^{1}

^{1}

^{2}

^{1}

This study was presented in 39th International Symposium on Intensive Care and Emergency Medicine in Brussel on 19th March, 2019.

There is a traditional assumption that to maximize stroke volume, the point beneath which the left ventricle (LV) is at its maximum diameter (P_max.LV) should be compressed. Thus, we aimed to derive and validate rules to estimate P_max.LV using anteroposterior chest radiography (chest_AP), which is performed for critically ill patients urgently needing determination of their personalized P_max.LV.

A retrospective, cross-sectional study was performed with non-cardiac arrest adults who underwent chest_AP within 1 hour of computed tomography (derivation:validation=3:2). On chest_AP, we defined cardiac diameter (CD), distance from right cardiac border to midline (RB), and cardiac height (CH) from the carina to the uppermost point of left hemi-diaphragm. Setting point zero (0, 0) at the midpoint of the xiphisternal joint and designating leftward and upward directions as positive on x- and y-axes, we located P_max.LV (x_max.LV, y_max.LV). The coefficients of the following mathematically inferred rules were sought: x_max.LV=α_{0}*CD-RB; y_max.LV=β_{0}*CH+γ_{0} (α_{0}: mean of [x_max.LV+RB]/CD; β_{0}, γ_{0}: representative coefficient and constant of linear regression model, respectively).

Among 360 cases (52.0±18.3 years, 102 females), we derived: x_max.LV=0.643*CD-RB and y_max.LV=55-0.390*CH. This estimated P_max.LV (19±11 mm) was as close as the averaged P_max.LV (19±11 mm, P=0.13) and closer than the three equidistant points representing the current guidelines (67±13, 56±10, and 77±17 mm; all P<0.001) to the reference identified on computed tomography. Thus, our findings were validated.

Personalized P_max.LV can be estimated using chest_AP. Further studies with actual cardiac arrest victims are needed to verify the safety and effectiveness of the rule.

Following the traditional assumption that the point (P_max.LV) beneath which the left ventricle (LV) is at its maximum diameter should be compressed to maximize stroke volume, it has been reported how to locate personalized P_max.LV (x_max.LV, y_max.LV) using parameters easily measurable on posteroanterior chest radiography.

We derived and validated rules to estimate P_max.LV using anteroposterior chest radiography, which is performed for critically-ill patients urgently needing determination of personalized P_max.LV, as follows: x_max.LV=0.643*cardiac diameter-right cardiac border and y_max.LV=55-0.390*cardiac height.

Optimum chest compression is critical for successful cardiopulmonary resuscitation (CPR). The current CPR guidelines have emphasized a specific rate (100–120 beats/min), depth (5–6 cm), full recoil after each compression, and minimizing pauses for compression [

This compression site for the general population has been determined by the studies that assumed that the stroke volume (SV) of the left ventricle (LV), a key point for successful CPR, would be maximized by compressing the theoretical optimum ‘point’ (P_max.LV) beneath which the LV is at its maximum diameter [

To apply the rules without the interruption of CPR, chest_PA should be performed before the cardiac arrest (CA). Due to the lack of previously investigated chest_PAs, these rules are applicable to only 71% and 38% of in-hospital CA (IHCA) and out-of-hospital CA (OHCA) victims, respectively [

In this study, we aimed to derive and validate the rules to determine personalized P_max.LV using the chest AP. We used the same methods as Cho et al. [

In this retrospective cross-sectional study, we included all consecutive non-CA adults aged ≥18 years who had undergone chest_AP within 1 hour of CT with sagittal reconstruction from 2012 to 2017 in CHA Bundang Medical Center. The institutional review board of the hospital approved this study protocol (2018-06-016-002). The requirement for informed consent was waived for this retrospective study.

