Chronic Scapholunate Interosseous Ligament Disruption: A Systematic Review and Meta-Analysis of Surgical Treatments

Abstract

Background: Although many techniques have been described, there is no clear optimal surgical treatment for chronic scapholunate interosseous ligament (SLIL) disruption. Methods: We identified 255 articles reporting outcomes of SLIL reconstruction. Of these, 40 studies (978 wrists) met eligibility requirements and reported sufficient data on radiographic outcomes to be included in the study. The mean and standard deviation of preoperative and follow-up assessments including scapholunate gap (SLG) and scapholunate angle (SLA) were used to calculate pooled standardized mean differences (SMD) with 95% confidence intervals (CIs). For other radiographic or clinical outcomes, there were not enough reported data to calculate a pooled effect size, and pooled nonstandardized comparisons were made. Results: The SMD between preoperative and postoperative SLA in tenodesis reconstruction was 0.7 (CI, 0.29 to 1.11, P = .001) and 0.04 (CI, –0.27 to 0.38, P = .8) for capsulodesis reconstruction. For SLG, tenodesis demonstrated an SMD of 1.1 (CI, 0.6 to 1.55, P < .001) compared with 0.1 (CI, –0.36 to 0.59, P = .6) for capsulodesis reconstruction. Tenodesis had a significant improvement compared with capsulodesis in SLA (P = .01) and SLG (P = .005). Tenodesis also showed improvement in grip strength and Disabilities of the Arm, Shoulder and Hand scores. Conclusions: Comparing preoperative and postoperative radiographic measurements, tenodesis reconstruction demonstrated significantly improved SLG and SLA relative to capsulodesis. Interpreted in the context of the limitations, existing data demonstrates some benefit of tenodesis reconstruction.

Introduction

Injury to scapholunate interosseous ligament (SLIL) may result in abnormal wrist kinematics and clinical symptoms limiting wrist function.^17,19 SLIL injury can be detected on plain radiographs of the wrist with scapholunate gap (SLG) widening or changes in the scapholunate angle (SLA).²⁰ If diagnosed early, SLIL disruption can be treated with direct primary repair.²⁷ If left untreated, chronic SLIL dissociation may lead to progressive degenerative arthritis of the radiocarpal and midcarpal joints.²⁹

Many techniques for SLIL reconstruction have been proposed. One group of techniques involves addressing SLIL injury with dorsal wrist capsulodesis (Figure 1).⁵ Another group of techniques involves using the flexor carpi radialis (FCR) for tenodesis.⁶ There are several variations on this technique (Figure 2).^3,23 Capsulodesis and tenodesis remain the most commonly used surgeries to reconstruct SLIL instability.³² Multiple other techniques have also been described, including temporary pin or screw placement, free tendon grafting, and use of synthetic suture reconstruction, along with multiple variations and combinations of tenodesis and capsulodesis.^{1,8,10,12,13,15,16,30}

Figure 1.

(a) In traditional Blatt capsulodesis, a strip of the dorsal wrist capsule is elevated off of its distal insertion, leaving the proximal insertion intact and crossing the radiocarpal joint. (b) The capsule strip in traditional Blatt capsulodesis is then fixated into a notch in the distal scaphoid dorsal cortex and fixated to the volar skin with pull-out wires tied over a button. (c) In the alternate technique of Berger capsulodesis, the proximal half of the dorsal intercarpal ligament is detached from its radial insertion and fixed to the lunate without crossing the radiocarpal joint. (d) In both techniques the subluxed scaphoid (arrow indicates rotation of the scaphoid in chronic scapholunate dissociation) is reduced to restore the anatomic appearance of the scapho-trapezial-trapezoidal joint. (e) In both techniques, the capsuloligamentous strip is secured to the scaphoid distal to the scaphoid’s axis of rotation with suture or bone anchor (arrow indicates vector of pull from capsulodesis repair).

Figure 2.

