The incidence of medial end clavicle fractures is higher than had previously been considered

Introduction

Fractures of the medial end of the clavicle can be difficult to diagnose on plain X-ray due to the superimposed cervical and thoracic spine. As a result, additional imaging, usually in the form of a CT scan, maybe required to diagnose certain fractures.

Previous X-ray epidemiological studies have found the incidence of medial end clavicle fractures to be between 2 and 3% of all clavicle fractures.^1–4 However, these studies were retrospective and, in certain cases, predated the introduction of digital X-ray systems that allow for data searches, to cross-reference with clinical notes, and for image manipulation and magnification. ⁵

Also, the patient populations, for these studies, were made up from patients that had been referred to their respective fracture clinics, based on an X-ray diagnosis from their Emergency Department (ED). As a result of this, there is a possibility that patients who had sustained a clavicle fracture that was not detected on the ED X-ray (missed or not visible) were not included in these studies.

At our hospital we have set up a system where the digital notes and X-rays of every patient that has been seen in the ED with a musculoskeletal injury, whether a fracture was detected on X-ray or not, are reviewed the next day by an Orthopaedic Consultant Surgeon at a Virtual Fracture Clinic (VFC). Patients with a fracture and patients without a visible fracture, but on whom there is a high index of concern, are then referred to an appropriate Orthopaedic Fracture Clinic.

We present the results of a study of the incidence and epidemiology of all clavicle fractures that presented at our institution over an 18-month period. We hypothesised that the incidence of medial-end clavicle fractures is higher than has previously been considered.

Materials and methods

We undertook a retrospective analysis of prospectively collected data on every patient that was seen in the ED of our hospital, who had sustained an injury to their shoulder girdle over an 18-month period. Only patients over the age of 16 years were included. Institutional review board approval was obtained for this study.

At our institution the electronic notes and digital X-ray images (EPIC, Epic Systems Corporation, Verona, WI) of every patient seen in the ED by a Health Professional (junior doctor or specialist nurse) are reviewed the day after presentation by a Consultant Orthopaedic Surgeon in a VFC. Prior to this all of the X-rays are reviewed by a Consultant Radiologist or a Higher Radiological Trainee, supervised by a Consultant, and the report is available at the VFC.

At the VFC, if a fracture is identified on the presenting digital X-ray, the patient is contacted by a VFC nurse and an appointment made for them to be seen at a Fracture Specific Clinic (Shoulder & Elbow, Hand, Hip & Knee, Foot & Ankle) by an appropriate Orthopaedic Surgeon in an appropriate time frame.

If a fracture is not seen, the patient is contacted by a VFC nurse on the phone, interviewed and given appropriate advice. If the VFC nurse or the patient still have concerns about the injury, an appointment is made for the patient to be seen at an appropriate Fracture Specific Clinic. All shoulder and elbow fractures and injuries are seen in a Shoulder and Elbow Fracture Clinic (SEFC) run by one of three Consultant Shoulder & Elbow Surgeons (GTS, LVR, NK).

Between 1 July 2017 and 31 December 2018 every patient seen in the SEFC, that had sustained a clavicle fracture, was included in the study. The majority of clavicle fracture patients had been identified on the presenting digital X-ray by the Health Professional at the time of their initial ED attendance. Additional clavicle fracture patients were identified on the presenting digital X-ray by the radiology department and by a Consultant Orthopaedic Surgeons at the time of the VFC.

A further group of clavicle fracture patients were identified where a fracture had not been noted on the presenting X-ray in the ED, by the radiology department or the VFC, but their symptoms were sufficiently severe that, when contacted by the VFC nurse, they were referred to the SEFC. Due to their on-going clinical symptoms these patients underwent a CT scan, where their fracture was identified.

Additionally, any patient that had been admitted directly as an emergency to the Trauma Unit whose injury had occurred within the catchment area of our hospital, rather than from the rest of our Major Trauma Centre catchment area, and who had sustained a clavicle fracture was identified. All of these patients had undergone a whole-body trauma CT scan on admission, from which their fracture had been identified.

Patients that had sustained a clavicle fracture associated with a dislocation of either the acromioclavicular joint or the sternoclavicular joint were excluded.

