Authors: Neil Krulewitz, Miles Lamberson, Ryan Walsh, Zachary Clark, Skyler Lentz, Lindsay Reardon
Categories: Original Article, Echocardiography, Transthoracic echocardiography, Resuscitation, CPR, Cardiac arrest, Emergency medicine
Source: Resuscitation Plus
Authors: Neil Krulewitz, Miles Lamberson, Ryan Walsh, Zachary Clark, Skyler Lentz, Lindsay Reardon
Standard CPR guidelines recommend chest compressions over the lower half of the sternum; however, this often results in compressions over the left ventricular outflow tract (LVOT) or proximal aorta, impeding blood flow. Improved outcomes have been noted when compressions are directed over the left ventricle.
We evaluated whether transthoracic echocardiography (TTE) can accurately identify the mid-left ventricle (mid-LV) and whether this approach aligns more closely with the true mid-LV than the standard American Heart Association (AHA) compression location based on computed tomography (CT) of the chest.
In this prospective observational study of adults undergoing chest CT, providers marked the AHA-recommended compression site and performed limited TTE to localize and mark the mid-LV. Radiopaque markers were placed at these locations and compared to CT-identified true mid-LV positions.
Among 65 patients, the mean distance from the AHA location to the true mid-LV based on chest CT was 74.2 mm. The distance from the ultrasound guided mid-LV marker to the true mid-LV was 64.6 mm. Ultrasound trained providers outperformed non-ultrasound trained providers in accuracy of localization. We found the AHA position to be cranial and medial to the true mid-LV. On CT, the most common structure beneath the AHA marker was the proximal ascending aorta (38.5%).
TTE-guided localization of the mid-LV is feasible and more accurate than the standard AHA landmark, particularly when performed by trained providers. A TTE-guided approach to mid-LV localization may optimize location of chest compression over the mid-LV, warranting further evaluation in resuscitation settings.
Closed-chest cardiopulmonary resuscitation (CPR) was first introduced in 1960 for patients suffering from cardiac arrest.^1^ The main objective of CPR is to circulate blood and oxygen to the brain, coronary arteries, and other vital organs in an attempt to restart intrinsic cardiac activity while preventing end-organ ischemia. Unfortunately, the prognosis for patients in cardiac arrest remains less than 10% survive to hospital discharge with good neurologic function.2, 3 Two theories have been proposed for the mechanism by which external chest compression provides antegrade blood flow. The cardiac pump theory postulates that direct compression of the ventricles creates greater pressure in the left ventricle than in the aorta and the right ventricle than the pulmonary artery, respectively, resulting in forward flow through the systemic and pulmonary vasculature.^4^ The thoracic pump theory posits that external thoracic compression increases intrathoracic pressure, thereby creating an arteriovenous pressure gradient which promotes blood flow through the heart, acting as a conduit, from the thoracic to systemic circulation.^4^ Recent studies have suggested a more likely role of direct ventricular compression in opening cardiac valves and directing forward flow of blood, as well as raised concern regarding optimal location of external compressions.4, 5, 6
The American Heart Association (AHA) guidelines for CPR recommend that a rescuer direct compressions at the lower half to lower third of the sternum in the middle of the chest between the nipples,^7^ while the European Resuscitation Council also recommends the lower half of the sternum in the center of the chest.^8^ Recent literature has suggested that this anatomic location may not be optimal for compression of the ventricles and generation of forward flow, and may even be detrimental as compression of the left ventricular outflow tract (LVOT) and ascending aorta occur frequently at this location, which may impede forward flow of blood.4, 5, 6 There is also a wide variation in anatomic position of these structures beneath the sternum, making a single location unlikely to be optimal in all patients.9, 10 A 2025 International Liaison Committee on Resuscitation report acknowledges that optimal hand position over the ventricles likely varies based on anatomy, however did not consider echocardiographic imaging in its recommendation.^8^
Utilization of transesophageal echocardiography during active CPR has bolstered the cardiac pump theory, and has demonstrated that when compressions are directed lower than the recommended midsternal location, more directly over the LV, increased LV compression and increased stroke volumes result.^6^ Further studies in swine have shown improved coronary perfusion pressure and increased rate of return of spontaneous circulation (ROSC) when compressions are directed over the LV rather than the LVOT.^11^ More recent literature has demonstrated that aortic valve compression during CPR as visualized by transesophageal echocardiography resulted in lower rates of ROSC compared to patients without aortic valve compression during CPR.^12^ Additionally, a recent study showed that compressions left of the sternum resulted in improved blood pressure during CPR when compared to compressions over the lower half of sternum.^13^
Widespread utilization of transesophageal echocardiography during CPR is not feasible in many institutions at this time. Our study therefore examines the feasibility of utilizing transthoracic echocardiography (TTE) to identify the mid-left ventricle (mid-LV). We hypothesize that TTE can accurately identify the area of mid-LV as confirmed by computed tomography (CT) scan of the chest. Secondly, we hypothesize that the AHA recommended location for compressions would identify an optimal location over the mid-LV less frequently than TTE localization. We also describe the anatomic structures most frequently located underlying this location based on CT.
