Address correspondence to Dr Christopher Partyka, MBBS, Emergency Department, Royal North Shore Hospital, Reserve Rd, St Leonards, NSW 2065, Australia. Address e-mail to [email protected].
Rib fractures typically result in severe pain that is notoriously difficult to manage, and early effective analgesia remains a crucial management priority. Recently, the SABRE (Serratus Anterior Plane Blocks for Rib Fractures in the Emergency Department) trial reported that the addition of a serratus anterior plane block (SAPB) to standard rib-fracture care improved analgesic outcomes. Cadaveric studies and case reports suggest that posteriorly located fractures may benefit from this technique. The true optimal anatomical location of injuries that may benefit from the SAPB is yet to be elucidated.
Spatial point pattern analysis examines the relationship between points in space (eg, randomly scattered or distributed based on associations with their attributed values of clinical interest, called marks). This approach has been used in the medical field for a variety of clinical explorations. We performed a post hoc analysis of the SABRE trial using spatial point pattern analysis to explore whether the anatomical location of rib fractures was associated with analgesic benefit after a SAPB.
METHODS
Ethics was approved by the South Western Sydney Local Health District Human Research Ethics Committee (2020/ETH02516). The SABRE trial was a multicenter pragmatic trial undertaken in 8 Australian emergency departments. The methods and results have been published. Patients in this cohort study were included from the SAPB arm of the trial if they received their SAPB, underwent computed tomography (CT) imaging of the thorax, and had primary outcome data (a composite pain score requiring a pain score reduction of 2 or more points and an absolute pain score of less than 4 out of 10 measured 4 hours post enrollment).
Anatomical locations of rib fractures were derived from thoracic CT images and were measured using a standardized technique (Supplemental Digital Content 1, Supplemental File, https://links.lww.com/AA/F4). The rib number and the measured angle of each fracture location provided an "x" and "y" location by which they could be plotted on an unfurled thoracic cage along with their coinciding primary outcome result. This created the window required to complete the spatial analysis.
To examine the spatial distribution of the primary outcome, an exploratory dot-plot was produced with rib fractures as single points and color used to delineate the primary outcome (called a mark). This facilitates visual inspection of point (and mark) distribution, identification of outliers, and potential clustering, although in areas of greater fracture density, this can be difficult to distinguish (overplotting). A plot of the kernel-smoothed log relative risk (RR) surface was created to display the relative density of marks with primary outcome met to those not met. To control for potential bias in areas with few fractures and 1 outcome that could give the appearance of a difference in RR, an adaptive bandwidth was used. A significant difference in surface log RR (P < .05) is indicated with line on the plot where it occurs. If no areas of significance were found, P values of.2 and.1 were plotted to check the function was working and indicate areas of potential benefit that had not reached significance. The log RR scale on the figure was transformed to an RR using the antilog of the original values for interpretability.
Statistical analyses were conducted using R Environment for Statistical Computing (v4.2.1) and are detailed in the Supplemental Digital Content 1, Supplemental File, https://links.lww.com/AA/F4.
RESULTS
Of the 105 participants in the SAPB arm of the SABRE trial, 78 (with a combined total of 356 individual rib fractures) met the inclusion criteria (Supplemental Digital Content 1, Supplemental Figure 1, https://links.lww.com/AA/F4). Patient characteristics of these patients are shown in the Table, grouped by those who did and did not meet the SABRE trial primary outcome.
Figure A presents an unfurled thoracic cage with each individual rib fracture represented by a single point colored by whether the corresponding patient reached the SABRE primary outcome threshold (red) or not (green). A density plot is overlaid and demonstrates the surface RR of not meeting the primary outcome. This plot did not show any anatomical regions of statistical significance impacting on the primary outcome measure. One region between ribs 5 to 8 on the right side, approximately 150° from the sternal center (posterolateral segment) correlated with a P value of.10 (dashed line), however this was not mirrored on the left hemithorax. To maximize the overall number of points for the primary outcome, the data in Figure A was amalgamated into 1 hemithorax (Figure B). Here, there is 1 region identified between the third and fourth ribs between 60° and 80° from the sternal center where the P value was 0.05.
Exploratory dot and smoothed density plots were created to examine each primary outcome constituent as continuous variables and are presented in Supplemental Digital Content 1, Supplemental Figure 2, https://links.lww.com/AA/F4.
DISCUSSION
Our spatial analysis did not show an association between the anatomical location of rib fractures and reduction in pain scores as measured in the SABRE trial. These real-world data suggest that any benefit from the SAPB could be accrued regardless of the anatomical location of the injury. Specifically, there was no evidence of a difference in RR present in the posterior rib segments which challenges the existing teachings that SAPB are not effective for fractures in this area.
When all fractures are superimposed onto a single representative hemithorax an area of statistical significance is identified. The higher density of fractures not receiving analgesic benefit from their SAPB in this area is likely due to the higher number of patients represented in this group.
Our study has limitations. This is a post hoc analysis of a larger clinical trial and was not powered specifically for spatial point pattern analysis. We also assume that rib fractures are the primary pain generator rather than overlying soft tissue injury. The findings remain exploratory and should prompt further prospective research. Each point in this analysis represents a single rib fracture, not an individual patient of which there was only 78. Although we analyze 356 rib fractures, it may be more prudent to have a larger sample to assess patient-centered outcomes.
In conclusion, the SAPB could be considered an analgesic adjunct for rib fractures regardless of the anatomical location of the injury, especially in the absence of an available alternate regional anesthesia technique.
ACKNOWLEDGMENTS
We thank the original investigators and clinicians who participated in the SABRE trial as well as the trial participants, relatives, and families for their contribution to this work.