Surgical management of acute, high-grade acromioclavicular joint separations: a systematic review

Abstract

Background

While a number of treatment options exist for repair of acute, high-grade acromioclavicular joint (ACJ) separation, none have emerged as the standard of care. The purpose of this study was to systematically review the literature on surgical treatment of acute, high-grade (Rockwood grades III-V) ACJ separations in order to compare outcomes between direct fixation and tendon graft ligament reconstruction.

Methods

A systematic review of the literature evaluating outcomes for acute ACJ separation treatment with direct fixation or free biologic tendon graft reconstruction was performed. The following databases were examined: the Cochrane Database of Systematic Reviews, the Cochrane Central Register of Controlled Trials, PubMed (1980-2021), and Embase (1980-2021). Studies were included if they reported a mean time to surgery as <6 weeks, contained >10 patients with a minimum 1-year follow-up, and reported clinical or radiographic outcomes.

Results

A total of 52 studies met the inclusion criteria. Seven studies reported outcomes following tendon graft ligament reconstruction (n = 128 patients). There were multiple methods of direct fixation. Thirty-three studies utilized suture button constructs (n = 1138), 16 studies used hook plates (n = 567), 2 studies used coracoclavicular screws (n = 94), 2 studies used suture fixation (n = 93), 2 studies used suture anchor (n = 55), 2 studies used suture cerclage fixation (n = 87), 1 used single multistrand titanium cable (n = 24), and 1 used K wire (n = 11). The mean follow-up Constant scores ranged from 77.5 to 97.1 in the fixation group compared to 90.3-96.6 in the tendon graft group. The mean visual analog scale scores ranged from 0 to 4.5 in the fixation group and 0.1-1 in the tendon graft group. Net CC distance ranged from 17.5 to 3.6 mm in the fixation group and 7.4-4 mm in the tendon graft group. The revision rates ranged from 0.0% to 18.18% in the direct fixation group and 5.88%-17% in the tendon graft group.

Conclusion

Direct fixation and tendon graft reconstruction for management of acute, high-grade ACJ separations have similar patient subjective and radiographic outcomes, as well as complication and revision rates at a minimum 1-year follow-up.

Injuries to the acromioclavicular (AC) joint (ACJ) are common, with nearly 12% of shoulder injuries involving this joint.⁶² ACJ separation accounts for 9% of all shoulder injuries and disproportionately affects young male athletes.¹⁰ Separation of the ACJ occurs after direct impact to the superior aspect of the shoulder while the arm is in adduction, a scenario seen more commonly in athletes or after a fall onto an outstretched hand.^13,17 A recent prospective study by Skjaker et al found that of the nearly 300 acromioclavicular-related injuries presenting to a single emergency department, 53% of injuries were sports-related (football, cycling, martial arts, alpine skiing, snowboarding, and ice hockey) and 82% of injuries were sustained by males.⁵⁰ In addition to increased prevalence, males were far more likely to experience injuries of greater severity.⁵⁰

While the epidemiology and pathophysiology of ACJ separations is well studied, controversy remains with regard to management. Patients sustaining Rockwood type I and II ACJ injuries (incomplete disruption of both the acromioclavicular and coracoclavicular [CC] ligaments) typically experience satisfactory outcomes with nonoperative management.^2,11,31 Although nonoperative management has shown recent success,^5,28 injuries of greater severity falling into Rockwood type IV-VI classifications typically require surgical correction.^2,31 There is no consensus on Type III AC separations.

While several treatment options exist for repair of acute, high-grade ACJ separation, none have emerged as the standard of care. Direct anatomic fixation techniques include, but are not limited to, CC screw fixation, suture fixation, Kirschner wire or hook plate (HP) fixation, and the use of cortical fixation devices including suture button constructs (SBCs).^7,26,35,55 Reported rates of complications, including hardware malfunction and migration, under-correction, loss of reduction (LOR), and coracoid and clavicular fractures, remain high.^7,26,35 Recently, anatomic reconstruction with free tendon graft has become increasingly popular due to reports of greater biomechanical and clinical results with a lower rate of complications.^25,26,32,44 While several reviews have compared surgical and conservative treatment for Rockwood type III AC joint injuries,^27,52 there has not been a systematic review that compares direct fixation techniques with tendon graft ligament reconstruction for patients who undergo surgical management of acute, high-grade (Rockwood III-V) AC joint injuries.

The purpose of this study was, therefore, to perform a systematic review of acute ACJ separation fixation techniques and compare clinical and radiographic outcomes as well as complication rates of direct fixation techniques with tendon graft reconstruction or augmentation.

Methods

Article identification and selection

A systematic review of the literature evaluating the outcomes and complications direct fixation of acute ACJ separation fixation, as well as reconstruction with free tendon graft, was performed using the Cochrane Database of Systematic Reviews, the Cochrane Central Register of Controlled Trials (1980-2021), PubMed (1980-2021), and Embase (1980-2021). The search queries were performed in July 2021 and conducted in accordance with the 2009 PRISMA (Preferred Reporting Items for Systematic Review and Meta-Analysis) statement.³⁶ This study was registered with the PROSPERO international prospective register of systematic reviews. The following search terms were included and combined with the Boolean operators AND/OR: “Acromioclavicular,” “Separation”, “Dislocation,” “Acute”, “reconstruction,” “anatomic fixation,” “fixation,” “treatment,” “Coracoclavicular ligament.”

The search results were compiled by 2 investigators before being combined. Duplicates were then removed. The inclusion criteria were as follows: (1) studies that examined clinical and radiographic outcomes after free tendon graft reconstruction or direct fixation of acute, high-grade (Rockwood grades III-V) ACJ separations, (2) acute ACJ separations defined by injuries treated <6 weeks from initial trauma, (3) greater than 10 patients, (4) minimum 1-year follow-up, (5) English language, (6) human subjects, and (7) level I-IV evidence. Exclusion criteria were as follows: (1) ACJ separations treated >6 weeks from initial trauma or undefined length of time between trauma and surgery, (2) synthetic ligament constructs, (3) concomitant preoperative scapular or clavicular fractures, (4) cadaveric studies, (5) level V evidence, and (6) review articles. Three authors (JHS, RWP, CC) examined the articles first by title, then by abstract, and finally by full text. If an abstract was not readily available, the full text was reviewed instead. If a discrepancy arose between reviewers, the principal investigator ultimately decided upon article inclusion or exclusion. Once an article was determined to meet the inclusion criteria, the citations of the article were thoroughly examined to verify that all relevant articles were identified.

