Perianchor Cyst Formation Around Biocomposite Biodegradable Suture Anchors After Rotator Cuff Repair
Background: Biodegradable anchors may lead to perianchor cyst formation or osteolysis. A new generation of anchors contain- ing osteoconductive material was recently presented, but there is currently no solid evidence that this concept decreases cyst formation around anchors.
Hypothesis: The null hypothesis was that the prevalence and severity of cyst formation around anchors would be similar for all 3 anchor types.
Study Design: Cohort study; Level of evidence, 2.
Methods: To evaluate differences between anchor behaviors postoperatively, this study included 2 groups of patients who underwent rotator cuff repair. In group 1 (n = 38), transosseous-equivalent rotator cuff repair was performed in all patients. At the time of repair, 2 different anchors (anchor A: 23% microstructured b-tricalcium phosphate plus 77% polylactic acid enan- tiomers [PLLA]; and anchor B: 30% hydroxyapatite plus 70% PLLA) were used for medial-row repair. Insertion locations (anterior or posterior) were randomly assigned. In group 2 (n = 38), the same procedure was performed; however, 1 of the anchors used in group 1 was changed (anchor A: as above; and anchor C: 30% b-tricalcium phosphate plus 70% faster absorbing polylactic-co-glycolic acid copolymer [85% PLLA plus 15% polyglycolic acid]). The presence and severity of fluid collection around the anchors was evaluated by magnetic resonance imaging at approximately 1 year after rotator cuff repair (12.7 6 0.9 months for group 1 and 12.6 6 1.8 months for group 2).
Results: In group 1, a fluid signal was observed in 14 patients (36.8%) for anchor A and in 12 patients (31.6%) for anchor B (P . .05). The severity of perianchor cyst formation was no different for the 2 anchors (respective fluid signal grades [0-4]: 24, 3, 9, 1, and 1 for anchor A; and 26, 4, 7, 1, and 0 for anchor B; P . .05). However, in group 2, cysts were observed in 19 patients (50%) for anchor A and in 3 patients (7.9%) for anchor C (P \ .001). For anchor C, only 2 cases of grade 1 and 1 case of grade 2 fluid collection were observed. Intergroup analysis of anchor A revealed no significant differences in the prevalence or severity of perianchor cyst formation (P . .05). Healing failure was observed in 12 patients (31.6%) in group 1 and 10 patients (26.3%) in group 2 (P . .05).
Conclusion: New-generation biodegradable suture anchors seem to decrease the severity of perianchor cyst formation. Future research is required to optimize the compositions and proportions of osteoconductive materials and polymers to improve adverse reactions. Nevertheless, controlling the properties of polymers and adding osteoconductive material both appear to enhance biocompatibility.
Keywords: biodegradable anchor; cyst formation; hydroxyapatite; tricalcium phosphate; rotator cuff repair
In previous studies, biodegradable anchors based solely on polymers are problematic with respect to absorption and perianchor cyst formation after rotator cuff repair.9,11,18 In a recent study with a large series, almost one-half of patients (46.4%) reported fluid collection around suture anchors at approximately 1 year postoperatively and 19.1% of patients demonstrated significant fluid collection, although there was no correlation between the grade of fluid collection and repair integrity.11 Cyst formation around biodegradable anchors is a significant problem; compared with metallic anchors, biodegradable anchors are commonly used to preserve bone stock after rotator cuff repair and to provide sufficient bone at the time of revision or forthcoming surgery. Therefore, efforts to min- imize this shortcoming of biodegradable anchors are required to ensure their continued use. Manufacturers are aware of this problem with polymer-based anchors, and a new generation of suture anchors containing osteo- conductive materials (b-tricalcium phosphate [TCP] and hydroxyapatite [HA]) was recently introduced. However, it has not been verified whether the addition of determine which anchor type is superior in this context. Our null hypothesis was that the prevalence and severity of perianchor cyst formation would be similar for all 3 anchor types.
