Point-of-care bedside ultrasound examination for the exclusion of clinically significant ankle and fifth metatarsal bone fractures; a single blinded prospective diagnostic cohort study

This was a single blinded prospective diagnostic cohort study, conducted between August 2015 and December 2017. The goal was to compare PoCUS to the reference standard of radiographic imaging. Approval by the local ethics committee was obtained and patients informed consent documented. The Standards for Reporting Diagnostic Accuracy (STARD) statement checklist was used for reporting [6].

This study was conducted in the ED of the University Medical Centre in Groningen, the Netherlands, a tertiary trauma centre receiving around 900 patients with a foot or ankle trauma per year. Patients presenting to the ED with a foot or ankle trauma were assessed according to standard clinical protocol. The triage nurses were trained to apply and document the OAR (Table 1). If positive, the patients were screened for eligibility [7].

Inclusion criteria were age ≥ 17 years, inversion or eversion trauma mechanism and a positive OAR. Exclusion criteria were negative OAR, open fractures, visible major dislocation, multi-trauma patients, previous fracture of the same ankle, degenerative ankle disease and a trauma ≥48 h, no informed consent, other fractures (calcaneus, tertiary malleolus), no blinding for the specific outcome of the OAR and the trauma mechanism, included after the closing date of the study [1, 5, 8‐10]. Patients of whom the records were not evaluable were ineligible, avulsion fractures were excluded. (Fig. 1).

There was inclusion of consecutive patients, done by all the emergency physicians (EP’s) or residents, as in usual clinical practice. Informed consent and PoCUS were performed by an EP or resident. The sonographer was blinded for the trauma mechanism and the specific outcome of the OAR, by not knowing specifically which rules of the OAR led to a positive finding. The patient would consecutively be assessed and treated, in accordance with current practice, by another ED doctor blinded for PoCUS outcome. Radiography of the ankle included anteroposterior and lateral radiographs of the ankle, and fifth metatarsal bone if appropriate.

All of the sonographers were either EP’s or residents with experience in PoCUS, trained with standardized sonography courses in emergency medicine. In total 23 sonographers were part of this study. These were 11 EP’s, 12 residents, and 1 EP expert sonographer. The expert sonographer was trained within the Ultrasound Leadership Academy Fellowship, a 12-month comprehensive course of PoCUS [11]. All sonographers were additionally trained by the expert in a 2-h theoretical and practical training specifically for PoCUS of the fifth metatarsal bone, distal tibia and fibula.

PoCUS was performed initially on a Zonare Z One ultrasound machine, but after mechanical failure a SonoSite X-porte was used. A 10-MHz linear probe was used for sonography of 3 regions of the ankle:

PoCUS was focussed on cortical disruption, indicating a fracture. PoCUS was not used to identify soft tissue or syndesmotic injuries of the ankle. The navicular bone was not scanned in this study with regards to low diagnostic value in previous research [10]. A 10-s sonography video of all patients was recorded to visualize the area of interest. Criteria for diagnosis of a fracture with PoCUS were cortical disruption or axial deviation of the bone surface as observed by the sonographer. See Fig. 2 for an example of a cortical disruption as seen by PoCUS. Fractures < 3 mm in width were considered non-significant avulsion fractures, as per previous studies [5].

All videos were secondarily assessed by the expert. Any differences between sonographers and expert sonographer were documented. The final evaluation of the radiographic images by the radiologist was considered the golden standard for the diagnosis of a fracture. The radiologist was blinded for the outcome of the PoCUS, but not for clinical findings that were documented. The result of the reference standard was not available to the sonographers or expert.

A sample size was calculated, using criteria based on previous research and desired study characteristics; a prevalence of patients with a fracture of 25%, a minimal acceptable sensitivity of 96% and minimal acceptable specificity of 97%, with a confidence interval of 5% for sensitivity. The calculated sample size for the determined sensitivity was 236 and for the specificity 62 patients [12].

Data analysis was done in SPSS Statistics 23 (IBM, Armonk, New York, USA). All demographics and clinical data were imported. Comparison between categorical variables was calculated by the chi-square test.

The measures of diagnostic accuracy were sensitivity and specificity of PoCUS for detection of fractures of the lateral and medial malleoli and fifth metatarsal bone compared to the golden standard radiographic images. Distal avulsion fractures of the fibula were not considered as a clinically relevant fracture, but mentioned [1].

This study was conducted in an academic tertiary hospital, which may result in bias especially for results in primary or secondary care. There were more ineligible patients than expected, due to missing data, non-assessable videos or because blinding was not guaranteed (thus when patients were included by the expert sonographer). The missing data caused a certain degree of bias. During the study period the ultrasound machine un-expectantly broke down resulting in a temporary pause in inclusion and images that could not be saved correctly or were lost in the process. Total blinding could not always be assured because of visible hematoma and swelling in foot and ankle trauma, however, this reflects common practice. In young patients under 20 years, the epiphyseal plate could still be present, in which case the hypoechoic appearance of the epiphyseal cartilage might appear as a cortical disruption when PoCUS is being performed by less experienced sonographers. However, there is a small probability of a still existing epiphyseal plate in the number of patients ≥17 years and under 20 years in this study.