Data collection

Data was collected by the same authors and recorded into a predetermined form. Variables for analysis included the following: patient demographics, time from injury-to-surgery, length of follow-up, complications, revisions, radiographic outcomes such as CC distance, clinical outcomes such as visual analog scale (VAS) and Constant scores (CS), and the surgical technique utilized. Means, standard deviations, and 95% confidence intervals were collected. The level of evidence for each study was assigned in accordance to the classification by Wright et al.⁶³ A meta-analysis was unable to be performed due to the heterogeneity of results and the significant discrepancy in number of studies in the tendon construct group compared to the direct fixation group (7 vs 45, respectively).

Results

The initial literature search identified 821 unique articles after removal of duplicate studies. After implementing inclusion and exclusion criteria, 637 studies were excluded by title screening and another 95 were excluded by abstract screening. Eighty-nine full-text articles were assessed for eligibility. Of these, 54 met the full inclusion criteria and were included in the final analysis (Fig. 1). Seven studies reported outcomes following tendon graft ligament reconstruction (n = 128 patients). There were multiple methods of direct fixation. Thirty-four studies utilized SBCs (n = 1138), 16 studies used HPs (n = 567), 2 studies used CC screws (n = 94), 2 study used suture fixation (n = 93), 2 studies used suture anchors (n = 55), 2 studies used suture cerclage fixation (n = 87), 1 used single multistrand titanium cable (n = 24), and 1 used K wire (n = 11). Twenty-one of the included studies were comparative studies that reported multiple techniques. Overall, there was 1 level I study, 4 level II studies, 10 level III studies, and 39 level IV studies. Table I summarizes all of the articles that were included for analysis. Study quality and risk for bias was assessed by 2 authors (JHS, RWP) using the MINORS (methodological index for non-randomized studies) instrument.⁵¹ Scores ranged from 9 to 19.

Figure 1.

Preferred Reporting Items for Systematic Review and Meta-Analysis diagram.

Table I.

Summary of included studies.

Author (technique)	Year	LOE	MINORS grade	No. of patients	Rockwood class	M	F	Mean age
Nie (HP)	2021	III	19	84	III=4, IV=46, V=34	∗	∗	36
Nie (SBC)	2021	III	-	28	III=2, IV=16, V=10	∗	∗	35.9
Panagopoulos (suture)	2021	IV	8	72	III=34, IV=14, V=24	67	5	37
Shen (HP)	2021	III	14	19	III-V	10	9	40.2
Shen (SBC)	2021	III	-	16	III-V	11	5	44.9
Yoo (HP)	2021	IV	17	10	III=5, V=5	7	3	44.4
Yoo (suture anchor)	2021	IV	-	12	III=5, V=7	8	4	42.8
Çarkçı (SBC)	2020	IV	10	36	III= 14, V=22	32	4	30.6
Kurtoglu (SBC)	2020	IV	13	25	III=12, V=13	20	5	30.7
Olivos-Meza (SBC)	2020	IV	8	52	III=21, IV=4, V=27	51	1	31.2
Rosslenbroich (SBC)	2020	IV	9	50	III=16, IV=2, V=32	47	3	38
Seo (HP w/cc repair)	2020	III	13	47	III=26, V=21	38	9	44.6
Seo (HP alone)	2020	III	-	73	III=34, V=39	65	8	46.9
Wang (SBC)	2020	IV	16	30	III=30	18	12	39.4
Wang (SBC)	2020	IV	-	30	III=30	12	18	42.2
Breuer (SBC)	2019	IV	10	51	III=8, IV=11, V=32	50	1	43
Lee (SBC)	2019	IV	15	35	III=8, IV=6, V=21	∗	∗	37
Lee (tendon construct)	2019	IV	-	12	III=1, IV=1. V=10	∗	∗	36
Seo JB (SBC, reduction-first)	2019	III	20	21	∗	19	5	51.6
Seo JB (SBC, tunnel-first)	2019	III	∗	11	∗	11	0	47.2
Zhao (SBC)	2019	IV	8	16	IV=7, V=9	10	6	50.3
Banffy (tendon construct)	2018	IV	9	17	III= 4, IV = 12, V= 1	15	2	41
f Abd Razak (HP)	2018	III	19	10	III-V	9	1	49.2
Bin Abd Razak (SBC)	2018	III	-	16	III-V	15	1	41.4
Issa (SBC)	2018	IV	10	19	III=15, IV = 4	16	3	34.4
Müller (HP)	2018	I	18	40	III-V	29	29	37.8
Stein (HP)	2018	II	17	27	III=3, IV=4, V=20	26	1	37.7
Stein (SBC)	2018	II	-	29	III=∗, IV=1, V=20	28	1	34.2
Xue (SBC)	2018	IV	12	25	V=25	15	10	43
Yin (HP)	2018	II	17	26	III=16, V=10	20	6	44.5
Zheng (CC screw + k wire)	2018	II	16	38	III=20, IV=13, V=5	22	16	40.6
Zheng (HP)	2018	II	-	43	III=25, IV=11, V=7	28	15	42.1
Zhu (tendon construct)	2018	IV	-	15	III=3, IV=2, V=7	12	3	49.7
Choi N (tendon construct)	2017	IV	8	30	III=20, V=10	27	3	42
Cisneros N (HP)	2017	IV	17	11	III=5, V=6	11	0	41
Porschke (SBC)	2017	IV	11	55	V=55	47	8	42
Shin (SBC)	2017	IV	10	21	III= 7, V=14	20	1	41.1
Zhang (SBC)	2017	IV	10	24	III=8, V=16	21	3	28.7
Cano-Martínez (SBC)	2016	IV	7	33	V=39	26	7	25
Cho (SBC + trans-articular pin)	2016	III	14	10	III=1, IV=2, V=7	9	1	38.2
Cho (SBC + trans-spine pin)	2016	III	∗	26	III=3, IV=4, V=13	23	3	41.8
Choi S (sa, sa+SBC)	2016	IV	9	43	IV=8, V=35	40	3	42.6
Metzlaff (HP)	2016	III	17	20	III=10, V=10	∗	∗	37.6
Metzlaff (SBC)	2016	III	-	24	III=14, V=10	∗	∗	37.6
Saier (SBC)	2016	IV	10	42	V=42	39	3	34.5
Ye (HP)	2016	II	19	23	III=23	14	9	33.4
Ye (tendon construct)	2016	II	-	23	III=23	16	7	34.3
Gao (HP)	2015	IV	16	24	V=24	19	5	36
Gao (titanium cable)	2015	IV	16	24	V=24	18	6	35.4
Katsenis (SBC)	2015	IV	11	50	IV=29, V=21	38	12	35.5
Kumar (HP)	2015	III	12	45	III=45	45	0	34.2
Pan (SBC)	2015	IV	7	22	III= 15, V=7	16	6	26
Rosslenbroich (SBC)	2015	IV	10	83	III= 28, IV=4, V=51	80	3	39
Yoon (HP)	2015	IV	15	24	V=24	19	5	38.8
Flinkkilä (SBC)	2014	IV	9	57	III= 15, V=42	51	6	40
Jensen (HP)	2014	III	15	30	III=12, V=18	28	2	39
Jensen (SBC)	2014	III	-	26	III=10, V=16	23	3	39
Spoliti (SBC)	2014	IV	8	19	III=10, IV=4, V=6	16	3	33
Beris (SBC)	2013	IV	7	12	III=8, IV=4	9	3	27.5
Horst (k wire)	2013	III	∗	11	III=11	10	1	32.2
Horst (SBC)	2013	III	13	16	III=16	15	1	38.9
Mardani-Kivi (tendon construct)	2013	III	13	18	III=5, V=13	16	2	33.4
Mardani-Kivi (suture)	2013	III	-	21	III=7, V=14	19	2	31.9
Venjakob (SBC)	2013	IV	11	23	III=3, IV=3, V=17	21	2	∗
von Heideken (HP)	2013	IV	10	22	V=22	17	5	40
Assaghir (cc screw)	2011	IV	11	56	III= 21; IV=9; V=26	48	8	32.25
Scheibel (SBC)	2011	IV	11	28	V=28	26	2	38.8
Yoo (tendon construct)	2011	IV	9	13	III=3, V=10	10	3	28.4
Salzmann (SBC)	2010	IV	10	23	III=3, IV=3, V=17	21	2	37.5
Greiner S (cerclage)	2009	IV	11	50	III=5, IV=1, V=44	43	7	35.3
Lädermann A (cerclage)	2010	IV	9	37	III=6, IV=12, V=19	35	2	33.6