METHODS
Inclusion and Exclusion Criteria
This study was performed after obtaining approval from our institutional review board (IRB), and informed consent was obtained from all patients involved (IRB No. B-1303/194- 003). Patients were enrolled prospectively from June 2012 to November 2013. The following inclusion criteria were applied: (1) full-thickness rotator cuff tear verified by preop- erative magnetic resonance imaging (MRI) and at the time of surgery, (2) arthroscopic transosseous-equivalent rotator cuff repair using 2 different types of new-generation biode- gradable suture anchors in the medial row, and (3) avail- ability for postoperative MRI at approximately 1 year postoperatively to evaluate anchor status and cuff integrity. Patients who underwent partial or single-row repair, revi- sion, any previous shoulder surgery, or open rotator cuff repair were excluded.
A total of 84 patients were recruited and they were allo- cated to 1 of 2 groups of 42 patients. Sample size analysis was performed using a 1-sample design. With the knowl- edge that the prevalence of fluid accumulation around solely the polymer anchor is approximately 50%, we postu- lated that new-generation anchors could reduce this prev- alence by one-half (25%) for a power of 0.9 and a type I error of 0.05. Our calculation revealed that 38 patients per group were needed; thus, we enrolled 42 patients to account for a possible dropout rate of 10%.
Two medial-row anchors were used in both study groups, as follows. For group 1, anchor A comprised 23% microstructured b-TCP plus 77% polylactic acid enan- tiomers (PLLA) (CrossFT; ConMed Linvatec) and anchor B consisted of 30% HA plus 70% PLLA (TWINFIX-HA; Smith & Nephew). For group 2, anchor A was the same as above and anchor C comprised 30% b-TCP plus 70% faster absorbing polylactic-co-glycolic acid (PLGA) copoly- mer (85% PLLA plus 15% polyglycolic acid [PGA]) (HEA- LIX ADVANCE; DePuy-Mitek) (Table 1). Because anchor A was used in both groups, 3 different types of new- generation anchors were included in this study. Four patients in each group were either lost to follow-up or refused postoperative MRI. Therefore, 38 patients in each group were included in this study (Table 2).
Surgical Procedures
All surgical procedures were conducted by a single surgeon. The operative technique was all-arthroscopic repair of a full-thickness rotator cuff tear using the transosseous- equivalent technique. Tear sizes were measured using a probe during the arthroscopic procedures and were classi- fied as follows according to anteroposterior tear involvement at the footprint: small (\1 cm), medium (1-3 cm), large (3-5 cm), or massive (.5 cm) (Table 2).8
Cuff footprints were prepared using a curette, rasp, and motorized shaver to eliminate soft tissue remnants. However, footprint cortices were undisturbed to maintain anchor stabil- ity. In accordance with the study protocol, 2 different types of anchors (anchors A and B for group 1, and anchors A and C for group 2) were inserted at the medial margin of the foot- print at the greater tuberosity for medial-row anchors. Deci- sions regarding the anterior or posterior positioning of anchors were made using a previously generated random table. After passing medial-row sutures through the rotator cuff and tying them, suture-bridge configuration repair was accomplished by inserting lateral-row anchors using medial- row suture strands (Figure 1).
Postoperative Imaging Evaluation
Routine postoperative MRI, in accordance with our clinic protocol, was performed at approximately 1 year postoper- atively (12.7 6 0.9 months for group 1 and 12.6 6 1.8 months for group 2).All postoperative MRI examinations were performed using Intera 1.5-T magnetic resonance scanners with a ded- icated shoulder coil (Philips Medical Systems). The patients were placed supine with the forearm in a semipronated (neutral) position. The imaging protocol included the follow- ing sequences: axial fast spin-echo proton density–weighted image with fat saturation (repetition time/echo time [TR/TE], 2500-5000/80-100 ms; slice thickness, 3 mm; slice gap, 0.3 mm; field of view, 14-15 cm; matrix, 256 3 204 pix- els; echo train length, 12-18), and oblique coronal and sagit- tal fast spin-echo T2-weighted images with fat saturation (TR/TE, 2500-5000/80-100 ms; slice thickness, 3 mm; slice gap, 0.3-1 mm; field of view, 14-15 cm; matrix, 256 3 204 pixels; echo train length, 12-18).