Our single center prospective observational study took place at an academic and tertiary care Emergency Department (ED) in an urban setting. The study protocol was reviewed and approved by the University’s IRB (STUDY00001948) as an expedited review utilizing a convenience sample. Patients were enrolled from July 2022 until June 2025. Patients 18 years of age or older in whom a CT chest was ordered as part of their clinical workup were eligible to be included in the study. Patients who were pregnant, incarcerated, or unable to provide informed consent themselves were excluded. Trained ED Research Coordinators screened patients with a pending Chest CT for eligibility. After confirming eligibility with an ED provider, written informed consent was obtained at the bedside prior to any study activities beginning.
The patient was positioned supine with their arms at their sides, mimicking chest CT and CPR positioning. An emergency medicine provider (see Table 1 below for stratification of provider training) identified and placed a 2.23 mm radiopaque marker (Suremark CT-23) on the skin surface where chest compressions would be performed based on AHA guidelines. Next, the provider performed a limited TTE, including parasternal long axis (Fig. 1A) and parasternal short axis views (Fig. 1B). The mid-LV was identified by Point of Care Ultrasound (POCUS) as being approximately the mid-point between the mitral annulus and the apex on parasternal long axis, and at the location of the papillary muscles on the parasternal short axis views. Effort was made during TTE to maintain angle of the probe as close to perpendicular to the horizontal as possible, as angle departure on ultrasound will result in the underlying structure not being directly deep to the probe. Once the mid-LV was identified by the provider, the provider marked the location with a skin marking pen. The site was then cleaned of ultrasound gel, and a 4.0 mm radiopaque marker (Spee-D-Mark SDM-BB40) was placed. The skin distance between the radiopaque markers was measured using a flexible tape measure.Table 1Participant exclusion and operator training level.FrequencyReason participant excluded (n = 14)Informed consent not completed prior to computed tomography scan6 (42%)Ultrasound not completed prior to computed tomography scan4 (28%)Computed tomography scan canceled after enrollment1 (8%)Radiopaque markers missing on computed tomography CT scan3 (22%)*Total excluded14 **Ultrasound operator training Level (n = 65)Ultrasound trained attending physician18 (28%)Non-ultrasound trained attending physician37 (57%)Resident physician9 (14%)Physician assistant1 (2%)*Total65Fig. 1(A) TTE Parasternal Long Axis View centered on mid-LV between mitral annulus and apex. LV = Left Ventricle; RV = Right Ventricle; LA = Left Atrium; Ao = Aortic outflow tract. (B) TTE Parasternal Short Axis view centered on papillary muscles. LV = Left Ventricle; RV = Right Ventricle. (C) CT chest axial cut with arrow indicating AHA radiopaque marker. (D) CT chest axial cut with arrow indicating US-guided mid-LV radiopaque maker.Abbreviations: CT – computed tomography; mid-LV – mid-left ventricle; TTE – transthoracic echocardiogram; US – ultrasound.
