrane Library, Google Scholar). The search terms were subjected to the following: "Superion," "X-STOP," "interspinous spacer," and "lumbar spinal stenosis" with various combinations of the operators "AND," "NOT," and "OR." There was a restriction of study design which was the randomized controlled trial (RCT) published between January 2000 and April 2017. The restricted language was English. The references cited in the relevant articles were also reviewed.

Inclusion criteria
All studies on treatment of lumbar spinal stenosis were reviewed. The criteria for inclusion of an article were (1) Superion compared with X-STOP for treatment of lumbar spinal stenosis, (2) a randomized controlled trial (RCT), and (3) studies with follow-up more than 6 months. Two authors assessed the potentially eligible studies independently; any disagreement was discussed and resolved by a third independent author.

Data extraction
The following data were extracted by two authors independently using a purpose-designed form: First author and year, study, design, region, details, intervention follow-up (months), and outcomes: 1) Zurich Claudication Questionnaire (ZCQ) patient satisfaction score 2) Axial pain severity 3) Extremity pain severity 4) Back-specific functional impairment 5) Reoperation 6) Complication Disagreement between the two reviewers was settled by the third reviewer. If any disagreements existed, a third author was consulted for a discussion until consensus was reached.

Quality assessment
The quality of the studies was independently assessed by the two authors according to a 12-item scale recommended by the Cochrane Back Review Group. If at least six of the 12 criteria, including randomization, allocation concealment, blinding (of the patients, assessors, and surgeons), similar baseline, selective reporting, loss to follow-up, patient compliance, similar timing, and intention-to-treat (ITT) analysis, were met without serious flaws, the studies were rated as having "low risk of bias." Otherwise, the studies were rated as having "high risk of bias."

Statistical analysis
The risk ratio (RR) and the corresponding 95% confidence interval (CI) were assessed for the dichotomous  outcomes, and the standardized mean difference (SMD) and 95% CI were assessed for the continuous outcomes. The chi-square test and Higgin's I 2 test were used to evaluate the heterogeneity. A p value less than 0.05 for the chi-square test or I 2 values exceeding 50% indicated substantial heterogeneity. A fixed-effect model was used if significantly statistical heterogeneity was absent; otherwise, a random-effect model was applied. Because of the limited number of included studies, we did not assess the possibility of publishing bias. We used Review Manager Software (RevMan Version 5.3, The Cochrane Collaboration) to conduct the statistical analysis.

Results
Flow chart for the inclusion of studies is shown in Fig. 1. RCTs with 1118 patients were included (Fig. 1). We recorded the characteristics of the 5 included trials, as well as the details of the clinical outcome measurement (Table 1).

Methodological study quality assessment
A summary of methodological domain assessment for each study is detailed in Fig 2.

Axial pain severity
Four studies reported axial pain severity of lumbar spinal stenosis after the operation. Because no heterogeneity existed among studies, the fixed model was used (Fig. 2). In forest plots, the overall effect indicated significant differences between the Superion and the X-STOP group (SMD: 0.03; 95% CI 0.15, 0.45; p < 0.0001, I 2 = 41%, p = 0.16) (Fig. 3).

Extremity pain severity
Four studies reported extremity pain severity postoperatively, and the random model was used due to heterogeneity in studies. The pooled results showed that the Superion group had no better improvement than X-STOP (SMD: 0.18; 95% CI − 0.06, 0.43; p = 0.14, I 2 = 62%, p = 0.05) (Fig. 4).

Back-specific functional impairment
Four studies reported back-specific functional impairment after the operation. No heterogeneity existed among studies, and the fixed model was used in metaanalysis. In the forest plots, the overall effect indicated no significant differences between the Superion and the X-STOP group (SMD: 0.04; 95% CI − 0.10, 0.19; p = 0.56, I 2 = 0%, p = 0.77) (Fig. 5).

Reoperation rate
The reoperation rate was reported in four studies. No heterogeneity existed among studies, and the fixed model was used in the meta-analysis. In the forest plots, the overall effect indicated no significant differences between the Superion and the X-STOP group (RR: 1.10; 95% CI 0.82, 1.48; p = 0.51, I 2 = 19%, p = 0.30) (Fig. 6).

ZCQ patient satisfaction score
Four studies reported the ZCQ patient satisfaction score postoperatively. Due to no heterogeneity among studies, the fixed model was used. The overall effect showed that there were significant differences between the Superion and the X-STOP group. (SMD: 0.23; 95% CI 0.08, 0.38; p = 0.002, I 2 = 0%, p = 0.61) (Fig. 7).

Complication
The complication was reported in two studies. No heterogeneity existed between two studies, and the fixed model was used in the meta-analysis. In the forest plots, the overall effect indicated no significant differences between the Superion and the X-STOP group (RR: 0.98; 95% CI 0.63, 1.53; p = 0.92, I 2 = 0%, p = 0.83) (Fig. 8).