Fluid signals around anchors were graded into 5 catego- ries using T2-weighted MRI: grade 0, perianchor fluid signal; grade 1, minimal perianchor fluid collection; grade 2, local collection of fluid; grade 3, fluid collection around the whole length of the anchor but of a diameter less than twice the anchor diameter; and grade 4, a diameter greater than grade 3 (Figure 2).11 Fluid signals were evaluated in multiple T2- weighted images to determine grades. When a perianchor fluid signal was considerable (grade 3 or 4) and the scan was considered oblique, we measured and compared diame- ters of the anchor and cyst for the largest anchor diameter. Integrity of repair was evaluated by a musculoskeletal radiologist not otherwise involved in the study. Two ortho- paedic surgeons evaluated the grade of perianchor fluid collection, with the final grade based on consensus by both. An intact cuff was defined as one showing mainte- nance of the insertion into the footprint, and an unhealed cuff was defined as a discontinuity at the footprint (Sugaya classification grades IV and V).23
Statistical Analyses
Statistical analysis was performed using the SPSS soft- ware package (version 21.0; IBM SPSS Statistics). Intra- group analysis was performed using the McNemar test and the Wilcoxon signed-rank test, and intergroup differ- ences were analyzed using the chi-square test or the Fisher exact test and Mann-Whitney test. Statistical significance was accepted for P \ .05.
RESULTS
In group 1 (n= 38), a fluid signal was observed in 14 patients (36.8%) for anchor A and in 12 patients (31.6%) for anchor B. The prevalence of fluid signal detection around anchors A and B was not statistically different (P = .774). Severities of cyst formation were similar for the 2 anchors (respective fluid signal grades [0-4]: 24, 3, 9, 1, and 1 for group A; and 26, 4, 7, 1 and 0 for group B; P = .521) (Table 3). However, in group 2 (n = 38), perianchor cyst prevalences were 19 patients (50%) for anchor A and 3 patients (7.9%) for anchor C, and these values were significantly different (P \ .001). For anchor C, only 2 cases of grade 1 and 1 case of grade 2 fluid collection were observed (respective fluid signal grades [0-4]: 35, 2, 1, 0, and 0). By contrast, 19, 9, 8, 1, and 1 cases of fluid collection were detected for anchor A in group 2, respectively (P \ .001) (Table 3). Intergroup analysis of anchor A showed no differences in the preva- lence of perianchor cyst formation or severity (P = .355 and .476, respectively). Healing failure of the rotator cuff tendon was observed in 12 patients in group 1 (31.6%) and 10 patients (26.3%) in group 2 (P = .613).
DISCUSSION
The results of this study demonstrate that the severity of perianchor cyst formation for new-generation biodegradable suture anchors for rotator cuff repair appears to be less than that of the previous generation of polymer-only anchors. In a previous study with polymer anchors (PLLA and poly- D,L-lactide from L-lactide and D-lactide), severe cyst forma- tion (grade 4) occurred in 8.6% of cases among 209 patients11; by contrast, rates of severe cyst formation were only 2.6% (2/ 76) for anchor A, 0% (0/38) for anchor B, and 0% (0/38) for anchor C in our study. However, the prevalence of a perian- chor fluid signal in the previous study (46.4%, 97/209) was not significantly different from the prevalence observed for anchors A (43.4%, 33/76) and B (31.6%, 12/38).11 The most interesting finding of the present study was the significantly lower rate of perianchor cyst formation of anchor C (7.9%, 3/ 38; P \ .001 vs anchor A). Anchors A and C both contain b- TCP (23% in anchor A and 30% in anchor C), but anchor C was composed of PLLA plus PGA and not solely PLLA.
