Ligamentization and Remnant Integration: Review and Analysis of Current Evidence and Implications for Scapholunate Reconstruction

Abstract

Background  Scapholunate interosseous ligament injuries are common but remain a therapeutic challenge. Current treatment modalities prioritize restoration of normal anatomy with reconstruction where appropriate. To date no reconstructive technique has been described that discusses the potential benefit of preservation of the scapholunate ligament remnant. Little is known about the “ligamentization” of grafts within the wrist. However, a growing body of knee literature suggests that remnant sparing may confer some benefit. In the absence of wrist specific studies, this literature must guide areas for potential augmentation of current surgical practices.

Objective  The purpose of this study was to perform a review of the process of ligamentization and a systematic review of the current literature on the possible role of ligament sparring and its effect on ligamentization.

Methods  A systematic search of the literature was performed to identify all the studies related to remnant sparing and the ligamentization of reconstructed tendons, regardless of graft type or joint involved from MEDLINE, EMBASE, and PubMed until February 1, 2016 using the following keywords: ligamentization, graft, remodelling, reconstruction, biomechan*, histolo∗, scapholunate ligament. Each selected study was evaluated for methodological quality and risk of bias according to a modified Systematic Review Center for Laboratory Animal Experimentation criteria.

Conclusions  The available literature suggests that ligament sparring demonstrated a trend toward improvements in vascularity, mechanoreceptors, and biomechanics that lessens in significance over time.

Clinical Relevance  This review suggests that remnant sparing may be one way to improve outcomes of scapholunate ligament reconstructive surgery.

Level of Evidence  This is a level I/II, review study.

Scapholunate interosseous ligament (SLIL) injuries are common. ¹ Despite the frequency of SLIL pathology, its treatment remains a challenge. ² Morbidity can result as the natural history of such injuries tends to be progressive, leading to osteoarthritis followed by salvage procedures including limited wrist fusion and proximal row carpectomy. Therefore, the pursuit of improved treatment modalities is a worthwhile endeavor.

Relevant Anatomy and Biomechanics

The SLIL is an intra-articular ligament that spans the scapholunate joint. Anatomically, the ligament can be divided into three distinct parts: dorsal, proximal, and palmar. ³ The dorsal subregion is the thickest of the segments at 3 mm. It lies between the proximal pole of the scaphoid and the dorsal portion of the lunate, spanning 5 mm. ⁴ In contrast, the palmar segment is similar in length and width but is only half as thick. The proximal segment is comprised largely of fibrocartilage and as such plays a relatively minor, albeit important, role in preserving joint kinetics. ²

The primary role of the SLIL is stabilization. ⁴ Both cadaveric and imaging studies have demonstrated that in the absence of a functional SLIL, scaphoid and lunate kinematics are altered, with the deleterious effect of this change worsened through repetitive movement. ^{5 6} Stabilization of the SLIL is aided by secondary stabilizers in the form of extrinsic ligaments. Traditionally, the SLIL has been treated as a single entity from a biomechanical perspective. ⁴ Recent studies have identified different mechanical properties of the three SLIL subregions: dorsal, proximal, and palmar ( Table 1 ).

Table 1. Biomechanical properties of subregions of the SLIL.

Linear load to failure, N	Berger et al 56	Logan et al 57	Shin et al 15	Hofstede et al 14	Cuénod et al 58	Nikolopoulos et al 59
Dorsal SLIL	260.3 ± 118.2	62	185 ± 87.0	141 ± 20	93.8 ± 32.8	83 ± 18
Palmar SLIL	117.9 ± 21.3	125	–	–	–	86 ± 16
Proximal SLIL	62.7 ± 32.2	–	–	–	–	–

Abbreviation: SLIL, scapholunate interosseous ligament.

