Rockwood type III is the most common type of acromioclavicular joint dislocation: A prospective cohort study investigating the incidence and epidemiology of acute acromioclavicular joint dislocations in an urban population

Abstract

Introduction

Acromioclavicular (AC) joint dislocations are common, but basic epidemiological features and sub-classification are not well investigated. The aim of the study was to investigate the incidence and epidemiology of acute AC joint dislocations in the capital region of Denmark.

Methods

All patients with acute AC joint dislocation admitted to the emergency departments at 3 University Hospitals serving a population of 549,225 residents were prospectively registered from January to December 2019. Patients with trauma to the shoulder, pain from the AC joint and increased coracoclavicular distance on radiographs were included and classified according to Rockwood's classification. Data on age, sex and mechanism of injury were registered.

Results

A total of 106 patients, male:female ratio 8.6:1, were included. Rockwood type III was most common accounting for 59/106(55.7%) of the injuries. The incidence was 19.3 per 100,000 person-years at risk (PYRS). The age distribution was bimodal peaking at the ages of 20–24 and 55–59 years. The most common mechanism of injury was sports, 80/106, with cycling accounting for 51/106.

Discussion/Conclusion

Rockwood type III was the most common type of AC joint dislocation constituting 55.7% of the injuries. The incidence was 19.3 per 100,000 PYRS. Young and middle-aged males were at highest risk and most injuries occurred during sports.

Introduction

Acromioclavicular (AC) joint dislocations are common injuries accounting for 9–12% of all injuries to the shoulder girdle, but data describing the epidemiology of the injury are scarce.^1–4 An AC joint dislocation may lead to weeks or even months of sick leave and restraint from work and sports, thus having an effect on people's health and leading to increased social costs. ⁵ The mechanism is often a direct trauma to the shoulder with the arm in an adducted position, but the injury also occurs indirectly by falling on the elbow or on a stretched arm.^1,3,6

AC joint dislocations are more common in men than in women with a ratio between 5:1 and 8:1 and the majority of the injuries are caused by sport traumas.^7–10

Most studies regarding the epidemiology of AC joint dislocations refer to a limited subgroup of elite athletes and only limited research is available describing the incidence and epidemiology of the injury in an urban population^7,9,11–17

The injury is usually classified according to the severity of radiographs. The most frequently used classification was described by Tossy et al. in 1963 and modified by Rockwood in 1984. It classifies the injury in type I-VI according to the displacement of the clavicle to acromion.^18,19

The most optimal treatment of AC joint injuries is a topic of debate worldwide. There is agreement on the treatment of Rockwood type I–II dislocations, which rehabilitates well with conservative treatment, and on the high-grade injuries of types IV, V and VI, which usually requires surgical intervention. However, the treatment of type III remains controversial.^6,20–23

The aim of this study was to investigate the incidence and epidemiology of acute AC joint dislocations in an urban population, including the distribution of injuries according to Rockwood’s classification, and to describe age, gender and mechanism of injury. We hypothesized that the incidence would be in the range of 8–45:100 000, that Rockwood types II and III would comprise the majority of injuries, that there would be an overweight of young to middle-aged males and that most injuries would occur during sports.

Materials and methods

The study was conducted as a prospective cohort study. All AC joint dislocations admitted to the emergency departments or orthopaedic departments at the three hospitals: Copenhagen University Hospital Hvidovre, Copenhagen University Hospital Amager and Copenhagen University Hospital Glostrup, were prospectively registered over a period of 1 year. The three including hospitals are the only available hospitals for acute injuries covering the catchment area of 549,225 residents by 1 January 2019. No private alternatives exist for acute injuries. If patients belonging to the catchment area were injured and seen at other hospitals in the country, they would have been referred for follow-up at University Hospital.

Population

The patients were identified by a weekly systematic evaluation of all radiographs of the AC joint and clavicle taken at the three hospitals.

Inclusion criteria were residents living within the catchment area with a trauma to the shoulder within the last 2 weeks resulting in a subsequent onset of pain from the AC joint and increased coracoclavicular distance on anteroposterior radiographs. Residents living in the catchment area who were primarily treated elsewhere for AC joint dislocations and subsequently referred for treatment were also eligible for inclusion in the study.