We excluded cases that required alteration in cardiopulmonary modifying medications between performing chest_AP and CT (anti-hypertensives, diuretics, inotropics, chronotropics, fluid loading, beta agonists, and anticholinergics) and those with any immeasurable parameters on either test. Cases with the following thoracic abnormalities were included as they were found not to affect the rules in the determination of P_max.LV in the previous study: >2-cm depth pleural effusion, >1-cm depth hemo-/pneumothorax, a destroyed lung, lobectomy, atelectasis, hiatal hernia, >5-mm depth pericardial effusion, pericardial tumor/cyst, thoracic aorta dissection/aneurysm, and widened mediastinum [

We defined anatomical structures on chest_AP, the chest surface, and CT as described by Cho et al. [

The reference point, (P_zero [0, 0, 0]), was defined as the midpoint of the xiphisternal joint on both the chest surface and CT. Leftward, cephalad, and into-the-thorax directions were designated as positive on x-, y-, and z-axes, respectively (

The midpoint of the LV where it showed the maximum diameter was identified using CT as follows: x_max.LV, y_max.LV, z_max.LV. P_max.LV, which was assumed to be located just vertically above that point on the chest surface (z=0), was then positioned at the following location: x_max.LV, y_max.LV, 0 (

Assuming that the relative location of P_max.LV within the heart does not change interpersonally, its proportional width and height compared with CD and CH would remain constant. Using this assumption, we have inferred the following rules to estimate x_max.LV and y_max.LV, where α_{0}, β_{0}, and γ_{0} were constants derived from the study population [

x_max.LV=α_{0}*CD-RB

y_max.LV=β_{0}*CH+γ_{0}

(For detailed mathematical inference, please see explanation S1 of reference number 12).

Using the derivation set, we measured x_max.LV and y_max.LV on CT; and CD, RB, and CH on chest_AP. Using these measurements, we defined α_{0} as the mean value of ‘(x_max.LV+RB)/CD’. β_{0} and γ_{0} were determined as the representative regression coefficient and constant, respectively, revealed on the linear regression analysis to express y_max.LV in terms of CH. We then investigated whether the assumptions of the linear regression analysis were met. By determining α_{0}, β_{0}, and γ_{0} mentioned as above, we could derive the rules to estimate x_max.LV and y_max.LV to locate P_max.LV.

We validated the derived rules to estimate P_max.LV in two ways (

Second, we checked the superiority of P_max.LV (estimated) over two kinds of candidate compression points. Thus, we compared the proximity of P_max.LV (CT_reference) to P_max.LV (estimated) with that of each candidate point. The first candidate was P_max.LV (averaged). We defined its coordinates as the mean value of x_max.LV and y_max.LV of all enrolled cases. The second candidate was three P_guidelines to represent the compression site recommended in the current CPR guidelines [

In order to compare the results of the current study with those from the previous study in which we determined P_max.LV using chest_PA, we compared the variables using unpaired t-tests, chi-square tests, and others [

One author (KS) measured all parameters in all patients. To verify interrater reliability, another (CS) measured CD, RB, CH; x_max.LV, y_max.LV, and y coordinates of the sternal top (y_sternum [top]) and bottom (y_sternum [bottom]) among 8% of cases that were randomly selected. The intra-class correlation coefficient was calculated for each variable.

Data on sex, age, height, weight, and comorbidities were obtained from patient medical records. Body mass index was calculated as ‘weight (kg)/height (m)^{2}’.

Demographic information, comorbidities, parameters on chest_AP and CT, and α_{0}, β_{0} and γ_{0} were shown in derivation, validation, and combined sets. Continuous variables were presented as mean± standard deviations and were compared using t-tests. Categorical variables were presented as proportions (%) and were compared using chi-squared tests.

We defined type 1 and 2 errors as <0.05 and <0.10, respectively. Expecting the correlation coefficient (r) to be >0.25 for the linear regression analysis to estimate the y_max.LV with CH, the sample size of the derivation set had to be ≥164. Allotting cases to derivation and validation sets in a 3:2 ratio and assuming a loss rate as 0.4, we reviewed ≥456 cases [

Statistical analyses were performed using IBM SPSS Statistics ver. 24 (IBM Corp., Armonk, NY, USA). Statistical significance was presumed when two-sided P-values were <0.05.

In total, 482 patients had chest_AP and CT performed within 1 hour of each other. Among them, 122 cases were excluded: 43 cases for receiving medication between chest_AP and CT and 79 cases for unmeasurable variables in either test. Finally, 360 cases (mean age±standard deviation 52.0±18.3 years, 102 [28.3%] females) were enrolled.