The SLIL is among the most richly innervated ligaments of the wrist, with innervation primarily through the posterior interosseous nerve. Secondary innervation is through the anterior interosseous nerve via the ligament of Testut. ⁷ Recent research has highlighted the importance of ligamentous mechanoreceptors for normal joint function. ⁸ In the wrist, variations in the density of Pacini and Ruffini receptors have been linked with the various roles of each ligament. ⁹ The SLIL has a high density of these receptors and is therefore likely to play a major role in maintaining normal proprioception. Of the SLIL segments, the palmar band plays the greatest role in proprioception. ¹⁰

Current Treatment Modalities

Crawford et al identified 11 reconstructive techniques considered to be appropriate for the treatment of various stages of scapholunate instability. ² Since their review, further techniques have been described or reported upon without the emergence of a gold standard. ^{11 12}

Graft choice, like reconstructive technique, is dependent on surgeon preference. Yet, the properties of the graft may have a notable impact on the outcome of the reconstruction. Graft choice is complicated and must take into account comparisons to the biomechanical properties of the native ligament in addition to factors such as expected functional outcome, graft incorporation, graft availability, and donor-site morbidity. ¹³

Due to the dominant role that the dorsal segment plays in stabilizing the scapholunate joint, some techniques prioritize dorsal segment reconstruction. Numerous studies have compared the biomechanical properties of grafts with the native dorsal SLIL. ^{14 15} However, these studies evaluate the properties of the graft taken from cadavers and therefore fail to take into account the changes that occur when a graft is placed in vivo, a process known as ligamentization.

Ligamentization

Ligamentization refers to the incorporation and remodelling of a graft following ligament reconstruction. First described by Amiel et al on the basis of morphological and biochemical differences between tendons and ligaments, it has since been extensively studied in the anterior cruciate ligament (ACL) literature. ^{16 17}

Numerous studies have evaluated the biological, histological, and biomechanical changes that occur during ligamentization. All tendons and ligaments are composed of dense connective tissue consisting of type I and III collagen, proteoglycans, and cells. However, the composition and arrangement of macromolecules differ and result in each structure's individual mechanical properties. Compared with tendons, ligaments are more metabolically active, contain cells with rounded nuclei, and have higher DNA content, more type III collagen, more proteoglycans, less total collagen, a different amount of nonreducible collagen cross-links, and a different distribution of collagen fibril diameters. ^{18 19}

Ligamentization follows a consistent and reproducible pattern starting with the early graft phase, then progressing to the remodelling phase and terminating with the maturation phase. ^{20 21} In the early phase, gene expression changes, resulting in angiogenesis and cell proliferation. ²² At maturation, the graft stabilizes, and no further biomechanical or biological changes are evident.

The process of ligamentization is better understood in the animal model. The maturation phase is estimated to occur after 12 weeks. ²³ In the human knee, one level of evidence 3 study and three level of evidence 4 studies were available documenting the ligamentization process of the ACL by arthroscopic biopsy. ²⁴ This process is considerably slower in humans; the maturation phase is estimated to occur after 9 to 18 months, implying that the peak of the graft biomechanical strength, in humans, is likely to occur much later. Successful ligamentization of the graft is crucial to the long-term efficacy of the ligament reconstruction, with the failure of ligamentization strongly associated with graft failure. ²⁵ Importantly, the free tendon graft can survive in an intraarticular environment, which is to say that it is histologically viable with evidence of neovascularization and without signs of necrosis.

The biological changes that mark ligamentization alter the biological and biomechanical behavior of the grafted tendon. As the donor tendon progressively loses its biological properties it gains “ligamentous” histological properties that closely resemble that of the replaced ligament, ²⁴ although ultrastructural differences remain. ^{26 27} Therefore, alterations to the properties of grafted tendons caused by ligamentization must be considered when critiquing tendon graft treatments. However, very little is known about the process of ligamentization outside of the knee. Recent advancements made in the description and evaluation of SLIL anatomy provide a good understanding of what a graft should mimic. ⁴ How such grafts behave in the wrist remains unknown. In the absence of primary studies, inferences can be made from the ACL literature about intra-articular ligamentization in the wrist.