(a) In tenodesis, a strip of flexor carpi radialis (FCR) tendon is detached proximally and passed volar to dorsal through a bone tunnel drilled into the scaphoid. (b) In traditional Brunelli tenodesis, the FCR strip is then passed across the radiocarpal joint and anchored to the distal radius (†); in another version of the modified Brunelli technique, the FCR strip passes over the scapholunate joint and is fixed to the lunate (*). (c) Van den Abbeele described a modification of the Brunelli technique where the FCR tendon strip is passed under and around the radiotriquetral ligament and sutured to itself. (d) In the modified Brunelli technique as described by Garcia-Elias, the strip of FCR is passed through the radiotriquetral ligament and then sutured to itself.

Although many procedures have been described, no consensus exists regarding optimal SLIL reconstructive technique. Comparisons between studies are difficult as methods of reporting outcomes vary widely. The aim of this study is to assess the existing data on SLIL reconstruction and evaluate whether one family of techniques demonstrates preferable clinical or radiographic outcomes compared with other techniques in the existing literature to date. To accomplish this, a systemic review of current literature was conducted with meta-analysis of data performed to test the null hypothesis that there are no radiographic or clinical outcome differences between SLIL surgical reconstructions.

Material and Methods

Registration

The study protocol was registered with PROSPERO, an international registry of systematic reviews (CRD42017072918).

Literature Search

We searched 3 electronic databases, PubMed, Educus, and CINAHL, for clinical studies on the outcome of procedures for SLIL reconstructive surgeries using these keywords: (((scapholunate ligament) OR scapholunate)) AND ((reconstruction surgery) OR (“Reconstructive Surgical Procedures”[Mesh])). This search yielded articles from 1978 to August 2017. Citations of included publications were then reviewed for other potential studies (Figure 3). Exclusion criteria were non-English-language publications, review articles, basic science studies, case reports with one patient only, biomechanical or cadaver studies, studies on lunate or perilunate dislocation injuries, studies with follow-up less than 3 months, descriptions of acute SL repair only (defined as less than 3 months from injury), or studies that described bone-retinaculum-bone or bone-ligament-bone autografting only.

Figure 3.

Flowchart of literature search resulting in 40 eligible articles representing 978 wrists.

Note. SLIL = scapholunate interosseous ligament.

Data Extraction

Preoperative and postoperative bibliographic data, surgical data, radiographic data, clinical outcomes, and functional outcomes were extracted in a predefined data coding sheet by 2 independent reviewers. Bibliography data included article name, study design, number of patients, number of wrists, sex ratio, mean age, duration of follow-up, patient attrition (withdrawn/drop out/loss of follow-up), smoking status, manual laborer status, medicolegal claim, mechanism of injury, dominant wrist involvement, static versus dynamic instability, SLIL injury severity, delay to treatment, and follow-up. Surgical data included use of arthroscopy, method of SL fixation, surgical approach, placement of bone tunnels, tendon harvest technique, reduction maneuvers, and closure techniques. Radiographic data included SLG, SLA, radiolunate angle, carpal height ratio, capitolunate angle, and radioscaphoid angle. Functional outcomes included wrist total range of motion (ROM); extension; flexion; ulnar deviation; radial deviation; grip strength; Wrightington Wrist Score; Disabilities of the Arm, Shoulder and Hand (DASH) score; and Mayo Wrist Score.^2,14,28 Clinical outcomes included visual analog scale (VAS) pain score, time to return to work, patient satisfaction, and subjective pain.

Data Analysis

The mean of preoperative and last follow-up postoperative outcome assessments with standard deviation (SD) were used to calculate pooled standardized mean difference (SMD) with 95% confidence interval (CI). If no SD was given for continuous data, the SD was estimated as half the mean value. When data were presented as 95% CI, the SD was calculated from the formula SD = 95% CI / 1.96 × √n. When the median and range was reported for continuous outcomes, the mean and SD were estimated by assuming that the mean is equivalent to the median and that the SD is a quarter of the range.²¹ To assess for publication bias, a funnel plot was drawn and evaluated visually for asymmetry while an Orwin’s fail-safe N was calculated to assess if missing publication affected the results.²⁴ Heterogeneity in studies was assessed using Q statistic and I² statistics. Due to significant heterogeneity of studies, random-effects models with an outcome measure of Hedges’ g, as bias-adjusted pooled SMD, with 95% CI was reported. Comprehensive Meta-Analysis, version 2.2064 (Biostat, Englewood, New Jersey) was used for data pooling.