We randomly assigned 237 (65.8%) and 123 (34.2%) cases to the derivation and the validation sets, respectively. There were no significant differences in sex, age, body habitus, comorbidities, and structural abnormalities between the sets (

Intra-class correlation coefficients to reveal the interrater reliability in measuring CD, RB, CH, x_max.LV, y_max.LV, y_sternum (top), and y_sternum (bottom) were all >0.95 (P<0.001, n=30) (

With the derivation set, α_{0} was 0.643±0.073, showing a normal distribution (Kolmogorov-Smirnov, P=0.20). As the mean of x_max.LV and y_max.LV was 52 and 11 mm, respectively, we located the P_max.LV (averaged) at (52 mm, 11 mm). The assumptions of the linear regression analysis were met to express y_max.LV in terms of CH. The regression analysis revealed the following:

y_max.LV=60-0.425*CH

(r=0.351, β_{0}: -0.425 [95% confidence interval, -0.556 to -0.294], P<0.001; γ_{0}: 60 [95% confidence interval, 45 to 75], P<0.001)

Using those representative statistics, we could estimate P_max.LV as follows:

x_max.LV=0.643*CD-RB (mm)

y_max.LV=60–0.425*CH (mm)

We validated these rules by applying them to the validation set (

Reanalyzing using the whole set of combined cases, we obtained the following modified rules to estimate P_max.LV on chest_AP (

x_max.LV=0.643*CD-RB (mm)

y_max.LV=55–0.390*CH (mm)

When these rules were applied to the whole set, the result of validation was similar (

Compared with the previous study that had estimated P_max.LV using chest_PA, the population for whom chest_AP was performed in the current study had fewer women and younger, taller, and heavier individuals (

The difference between the estimated and reference x_max.LV and y_max.LV did not significantly differ from that of the previous study. However, all the distances from P_max.LV (CT_reference) to P_max.LV (estimated), P_max.LV (averaged), and P_guidelines were larger than those revealed in the previous study based on chest_PA, with the exception of P_guideline (top).

This study revealed that the theoretical P_max.LV could be located with parameters easily measurable on chest_AP. We derived and validated its estimating rule as follows: x_max.LV=0.643*CD-RB and y_max.LV=55–0.390*CH (mm). For P_max.LV (CT_reference), this P_max.LV (estimated) was as close as P_max.LV (averaged) and closer than any of the three P_guidelines.

To our knowledge, this is the first study to estimate the theoretical P_max.LV using parameters measured on chest_AP. If the clinical effectiveness and safety of this estimation rule is verified in actual CPR, this rule in addition to that using chest_PA might help clinicians locate P_max.LV. Additionally, it might assist in guiding personalized optimum chest compression in up to 60% of OHCA and 100% of IHCA patients with easily available radiography which would have been checked using either chest_PA or chest_AP prior to CA.

Due to its similar study design, this study enhances the merits of the one previously conducted [

This study showed some differences compared with the previous one, which were not confined to demographic differences that we could not control for (

Initially, we expected that the applicability of the rule to estimate P_max.LV would rise from 71% and 38% to 100% and 60% for IHCA and OHCA, respectively, if P_max.LV became estimable using the previously investigated chest_AP and/or chest_PA. However, we excluded 79 (16.4%) cases among the 482 eligible patients for unmeasurable variables in either chest_AP or CT, thus falling short of that expectation. We think that P_max.LV (averaged) could be used to guide CPR when the estimation rules are not applicable. However, it remains unclear which P_max.LV (averaged) should be adopted; (i.e., the range determined in the previous study [50 mm, -7 mm] or that of the current study [52 mm, 11 mm]).

Comparing the previous study to the current one, x_max.LV was almost the same (50 vs. 52 mm); however, the y_max.LV differed (-7 vs. 10 mm). Both P_max.LVs have been defined on CT, the reference, rather than on chest_PA or chest_AP. Therefore, the difference in y_max.LV would result from the different phase of respiratory cycle during CT rather than the caudal cardiac movement due to the pull of gravity. In the previous study, patients underwent chest_PA; thus, they must have breathed fully during CT. However, those in the current study underwent chest_AP instead of the standard chest_PA, meaning that most of them could not have breathed fully holding their breaths at the end of full inspiration during CT. Their respiratory cycle might have resided even on the expiration phase. The difference in the y_max.LV can be explained by the fact that the LV might move cephalad up to 5 cm at the end of full expiration [

We think the actual y_max.LV should be determined considering the phase (whether inspiratory or expiratory) of the respiratory cycle during CPR [

This study shares some similar limitations with the one previously conducted [

In conclusion, personalized theoretical P_max.LV, which does not differ from its averaged location and is superior to the locations recommended by current CPR guidelines, may be located with parameters easily measurable on chest_AP. Further clinical studies are needed to verify its effectiveness and safety in addition to modifying its coordinates considering the cardiac movement during CPR by the altering lung-to-heart spatial relationship.