M, Male; F, Female; LOE, Level of evidence; SBC, suture button construct; HP, hook-plate fixation; CC screw, coracoclavicular screw.

^∗Not Reported.

Constant-Murley outcome score

There were 44 studies (16 comparative) that reported outcomes using the Constant Murley score (n = 1869 patients). The breakdown was as follows: Thirty-one studies reported on suture button fixation constructs (n = 1008 patients), 17 studies reported on HP fixation (n = 538 patients), 2 studies reported on suture fixation (n = 93 patients), 2 studies reported on cerclage fixation, 1 study used a multistrand titanium cable (n = 24 patients), 2 studies used suture anchors (n = 55 patients), and 1 study reported on CC screw fixation (n = 38 patients) (Table II). Four studies reported on tendon graft ligament reconstruction (n = 69 patients) (Table III). The mean follow-up ranged from 12 to 92.5 months,^42,53,65,73 and time-to-surgery ranged from 2.6 to 15 days.^15,73 CS ranged from 82.8 to 97.1 in the SBC group, 77.5-93.7 in the HP group, 91-92.1 in the suture fixation group, 91.7-96 in the cerclage fixation group, and 90.3-96.6 in the tendon graft group.^{4,18,24,29,39,40,47,65,68,70,71} Only one study in the CC screw group reported CS (90.6).⁷²

Table II.

Clinical outcomes of direct fixation techniques.