Exclusion criteria were patients not being residents in the catchment area.

Radiographic evaluation

Prior to inclusion radiographic projections of the shoulder and AC joint were standardized according to written instructions across the three hospitals. The standard projection for the AC joint was an AP view of the AC joint with no stress applied and with the contralateral side as a comparison or a panorama (bilateral Zanca) view with a 15° cephalic tilt of the beam. ³ The standard view of the clavicle was an AP radiograph. The injuries were graded as Rockwood types II–VI according to the percentual increase in the coracoclavicular distance as described in Table 1. All radiographs were graded with a radiograph of the contralateral shoulder as a comparison.

Table 1.

Differentiation of acromioclavicular (AC) joint dislocations according to the Rockwood classification.^4,24

Rockwood’s classification	AC joint displacement on anteroposterior radiographs
Type I	A sprain to the AC joint. No displacement or widening of the joint.
Type II	A slight upward displacement of the lateral tip of the clavicle is present. Radiographs show an increase in the coracoclavicular distance of <25% compared to the uninjured side.
Type III	Radiographs reveal a 100% displacement of the clavicle to acromion and an increase in the coracoclavicular distance of 25–100% compared to the uninjured side.
Type IV	The lateral tip of the clavicle is displaced posteriorly to the acromion.
Type V	Radiographs reveal an upward displacement of the clavicle to acromion of >100% resulting in an increase in the coracoclavicular distance of 100–300% compared to the uninjured side.
Type VI	The lateral tip of the clavicle is displaced inferiorly to the coracoid process.

After 1 year, when data collection was completed, a second investigator evaluated the same radiographs and inte-rrater reliability was investigated. Before further analysis potential disagreement in the classification was settled by discussion.

Variables

The variables of interest were: age, sex, city district, time of injury, affected side and mechanism of injury. All AC joint dislocations were graded according to Rockwood’s classification types II–VI based on the criteria listed in Table 1. Rockwood type I was not included as it represents a sprain to the joint rather than a true dislocation.

Data were collected and managed using Research Electronic Data Capture (REDCap) tools hosted at Regionsgården, Kongens Vænge 2, 3400 Hillerød, Denmark. REDCap is a secure, web-based software platform designed to support data capture for research studies, providing (1) an intuitive interface for validated data capture; (2) audit trails for tracking data manipulation and export procedures; (3) automated export procedures for seamless data downloads to common statistical packages and (4) procedures for data integration and interoperability with external sources.^25,26

Statistics

The overall incidence as well as age- and sex-specific incidences were calculated as the number of AC joint dislocations divided by the person-years at risk (PYRS). As the inclusion period was 1 year each person was estimated to contribute 1 year at risk. No censoring for death was performed.

Between-group analyses were performed with Student’s t-test for continuous and normally distributed data, Mann–Whitney U test for skewed data and the chi-square test for categorical data.

Inter-rater reliability of the radiographic evaluations was investigated by Kappa statistic.

IBM SPSS Statistics Version 25 was used for the statistical analysis. Results were considered statistically significant if p < 0.05.

Results

From 1 January to 31 December 2019, 106 patients were diagnosed with an AC joint dislocation corresponding to an overall incidence of 19.3 per 100 000 PYRS. The incidence was 34.9 per 100,000 PYRS for males and 4 per 100,000 PYRS for females.

Median age was 40 (interquartile range (IQR): 27–52, minimum 19 maximum 77). 95/106 (89.6%) persons were male and 11/106 (10.4%) were female giving a male-female ratio of 8.6:1 (p < 0.01). There was no statistically significant difference in median age for males 41 years (IQR: 27–53, minimum 19 maximum 77) and females 37 years (IQR: 23–48, minimum 21 maximum 53)) (p = 0.31). For males, a bimodal age distribution was seen peaking at the age of 20–24 (69.4 per 100,000 PYRS) and 55–59 years (86.2 per 100,000 PYRS). For females, the incidence rates were lower and appeared more evenly distributed (Figure 1).

Figure 1.

Incidence of AC joint dislocations by age and sex.

AC: acromioclavicular; PYRS: person-years at risk.

Laterality of the injury was evenly distributed with 56/106 (52.8%) cases on the right side and 50/106 (47.2%) on the left side (p = 0.62).