Biomechanical Changes

The mechanical strength of all biological ACL reconstruction grafts goes through a period of decreasing strength postoperatively. The magnitude and duration of this decrease are generally thought to be smaller with autograft compared with allograft. Studies up to a duration of 2 years have demonstrated that a full return to normal ACL properties is not achieved in the healing phase. ²⁸ In an animal model, a significant reduction in graft-structural mechanical and material properties occurs as early as 7 weeks after graft surgery. A gradual return of these properties occurs over a 12-month period, but they remain significantly reduced compared with the graft at the time of harvest. ^{29 30}

The Anterior Cruciate Ligament as a Model

The native ACL and SLIL have several features in common. Functionally, both ligaments are thought to be the primary source of proprioception and stability for their respective joint. ^{7 31} This is underpinned by a similar density of proprioceptive receptors and associated innervation. Histologically, the ligaments also share properties. ^{4 32}

The most commonly used graft sources to reconstruct these ligaments also share characteristics. The semitendinosus, patellar, and flexor carpi radialis (FCR) tendons are all extrasynovial. This is important as intrasynovial and extrasynovial tendons have different cellular, biological, and mechanical properties. ³³ As the process of ligamentization alters these properties, it is assumed that studies evaluating the ligamentization of semitendinosus grafts will be more readily predictive of FCR ligamentization than an intrasynovial tendon.

Although not demonstrated in the wrist, bone–patellar–bone and hamstring grafts, both of which are extrasynovial tendons, have been shown to undergo a similar ligamentization process when grafted into the knee. ²⁴ Important to the ligamentization process is exposure to synovial fluid, ¹⁷ with very little known about extra-articular ligamentization.

The biomechanical properties of hamstring and bone–patellar–bone grafts have a tensile strength that closely replicates that of a native ACL. ³² Although no FCR tensile strength studies exist in the literature with specific reference to SLIL reconstruction, Grutter and Petersen demonstrated that an FCR graft used in acromioclavicular ligament reconstruction has a tensile strength in excess of the native SLIL. ³⁴

The Value of Remnant Sparing

There are many studies evaluating ligamentization and the value of remnant sparing in the ACL literature. Positive clinical outcomes have been reported, ³⁵ but it is difficult to draw inferences from such studies due to the differing demands of the ACL compared with the SLIL. Biological studies evaluating the effect of preservation of the ligament remnant on ligamentization provide a more directly translatable source of information. There are five recent animal model studies that evaluate the biological effect of remnant sparing on complete ACL reconstructions ( Table 2 ). ^{36 37 38 39 40} The search strategy for identifying these papers is outlined in Supplementary Fig. S1 , available in the online version.

Table 2. Characteristics of studies evaluating remnant sparing and ACL reconstruction.