Discussion
The results of this meta-analysis demonstrated that the Superion spacer provides the similar clinical outcome in extremity pain severity, back-specific functional impairment, reoperation rate, and complication back function. However, Superion showed a better outcome than X-STOP in axial pain severity and ZCQ patient satisfaction score. The similarity mechanism in two types of ISP may account for the pooled outcome. The mechanism of various ISPs in current spine field includes five points [17][18][19]: (1) enlargement of the spinal canal area, (2) increase of the neural foramina area, (3) unloading of the posterior annulus and intradiscal pressure, (4) distraction of interspinous distance, and (5) strength of the spinous processes.
Since the 1950s, the first ISP was designed and applied by Konwles [20] in clinical with many defects. For three decades of progress, the first modern ISP, Wallis spacer, was developed by Senegas [21] in the early 1980s. Moreover, as a new procedure, the ISP becomes an alternative option for the patient and makes a profound influence in treating LSS. Now, designs of ISP vary from standalone spacers represented by the X-STOP to "dynamic" spacers represented by the Coflex, respectively. The X-STOP was approved by the US Food and Drug Administration based on laboratory, mechanical, cadaver, and clinical studies in 2005 for treatment of patients aged above 50 suffering from neurogenic intermittent claudication secondary to a confirmed diagnosis of LSS. Additionally, many studies have shown favorable outcomes compared to decompression surgery (with or without fusion) [22,23]. According to a recent report, X-STOP was ceased to sale and distribution by its manufacturer (Medtronic, Inc., Minneapolis, MN, USA) in 2015, leaving the Superion as the de facto clinical option for surgeons and their patients. From experiment to clinical practice paradigm for remedy of LSS, the Superion as the second generation of stand-alone ISP was approved by the US Food and Drug Administration for commercial distribution in the USA on May 20, 2015. Both Superion and X-STOP spacers, albeit similarities in the mechanism of action and clinical outcomes, have distinct differences in device design and surgical technique. Superion spacer (size range from 8 to 16 mm) is inserted through a cannula placed between adjacent spinous processes without dissection of the spinal musculature. Compared to the X-STOP spacer with large surgical exposure, the minimally invasive nature of the Superion spacer contributes to smaller blood loss, less operative time, and shorter hospitalization. In addition to its small incision, the Superion procedure avoids generating large scar tissue around symptomatic levels which may reduce the intricacy of future revision, removal, or further decompression surgery [24]. The data from previous 2-year follow-up studies suggested that both DI and the Superion procedure provide effective and durable symptom relief of claudication symptom. For back-pain severity, the Superion group and DI group obtain an average percentage improvement of 65 and 52%. For the leg pain severity, the average percentage change with the spacer is 70%, and laminectomy is 62% in comparison to [8]. Although patients with IPS Fig. 6 The reoperation rate of the Superion and the X-STOP for the treatment of LSS Fig. 7 ZCQ patient satisfaction score of the Superion and the X-STOP for the treatment of LSS effectively ameliorate mild to moderate symptom of LSS in midterm and with the relatively small trauma of surgery, the cost-effectiveness issue for IPS should be considered. Parker et al. [25] use Markov model to evaluate three strategies of care for LSS. The study shows that CC has the lowest cost at $10,540 and the lowest quality-adjusted life year increase, while ISP and DI were nearly identical at about $13,950 and also at qualityadjusted life years. They indicated that surgical care, especially in ISP procedure, provides superior value versus sustained CC in treating LSS. However, another study reveals that for the 1-level procedure, a mean cost of $9291 for DI, $7900 for ISP, and $3478 for CC, and for 2-level $13,429 was a cost for ISP, and the other two treatments remain unchanged [26]. Despite the merits of the ISP were mentioned above, it is noteworthy that not only complications associated with process fracture, dislodgements, and migrations of the spacer, heterotopic ossification [27], brought barriers to its safety but also the exceeding indication restricted its clinical application.
It is noteworthy that both IPD and decompressive surgery are current treatment strategies for lumbar spinal stenosis. However, which procedure can bring patients better outcomes is still pending. Moojen et al [28,29] showed that, at 2 years, the success rate according to the ZCQ for the IPD group [69% (95% CI 57-78%)] did not show a significant difference compared with decompression [60% (95% CI 48-71%) p = 0.2]. Furthermore, longterm VAS back pain was significantly higher [36 mm on a 100-mm scale (95% CI 24-48)] in the IPD group compared to the decompression group [28 mm (95% CI 23-34) p = 0.04]. But the IPD group caused higher reoperation rate (29%) than that in the decompression group (8%) in the early post-surgical period. In 2013, Strömqvist et al. [22] conducted a trial to compare the X-STOP with conventional decompression in patients with neurogenic intermittent claudication. The results showed that the clinical improvement of two procedures were similar at 6, 12, and 24 months follow-up. Similarly, patients suffered from a higher rate of reoperations in the X-STOP group (26%) than decompression surgery (6%) at endpoint. Intriguingly, in two meta-analysis reviews [30,31], we found that there is no significant difference between the IPD and decompression surgery for effective indicator such as Visual Analog Scale, Oswestry Disability Index, and Roland Disability Questionnaire and Complications. In addition, for the reoperation rates, both studies draw the same conclusion that the decompression group is lower than the IPD group.
This meta-analysis has a few limitations. According to our research results and inclusion criteria, five RCT studies are included. A few RCTs and incomplete data may reduce the quality of evidence and strength of analysis. Blinding of patients and surgeons are difficult to evaluate surgical effect in a clinical trial. Inadequate blinding is reported to generate 15% overestimation of treatment effect.

Conclusions
Based on the above study from 18 to 24 months of clinical pooled outcome, we concluded that both the Superion and X-STOP interspinous spacers can relieve symptoms of LSS. In addition, the pooled results indicate that Superion shows advantage than X-STOP in axial pain severity and ZCQ patient satisfaction score, whereas similar in other outcomes. The Superion spacer may represent a promising spacer for patient with LSS owing to its nature of minimally invasive, easily operable, and less damage. Because of the advantage of IPS technique, it would win a wide place in the future degenerative lumbar microsurgery.