Radiographic classification of the fractures was made according to the system described by Allman. ⁶ This system identifies three groups of fractures (Group 1 – fractures of the middle third; Group 2 – fractures of the lateral third, distal to the conoid tubercle; Group 3 – fractures of the medial third). The severity of the fracture is classified into three subgroups (Subgroup a – undisplaced, Subgroup b – displaced and Subgroup c – comminuted). We classified displaced fractures according to Postacchini et al., ¹ where a fracture was considered to be displaced where the distance between the inferior border of one fragment and that of the corresponding border of the other fragment at the fracture site exceeded 3 mm.

For each patient we recorded their gender, age (by decade) and categorised their mechanism of injury as either a Fall, Sports Injury, Cycling Injury, Road Traffic Accident (RTA) or other.

Results

Over the 18-month study period 558 clavicle fractures were identified in 555 patients. Of these 132 were in females and 423 were in male patients. Three-hundred and sixty patients (64.5%) had an Allman Group 1 Fracture, 139 (24.9%) had an Allman Group 2 Fracture and 59 (10.6%) had an Allman Group 3 Fracture. The severity of the Group 2 (Lateral) and Group 3 (Medial) fractures was similar with approximately 60% Subgroup a, 25% Subgroup b and 15% Subgroup c (Table 1).

OPEN IN VIEWER

Table 1.

Incidence of fractures.

	Total	Subgroup a	Subgroup b	Subgroup c
Grp 1 (Middle)	360 (64.5%)	117 (32.5%)	147 (40.8%)	96 (26.7%)
Grp 2 (Lateral)	139 (24.9%)	81 (58.4%)	36 (25.7%)	22 (15.9%)
Grp 3 (Medial)	59 (10.6%)	36 (61.0%)	14 (23.7%)	9 (15.3%)

The median decade for Group 1 fractures was the third decade with most fractures occurring in the second decade (mode) and gradually decreasing to the ninth decade. The median decade for Group 2 fractures was the fifth decade with the incidence of fractures rising up to the fifth decade and then declining to the ninth decade. Group 3 fractures occurred from the second decade onwards and then remained at a fairly consistent level up to the eighth decade (Table 2).

OPEN IN VIEWER

Table 2.

Incidence of fracture position by age (decades).

		10–19	20–29	30–39	40–49	50–59	60–69	70–79	80–89	91+
Grp 1 (middle)	360	105	93	63	29	30	21	10	9	0
Grp 2 (lateral)	139	7	23	22	31	19	14	12	9	2
Grp 3 (medial)	59	11	6	9	7	11	6	9	0	0

Almost all of the Group 1 and Group 2 fractures (89.2 and 96.7%) were identified on the initial X-ray by a Health Professional in the ED. The other additional fractures were identified either by the radiologist, at the VFC or the Fracture Clinic (10.1 and 2.5%) or occurred in polytrauma patients (0.7 and 0.8%). Only half (52.5%) of the Group 3 fractures were identified in the ED with 22% being identified the next day at the VFC. In 10%, no fracture was identified on the X-ray and a diagnosis was made after a CT scan. Fifteen percent of the Group 3 fractures occurred in polytrauma patients (Table 3).

OPEN IN VIEWER

Table 3.

Source of diagnosis.

	Total	ED	Radiologist	VFC	Clinic (CT)	Polytrauma (CT)
Grp 1 (middle)	360	348 (96.7%)	9 (2.5%)	0	0	3 (0.8%)
Grp 2 (lateral)	139	124 (89.2%)	9 (6.5%)	4 (2.9%)	1 (0.7%)	1 (0.7%)
Grp 3 (medial)	59	31 (52.5%)	9 (15.3%)	4 (6.7%)	6 (10.2%)	9 (15.3%)

The majority of Group 1 and Group 2 fractures resulted from falls (42.4 and 40.3%). However, the majority of Group 3 fractures were the result of RTAs (47.7%) (Table 4).

OPEN IN VIEWER

Table 4.

Mechanism of injury.

	Total	Fall	Sport	Bike	RTA	Other
Grp 1 (middle)	360	145 (40.3%)	77 (21.4%)	54 (15.0%)	60 (16.7%)	24 (6.7%)
Grp 2 (lateral)	139	59 (42.4%)	19 (13.7%)	36 (25.9%)	11 (7.9%)	14 (10.1%)
Grp 3 (medial)	59	9 (14.3%)	11(19.0%)	11 (19.0%)	28 (47.7%)	0

Discussion

The main findings of this study suggest that the incidence of medial end, Allman Group 3, clavicle fractures is higher than has previously been reported. Additionally, identification of the fracture on plain X-ray can be difficult, particularly for less experienced Health Professionals.