	Takahashi et al 38	Sun et al 37	Wu et al 39	Zhang et al 40	Song et al 36
Study type	Laboratory-controlled intervention study	Laboratory-controlled intervention study	Laboratory-controlled intervention study	Laboratory-controlled intervention study	Laboratory-controlled intervention study
Species	Sheep	New Zealand rabbits	New Zealand rabbits	New Zealand rabbits	New Zealand rabbits
Primary surgery	ACL reconstruction	ACL reconstruction	ACL reconstruction	ACL reconstruction	ACL reconstruction
Graft origin	Autograft	Autograft	Autograft	Autograft	Autograft
Graft type	Semitendinosus tendon	Achilles tendon	Achilles tendon	Achilles tendon	Semitendinosus tendon
Groups	Two groups of 21 sheep, unilateral surgery	40 rabbits, bilateral surgery	52 rabbits in six groups, bilateral surgery	75 rabbits in four groups, bilateral surgery	60 rabbits in two groups, unilateral surgery
Intervention	RP	RP	RP	RP (sleeve technique), RP (tensioning technique)	RP
Comparison	RR	RR	RR	Control, conventional ACL reconstruction	RR
Study intervals	4 and 12 wk	4, 8, and 12 wk	6, 12, 18, and 24 wk	3, 6, 12 wk	24 wk
Biomechanical assessment	Tensile test of femur–graft–tibia complex; Drawer testing	Tensile test of femur–graft–tibia complex	Tensile test of femur–graft–tibia complex	Tensile test of femur–graft–tibia complex	Tensile test of ACL remnants and graft
Proprioception assessment	S-100 immunohistochemistry demarcating type 1, 2, and 3 proprioceptive organs	Not assessed	Not assessed	S-100 immunohistochemistry; GAP-43 mRNA expression; Tissue NT-3 mRNA expression	NF immunohistochemistry
Vasculature assessment	α-Smooth muscle actin immunohistochemistry	Micro-angiography, histological score (not included)	Laser Doppler, CD34 immunohistochemistry	VEGF mRNA expression; CD34 immunohistochemistry	VEGF immunohistochemistry
Other key assessments	Histological observations and tissue dimensions	Histological score of graft and tendon–bone interface based on vessel number, fibroblast number, collagen fiber density, collagen fiber orientation	Collagen type 3 immunohistochemistry, expression; Tendon–bone integration	Histological observations, cell proliferation	Not applicable
Overall conclusion of authors	Preservation of the ACL remnant tissue may improve graft healing	Remnant preservation enhanced ACL graft healing with improved biomechanical properties	Remnant preservation may benefit healing of the tendon graft	No significant differences noted between remnant preserving and conventional ALC reconstruction methods	Remnant repairing may not replace conventional ACL repair

Abbreviations: ACL, anterior cruciate ligament; GAP-43, growth-associated protein 43; mRNA, messenger RNA; NF, neurofilament; NT-3, neurotrophin 3; RP, remnant ligament preserved; RR, remnant ligament resected; VEGF, vascular endothelial growth factor.

Due to a lack of uniformity in measurements between studies, quantitative comparison cannot be undertaken. Each study evaluated biomechanics and vascularity, and three studies assessed proprioception. Additional measures were not comparable between studies and are therefore beyond the scope of this review.

Initially, the findings of the studies appear inconsistent, with only three papers supportive of remnant sparing. Yet, a closer evaluation reveals some marked consistency, especially when accounting for harmonized time points, as demonstrated in Fig. 1 .

Fig. 1.

Timeline of biomechanical, proprioception, and vascular assessments across the identified studies.

Of note is the trend toward significant differences in vascularity, mechanoreceptors, and biomechanics that lessens in significance over time ( Table 3 ). This is particularly true of vascularity, which was reported by three studies to be significantly increased at early time points (4, 8, and 12 weeks) before converging by 18 weeks. Takahashi et al ³⁸ also demonstrated a significant increase in Ruffini, Pacinian, and Golgi tendon receptors in the remnant-preserved (RP) group at 4 and 12 weeks, with the total number of mechanoreceptors at 12 weeks comparable to native ACL. Song et al ³⁶ noted no significant difference at 24 weeks, which may indicate a convergence of the remnant-resected and RP groups at a later time point.

Table 3. Key findings of each study.