Rockwood type III was the most common type of AC joint dislocation accounting for 59/106 (55.7%) of the injuries, although type II was more common in women accounting for 7/11 (63.6%) (p = 0.2) (Figure 2).

Figure 2.

Incidence of AC joint dislocations by sex and the Rockwood classification.

AC: acromioclavicular; PYRS: person-years at risk.

Rockwood type II accounted for 43/106 (40.6%) and type V for 4/106 (3.8%). Type I was not included in the study and there were no type IV or VI dislocations (Figure 2).

The two raters agreed on the classification in 92/106 cases. In 14/106 cases, there was disagreement if the AC joint dislocation was a type II or III. After discussion, nine cases were classified as type II and five as type III. The inter-rater reliability was calculated with a kappa value of k = 0.76 (p < 0.01), 95% CI (0.65, 0.87), which represents a substantial level of agreement. ²⁷ After discussion, there was 100% agreement between the two raters.

The most common mechanism of injury was sports accounting for 80/106 (75.5%) with cycling accounting for almost half of all cases 51/106 (48.1%) (Figure 3). Soccer and cycling contributed equally for people aged 18–29 years, whereas cycling was the most common mechanism of injury in the middle-aged population (30–59 years) (Figure 4).

Figure 3.

Causes of acromioclavicular (AC) joint dislocations by sex.

Figure 4.

Mechanism of injury by age group.

The incidence of AC joint dislocations was lowest in March and peaked in May (Figure 5).

Figure 5.

Incidence of AC joint dislocations by month.

AC: acromioclavicular; PYRS: person-years at risk.

Discussion

The overall incidence was 19.3 per 100,000 PYRS with 34.9 per 100,000 PYRS for males and 4 per 100,000 PYRS for females. In comparison, the overall incidence of primary shoulder dislocations is reported to be 23.9 per 100,000 PYRS. ²⁸

This result is in accordance with results from previous studies reporting incidences of 8–45:100 000.^7–10,15,16 Two prospective studies published in 2020 report incidences similar to ours; in a general population in Helsingborg, Sweden, Nordin et al. found an incidence of 20 per 100,000 person-years, with type I constituting 13% of the injuries, ¹⁵ and Skjaker et al. described the incidence of shoulder injuries in Oslo and found an incidence of 45 per 100,000 person-years for AC joint dislocations when including Rockwood type I that represented two-thirds of the injuries. ¹⁶ Excluding type I, as in our study, the incidence was approximately 15 per 100,000 PYRS much like the incidence found in the study by Nordin et al. and in the present study. The comparable results in the three studies support the probability of the found incidence rates to be valid for urban populations in the Nordic countries. Copenhagen is considered one of the most bike-friendly cities in the world with around 90% of the population owning a bike. ²⁹ The incidence of bicycle accidents found in this study might be higher than in cities with fewer people owning a bike. As such, one should be careful to generalize the incidence rates to other parts of the world.

Rockwood type III AC joint dislocation was the most common type of injury to the AC joint comprising more than half (55.7%) of all injuries in the present study. The treatment of type III is controversial, and the result stresses the importance of further research investigating the best treatment for this common type of injury.

Two previous studies support the findings of the present study. In 2013, a retrospective database study by Chillemi et al. found type III accounting for 40% of the injuries, and in 2020 Nordin et al. found type III accounting for 33% in a prospective study.^7,15

The classification of the injury in this study was based upon the coracoclavicular difference, measured on either a panorama radiograph with both AC joints in the same picture or bilateral AP radiographs of the injured and the uninjured side. In Rockwood type I injuries, the AC ligaments are sprained with no dislocation of the joint and these injuries cannot be detected on radiographs. ³⁰

As type I is not a true dislocation and difficult to detect it may contribute to an over or under diagnosis of AC joint dislocations. Previous studies report incidence rates for type I from 13% to 68% underlining the questionable reliability in diagnosing type I injuries.^15,16 Based on this, type I injuries were not included in the present study.