Fractures of the medial end of the clavicle have previously been reported as rare. Nordqvist and Petersson ³ found an incidence of 3% of medial end clavicle fractures in a series 2035 clavicle fracture X-rays, sampled from the Diagnostic Radiology Department in Malmo over 35 years. In a series of 1000 clavicle fracture X-rays, from patients seen in the Fracture Clinic, Robinson ² found that 2.8% of the fractures involved the medial end. Similarly, Postacchini et al. ¹ found that medial end clavicle fractures occurred in 2% of clavicle fractures in a series of 535 X-rays of patients treated in their hospital over an 11-year period. All of these studies were based on conventional X-rays.

More recent studies have focused on a subgroup of polytrauma patients. In a study by Throckmorton and Kuhn ⁷ of 614 clavicle fractures, referred to a Level 1 Trauma Centre over a five-year period, they found the incidence of medial end clavicle fractures to be 9.7%. Of the 55 medial end clavicle fractures, 45 (82%) were identified on an initial CT trauma series and 10 (18%) were identified on a plain film. Subsequent plain radiographs were taken on 43 of the CT patients and in 12 (22%) the fracture could not be identified. The authors felt that the higher incidence of medial end fractures that they had reported in comparison to previous studies may be a result of their more liberal use of CT scans. Salipas et al. ⁸ reported on a study of medial end clavicle fractures that had been sustained by patients admitted to their Level 1 Trauma Unit with a minimum Injury Severity Score of 15. Over a 5-year period they identified 440 clavicle fractures, of which 68 (15.4%) were of the medial end.

When devising our VFC, we set up as robust a system as possible so that the X-ray and notes of every patient that presented to our ED with a musculoskeletal injury could be reviewed by a radiologist and Consultant Orthopaedic Surgeon the next day. Using our VFC we have endeavoured to identify and treat every patient who presented to our hospital that had sustained a clavicle fracture over the study period. Our study is retrospective but, unlike the previously mentioned studies, the data were prospectively collected on patients with a clavicle fracture that were seen in our hospital ED rather than just those seen in the Fracture Clinic. Also, we used digital imaging, which has been shown to be more accurate than conventional imaging, with regards to fracture identification. ⁵

Nearly 90% of Allman Group 1 and nearly 100% of Allman Group 2 fractures were identified in the ED by either a junior doctor or specialist nurse. However, only just over 50% of Allman Group 3 fractures were identified in the ED. About 20% were identified in the VFC and 15% were sustained by polytrauma patients. Interestingly, about 10% of the fractures could not be identified on a plain X-ray and were only diagnosed after a CT scan.

Medial end clavicle fractures are notoriously difficult to identify on plain X-rays.^7,9 The results of our study appear to confirm this and may explain why the incidence of 10.6%, that we found, is significantly higher than that previously reported in other plain X-ray studies.^1–4 Whilst we used CT scans to diagnose polytrauma and more clinically challenging patients, this only accounted for 25% of the cases.

We found the demographics of medial end clavicle fractures to differ from other fracture types. About 65% of medial end clavicle fractures occurred as the result of a bicycle or motor vehicle accident, whilst about 65% of middle and lateral end clavicle fractures occurred as the result of a lower energy fall or sports injury. The incidence of medial end clavicle fractures appeared to be the same from the second to the eighth decade, whilst middle third fractures peaked in the second decade and lateral end fractures in the fifth decade and then both declined. Whilst the studies of Throckmorton and Kuhn ⁷ and Salipas et al. ⁸ have highlighted the increased incidence of medial end clavicle fractures in polytrauma patients, the results of this study suggest that the incidence of medial end clavicle fractures is also increased in patients with higher energy injuries in the general population.

There are limitations to this study. The data were reviewed retrospectively, and it is possible that some patients who had sustained a clavicle fracture were still missed. Also, we have described the incidence of clavicle fractures for our Hospital's specific catchment area. It is likely that the demographics, with regards to population age, density and infrastructure (rural, built up, specific industries), are likely to differ between other areas.

Conclusion

The results of this study suggest that the incidence of medial end clavicle fractures in the general population is higher than had previously been considered. They can be difficult to diagnose by less experienced Health Practitioners on plain radiographs and a low threshold for more senior review or additional imaging should be considered.