The main reason to use a biodegradable anchor is that it preserves bone stock for possible revision or reoperation; another benefit is that when MRI is used postoperatively to assess soft tissue status, the biodegradable anchor also shows fewer artifacts than metal anchors. By contrast, anchors composed solely of polymers do not satisfy the main objective of their use.9,11,18,21 Therefore, the addition of biocomposite materials to polymers has been an impor- tant research priority in the biodegradable implant market. In knee surgery, composite interference screws containing HA or TCP are used to improve osteoconductivity and offset the acid degradation products of PLLA scaffolds, and they have shown better results than polymer-only interference screws in terms of osteolysis, cyst formation, and tunnel widening.1,3,12,14,17 Alloplastic calcium phosphates have long been used for bone reconstructive procedures. TCP and HA are the best-known calcium phosphates. HA has a Ca/P ratio of 1.67, which matches the inorganic phase of the crystallographic structure of bone. Generally, HA-based bone substitute materials are considered nonab- sorbable or have much lower degradation rates than b-TCP.4,10 On the other hand, TCP has a Ca/P ratio of 1.5, which is closer to the composition of the amorphous biolog- ical precursor of bone.15 Furthermore, in bone defects with high osteogenic potential, TCP is substituted by newly formed bone at a high rate.5,10 This may be advantageous when bone defects are located within the confines of the original skeletal envelope, such as cystic cavities, in which later implant placement is planned. However, for more demanding defect types, the rate of degradation may be too high to ensure sufficient mechanical support and strength at the augmented site16,24; in such cases, a substi- tute material with no or very limited degradation is pre- ferred to preserve augmented volume.6 Therefore, b-TCP theoretically appears to be appropriate when it is used in combination with polymer as an anchor material. However, no information is currently available regarding the relative merits of b-TCP/polymer rotator cuff anchors. On the basis of our results, it appears that adding the same type of calcium phosphate could yield different results in cyst formation; thus, b-TCP/polymer ratios7 and changes in polymer variables could influence mechanical properties and degradation rates, as could the addition of PGA. Inter- est has declined in the use of 100% PGA anchors because they degrade rapidly and lose strength soon after implanta- tion20; thus, PLLAs are now the main materials used to pro- duce bioabsorbable anchors.22 However, PLLA degrades slowly and generates adverse biologic reactions.2,11,13,19,25 As our results show, although TCP proportions are slightly different, the main difference between anchors A and C is the polymer composition (anchor C contained PGA and PLLA). We suggest that the compositional make-up requires further study.
This study has several limitations. First, anchors B and C were compared indirectly, because 3 different types of anchor could not be used in a single patient. For a suture bridge–type double-row repair, 2 medial-row anchors are usually used; thus, 2 anchors per group were used. Never- theless, intergroup analysis showed that the prevalence and severity of cyst formation for anchor A was similar in the 2 study groups, which supports the validity of our study design regarding the indirect comparison between anchors B and C. Second, patient allocation to groups was not ran- domized, because group 2 was enrolled after group 1; thus, some baseline characteristics were significantly differ- ent between the 2 groups (Table 2). However, the influences of sex distribution and tear size differences on perianchor fluid collection are likely to be small, and the similarity between anchor A results in the 2 groups supports these presumptions. Third, this was a short-term cross-sectional study, because postoperative MRI was performed at approx- imately 1 year after surgery. Therefore, the long-term courses and natural history of perianchor cysts are undeter- mined. We suggest that a follow-up study be undertaken to understand the long-term implications of cysts and anchor absorption. Fourth, we did not include all types of composite bioabsorbable anchors available on the market; thus, our results do not include those of all new-generation anchors. Studies on this topic and on different types of biocomposite anchors are ongoing. Finally, we assessed postoperative anchor status by MRI and not by computed tomography (CT), and osteoconductivity is better evaluated by CT. How- ever, we considered the addition of a postoperative CT scan for evaluating anchor osteoconductivity unethical.
The prevalence of fluid collection around new- generation biodegradable anchors is still not negligible, but the severity of perianchor fluid collection appears to be less than that around solely polymer-based anchors. Of the 3 anchors included in this study, anchor C (30% b-TCP plus 70% faster absorbing PLGA copolymer) had the best performance in terms of perianchor cyst formation. It is currently uncertain whether the proportions of b-TCP or the properties of polymers are responsible for this differ- ence. Further studies are required to clarify this issue.
CONCLUSION
The new generation of biodegradable suture anchors CP 43 seems to decrease the severity of perianchor cyst formation.