No potential conflict of interest relevant to this article was reported.

Identification of anatomical parameters on anteroposterior chest radiography, the zero point, and the theoretical optimum chest compression point. (A) Definition of anatomical parameters on anteroposterior chest radiography. Three vertical lines are drawn (white solid ones). The midline starts from the spinous process of the highest visible cervical vertebra and ends at the midpoint between the pedicles of the lowest visible thoracic vertebra. Its x coordinate is defined as zero. Then two parallel lines are drawn, which touch the right and left cardiac borders tangentially. The cardiac diameter (CD) is defined as the distance between these two lines. The distance from the midline to the line touching the right cardiac border is designated as RB. Thereafter, two horizontal lines (black dotted ones) are drawn. The upper one touches the bottom of the carina, where the lowest surfaces of the two main bronchi meet. The lower line contains the uppermost point of the left hemi-diaphragm. The distance between these two horizontal lines is designated as the cardiac height (CH). (B) Clinical and (C) radiographic identification of the zero point (P_zero). The midpoint of the xiphisternal joint, where both the costal margins, sternal body, and xiphoid process meet, has been selected as the P_zero with its own coordinate of (0, 0, 0). From P_zero, horizontal, vertical, and into-the-thoracic vertical lines, which form right angles with one another, were drawn as x, y and z axes, respectively. Leftward, upward, and into-the-thorax directions were designated as positive. (D) Identification of the theoretical optimum chest compression point on computed tomography (P_max.LV [CT_reference]). First, the midpoint of the left ventricle (LV), where the LV shows its maximum diameter, is identified by navigating through the sagittal sections of computed tomography (See

Location of estimated and averaged theoretical optimum chest compression points and three representative points recommended by current cardiopulmonary resuscitation guidelines to demonstrate their closeness to the reference theoretical optimum chest compression point identified on computed tomography (CT). The estimated P_max.LV (P_max.LV [estimated]) was validated in two ways. First, to check its precision, its x and y coordinates were compared with those of P_max.LV (CT_reference), the reference values measured on CT. Secondly, we assessed its superiority over the other candidate compression points by showing how close it is to P_max.LV (CT_reference). We compared its distance to P_max.LV (CT_reference) (white double-headed solid arrow) with the distance from P_max.LV (CT_reference) (1) to the averaged P_max.LV (P_max.LV [averaged]), which was defined by the averaged value of the x and y coordinates for the study population (white double-headed dashed arrow) and (2) to three points representing the lower sternal half where the current cardiopulmonary resuscitation guidelines recommend to compress (black double headed solid arrows): P_guideline (top), P_guideline (middle), and P_guideline (bottom). These three P_guidelines were identified along the vertical midline of the lower sternal half at equal distances designated as top, middle, and bottom with their y coordinates: y_guideline (top), y_guideline (middle), and y_guideline (bottom), which equals y_sternum (bottom), respectively. By designating the uppermost and lowest y coordinate of the whole sternum as y_sternum (top) and y_sternum (bottom), respectively, we could calculate y_guideline (top) and y_guideline (middle) as ‘y_sternum (bottom)+(whole sternal length)/2’ and ‘y_sternum (bottom)+(whole sternal length)/4’, where ‘whole sternal length’ equals ‘y_sternum (top)-y_sternum (bottom).’ Rectangle: P_zero; star: P_max.LV (CT_reference); cross: P_max.LV (estimated); white circle: P_max.LV (averaged); black circles: P_guidelines (top, middle, and bottom); ①–⑤: distances from P_max.LV (CT_reference) to P_max.LV (estimated) (①), P_max.LV (averaged) (②), P_guideline (top) (③), P_guideline (middle) (④) and P_guideline (bottom) (⑤). Adapted from Cho S et al. Resuscitation 2018;128:97-105, with permission from Elsevier [