Chest CTs interpreted by an attending cardiothoracic radiologist with 15 years of experience were used as the reference standard for localization of the mid-LV. Both enhanced and unenhanced, gated and non-gated, full-coverage CT chest protocols (thoracic inlet through diaphragm) were included in the study and were performed on Philips iCT 256- and 128-slice scanners (Philips Healthcare, Amsterdam, the Netherlands). Multiplanar reconstruction (MPR) was employed to standardize LV orientation (Visage Imaging, Melbourne, Australia). A long-axis view was first established by positioning a reference line through the expected location of the mitral valve (atrioventricular sulcus) and extending it to the LV apex. Short-axis images were then generated by placing a perpendicular reference line through the mitral valve plane to the LV apex. The basal level was defined as the most superior short-axis slice immediately below the atrioventricular sulcus, and the apical level was defined as the most inferior slice demonstrating myocardial contour adjacent to epicardial fat on the left. The mid-LV level was defined as the point equidistant between the mitral valve plane and the LV apex. Once identified using MPR, this mid-LV point was localized on true axial images. Using the localizer, the sagittal plane of this point on the skin (x-axis) was recorded, along with the corresponding slice plane (z-axis). Due to differences in protocol reconstruction, slice thickness and spacing for both axial and sagittal reconstructions were recorded. Both radiopaque markers were subsequently identified on the axial CT images, and the slice and sagittal planes were recorded (Fig. 1C, D). For the AHA radiopaque marker, the underlying structure in the craniocaudal (z-axis) dimension relative to the left ventricle (LVOT, basal, mid, apical), aortic valve, or aorta (root, proximal ascending, mid ascending, distal ascending, arch) was recorded.
Data were collected using an electronic data capture tool, REDCap. CT results were later extracted by an attending cardiothoracic radiologist. Calculated fields in the electronic data capture calculated the distance between all points using the formulas shown in Fig. 2. Fig. 2 depicts the relationship of points, and how the Pythagorean theorem was applied to calculate Euclidean distance between AHA radiopaque marker, US-guided radiopaque marker and true mid-LV.Fig. 2Locations and measurements.Illustration shows relationship between AHA radiopaque marker and true mid-LV. The same calculations were repeated for US-guided radiopaque marker and true mid-LV.Abbreviations: AHA – American Heart Association; ROM – radiopaque marker; mid-LV – mid-left ventricle; US – ultrasound.
Data were analyzed in StataSE 18 for Mac (StataCorp, College Station, TX). A paired t-test was used to compare distances between the AHA radiopaque marker location and the true mid-LV and US-guided mid-LV radiopaque marker and the true mid-LV with 95% confidence intervals (CIs) and p-values reported. Visualizations completed in RStudio for Mac (Posit, Boston, MA) with the tidyverse package.
A total of 79 patients were enrolled, with 65 valid cases included in the final analysis (Table 1). Patients were only excluded in the event that informed consent could not be obtained, CT imaging was not completed, or radiopaque markers were accidentally removed prior to CT imaging (see Table 1 for details). The providers that performed the localization of mid-LV by TTE and placement of radiopaque markers are in Table 1. Average age of study subjects was 63 years old. Among the 65 paired observations, mean distance from AHA sternal radiopaque marker placed on the skin surface to true mid-LV based on chest CT was 74.2 mm. The mean distance from US-guided localization of mid-LV radiopaque marker placed on the skin surface to true mid-LV based on chest CT was 64.6 mm. The difference between these two measurements was 9.5 mm (95% CI 2.2–16.9, p = 0.012). Fig. 3 visually displays the distribution of distance among the cohort.Fig. 3Distances between chest radiopaque marker and true mid-LV as identified on chest CT using sternal AHA position and ultrasound-guided localization in the Euclidean and cranial-caudal planes. The distribution of data is shown as a violin plot, with the median shown by the thick horizontal line, interquartile range thin horizontal lines and whiskers represent the largest and smallest values within 1.5 times the interquartile range.Abbreviations: AHA – American Heart Association; CT – computed tomography; mid-LV – mid-left ventricle; US – ultrasound.