Results

Search Results

After removal of duplicates, we identified 255 articles, of which 48 were eligible to review. Of these, 2 articles were excluded for insufficient quantitative data reporting, and 6 studies discussing screw fixation and suture anchor repair were excluded as they had insufficient number of patients for inclusion in a meta-analysis. We identified techniques as either “capsulodesis reconstruction” or “tenodesis reconstruction” based on descriptions of surgical technique. Forty studies (978 wrists) reported radiographic or clinical outcomes of SLIL reconstructions and were included in the present systematic review.

Radiographic and Clinical Outcomes

Twenty-three studies (526 wrists) reported sufficient data on radiologic outcome to include in meta-analysis. There was great variability among studies with respect to reported clinical outcome measures; therefore, there were too few data that were amenable to pooled analysis. Thus we were unable to perform a meta-analysis on clinical or functional outcomes.

Heterogeneity and Publication Bias Assessment

There was a significant asymmetry in the funnel plot (Supplemental Figure 1) by Begg and Mazumdar rank correlation test indicating potential missing studies with negative results (Kendall’s tau = 0.34, P = .02).⁴ Orwin’s fail-safe N indicates at least 23 studies are needed to bring Hedges’ g under trivial cutoff. This means the overall effect size is likely overestimated in the current analysis. However, the effect remains statistically significant.

Patient Characteristics

A total of 972 patients (978 wrists) were included across publications. Of these, 71.5% were male with an average age of 40 years (range, 29-50 years), and 67.3% had involvement of their dominant wrist. The average length of postoperative immobilization was 7.1 weeks (range, 4-10 weeks), and the mean length of follow-up was 39.1 months (range, 5-165.6 months) with 39.9 month follow-up for tenodesis and 37.2 month follow-up for capsulodesis. Fourteen of the included studies (35%) used wrist arthroscopy to confirm the diagnosis of SLIL injury, assess the degree of damage, and/or rule out arthritis.

Scapholunate Angle

Thirteen studies reported SLA for tenodesis reconstruction and 11 for capsulodesis reconstruction (Table 1). The SMD between preoperative and last follow-up SLA in tenodesis reconstruction was 0.7 (CI, 0.29 to 1.11, P = .001) and 0.04 (CI, –0.27 to 0.38, P = .8) for capsulodesis reconstruction (Figure 4). Tenodesis had a significantly improved SLA compared with capsulodesis (P = .01).

Table 1.

Scapholunate Gap and Angle: Standardized Mean Difference Between Preoperative and Last Follow-up Postoperative Scapholunate Angle Within Treatment Groups for Capsulodesis and Tenodesis.

(a)	No. of studies	Heterogeneity (I2)	Pre-op SLA (degree)	Post-op SLA (degree)	Hedges’ g (95% CI)	P value
Capsulodesis	11	82%	59.9	58.8	0.04 (-0.27-0.38)	.8
Tenodesis	13	83%	73.9	61.0	0.7 (0.29-1.11)	.001*
Comparison between procedures						.01*
(b)	No. of studies	Heterogeneity (I2)	Pre-op SLG (mm)	Post-op SLG (mm)	Hedges’ g (95% CI)	P value
Capsulodesis	10	89%	3.4	3.1	0.13 (-0.36-0.59)	.6
Tenodesis	13	85%	4.4	2.9	1.08 (0.6-1.55)	<.001*
Comparison between procedures						.005*

Note. When compared between groups, tenodesis resulted in a greater standardized mean difference correction of (a) scapholunate angle (SLA) (P = .01) and (b) scapholunate gap (SLG) (P = .005). CI = confidence interval.

^*P < .05.

Figure 4.

Studies on tenodesis (asterisks) and capsulodesis (circles) demonstrating standardized mean difference of preoperative to last follow-up postoperative scapholunate angle (SLA) with 95% confidence interval (CI). Tenodesis demonstrated a significant improvement in SLA compared with capsulodesis (P = .01).

Scapholunate Gap

Thirteen studies reported SLG for tenodesis reconstruction and 10 for capsulodesis reconstruction (Table 1). The SMD between preoperative and last follow-up SLG in tenodesis reconstruction was 1.1 (CI, 0.6 to 1.55, P < .001) and 0.13 (CI, –0.36 to 0.59, P = .6) for capsulodesis reconstruction (Figure 5). Tenodesis had a significantly improved SLG compared with capsulodesis (P = .005).