Estimation of the theoretical optimum chest compression point from anteroposterior and posteroanterior chest radiography and the difference in cardiac height (CH) and y_max.LV rooted in their technical differences. (A) Estimation of the theoretical optimum chest compression point from anteroposterior chest radiography (chest_AP). (B) Estimation of the theoretical optimum chest compression point from posteroanterior chest radiography (chest_PA) and the difference in its CH and y_max.LV from that of anteroposterior chest radiography caused by its technical differences. When undergoing chest_PA, patients stand up and breathe in to the fullest extent. As they stand up, their heart with its y_max.LV is pulled downward by gravity and the diaphragm is also pushed downward. As they breathe in fully, the lungs inflate to the maximum capacity and push the diaphragm caudally again. These technical differences in chest_PA cause larger CH and lower y_max.LV than those caused by technical differences in chest_AP, which is taken with the critically ill patient usually in a supine position without complete control of the respiratory cycle. CD, cardiac diameter; RB, right cardiac border. Dotted ellipses: proximal end of the main bronchi; triangle: the bottom of the carina; diamond: the uppermost point of the left hemi-diaphragm; star: P_max.LV; circles: the right and left cardiac borders the vertical tangential lines meet the heart; rectangle: P_zero. Adapted from Cho S et al. Resuscitation 2018;128:97-105, with permission from Elsevier [

Reasons for the difference in the right (RB) and left (LB) cardiac border and the cardiac diameter (CD), which is the sum of RB and LB, when measured on anteroposterior chest radiography (chest_AP) compared with those measured on posteroanterior chest radiography (chest_PA) demonstrated by simulating both techniques using an image obtained via computed tomography. (A) Simulation of chest_PA. The radiation beam, which starts 180 cm away from the receiving plate, meets the heart located posteriorly. (B) Simulation of chest_AP (1). The radiation beam meets the heart being located anteriorly within the thorax and causes a larger image of the heart and thus a larger RB and CD on the receiving plate than those of chest_PA. (C) Simulation of chest_AP (2). The radiation beam travels a variable but consistently shorter distance from the source to the receiving plate than that in chest_PA: 122±7 ranging from 110–132 cm, which is definitely shorter than 180 cm, the fixed distance to perform chest_PA. Therefore, it causes a larger cardiac image, RB, and CD again on the receiving plate compared with those of chest_PA. Dotted arrows: radiation beam to the right and left cardiac borders; Dotted line: reference line to define x=0; Thick baseline: the bottom line to which the source of the radiation beam belongs. For easy explanation, the points the radiation beam meets on the right and left cardiac borders have been assumed to lie on the same plane. Adapted from Chon SB et al. J Korean Med Sci 2011;26:1446-53, with permission from Korean Academy of Medical Sciences [

Demographic and radiographic characteristics of derivation and validation sets

Information | Derivation (n=237) | Validation (n=123) | P-value | Total (n=360) |
---|---|---|---|---|