Author (technique)	Year	No. of patients	Post-op Constant	Post-op VAS
Suture-button construct fixation
Olivos-Meza A	2020	52	97.1	†
Zhao Y	2019	16	97.1	0
Issa SP	2018	19	96.2 ± 5.1	†
Shen G	2021	16	95.7 ± 7.3	0.4 ± 0.6
Shin SJ	2017	21	95.4 ± 3.4	0.4 ± 0.6
Stein T	2018	29	95.3 ± 4.4	†
Breuer R	2019	51	95 ± 8.8	0.9 ± 0.1
Xue C	2018	25	95 ± 2.9	0 ± 0.5
Beris A	2013	12	94.8 (84-100)	0.2 (0-2)
Rosslenbroich SB	2015	83	94.7 (range, 61-100)	†
Rosslenbroich SB	2020	50	94.4 ± 8.2	†
Salzmann GM	2010	23	94.3 ± 3.2	†
Cano-Martínez JA	2016	33	94.1 ± 5.5	†
Saier T	2016	42	94 ± 4	†
Wang YC	2020	30	93.7 ± 1.8	†
Metzlaff S	2016	24	93.6 ± 3.4	†
Wang YC	2020	30	93.3 ± 1.6	†
Pan Z	2015	22	93.1 ± 2.4	†
Katsenis DL	2015	50	93.0 (range, 84-100)	†
Çarkçı E	2020	36	92 ± 4.6	†
Scheibel M	2011	28	91.5	†
Venjakob AJ	2013	23	91.5 ± 4.7	4.5 ± 1.9
Choi S	2016	43	91.2 (74-100)	†
Spoliti M	2014	19	89.7 (range, 64-99)	†
Nie S	2021	28	89.3 ± 4.2	1.2 ± 0.6
Jensen G	2014	26	89∗ (range, 52-100)	1.3 ± 1.8
Flinkkilä TE	2014	57	88 ± 10	†
Bin Abd Razak	2018	16	87.6 ± 11.7	1 ± 1.7
Kurtoglu A	2020	25	87.2 (3.2)	†
Porschke F	2017	55	85∗ (range, 44-100)	†
Zhang LF	2019	24	82.8 (4.89)	†
Hook plate fixation
Shen G	2021	19	93.7 ± 6.6	0.7 ± 0.6
Metzlaff S	2016	20	92.8 ± 3.8	†
Kumar N	2015	45	91.8 (95% CI, 88.5-93.1)	†
Cisneros N	2017	11	91.4 ± 6.8	1.45 ± 1.5
Zheng J	2018	43	91.3 ± 4.5	2 ± 1
von Heideken J	2013	22	90∗ (IQR, 86-92)	†
Yoon JP	2015	24	90.2 ± 9.9	1.6 ± 1.5
Stein T	2018	27	90.2 ± 7.8	†
Jensen G	2014	30	88∗ (range, 22-99)	1.7 ± 2.3
Gao YS	2015	24	86.1 ± 5.7	†
Seo JB	2020	47	84 ± 10.6	†
Seo JB	2020	73	83.8 ± 9.8	†
Nie S	2021	84	83.3 ± 8.8	1.8 ± 1.1
Gang Y	2016	23	80.4 ± 11.5	†
Yin J	2018	26	79.3 ± 5	4.1 ± 1.5
Bin Abd Razak	2018	10	77.5 ± 12.3	1 ± 0.7
Yoo YS	2021	10	†	1.1 ± 1.5
Suture Fixation
Panagopoulos A	2021	72	92.1 ± 7.3	†
Mardani-Kivi	2013	21	91 ± 1	0.01 ± 0.2
Cerclage Fixation
Greiner S	2009	50	91.7 ± 8.7	†
Lädermann A	2010	37	96 ± 7.7	0.08 ± 1.5
Coracoclavicular screw fixation
Zheng J	2018	38	90.6 ± 4.8	1.2 ± 1.1
Multistrand Titanium Cable
Gao YS	2015	24	97.5	†
Suture Anchor
Yoo YS	2021	12	†	1 ± 1.3

VAS, visual analog scale; SBC, suture button construct.

^∗Median.

^†Not reported.

Table III.

Clinical outcomes of tendon graft ligament reconstruction techniques.

Author (technique)	Year	No. of patients	Post-op Constant	Post-op VAS
Banffy MB	2018	17	∗	1 ± 1.3
Yoo Y	2011	13	96.6 (range, 90-100)	∗
Zhu Y	2018	15	94 ± 10	0.4 ± 1
Mardani-Kivi M	2013	18	92 ± 2	0.1 ± 0.8
Gang Y	2016	23	90.3 ± 5.4	∗

VAS, visual analog scale.

^∗Not reported.

Using an anchor-based method, Stein et al determined that the minimum clinically important difference in CS following surgical treatment for acute, high-grade AC joint separations was 16.6.⁵⁴ Direct comparison of mean values was not possible due to the limited number of studies in the tendon graft group.

Visual analog pain score

Seventeen of the included studies provided data for postoperative results using the VAS for pain (n = 649 patients). Of the direct fixation methods, there were 10 studies in the SBC group (n = 234), 9 studies in the HP group (n = 257 patients), 1 study in the CC screw group (n = 38 patients), 2 studies in the cerclage group (n = 87 patients), 1 study in the suture group (n = 21 patients), and 1 study using suture anchors (n = 12 patients) (Table II). There were 3 studies in the tendon graft ligament reconstruction group (n = 50 patients) (Table III). The mean follow-up ranged from 12 to 58 months.^60,73 The mean time-to-surgery ranged from 4.3 to 15 days.^72,73 Postoperative VAS ranged from 0 to 4.5 in the SBC group, 0.7-4.1 in the HP group, and 0.1-1 in the tendon graft group.^{2,29,47,60,64,66} Zheng et al found postoperative VAS to be 1.2 after CC screw fixation, Mardani-Kivi reported a mean VAS of 0.01 after suture fixation, and Yoo reported a mean VAS of 1 following fixation with suture anchors.^29,67,72

CC distance

Thirty-three studies reported CC distance as an outcome (n = 1375 patients). There were 25 studies in the SBC group (n = 734 patients), 7 studies in the HP group (n = 260 patients), 2 studies in the CC screw group (n = 94 patients), 1 study using suture fixation (n = 72 patients), 2 studies using cerclage fixation (n = 87 patients), 2 studies using suture anchors (n = 55 patients), and 1 study using k wire (n = 11 patients) (Table IV). There were 3 studies in the tendon graft ligament reconstruction group (n = 62 patients) (Table V). The mean follow-up ranged from 12 to 186 months.^9,73 The mean time to surgery ranged from 3.5 to 15 days.^1,6,73

Table IV.

Radiographic outcomes of direct fixation techniques. Data sorted by net CC distance.