	Takahashi et al 38	Sun et al 37	Wu et al 39	Zhang et al 40	Song et al 36
Biomechanics at respective study intervals	Significant improvement in anterior translation at 60-degree knee flexion ( p  = 0.02); significantly improved initial joint stiffness at 60 and 90 degrees ( p  = 0.0328, p  = 0.0369) in RP group at 12 wk; no significant differences in maximum load, linear stiffness or elongation at failure at 4 or 12 wk	Significant improvement in stiffness ( p  < 0.001), ultimate failure load ( p  = 0.017), elongation at failure ( p  = 0.004), and yield load ( p  = 0.025) in the RP group at 8 wk	Significant improvement in stiffness ( p  < 0.05), ultimate failure load ( p  < 0.05), elongation at failure ( p  < 0.01), and yield load ( p  < 0.01) in the RP group at 24 wk	No significant difference in ultimate failure load, or graft elongation at failure at 24 wk	No significant differences in ultimate failure load, graft elongation at failure, or stiffness at 24 wk
Proprioception at respective study intervals	Increase in Ruffini, Pacinian, and Golgi tendon receptors in the RP group, no proprioceptors evident in RR group at 4 wk	Not assessed	Not assessed	No proprioceptors evident; no significant difference in NT-3 mRNA expression; no significant difference in GAP-43 mRNA expression at 3 wk	No significant difference in the number of NF-positive mechanoreceptors at 24 wk
	Significant increase in total number of Ruffini, Pacinian, and Golgi tendon receptors, and total number of proprioceptors ( p  < 0.001, p  = 0.0072, p  < 0.001, p  = 0.0009) throughout the study period (terminated at 24 wk)			No significant difference in the number of proprioceptors; no significant difference in NT-3 mRNA expression; no significant difference in GAP-43 mRNA expression at 6 wk
				No significant difference in the number of proprioceptors; no significant difference in NT-3 mRNA expression; no significant difference in GAP-43 mRNA expression at 12 wk
Vascularity at respective study intervals	Significant increase in the number of blood vessels in the RP group ( p  = 0.0212) at 4 wk	Significant increase in vascular density on graft surface in the RP group at 4 wk ( p  = 0.002)	Significant increase in blood flow of the graft in the RP group ( p  < 0.05); significant increase in CD34-positive cells in the RP group ( p  < 0.05) at 6 wk	No significant difference in VEGF mRNA expression; no significant difference in the number of CD34-positive cells at 3 wk	No significant difference in the number of VEGF-positive vessels at 24 wk
			Significant increase in blood flow of the graft in the RP group ( p  < 0.05); significant increase in CD34-positive cells in the RP group ( p  < 0.05) at 12 wk	No significant difference in VEGF mRNA expression; no significant difference in the number of CD34-positive cells at 6 wk
		Significant increase in vascular density on graft surface in the RP group at 8 wk ( p  = 0.020)	No significant difference in blood flow of the graft or CD34-positive cells at 18 wk	No significant difference in VEGF mRNA expression; no significant difference in the number of CD34-positive cells at 12 wk
	No significant difference in the number of blood vessels at 12 qk	No significant difference in vascular density of graft surface at 12 wk	No significant difference in blood flow of the graft or CD34-positive cells at 24 wk; Significant increase in blood flow of the graft overall ( p  < 0.01); significant increase in mean number of CD34-positive cells in the RP group ( p  < 0.05) throughout the study period (terminated at 24 wk)

Abbreviations: GAP-43, growth-associated protein 43; mRNA, messenger RNA; NF, neurofilament; NT-3, neurotrophin 3; RP, remnant ligament preserved; RR, remnant ligament resected; VEGF, vascular endothelial growth factor.

Significant differences were noted in the biomechanics of the graft, but these results were mixed and the evidence was therefore less convincing. Sun et al ³⁷ found significant improvements in the RP group at 8 weeks, as did Wu et al ³⁹ at 24 weeks, but the other studies found no difference at 12 and 24 weeks. Notably, remnant integration did not result in poorer graft performance.

The results presented in Zhang et al ⁴⁰ run contrary to other findings, with no significant differences at any time point. Although not directly comparable, a similar study by Xie et al ²² reported significant increases in VEGF (vascular endothelial growth factor), NT-3 (neurotrophin 3), and GAP-43 (growth-associated protein 43) expression at various intervals in contrast with the findings of Zhang et al, although different techniques were utilized. Of note, Zhang et al evaluated Achilles grafts, whereas Xie et al used semitendinosus grafts, as did Takahashi et al. ³⁸

Clinical Relevance to SLIL Injuries and Future Applications

The most important finding from this review in the context of SLIL reconstruction is the trend toward enhanced early ligamentization in grafts that incorporate the remnant stump. Remnant preservation is not currently performed in SLIL reconstruction and may provide an avenue for enhanced therapies. Although this review was limited to complete ruptures, similar findings have been demonstrated in cases of partial rupture in the ACL literature. ⁴¹ Potential benefits must be considered in the context of any added operative technical difficulty and increased operative time. Nevertheless, several inferences can be proposed.

