SGEM#332: Think Outside the Cardiac Box

Date: May 20th, 2021 Guest Skeptic: Dr. Robert Edmonds is an emergency medicine physician in the Air Force in Dayton, Ohio, and a University of Missouri-Kansas City residency alumni from 2016. Reference: Jhunjhunwala et al. Reassessing the cardiac box: A comprehensive evaluation of the relationship between thoracic gunshot wounds and cardiac injury. Journal of Trauma and Acute Care Surgery. September 2017 DISCLAIMER: THE VIEWS AND OPINIONS OF THIS PODCAST DO NOT REPRESENT THE UNITED STATES GOVERNMENT OR THE US AIR FORCE. This SGEM episode was recored live for the Truman Medical Centers Multidisciplinary Trauma Conference. We did the session over zoom as an SGEM Journal Club. If you would like a copy of the slides from the presentation you can download them free open access at this LINK. Case: You receive a call on the Biocom for an incoming Type A trauma, three minutes out.  The patient is an adult male with a gunshot wound to the chest, and they’re combative with emergency medical services (EMS). Upon arrival in the emergency department (ED), the patient is incoherently speaking, has a pulse of 135 beats per minute, blood pressure of 85/50 mm Hg, and an obvious open wound in their left mid-axillary line at the level of the nipple. Background: Penetrating trauma is a major disease burden in the United States, and gunshot wounds cause 30,000 deaths annually [1] . As a country, penetrating trauma accounts for about 10% of all trauma cases [2] , but at some trauma centers it can reach much higher numbers. Here at the Truman Medical Center the average penetrating trauma for gunshot wounds alone represents ~19% of all traumas.  Naturally, patients with a direct cardiac injury from a gunshot wound (GSW) require prompt identification and management, so tools have sprung into existence to attempt to risk stratify patients at a higher risk of an underlying cardiac injury. One of the more common tools is the “cardiac box”. This three-dimensional area is at the highest risk of cardiac injury. The anatomical area is defined anteriorly as between the clavicle and xyphoid, and between the bilateral midclavicular lines. Per the authors, “The dogma of the cardiac box is largely based on small studies with primarily stab wounds. The underlying issue is that stab wounds are low kinetic energy and result from instruments with a fixed length. Thus, most stab wounds usually only result in a cardiac injury if the entrance is in very close proximity to the heart or there is a long weapon. Because these studies did include gunshots, the concept of the “box” was ultimately uniformly applied to all mechanisms. Injuries from high kinetic energy projectiles, however, can cause cardiac injury from entrance wounds to any area of the torso, especially the thorax.” Although it may be obvious to some that injury outside the cardiac box doesn’t rule out injury to the heart, the existence of such a tool colors our language and shifts the perceived risk in the clinician’s head.  According to a recent study in the Journal of Surgical Research [3] , 44% of all penetrating thoracic trauma patients presented to a non-trauma center (not a level 1 or level 2 ACS defined trauma center). For clinicians in these settings, use of the “cardiac box” nomenclature can have a significant impact on the perceived injuries when communicating with an on-call surgeon or when transferring the patient to another facility. If the injury is outside the cardiac box, it can be perceived as less concerning and may give the treating team a false sense of security. Clinical Question: Are the anatomic borders of the cardiac box adequate to predict cardiac injury from gunshot wounds? Reference: Jhunjhunwala et al. Reassessing the cardiac box: A comprehensive evaluation of the relationship between thoracic gunshot wounds and cardiac injury. Journal of Trauma and Acute Care Surgery. September 2017 Population: All patients in the Fulton County (Georgia) Medical Examiner’s autopsy registry who sustained a penetrating torso gunshot wound from January 2011 to December 2013. Excluded: Injuries above the clavicles and below the xyphoid Exposure: Wounds that were in the cardiac box- “defined as the two-dimensional plane covering the anterior surface of the thorax from the level of the clavicle to the tip of the xiphoid… and between the midclavicular lines (laterally).” The authors also included this same region projected onto the posterior thorax as well. Comparison: Patients with wounds outside the “cardiac box” Outcome: Primary Outcome: Cardiac injury Safety Outcome: None (this data all came from autopsy reports) Authors’ Conclusions: “For GSWs, the current cardiac box is inadequate to discriminate whether a gunshot wound will cause a cardiac injury. As expected, entrance wounds nearest to the heart are the most likely to result in cardiac