Author (technique)	Year	No. of patients	Pre-op CC distance (mm)	Final CC distance (mm)	NET CC distance (FINAL – PRE)
Suture-button construct fixation
Bin Abd Razak ∗	2018	16	29.3 ± 11.1	11.8 ± 1.7	17.5
Kurtoglu A (SBC)	2020	25	24.9	9.3	15.6
Xue C	2018	25	23 ± 5.4	8 ± 0.9	15
Stein T ∗	2018	29	31.89 ± 8.3	16.92 ± 4.2	14.97
Wang YC	2020	30	23.5 ± 2.1	11.4 ± 1.13	12.1
Wang YC	2020	30	23.57 ± 2.7	11.47 ± 1.2	12.1
Breuer R ∗	2019	51	23 ± 5.49	11.25 ± 3.14	11.75
Seo JB (SBC, reduction-first)	2019	21	16.1	6	10.1
Salzmann GM	2010	23	20.56 ± 4.6	10.56 ± 3.6	10
Zhang LF (SBC)	2019	24	21	11.43	9.57
Seo JB (SBC, tunnel-first)	2019	11	18.8	9.7	9.1
Scheibel M (SBC)	2011	28	22.1	13.6	8.5
Olivos-Meza A∗ (SBC)	2020	52	19.3 (5.2)	10.83 (3.2)	8.47
Shin SJ	2017	21	17.5 ± 5.2	9.5 ± 2.2	8
Çarkçı E	2020	36	18.7 ± 3.5	11 ± 2.7	7.7
Cho CH (SBC)	2016	10	15.9	9.1	6.8
Cho CH (SBC)	2016	26	14.5	8.8	5.7
Horst K (SBC)	2013	16	17.11 (5.3)	11.55 (1.77)	5.56
Flinkkilä TE	2014	57	22 ± 5	17 ± 4	5
Issa SP	2018	19	16.1 ± 3.1	11.5 ± 3.6	4.6
Katsenis DL	2015	50	12.8 (range, 9.5-15)	9.2 (range, 8.7-10.2)	3.8
Cano-Martínez JA	2016	33	†	8.24 ± 0.7	†
Jensen G	2014	26	†	23.8 (range, 18.5-32)	†
Metzlaff S ∗	2016	24	†	13.2 (range, 11.7-24)	†
Venjakob AJ ∗	2013	23	20.5 ± 4.6		†
Beris A (SBC)	2013	12	†	10.5	†
Zhao Y (SBC)	2019	16	†	10.13125	†
Hook plate fixation
Bin Abd Razak ∗	2018	10	30.6 ± 12.4	13.6 ± 4.8	17
Stein T ∗	2018	27	33.4 ± 8.2	19.4 ± 6.4	14
Zheng J ∗	2018	43	22.1 ± 3	12.2 ± 1.9	9.9
Seo JB	2020	73	15.5 ± 4.6	9.1 ± 3.3	6.4
Seo JB	2020	47	14.4 ± 5.7	9 ± 2.8	5.4
Jensen G	2014	30	†	23 (range, 15.6-32.2)	†
Metzlaff S ∗	2016	20	†	14.1 (range, 12.1-23)	†
Yoo Y	2021	10	†	11.27 ± 2.59	†
Coracoclavicular screw fixation
Zheng J ∗	2018	38	22.7 ± 3.1	12.7 ± 1.8	10
Assaghir YM	2011	56	†	10.3 ± 1.2	†
Suture Fixation
Panagopoulos A	2021	72	19.3 ± 2.1	6.4 ± 0.9	12.9
Cerclage Fixation
Greiner S	2009	50	†	2.2 ± 2.8	†
Lädermann A	2010	37	†	2.8 ± 3.1	†
K-wire
Horst K (k wire)	2013	11	17.16 (6.04)	13.52 (2.69)	3.64
Suture Anchor
Yoo YS	2021	12	†	7.9 ± 2.5	†
Choi S (sa, sa+SBC)	2016	43	19.7 (5.2)	8.8 (2.4)	10.9

CC, coracoclavicular; SBC, suture button construct.

^∗=identifies studies that did NOT utilize contralateral shoulder comparison for CC distance measurements.

^†Not reported.

Table V.

Radiographic outcomes of tendon graft ligament reconstruction techniques.

Author (technique)	Year	No. of patients	Pre-op CC distance (mm)	Final CC distance (mm)	NET CC distance (FINAL – PRE)
Banffy MB ∗	2018	17	37.4 ± 6.4	30 ± 3.1	7.4
Choi NH	2017	30	15.5 ± 3.7	10.6 ± 3.3	4.9
Zhu Y	2018	15	15 ± 4	11 ± 3	4

CC, coracoclavicular.

^∗=identifies studies that did NOT utilize contralateral shoulder comparison for CC distance measurement.

Postoperative CC distance at the final follow-up ranged from 8 to 30 mm.^2,64 Net CC distance was calculated by subtracting the preoperative CC distance from the CC distance at the final follow-up and it ranged from 17.5 to 3.8 mm in the SBC group, 17 to 5.4 mm in the HP group, and 7.4 to 4 mm in the tendon graft group.^2,4,22,45,73 Net CC distance was reported by only one study in the CC screw (10 mm), suture fixation (12.9 mm), k wire (3.64 mm), and suture anchor (10.9 mm) groups.^20,40,67,72

Revisions and complications

A total of 21 studies provided revision data (n = 904 patients). There were 14 studies from the SBC group (n = 558 patients), 5 studies from the HP group (n = 140 patients), 2 studies from the cerclage group (n = 87 patients), and 1 study from the suture group (n = 72 patients) (Table VI). There were 2 studies from the tendon graft ligament reconstruction group (n = 47 patients) (Table VII).

Table VI.

Revision rate of direct fixation techniques.