Figure 5.

Studies on tenodesis (asterisks) and capsulodesis (circles) demonstrating standardized mean difference of preoperative to last follow-up postoperative scapholunate gap (SLG) with 95% confidence interval (CI). Tenodesis demonstrated a significant improvement in SLG compared to capsulodesis (P < .001).

Clinical Outcomes

There were insufficient data to report a pooled effect size for clinical outcomes. However, in comparing preoperative with postoperative results, tenodesis demonstrated significant improvement in VAS and DASH scores with a 3.4 (P < .001) and 27.3 (P < .01) point improvement, respectively (Table 2). Capsulodesis showed a reduction in VAS of 3.2 points (P = .5) and DASH of 20.2 points (P = .7). Both groups demonstrated similar percentages of patients able to return to work, with 87% of patients with tenodesis and 86% of patients with capsulodesis able to return work at latest follow-up. Too few studies reported Wrightington Wrist Score or Mayo Wrist Score outcomes to calculate significant differences between groups.

Table 2.

Clinical and Functional Outcomes: Preoperative and Postoperative Clinical and Functional Outcomes of Tenodesis and Capsulodesis With P Values Showing Significant Postoperative Change Within Treatment Groups.

	Tenodesis			Capsulodesis
	Pre-op	Post-op	P value	Pre-op	Post-op	P value
SLA	73.9	61	.01*	59.9	58.8	.8
SLG	4.4	2.9	.001*	3.4	3.2	.5
VAS	5.5	2.1	.001*	6.4	3.2	.05*
DASH	47.5	20.2	.002*	39	18.8	.7
Wrist extension (degree)	60	53.4	.04*	57.6	50.1	.05
Wrist flexion (degree)	54.3	46.5	.06	58.8	48.3	.001*
Ulnar deviation (degree)	30.6	30.2	.9	33.6	29.7	.4
Radial deviation (degree)	20.7	20.1	.7	17.9	15.8	.4
Grip strength (%CL)	60.2	77	.001*	68.3	79.4	.2
Return to work (%)		87	—		86	—
Infection (%)		3	—		5	—
Secondary surgery (%)		6	—		4	—

Note. SLA = scapholunate angle; SLG = scapholunate gap; VAS = visual analog scale (1-10); DASH = Disabilities of the Arm, Shoulder and Hand score (0-100); %CL = percentage of contralateral side.

^*P < .05.

Functional Outcomes

Standardized data meta-analysis could not be performed for functional outcomes due to insufficient data. Comparing preoperative with last follow-up postoperative results demonstrated a significant reduction in wrist flexion of 10.5° (P < .001) for capsulodesis patients and a significant reduction of wrist extension of 6.6° (P = .04) for tenodesis patients (Table 2). There was no significant change in either group from preoperative to postoperative radial deviation or ulnar deviation. When comparing the change in wrist ROM between tenodesis and capsulodesis groups, there was no significant difference between groups (Supplemental Table 1). Likewise, when wrist ROM was assessed within tenodesis and capsulodesis groups based on whether the technique described crossed the radiocarpal joint, there was no difference between subgroups. Grip strength improved significantly by 16.8% of the contralateral hand for tenodesis patients (P < .001) and by 11.1% of the contralateral hand for capsulodesis patients (P = .2) (Table 2).

Static or Dynamic SLIL Instability

Of the articles describing tenodesis repair, 7 treated static SLIL instability only, 3 treated dynamic instability only, and 2 treated both static and dynamic instability. Of the articles describing capsulodesis repair, 4 treated static SLIL instability only, and 2 treated dynamic instability only. The remaining articles either did not differentiate between static and dynamic SLIL instability or treated both without adequately defining the treatment outcomes between groups. There was no significant difference in SLG, SLA, or VAS between tenodesis performed for static or dynamic SLIL instability (P = .2, P = .4, P = .4 respectively). There was insufficient data to calculate other outcomes between tenodesis and capsulodesis performed for static versus dynamic SLIL instability.