injury, but, from a clinical standpoint, it is best to think outside the “box” for GSWs to the thorax.” Quality Checklist for Observational Study: Did the study address a clearly focused issue? Yes Did the authors use an appropriate method to answer their question? Yes/No Was the cohort recruited in an acceptable way? Yes Was the exposure accurately measured to minimize bias? Yes Was the outcome accurately measured to minimize bias? Yes Have the authors identified all-important confounding factors? Yes Was the follow up of subjects complete enough? Yes How precise are the results? Very Do you believe the results? Yes Can the results be applied to the local population? Unsure Do the results of this study fit with other available evidence? Yes Results: The authors identified 263 patients with 735 penetrating torso injuries over three years. They reported 620 (84%) of the injuries were from gunshot wounds (GSW) with a mean of 2.6 GSW/patient. After exclusions, 320 GSWs were included in the study. The mean age was 34% and 87% were male. Half of the patients were pronounced dead at the scene and 46% were dead on arrival. Key Results: The relative risk of having a cardiac injury from a GSW in the cardiac box is low and not statistically different from those outside the cardiac box. Primary Outcome: Cardiac Injury 80/257 (31%) in the box and (14/67) 21% outside the box RR = 0.96; p=0.82 We had a hard time confirming their numbers when we tried to calculate the RR. Also their percentage of GSW was 80% in the abstract and 84% in the result section. These discrepancies and issues made us less confident in the study. 1. Autopsy: The authors touch on the fact that their study was an autopsy study. This would exclude all the patients who suffered survivable trauma. In their patient demographics they list the frequency of dead at scene as 50% and dead-on arrival as 46% of all patients. This has huge implications as the patient population is massively different than the patients who survive long enough to enter the care of the hospital team. Although limiting the study to autopsy did provide opportunity to definitively examine for cardiac injury and to precisely evaluate for the location of wounds, the exclusion of patients with survivable injuries presents a challenge to the generalizability of the data.  It is unknown whether the sort of patient who survives long enough to receive care may have an injury pattern more consistent with the traditional cardiac box, and perhaps the tool performs better.  The authors mention that in their “experience with high volumes of penetrating cardiac injuries, these findings parallel clinical practice and experience,” but this is opinion rather than evidence. 2. Death Bias: This will expand on point #1. We often see survival bias but in this case the opposite is true. The bias is towards those who died. It is a form of selection bias. We do not know the injury pattern (inside or outside the cardiac box) for those patients who had GSWs, did not present to the hospital or presented to hospital and survived. 3. Power: The study doesn’t mention how they determined how many patients to enroll in the study. To adequately power a study to detect a difference between two groups, generally a calculation is performed beforehand (a priori) based on the estimated difference between the exposure group and the control group. The authors don’t comment on a power calculation and don’t mention their pre-trial expectation of a difference in cardiac injury based on anatomic location. Performing a power calculation a priori is based on two variables. One is setting the delta (difference) you assume there is between the two groups. In this case it would be cardiac injuries that are from penetrating injuries inside the cardiac box to those outside the cardiac box. The second is the confidence interval that is traditionally set at 95%. This can be one tailed (5%) or two-tailed (2.5%). Doing a power calculation after data has been collected would be wrong and potentially misleading. It is called observed power or post-hoc power. All it does is restate the p-value. Since observed power is a direct function of the p value it does not provide any additional useful information. Unfortunately, you will still see some journals request a post-hoc power calculation mistakenly thinking it will help them distinguish between “true negatives” from “false negatives”. In other word, to differentiate between the lack of observed effect is due to an ineffective treatment or the study shows no effect because it is too small. If you understand how that is not the case you can see how doing these post-hoc power calculations can be misleading, misinterpreted, and ultimately not helpful [4, 5]. 4. 2D vs. 3D: The authors used a two-dimensional definition of the cardiac box looking at the anterior and posterior surface of the thorax. However,