Demographics | ||||

Female, n (%) | 68 (28.7) | 34 (27.6) | 0.83 | 102 (28.3) |

Age (yr) | 51.8 ± 17.5 | 52.6 ± 19.7 | 0.72 | 52.0 ± 18.3 |

Height (cm) | 166.5 ± 8.8 | 166.1 ± 8.0 | 0.70 | 166.3 ± 8.5 |

Weight (kg) | 66.1 ± 13.4 | 63.8 ± 10.7 | 0.089 | 65.3 ± 12.6 |

BMI (kg/m^{2}) |
23.7 ± 3.8 | 23.0 ± 3.1 | 0.077 | 23.5 ± 3.6 |

Chest_AP (mm) | ||||

CD | 156 ± 17 | 154 ± 19 | 0.277 | 155 ± 18 |

RB | 48 ± 12 | 46 ± 11 | 0.10 | 48 ± 12 |

CH | 114 ± 18 | 117 ± 18 | 0.13 | 116 ± 18 |

Chest CT (mm) | ||||

y_sternal top | 142 ± 15 | 141 ± 14 | 0.59 | 142 ± 15 |

y_sternal bottom | -43 ± 12 | -44 ± 13 | 0.87 | -43 ± 12 |

x_max. LV | 52 ± 10 | 52 ± 10 | 0.89 | 52 ± 10 |

y_max. LV | 11 ± 20 | 9 ± 20 | 0.36 | 10 ± 20 |

Derived constants | ||||

α | 0.643 ± 0.073 | NA | NA | 0.643 ± 0.080 |

β (95% CI) | -0.425 (-0.556, -0.294) | NA | NA | -0.390 (-0.498, -0.282) |

γ (95% CI) | 60 (45, 75) | NA | NA | 55 (43, 68) |

Estimated value minus CT_reference value (mm)^{a)} |
||||

x_max.LV | NA | 1 ± 13 | NA | 0 ± 12 |

y_max.LV | NA | 1 ± 19 | NA | 0 ± 18 |

Distance from P_max.LV (CT_reference) to (mm)^{a)} |
||||

P_max.LV (estimated) | NA | 20 ± 11 | NA | 19 ± 11 |

P_max.LV (averaged) | NA | 19 ± 11 | 0.13^{b)} |
19 ± 11 |

P_guideline (top) | NA | 67 ± 13 | < 0.001^{b)} |
67 ± 13 |

P_guideline (middle) | NA | 56 ± 11 | < 0.001^{b)} |
56 ± 10 |

P_guideline (bottom) | NA | 76 ± 18 | < 0.001^{b)} |
77 ± 17 |

BMI, body mass index; AP, anteroposterior; CD, cardiac diameter; RB, the distance from the thoracic midline to the parallel line touching the right cardiac border tangentially; CH, cardiac height; CT, computed tomography; LV, left ventricle; NA, not available; CI, confidence interval; CT_reference value, reference values measured on CT; P_max.LV, the point compression of which is presumed to maximize the stroke volume of the left ventricle with its coordinate of (x_max.LV, y_max.LV, 0); P_guideline, the points along the lower sternal half with (top) at its top, (middle) at its middle and (bottom) at its bottom, respectively; x_‘A’, x coordinate of point ‘A’; y_‘A’, y coordinate of point ‘A.’

For comparison within the validation set, the coordinate of P_max.LV (estimated), P_max.LV (averaged), P_guideline (top), P_guideline (middle) and P_guideline (bottom) were designated as follows using α_{0}, β_{0}, and γ_{0} calculated from the derivation set: (0.643*CD-RB, 60-0.425*CH), (52, 11), (0, y_sternal bottom+sternal length/2), (0, y_sternal bottom+sternal length/4) and (0, y_sternal bottom), respectively (where, sternal length=y_sternal top-y_sternal bottom). For comparison within the total cases, P_max.LV (estimated) were located at (0.643*CD-RB, 55-0.390*CH) using the new α_{0}, β_{0}, and γ_{0} derived from the combined set. P_max.LV (averaged) was set at (52, 10), accordingly, while the three P_guidelines were defined in the same way as above.

Compared with the distance from P_max.LV (CT_reference) to P_max.LV (estimated) by paired t-test.

Comorbidity and structural abnormality of derivation and validation sets

Status | Derivation (n = 237) | Validation (n = 123) | P-value | Total (n = 360) |
---|---|---|---|---|

Comorbidity, n (%) | ||||

Hypertension | 36 (15.2) | 27 (22.0) | 0.11 | 63 (17.5) |

Diabetes mellitus | 26 (11.0) | 10 (8.1) | 0.39 | 36 (10.0) |

Hyperlipidemia | 3 (1.3) | 3 (2.4) | 0.42 | 6 (1.7) |

Ischemic heart disease | 6 (2.5) | 3 (2.4) | > 0.99 | 9 (2.5) |

Heart failure | 2 (0.8) | 0 (0.0) | 0.55 | 2 (0.6) |

Obstructive lung disease | 5 (2.1) | 2 (1.6) | > 0.99 | 7 (1.9) |

Restrictive lung disease | 0 (0.0) | 0 (0.0) | NA | 0 (0.0) |

Pulmonary embolism | 3 (1.3) | 1 (0.8) | > 0.99 | 4 (1.1) |

Chronic kidney disease | 4 (1.7) | 2 (1.6) | > 0.99 | 6 (1.7) |

Chronic liver disease | 3 (1.3) | 3 (2.4) | 0.42 | 6 (1.7) |

Stroke | 9 (3.8) | 10 (8.1) | 0.081 | 19 (5.3) |

Malignancy | 7 (3.0) | 7 (5.7) | 0.25 | 14 (3.9) |

Structural abnormality, n (%) | ||||

Hemo-/pneumothorax > 1-cm depth | 17 (7.2) | 9 (7.3) | 0.96 | 26 (7.2) |

Atelectasis/lobectomy | 11 (4.6) | 5 (4.1) | 0.80 | 16 (4.4) |

Pleural effusion/empyema > 2-cm depth | 8 (3.4) | 6 (4.9) | 0.57 | 14 (3.9) |

Wide mediastinum | 6 (2.5) | 2 (1.6) | 0.72 | 8 (2.2) |

Aortic dissection | 4 (1.7) | 0 (0.0) | 0.30 | 4 (1.1) |

NA, not available.