No Rockwood type IV dislocations were registered. The ability to identify Rockwood type IV dislocations solely based on anteroposterior radiographs is limited. Axillary projections improve the likelihood of identifying posterior clavicular translation in the AC joint but the accuracy remains poor why a clinical examination is the best way to exclude type IV dislocation.^31,32 49 of the 106 patients were clinically assessed by the investigators and did not present with type IV dislocation. It is a limitation of the study that not all patients were clinically assessed by the investigators, which could lead to an underestimation of type IV injuries. However, previous similar studies only reported type IV injuries in 1/105 patients ⁷ and 2/157 patients, ¹⁵ respectively.

Rockwood type V was under-represented in the present study with only four cases (3.8%) compared to the previous studies reporting 20% ¹⁵ and 21%. ⁷ Nordin et al. used bilateral radiographs with both AC joints exposed whereas Chillemi et al. based the classification upon anteroposterior and axillary radiographs of the affected shoulder only. The classification was therefore based on the degree of dislocation of the clavicle to the acromion, with no healthy shoulder as a comparison. When the coracoclavicular difference is not measured, a type III can present as a type II or V, potentially explaining the difference between the present study and the results presented by Chillemi's group. Nordin et al. defined type V as ‘>100% increased CC-difference’, which is similar to our definition, and the reason for the difference in the results cannot be determined.

There was an overrepresentation of males to females at a ratio of 8.6:1, which is in accordance with the ratio of 5:1–8.5:1 that has been described in previous studies.^7,8,10,16

An explanation hereof may be males having a more risk-taking behaviour ³³ and males more often taking part in high speed and high impact activities. The two main activities leading to AC joint dislocations were soccer and bicycling. Soccer is played by 10.1% of Danish men compared to 2.3% of women,^34,35 and road cycling on a regular basis is performed by 11% of Danish men compared to 5% of women. ³⁶

In the present study three-quarters of the injuries occurred during sports with cycling accounting for half of all injuries. These observations support the findings of other European studies describing the mechanism of injury for AC joint dislocations in a European urban population.^7–9,16 However, injuries during cycling were not divided according to the purpose of the ride, and it is expected that a considerable percentage of the cycling accidents occurred in traffic.

The sport causing AC injuries varies with cultural and regional preferences. In the United States, collision sports like rugby and American football are very popular and injuries to the AC joint are reported to account for 29–41% of all injuries to the shoulder among athletes,^{12–14,17,37} whereas cycling and soccer are more popular sports in Europe.^7,16

The bimodal age distribution for men peaking in the twenties and the fifties deviates from previous studies only reporting a peak for men in their twenties.^7,10,16 This might be explained by the interest in bicycling as a recreational sport in the age groups of 50–69 years. ³⁶ One study from Hindle et al. also reported a bimodal age distribution with peaks in the ages of 15–24 and 35–54 years, ⁸ but the mechanism of injury was not reported.

The strength of the present study was a prospective design with a systematic evaluation of all radiographs of the AC joint and clavicle for a population of >500,000 residents. Due to the structure of the healthcare system, there is a very small risk of missing patients as no private alternative exists for acute injuries. Standard radiological projections were used and the inter-rater reliability of the raters was investigated resulting in a low risk of misclassification.

The study is limited by the hospital setup as only patients who attended an emergency department or were referred for follow-up were included. Patients, where no radiograph of the AC joint or the clavicle was taken, will have been missed for inclusion. Also, patients injured outside Copenhagen who did not attend a follow-up in their home hospital will have been missed giving a risk of underestimation of the true incidence.

This study provides detailed information about the epidemiology of AC joint dislocations in an urban population. It is the first prospective study to estimate the incidence of the injury based on a systematical evaluation of radiographs from emergency departments and orthopaedic departments in a specific catchment area. The study is one of the first to describe the distribution of AC joint dislocations according to Rockwood’s classification and clearly demonstrates that the debated type III is the most common type whereas type V injuries might be overreported in previous studies. This finding is of interest for the orthopaedic surgeon as studies indicate that type III dislocations might rehabilitate well without surgery. If so, the present findings suggest that the vast majority of patients can do without surgery.

Conclusions

Rockwood type III was the most common type of AC joint dislocation constituting 55.7% of the injuries. The overall incidence was 19.3 per 100,000 PYRS; 34.9 per 100,000 PYRS for males and 4.0 per 100,000 PYRS for females. Young and middle-aged males were at highest risk. Seventy-five percent of the injuries occurred during sports, most frequently in cycling.