Author (technique)	Year	No. of patients	Revisions (number)	Revision rate	Complications (number)
Suture-button construct fixation
Shen G	2021	16	0	0.00%	Redislocation (1)
Wang YC	2020	30	0	0.00%	No complications
Scheibel M (SBC)	2011	28	0	0.00%	No complications
Cho CH (SBC)	2016	10	0	0.00%	Pin tract infection and clavicle hole fracture (1), reduction loss by severe subsidence (1), shoulder stiffness (1)
Cho CH (SBC)	2016	26	0	0.00%	Pin tract infection and clavicle hole fracture (1), reduction loss by severe subsidence (1), shoulder stiffness (1)
Zhao Y (SBC)	2019	16	0	0.00%	∗
Cano-Martínez JA	2016	33	1	3.03%	Superficial infection (1), hardware symptoms (3), hypertrophic scar (5)
Wang YC	2020	30	1	3.33%	Redislocation (1)
Breuer R	2019	51	2	3.90%	No complications
Metzlaff S	2016	24	1	4.17%	∗
Salzmann GM	2010	23	2	8.70%	Infection (1), coracoid fracture (1), reduction loss (1)
Rosslenbroich SB	2015	83	8	9.64%	Infection (1), recurrent dislocation (8)
Flinkkilä TE	2014	57	7	12.28%	Failure of fixation (9)
Rosslenbroich SB	2020	50	9	18.00%	Infection (1),r
Porschke F	2017	55	10	18.18%	Infection (2), impaired wound healing (3), re-instability (4), hardware irritation (1)
Jensen G	2014	26	3	12.00%	CC ligament calcifications (4), recurrent vertical and horizontal instability (3)
Hook plate fixation
Von Heideken J	2013	22	0	0.00%	No complications
Kumar N	2015	45	0	0.00%	No complications
Shen G	2021	19	0	0.00%	Acromial erosion (1)
Yoon JP	2015	24	0	0.00%	Mild subacromial erosion (6), cevere subacromial erosion (3)
Jensen G	2014	30	4	13.00%	Post-traumatic stiffness (1); stiffness in combination with symptomatic ACJ osteoarthritis (2), fracture of the acromion(1)
Suture Fixation
Panagopoulos A	2021	72	2	2.77%	Mild loss of reduction (9), dislocation recurrence (2), superficial infection (1), lateral clavicle fracture (1), persistent ACJ tenderness (2)
Cerclage Fixation
Greiner S	2009	50	2	4%	Superficial infection (1), extensive CC calcification (1), complete secondary dislocation (2)
Lädermann A	2010	37	1	2.77%	No complications

SBC, suture button construct; CC, coracoclavicular; ACJ, Acromioclavicular Joint.

^∗Not reported.

Table VII.

Revision rate of tendon graft ligament reconstruction techniques.

Author (technique)	Year	No. of patients	Revisions (number)	Revision rate	Complications (number)
Banffy MB	2018	17	1	5.88%	Suture knot stack (3), failure of reduction (1)
Choi NH	2017	30	5	17.00%	Superficial infection (1), Distal clavicle fractures (3), CC interspace ossification (1), clavicular erosion (1)

CC, coracoclavicular.

Revision rate was calculated by dividing the number of revisions by the number of patients and it ranged from 0% to 18.18% overall, 0%-18.18% in the SBC group, 0%-13% in the HP group, and 5%-17% in the tendon graft reconstruction group.^{2,9,19,21,23,41,47,61,69}

Complication data was also recorded. In the SBC group, 13 included studies reported LOR as a complication.^{7,8,12,20,33,39,43,44,48,53,60,61,70} Rates of LOR ranged from 3.3% to 54.4%.^12,61 Six included studies reported infection, with rates ranging from 1.2% to 4.3%.^6,18,41, 42, 43, 44 In the HP group, acromial/subacromial erosion was the most commonly reported complication with rates ranging from 5.3% to 46.1%.^9,68,73 Data from the CC screw group was limited as there were only 2 included studies. Zheng et al reported 1 case of soft tissue irritation and 3 recurrent ACJ separations from a cohort of 38 patients.⁷² Assaghir et al followed a cohort of 56 patients and reported 1 case of AC osteoarthritis, 1 case of mild calcification localized anterior to the clavicle, 5 mild CC calcifications, and 1 AC calcification.¹ Periclavicular ossification was the most commonly reported complication in the suture fixation group.⁴³ In the tendon graft reconstruction group, 4 studies reported LOR as a complication, with rates ranging from 23% to 56%.^9,25,67,73 Only 1 study reported infection as a complication (1/30 patients, 3.3%).⁹ Within the 2 studies in the cerclage fixation group, Greiner et al¹⁸ reported 1 case of superficial wound infection, one extensive CC calcification and 2 complete secondary redislocations in a cohort of 50 patients. The second study within the cerclage group reported 1 case of skin irritation from suture knots that required revision and another who developed a transitory postoperative plexus lesion in a cohort of 37 patients.²⁴

Comparative studies

The results of the included comparative studies are included in Table VIII. Of the 6 studies that compared HP to SBC, 3 reported that SBC fixation resulted in higher CS at the final follow-up^4,38,54 and 3 found no difference between groups.^21,33,47 Additionally, 4 of these 6 studies reported VAS scores.^4,21,38,47 Three found no difference in VAS scores when comparing HP to SBC,^4,21,47 but Nie et al found that the SBC group had better VAS scores at final follow-up (1.2 ± 0.6 vs. 1.8 ± 1.1, respectively).³⁸ There were also 4 studies that compared CC distance between HP and SBC. Stein et al calculated the relative CC distance (100/AC distance × CC distance) and found that both groups had equal LOR at final 2-year follow-up.⁵⁴ Similarly, Metzlaff et al and Jensen et al found no significant difference between groups at final follow-up.^21,33 Conversely, Bin Abd Razak et al found that SBC fixation produced significantly better CC distance at final follow-up.⁴ Zheng et al compared HP fixation to CC screw fixation (supplemented with K-wire) and found that, although there was no difference in CS at final follow-up, the CC screw group had shorter operative times, smaller incisions, less blood loss, shorter hospitalizations, and better pain scores.⁷² Among the other comparative studies, Seo et al compared “tunnel first” and “reduction first” methods of SBC fixation and found that the “reduction first” method produced better clinical (Korean Shoulder Scoring System) and radiographic (CC distance, CC tunnel width) outcomes.⁴⁶ Cho et al compared temporary fixation with trans-articular pins to temporary fixation with trans-spinal pins and found no significant difference between groups.⁸ Finally, Horst et al compared SBC fixation to K wire fixation and found no significant difference in CC distance.²⁰

Table VIII.

Outcomes of comparative studies.