Stratifying by US operator training, US fellowship trained emergency physicians performed 18/65 TTEs. Among this group, the mean distance from US-guided localization of the mid-LV radiopaque marker placed on the skin surface to true mid-LV distance was 46.7 mm. Non-US trained providers (non-US trained physicians, resident physicians and Physician Assistants) performed 47/65 mid-LV localizations by TTE and had a mean distance of 71.5 mm – a difference in means between provider groups of 24.8 mm (95% CI 7.7–41.8, p = 0.005).
We compared cranio-caudal and medial–lateral distances between the AHA sternal radiopaque marker placed on the skin surface and true mid-LV based on chest CT, as well as US-guided localization of mid-LV radiopaque marker placed on the skin surface to true mid-LV based on chest CT. The true mid-LV was 50.7 mm lateral and 48.0 mm caudal to the AHA guidelines radiopaque marker. The true mid-LV was 57.8 mm lateral and 7.9 mm caudal to the US-guided radiopaque marker. Fig. 3 visually displays the distribution of distances. The mean distance as measured on the skin surface between the AHA and the US-guided radiopaque markers was 58.5 mm. The most common structure found under the AHA location on CT was the proximal ascending aorta (38.5%, 25/65) (Table 2).Table 2Structures underlying AHA location based on computed tomography chest scan.**Structure***Number of studies (%)Left ventricle9 (14%)Left ventricular outflow tract1 (2%)Aortic valve8 (12%)Aortic root7 (11%)Proximal ascending aorta25 (38%)Middle ascending aorta6 (9%)Distal ascending aorta5 (8%)Aortic arch4 (6%)*Total65
In this study, we evaluated the accuracy of TTE in identifying the mid-LV as compared to the position recommended by the AHA. Our results demonstrated closer placement to the mid-LV when using TTE as compared with the AHA location, with even greater accuracy achieved by ultrasound fellowship-trained emergency physicians.
Traditional CPR as recommended by the AHA guidelines emphasizes compressions over the lower half to lower third of the sternum. However, our study validates findings in previous literature, demonstrating that in a high percentage of patients, the structures underlying the AHA guidelines area are most often the proximal ascending aorta, aortic valve, and LVOT. In our study, the LV underlied the AHA location in only 14% of patients. A growing body of evidence suggests that compressions over the LVOT, aortic valve, or aortic root result in worse outcomes and lower rates of ROSC.5, 6, 12 Recent literature has also shown significant improvement in ROSC when the aortic valve and LVOT are not compressed.6, 12 Our findings therefore suggest that TTE may help clinicians identify a more favorable location for chest compressions.
Our study found that, on average, the mid-LV was more lateral and caudal than the AHA guidelines location, which may corroborate recent literature which has suggested improved hemodynamics with compressions left of sternum.^13^ In practical terms – the CT-identified true mid-LV was 7.4 cm caudal and lateral to the AHA guidelines location for chest compression, while ultrasound-guided localization of the mid-LV was considerably closer, particularly in the hands of POCUS trained emergency providers. Interestingly, the mean distance from the US-guided mid-LV marker to the true mid-LV based on chest CT was still 6.5 cm amongst the entire cohort. The mean distance as measured on the skin surface between the AHA marker and the US-guided mid-LV marker was 5.8 cm. This suggests that probe angulation likely resulted in some of the difference between the location of the US-guided mid-LV marker and the true mid-LV.
Training for identification of the mid-LV was minimal, which allows for generalizability of our results, but likely at the cost of accuracy. Emergency medicine providers without advanced US training had skills in basic echocardiographic image acquisition. They performed the methodology in the study with only instructions received in an email detailing how to obtain the views, and place the markers. Notably, our study found that providers with prior advanced US training were able to locate the mid-LV with significantly better accuracy, which highlights the value of focused cardiac POCUS education. We suspect that more protocolized education and practice may improve accuracy of mid-LV identification. As POCUS expertise becomes increasingly common in emergency medicine and critical care, integrating training in rapid mid-LV identification into standard curricula may enhance CPR quality, particularly in settings where transesophageal echocardiography is not available.