First, remnant integration may lead to accelerated and enhanced healing, as measured through graft vascularity, particularly in the first 12 weeks postsurgery. Possible associated positive effects include improved biomechanics, although the evidence is mixed. This may lead to altered rehabilitation pathways, along with a faster return to activities. In the absence of an established timeframe for ligamentization in the wrist, definite conclusions are difficult to draw. Current ACL evidence suggests that maturation takes 9 months, although reconstructed ligaments are usually exposed to stress well before that point. ²⁴ MRI evaluation of ligamentization may provide one modality for assessing healing noninvasively. ⁴²

Second, remnant integration may offer improved proprioception compared with current reconstruction techniques, especially early in the healing process. The SLIL plays a crucial role in proprioception in the wrist, and the subsequent pathology may be attributed to a lack of adequate proprioception in SLIL deficient wrists. Currently, a range of tendon grafts are used in SLIL reconstruction; however, this is done in the absence of knowledge as to how these grafts heal. The ACL stump is known to be dense in mechanoreceptors, although this diminishes over time after ligament rupture. ^{43 44} Comparable studies do not exist for the SLIL ligament.

Nonetheless, the proliferation of mechanoreceptors demonstrated by Takahashi et al in hamstring grafts may suggest that the remnant plays an important role in enhancing the proliferation of these proprioceptors. ³⁸ This response may be tempered by week 24 as indicated by Song et al. ³⁶ In comparison, the sole study that evaluated proprioceptors in Achilles grafts failed to demonstrate a significant difference at any time point. ⁴⁰ This discrepancy is particularly important in the context of a study by Kim et al ⁴⁵ that demonstrated a lack of mechanoreceptors in Achilles allografts used in nonremnant-sparing ACL reconstructions. A further study by Chun et al ⁴⁶ evaluating biopsies of Achilles allografts used in remnant-sparing ACL reconstructions also reported fewer mechanoreceptors than expected. When these three studies are taken into account, it may suggest a fundamental inability of specific graft types to undergo certain aspects of ligamentization.

The resolution of this question could have ramifications for SLIL reconstructive surgery. First, some grafts may develop insufficient mechanoreceptors. Alternatively, formation may be increased in certain graft types. Either way, graft choice could be refined based on this criterion. Second, mechanoreceptor proliferation or the absence of it may serve to emphasize the importance of reconstructing the palmar portion of the SLIL, the technique for which has been previously described but is not uniformly performed. This is perhaps due to deference to the presumed significance of the structural importance of the dorsal segment. The inverse relationship between the number of mechanoreceptors and time since injury may partially explain the difference in results between acutely reconstructed SLILs and those reconstructed after delayed diagnosis. A decreasing ability of the remnant to support revascularization as injury chronicity increases may also be an alternate factor explaining the difference in acute and chronic reconstruction outcomes.

It is worth noting that these findings from the ACL literature are most likely to be replicated in the wrist where there is overlap in surgical techniques. Not all SLIL reconstruction methods are entirely intra-articular, which will likely affect ligamentization. Of the methods for SLIL reconstruction, and acknowledging associated concurrent injuries to primary and secondary stabilizers, the reconstruction techniques that are anatomically based may offer the best opportunity to replicate not only the static alignment of the scaphoid and lunate but also scapholunate kinematics. This could be performed using the technique described by Athlani et al, ⁴⁷ which includes reconstruction of the dorsal component of the SLIL and the dorsal intercarpal ligament. Alternatively, anatomical dorsal and volar SLIL reconstruction have been described by others. ^{48 49 50 51 52 53 54 55}

However, the process of ligamentization, the ligamentization potential of different commonly used graft types, and the biomechanical characteristics of the graft material used may yet fundamentally determine the outcome of the reconstructive techniques.