Author	LOE	Comparison	Outcomes
Nie S (2021)	III	HP vs. SBC	SBC group significantly less skin incision, hospitalization, and blood loss SBC group significantly better VAS and Constant
Shen G (2021)	III	HP vs. SBC	No significant difference in VAS, UCLA, or Constant
Yoo Y (2021)	IV	HP vs. multiple suture anchors	Suture anchors better ASES, SPADI, QuickDASH, and CC distance
Seo JB (2020)	III	HP vs. HP + CC ligament repair	No difference in ASES, Constant, or CC distance No difference in timing of implant removal HP + CC repair had lower CC distance ratio than HP alone
Wang YC (2020)	IV	SBC: Twin Tail Tightrope (TTR) vs. SBC: Endobutton	TTR group smaller incision length, shorter operative times, less blood loss No difference in Constant or CC distance
Lee (2019)	IV	Tendon construct vs. SBC	No significant difference in ASES, SANE, or loss of reduction
Seo (2019)	III	SBC (tunnel first) vs. SBC (reduction-first)	KSS better in reduction-first Less loss of coracoclavicular distance ratio in reduction-first Smaller coracoclavicular tunnel width in reduction-first group No difference in ASES
Bin Abd Razak (2018)	III	HP vs. SBC	Longer operative time in SBC group SBC group better Constant and CC distance at 1 year No difference in VAS
Stein T (2018)	II	HP vs. SBC	SBC group better Taft, Constant, and NAS No difference in CC distance
Yin J (2018)	II	HP w/double-tunnel CC reconstruction vs. HP alone	Reconstruction group longer incision length and operative time Reconstruction group better VAS, Constant, ASES, and satisfaction rate
Zheng J (2018)	II	HP vs CC Screw (supplemented with K wire)	CC screw shorter operative time, smaller incision, less blood loss, shorter hospitalization CC screw better VAS scores No difference in Constant or CC distance
Cisneros N (2017)	IV	HP vs. nonoperative management	No difference in SF-36 (mental), SF-36 (physical), VAS, DASH, Constant, or satisfaction 100% of non-operative patients had complete separation at final follow-up, compared to 36% in the hook plate group
Cho (2016)	III	SBC (trans-spinal pin) vs. SBC (trans-articular pin)	No difference in ASES, UCLA, or subjective shoulder score
Metzlaff S (2016)	III	HP vs. SBC	No significant difference in Constant, Taft, ACJI, or CC-distance
Ye G (2016)	II	HP vs tendon graft construct (semitendinosus + suture button)	Tendon graft reconstruction group better Constant
Gao Y (2015)	IV	HP vs. single multistrand titanium cable (MSTC)	MSTC group better Constant
Yoon J (2015)	IV	HP vs. CC ligament reconstruction	No difference in VAS or Constant HP greater reduction in CC distance
Jensen G (2014)	III	HP vs. SBC	No significant difference in VAS, SST, Constant, Taft, or CC distance
Horst (2013)	III	SBC vs. k Wire	No difference in CC distance
Mardani-Kivi M (2013)	III	Semitendinosus (ST) autograft vs. suture fixation	No difference in Constant scores No difference in VAS scores during rest or routine daily activity ST autograft group better VAS during intense activity
von Heideken J (2013)	IV	Acute vs. delayed HP	No difference in Constant Acute group better SPADI, QuickDASH, SSV Acute group less pain at rest and during movement

SBC, suture button construct; HP, hook-plate fixation; CC screw, coracoclavicular screw; VAS, visual analog score; SST, simple shoulder test; AC, acromioclavicular; ACJI, AC joint instability score; UCLA, UCLA shoulder score; NAS, numeric analog scale; ASES, American Shoulder and Elbow Surgeons score; SSV, subjective shoulder value; LOE, Level of evidence; MSTC, multistrand titanium cable; SPADI, Shoulder Pain and Disability Index; ST, Semitendinosus ; TTR, Twin Tail Tightrope ; KSS, Korean Shoulder Scoring System; DASH, Disabilities of the Arm, Shoulder and Hand.

Two studies compared direct fixation to reconstruction with tendon grafts: Ye et al compared tendon graft reconstruction using semitendinosus tendon + suture button to HP fixation and found that the tendon construct group reported better CS at final follow-up (90.3 vs. 80.4).⁶⁵ Lee et al compared tendon graft ligament reconstruction to suture button fixation and found no significant difference in American Shoulder and Elbow Surgeons score (ASES), Single Assessment Numeric Evaluation (SANE), or LOR.²⁵

Discussion

The most important finding of this systematic review was that direct anatomic fixation of acute, high-grade ACJ separation appears to have similar patient subjective and radiographic outcomes to reconstruction with a free tendon graft at a minimum 1-year follow-up.

The goal of treatment for patients with high-grade ACJ separations is to restore CC distance to allow for healing of the ruptured ligaments and to restore normal kinematics. This may be achieved by a variety of surgical techniques, with the optimal method remaining a topic of controversy. Fixation of the ACJ has been described with the use of HP and screw constructs, more recently with the use of SBCs. Suture fixation and many other fixation techniques have been used as well. These fixation techniques, however, are all associated with a myriad of technique-specific complications.^34,49,59,65 Recent literature has emphasized the need for anatomic CC ligament reconstructions to best reproduce the native biomechanics of the shoulder.¹³ Studies have demonstrated that anatomic reconstructions with semitendinosus grafts can reproduce intact state biomechanics while also demonstrating high loads to failure.^3,14 In addition, it has been suggested that reconstruction with biological free tendons may be advantageous over synthetic materials by providing biological fixation and longevity due to secondary vascularization.⁵⁸ As tendon graft constructs have grown in popularity for the treatment of chronic ACJ separations, it has yet to be determined whether tendon grafts are necessary in the acute setting. Notably, utilization of free tendon grafts may be a conduit for increased fracture risk due to the requirement of wider bone tunnels for graft passage.^13,17,30,32 This review attempted to compare outcomes and complications in patients who underwent ACJ direct fixation using either SBC, HP, or CC screw to those who underwent tendon graft reconstruction.