It is not well understood if moving the location of external chest compressions changes which cardiac structures are compressed. It is unclear if compressions applied directly above a location translate to compression of the directly underlying structure, or if instead the sternum and ribs work as a hinge, distributing forces. While TTE significantly outperformed the AHA guidelines location in the cranial-caudal plane (7.9 mm vs 48.0 mm distance to true mid-LV), it did not perform as well in the medial–lateral axis. The true mid-LV was significantly more lateral (57.8 mm) than the US-guided marker. We suspect that this is because most providers are accustomed to obtaining parasternal images very close to the sternum, which may require further education to change this practice. Automatic external compression devices are often difficult to maneuver on the chest and cannot easily be moved laterally, therefore movement of compressions may be limited in practice within a cranio-caudal plane. While we anticipate that compressions will be most often moved in a caudal direction, we do not anticipate a significant increase in abdominal injury, although it is possible that moving off midline will increase the risk of rib injuries. More study is required to determine this risk. We suspect that the potential benefit previously demonstrated by transesophageal echocardiography of compressions directed over the mid-LV is likely to outweigh these risks in the setting of cardiac arrest.
Our study results have particular relevance in the prehospital setting. The increasing availability of portable ultrasound devices may enable prehospital providers to use TTE to guide compression location in the field. Implementing basic cardiac ultrasound protocols in paramedic education could support more tailored resuscitation even before hospital arrival. More studies are needed to determine if the addition of US-guided mid-LV identification would be beneficial in a prehospital setting.
While our prospective study demonstrates the potential of TTE to more accurately locate the mid-LV, several limitations warrant discussion. While operators took steps to minimize ultrasound probe angulation and variability, factors such as body habitus, image quality, and user experience may still impact accuracy. Difficulty in obtaining adequate cardiac visualization secondary to body habitus may result in angulation of the probe to obtain better views, which may impact accuracy.
Furthermore, although TTE was superior in targeting the mid-LV compared to the AHA guideline, our population of patients did not have cardiac arrest, and thus we are unable to assess whether this would translate into improved hemodynamic performance or clinical outcomes. These questions remain for future research. We do not foresee significant changes in structural heart location during arrest, and images can be obtained in the pre-arrest setting or during arrest.
Additionally, we did not record time to image acquisition, so the impact of the time taken to localize the mid-LV by TTE on the resuscitation is unknown. While TTE has been shown previously to lengthen pulse check duration, conceivably a provider could identify the mid-LV in the pre- or peri-arrest period and mark the site for compressions, thereby eliminating the need for increased pause duration in compressions.^14^ This may be less feasible for prehospital providers arriving to patients in out of hospital cardiac arrest.
The inclusion of patients receiving non-contrast CT images requires assumptions regarding LV mass, wall thickness and position of the interventricular septum due to inability to visualize the intraventricular cavity. Although we acknowledge this limitation, it is unlikely this would affect the mid-LV measurement by more than a few millimeters in any one plane and therefore not substantially impact the results. Similarly, the use of both gated and non-gated CT protocols introduces cardiac motion artifacts and differences in cardiac cycle although the impact on mid-LV location would be unlikely to change the significance of our results.
Our findings suggest that TTE-guided localization of the mid-LV is feasible and likely superior to the location recommended by the AHA guidelines. TTE may prove a practical tool to guide optimal location of chest compressions during or prior to CPR. With proper training, this approach could be implemented across a range of care settings. Future studies should focus on validating this approach in active cardiac arrest, comparison with transesophageal echocardiography to assess quality of compressions based on TTE location, assessing ease of use in real-world clinical environments, and evaluating the impact on resuscitation outcomes.
Neil Krulewitz: Writing – review & editing, Writing – original draft, Supervision, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Miles Lamberson: Writing – review & editing, Writing – original draft, Formal analysis, Data curation, Conceptualization. Ryan Walsh: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Zachary Clark: Writing – review & editing, Project administration, Investigation. Skyler Lentz: Writing – review & editing, Formal analysis. Lindsay Reardon: Writing – review & editing, Formal analysis.
The authors declare that they have no conflicts of interest. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.