Complications

Among patients in studies that listed complications, 3% of tenodesis patients and 5% of capsulodesis patients experienced infection (P = .1). Secondary surgery was required for 6% of tenodesis patients and 4% of capsulodesis patients (P = .5).

Discussion

Despite many published reports on methods of scapholunate ligament reconstruction, treatment of chronic scapholunate instability remains challenging. Our meta-analysis of published techniques demonstrated better overall radiographic outcomes with tenodesis compared with capsulodesis after similar durations of follow-up. In our meta-analysis, the SLG was 1 mm narrower in tenodesis procedures when compared with capsulodesis reconstruction. Similarly, tenodesis resulted in about a 12° narrower SLA compared with capsulodesis.

This meta-analysis should be taken in context of its strengths and limitations. While often easier to measure and report, radiographic outcomes do not necessarily correlate with clinical or functional results. Many papers did not publish preoperative data, and as a result, they were excluded from meta-analysis. In addition, many studies used nonstandard or inconsistent scales to report clinical and functional outcomes, which made direct comparison with existing literature and formal meta-analysis impossible. Despite the multiple studies published on this topic, our analysis indicated that an additional 23 studies would be needed to bring the effect size under a trivial cutoff. The current meta-analysis demonstrates significant asymmetry due to these missing studies, which means that the actual effect size may be smaller than that which we observed. However, despite the limitations of the existing data, radiographic results reached statistical significance.

Although we were unable to assess standardized effect size for clinical outcomes, in comparing preoperative and postoperative results, tenodesis appeared to have favorable outcomes in reducing pain, improving hand functionality and increasing grip strength as evidenced by significant improvements in VAS, DASH score, and grip strength as compared with capsulodesis (Table 2). Both capsulodesis and tenodesis resulted in decreased wrist flexion of around 10% to 15% and decreased wrist extension of approximately 10%, with no significant difference between the two reconstructive techniques (Supplemental Table 1). Patients should be counseled to expect decreased wrist ROM when undergoing either tenodesis or capsulodesis scapholunate ligament reconstruction.

Three retrospective studies directly compared the results of capsulodesis and tenodesis, though only 2 included a sufficient number of patients to permit statistical analysis.^22,25,26 Interestingly, all 3 studies draw disparate conclusions. In a study of 29 patients, Moran et al found similar clinical and radiographic results among 2 groups of patients who underwent dorsal capsulodesis or tenodesis reconstruction.²² Saffar et al found that 6 patients treated with tenodesis had worse radiographic findings and grip strength compared with 7 patients treated with capsulodesis; however, the small number of patients precluded statistical analysis of these outcomes.²⁶ Rohman et al assessed 50 patients with chronic scapholunate instability and found that patients with tenodesis had significantly improved radiographic outcomes and a trend toward lower overall failure rate.²⁵

Multiple in vivo studies suggest that scapholunate ligament diastasis occurs after injury of the scapholunate ligament in conjunction with an injury to a secondary stabilizer of the wrist.^11,18,31 Capsulodesis procedures attempt to augment ligament repair with a dorsal capsular plication, while tenodesis procedures attempt to introduce a nonanatomic stabilizer to address disordered wrist kinematics. Although these data cannot provide insights into actual kinematics, the study suggests in general the tenodesis concept may be more successful in improving carpal relationships.

In a recent survey of hand surgeon preferences, 15% of responding surgeons preferred a Blatt style capsulodesis compared with 3% who preferred tenodesis.³² Some surgeons feel that tenodesis is a more technically challenging procedure, which could explain its lack of current widespread use.^7,26 Another concern regarding tenodesis is that tendon is stiffer than native scapholunate ligament (with a greater Young’s elastic modulus); thus, introducing tendon to reconstruct the scapholunate ligament will cause undue tension on scaphoid movement and restrict wrist motion.^7,9

Conclusion

Comparing preoperative to follow-up radiographic measurements, tenodesis reconstruction demonstrated significantly improved SLG and SLA relative to capsulodesis. Meta-analysis of functional or clinical outcomes was not possible with the existing data within published studies due to insufficient data reporting and high heterogeneity of reported outcomes. Future study of SLIL dissociation will benefit from more uniform reporting of functional and clinical results.