Thirty-three of the included studies reported outcomes for the treatment of acute ACJ separations with SBC. These constructs have demonstrated high load to failure in biomechanical studies,^3,56 and may be implanted arthroscopically, decreasing the morbidity associated with an open procedure. Of the included comparative studies, 15 used HP fixation as a comparison group. Ye et al compared results of HP fixation to treatment with suture button fixation augmented with autologous semitendinosus graft.⁶⁵ They found that the tendon graft reconstruction group had superior functional outcomes and fewer complications at final follow-up.⁶⁵ Several additional studies compared variations of HP fixation. Seo at al compared HP fixation alone to HP fixation with direct CC ligament repair and found that although the CC ligament repair group showed better maintenance of reduction, there was no difference in functional outcomes.⁴⁵ Yoon et al compared HP fixation to CC ligament reconstruction and found no difference in functional outcomes, but noted that the HP group had better maintenance of reduction.⁶⁹ Conversely, Yin et al compared HP fixation alone to HP fixation with double-tunnel CC ligament reconstruction and found that the reconstruction group reported significantly better pain and function scores.⁶⁶ Interestingly, Cisneros et al compared HP fixation to non-operative management in patients with grade III and V injuries and found that there was no difference in outcome scores between groups.³⁷ Further, multiple systematic reviews have compared surgical and conservative treatment of Rockwood grade III injuries without definitive evidence supporting one treatment strategy over the other.^27,52 Smith et al found that operative treatment led to better cosmetic outcomes but noted that there was otherwise a paucity of quality literature to recommend one treatment modality over the other.⁵² Longo et al found that operative patients seemed to have less persistent pain than non-operative patients, but similarly reported there was insufficient evidence to make a justifiable recommendation.²⁷

There was heterogeneity in subjective and radiographic outcome parameters among the 7 studies that used free biological tendon grafts.^{2,9,25,29,65,68,73} Six used semitendinosus grafts,^{2,9,25,29,65,68} and 1 used a peroneus longus graft.⁷³ Choi et al⁹ utilized a single-tunnel CC ligament reconstruction technique with autologous semitendinosus graft and found that, although surgical failure (30% failure rate and 17% revision rate) and LOR (46.7% LOR rate) were not associated with significantly worse outcomes, complications such as coracoid fracture, infection, and ossification were associated with significantly worse ASES and UCLA shoulder scores. These complications have led to the suggestion of more ubiquitous utilization of SBC as they involve smaller tunnels and may avoid donor site morbidity and excess cost.^16,17 In the only study to use an ipsilateral palmaris longus tendon autograft and SBC, Takase et al⁵⁷ reported a mean UCLA shoulder score of 28.4 with 13.6% of patients experiencing bone tunnel widening following an arthroscopic anatomic reconstruction of the CC ligaments. Both studies that utilized a semitendinosus allograft reported excellent ASES and SANE scores.^2,25 However, Banffy et al also used an additional suture-button to reconstruct the CC ligaments; only 1 patient required revision surgery. In the only study comparing semitendinosus allograft and arthroscopic suture-button fixation, Lee et al reported no significant differences in postoperative ASES scores, SANE scores, or LOR between the 2 groups. However, this study was limited by a small sample size and did not report radiographic outcomes.

Treatment of type III injuries remains controversial and depends largely on the patients’ activity level. Our study found that the most commonly reported indication for surgical intervention in Rockwood grade III injuries was patient life style, with studies typically choosing an operative approach for young, active overhead workers, manual laborers, or athletes, all with higher requirements for functional recovery.^{9,23,33,40,47,61,68,72}

The findings of this review support the recommendation for fixation of acute high-grade AC separations considering there is no clear advantage shown from the addition of a tendon graft. We cannot advocate for a particular type of acute fixation because all methods (SBC, HP, suture, cerclage, CC screw, multistrand titanium cable, K-wire, and suture anchor) showed similar outcomes.

Overall, future prospective, comparative studies with consistent outcome parameters and reconstruction techniques (ie, number of tunnels, type of button device or graft, etc.) are necessary to delineate the optimal surgical management for acute, high-grade AC joint separations as there were no randomized control trials directly comparing tendon constructs to suture button fixation and/or CC screw fixation.

The present review is not without limitations. Surgical correction of acute, high-grade ACJ separations was defined as having been repaired within 6 weeks of the initial injury. Therefore, extrapolating results to all acute cases depend upon one’s definition of an acute ACJ separation. Within both the tendon graft and direct fixation groups, there was variability in intraoperative techniques including number and width of tunnels, types and number of implants, and types of grafts utilized for reconstruction. There was also an inconsistent definition of LOR and failure among studies which may have influenced the results. A lack of studies measuring outcomes at predetermined time points made changes over time difficult to assess. In addition, heterogeneity among subjective and radiologic outcome parameters did not allow for direct comparisons across treatment groups. Furthermore, the majority of included studies were level IV evidence, making it difficult to reach a definitive clinical recommendation due to the lack of randomized control trials and limited comparative trials. Future comparative studies with standardized clinical outcome parameters and radiograph measurements are needed to elucidate the optimal treatment technique for these injuries.

Conclusion

Direct fixation of acute, high-grade ACJ separation appears to have similar patient subjective and radiographic outcomes, as well as complication and revision rates, to reconstruct with a free tendon graft at a minimum 1-year follow-up.

Disclaimers:

Funding: No funding was disclosed by the authors.

Conflict of interest: Fotios Tjoumakaris. a. AAOS: Board or committee member. b. American Board of Orthopedic Surgery, Inc.: Board or committee member. c. American Orthopedic Society for Sports Medicine: Board or committee member. d. Trice Medical, Inc.: Stock or stock Option. Kevin B Freedman. a. American Orthopedic Society for Sports Medicine: Board or committee member. b. DePuy, A Johnson & Johnson Company: Paid consultant. c. Vericel: Paid consultant. The other authors, their immediate families, and any research foundation with which they are affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.