Intra-class correlation coefficient to measure key variables on anteroposterior chest radiography and computed tomography

Variable | Intra-class correlation coefficient | P-value |
---|---|---|

On anteroposterior chest radiography | ||

Cardiac diameter | 0.98 | < 0.001 |

Right cardiac border | 0.98 | < 0.001 |

Cardiac height | 0.99 | < 0.001 |

On computed tomography | ||

x_max.LV | 0.98 | < 0.001 |

y_max.LV | 0.97 | < 0.001 |

y_sternum (top) | 0.97 | < 0.001 |

y_sternum (bottom) | 0.96 | < 0.001 |

LV, left ventricle.

Comparison with the previous study which aimed to estimate the theoretical optimum chest compression point using chest_PA

Information | Previous study performed with chest_PA (n=266) | Current study performed with chest_AP (n=360) | P-value |
---|---|---|---|

Demographics | |||

Female, n (%) | 120 (45.1) | 102 (28.3) | < 0.001^{**} |

Age (yr) | 57.6 ± 16.4 | 52.0 ± 18.3 | < 0.001^{**} |

Height (cm) | 162.6 ± 8.7 | 166.3 ± 8.5 | < 0.001^{**} |

Weight (kg) | 60.9 ± 10.9 | 65.3 ± 12.6 | < 0.001^{**} |

BMI (kg/m^{2}) |
23.0 ± 3.7 | 23.5 ± 3.6 | 0.092 |

Chest X-ray (mm) | |||

CD | 142 ± 18 | 155 ± 18 | < 0.001^{**} |

RB | 44 ± 11 | 48 ± 12 | < 0.001^{**} |

CH | 131 ± 20 | 116 ± 18 | < 0.001^{**} |

Chest CT (mm) | |||

y_sternal top | 139 ± 15 | 142 ± 15 | 0.024^{*} |

y_sternal bottom | -42 ± 11 | -43 ± 12 | 0.084 |

x_max.LV | 50 ± 10 | 52 ± 10 | 0.067 |

y_max.LV | -7 ± 17 | 10 ± 20 | < 0.001^{**} |

Derived constants | |||

α | 0.664 ± 0.069 | 0.643 ± 0.080 | < 0.001^{**} |

β (95% CI) | -0.356 (-0.446, -0.266) | -0.390 (-0.498, -0.282) | 0.64 |

γ (95% CI) | 40 (28, 51) | 55 (43, 68) | < 0.001^{**} |

Estimated value minus CT_reference value (mm) | |||

x_max.LV | 0 ± 10 | 0 ± 12 | 0.93 |

y_max.LV | 0 ± 15 | 0 ± 18 | 0.73 |

Distance from P_max.LV (CT_reference) to (mm) | |||

P_max.LV (estimated) | 15 ± 9 | 19 ± 11 | < 0.001^{**} |

P_max.LV (averaged) | 17 ± 10 | 19 ± 11 | 0.014^{*} |

P_guideline (top) | 76 ± 13 | 67 ± 13 | < 0.001^{**} |

P_guideline (middle) | 54 ± 11 | 56 ± 10 | 0.027^{*} |

P_guideline (bottom) | 63 ± 13 | 77 ± 17 | < 0.001^{**} |

chest_PA, posteroanterior chest radiography; chest_AP, anteroposterior chest radiography; BMI, body mass index; CD, cardiac diameter; RB, the distance from the thoracic midline to the parallel line touching the right cardiac border tangentially; CH, cardiac height; CT, computed tomography; LV, left ventricle; CI, confidence interval; CT_reference value, reference values measured on CT; P_guideline, the points along the lower sternal half with (top) at its top, (middle) at its middle and (bottom) at its bottom, respectively; P_max.LV, the point compression of which is presumed to maximize the stroke volume of the left ventricle with its coordinate of (x_max.LV, y_max.LV, 0); x_‘A’, x coordinate of point ‘A’; y_‘A’, y coordinate of point ‘A.’ Data on the left column have been retrieved from Cho S et al. Resuscitation 2018;128:97-105.12.

P<0.05,

P<0.01.