Clavicle shortening is considered one of the main parameters to make a surgical referral, however, there is no consensus on the exact value of this measure. Lazarides et al. [25] considered bad prognosis for the conservative treatment fractures with shortenings greater than 1,4cm in women and 1,8cm in men. Hill et al. [9], consider shortening values greater than 2,0 cm to indicate surgery and they don’t mention gender differences in their study. Postattine et al. [32] consider shortening as the fracture overlap percentage, being referred to surgery those who present a percentage of overlap greater than 13–15% of the total length. And De Giorgis et al. [33] parameter is a shortening percentage of 9,7% in comparison to the integral clavicle length.
Regarding the variation of the clavicle length among individuals, Daruwalla et al. [20] showed, through integral clavicles tomographic measurement, that there is a variation of 129.4 mm to 161.2 mm in that bone’s size, and King et al. [21] found a length variation of 121.5 to 183.3 mm. In our study, we observed a variation from 132mm to 169mm and, therefore, using an absolute but not relative percentage shortening value may cause different effects according to the clavicle size.
As well as Lazarides et al. [25], we measured the clavicle shortening from the contra lateral clavicle length measure, as a parameter of normal length. Some authors measure the shortening from the edges of the fractured fragments in a straight-line projection [16–18, 30]. That methodology may not be precise due to the “S” shape [11, 15] characteristic of that bone and to the comminution of many clavicle fractures, associated with images overlap of a simple radiography [23].
There is no consensus over X-ray tilt. We found, in our literature search, a wide angular variation in the incidence of these rays. Studies show exams performed at the following angles: 0º, 10º cephalic, 15º caudal, 20º caudal, 20º cephalic, 30º cephalic, 45º cephalic and 45º caudal [1–31], in neither of which there is an explanation for this great angular variation. In our study, we standardized inclinations to 0º (due to its vast use by most authors) and 10º cephalic (for being the incidence that better evaluates the limits of the acromioclavicular articulation for a precise assessment of that bone’s lateral boundary).
Currently, CT exam is considered the gold standard method [16, 18, 20–25]. It allows a detailed assessment of the bone fragments, as well as a three-dimensional assessment through one simple exam. Omid et al. [24] show the superiority of a CT over a simple radiography in the evaluation of the fractured clavicle shortening. However, one of the main questions regarding the comparative radiography and tomography studies is in relation to the positioning of the patient. When lying down, the gravitational force vector relative to the patient is different from a standing position, and that could affect the measurement of the fractured clavicle length. Onizuka et al. [27] made a similar assessment of the gravitational impact, comparing only AP X-rays of the fractured clavicle with a 15° tilt performed with a patient standing and then lying. The authors obtained a significant difference in the angular measurement and vertical deviation, however, as we did in our study, they observed no change in length and shortening.
In this study, we searched for the radiographic incidence that presents that presents the most reliable length measurements of the fractured clavicle in comparison with the measurements obtained in CT. Lying PA incidence was the one that showed values closest to the tomographic exam. Other studies make similar assessments [16, 18, 24], however, none of the studies in our survey included the AP and PA incidences in patients standing and then lying. Different from the results obtained by Smekel et al. [16], our study did not show a significant statistical difference between AP and PA incidences. Despite the standing PA exam being the most reliable CT to the clavicle length measurements, it was also the one in which patients complained of much pain during the examination. As there is no statistically significant difference between the exams, we do not recommend this incidence performance in order to avoid unnecessary discomfort to the patients.
We also did not assess the relation between fracture time and shortening pattern. In our evaluation, we included patients whose fractures had progressed less than 3 weeks, and according to Onizuka et al. [27], it is in this interval most deviations occur. After 21 days, the presence of fibrous scar tissue stabilizes the fracture and prevents the displacement of fragments even when the patient changes position.
The limitations of our study were: limited number of assessments, heterogeneous distribution between male and female individuals, we did not consider spinal and postural disorders, nor the biotype of individuals that can affect the clavicle positioning during a radiography performance. Another limitation was the execution of the exams by different radiography technicians. Despite the radiographic technique being standardized in relation to the patient positioning during examination, the execution of a good quality exam depends on experienced technicians. In our study, we had to repeat the radiographic exam in three patients because of unsatisfactory exams that did not include both clavicles in their entirety.
In case there was a significant statistical difference in our study, we would be able to standardize a single radiographic method to the assessment of the fractured clavicle length and facilitate conduct taking.