Serious Injuries Secondary to Cardiopulmonary Resuscitation: Incidence and Associated Factors

Emergencias 2019;31:327-334

ORIGINAL ARTICLE Serious injuries secondary to cardiopulmonary resuscitation: incidence and associated factors

Youcef Azeli1-3, Eneko Barbería4,5, María Jiménez-Herrera6, Alberto Ameijide7, Christer Axelsson8, Alfredo Bardají3,5,9

Objective. To determine the incidence of serious rib cage damage (SRD) and serious visceral damage (SVD) Authors’ affiliation: 1Emergency Medical System secondary to cardiopulmonary resuscitation (CPR) and to explore associated factors. of Catalonia, L’Hospitalet de Methods. We analyzed data from the prospective registry of cases of sudden death in Tarragona, Spain (the ReCaPTa Llobregat, Barcelona, Spain. 2Emergency Department, Hospital study). Cases were collected from multiple surveillance sources. In this study we included the cases of autopsied Universitari Sant Joan de Reus, nonsurvivors after attempted manual CPR between April 2014 and May 2016. A specific protocol to detect injuries Reus, Tarragona, Spain. 3Institut d’Investigació Sanitaria secondary to CPR was used during the autopsies. Pere Virgili, Tarragona, Spain. Results. We analyzed 109 cases. The mean age at death was 63 years and 32.1% were women. SRD were found in 4Institut de Medicina Legal y Forense de Catalunya, Tarragona, 63.3% and SVD in 14.7%. The group with SRD were significantly older (63vs 59 years, P=.031) and included higher Spain. percentages of persons with a chest circumference over 101 cm (56.5 vs 30%, P=.016) and a waist circumference 5Universitat Rovira i Virgili, over 100 cm (62.3 vs 37.5%, P=.017). A multivariable analysis confirmed chest circumference over 101 cm as the Tarragona, Spain. 6Department of Nursing, Rovira only risk factor for SRD (odds ratio [OR], 2.45; 95% CI, 1.03–5.84) and female sex as the only risk factor for SVD (OR, i Virgili University, Tarragona, 5.02; 95% CI, 1.18–21.25). Spain. 7Biostatistics Unit, Fundació Lliga Conclusion. Women and any patient with a chest circumference greater than 101 cm are at greater risk for serious per a la Investigació i Prevenció injuries related to CPR. del Càncer, Reus, Tarragona, Spain. Keywords: Cardiopulmonary resuscitation. Rib fractures. Thoracic injuries. 8University of Borås, Borås, Västra Götaland, Sweden. 9Cardiology Department, Joan XXIII University Hospital, Incidencia y variables asociadas a lesiones graves secundarias a reanimación Tarragona, Spain. cardiopulmonar Corresponding author: Youcef Azeli Objetivo. Determinar la incidencia de lesiones torácicas óseas graves (LTOG) y de lesiones viscerales graves (LVG) se- Sistema d’Emergències Mèdiques cundarias a reanimación cardiopulmonar (RCP) y estudiar las variables asociadas a ellas. de Catalunya 112 Reus, Carrer dels Pagesos, 2 Método. Se analizaron los datos del proyecto ReCaPTa, estudio prospectivo de la muerte súbita, con varias fuentes de 43204 Reus, Tarragona, Spain información. Se incluyeron los pacientes no supervivientes a un intento de RCP manual, entre abril de 2014 y mayo E-mail: de 2016, en los que se aplicó un protocolo autópsico específico para el estudio de las lesiones secundarias a la RCP. [email protected]

Resultados. Se analizaron 109 sujetos, con una mediana de edad de 63 años y un 32,1% de mujeres. El 63,3% pre- Article information: sentaron LTOG y el 14,7%, LVG. El grupo con LTOG fue de mayor edad (63 vs 59 años, p = 0,031) y se asoció a una Received: 11-5-2019 mayor proporción de pacientes con un perímetro torácico > 101 cm (56,5 vs 30%, p = 0,016) y con un perímetro Accepted: 27-7-2019 Online: 12-9-2019 abdominal > 100 cm (62,3 vs 37,5%, p = 0,017). En un análisis multivariado el perímetro torácico > 101 cm fue el único factor de riesgo para LTOG (OR 2,45; IC 95% 1,03-5,84) y el sexo femenino fue el único factor de riesgo para Editor in charge: LVG (OR 5,02; IC 95% 1,18-21,25). Guillermo Burillo Putze Conclusión. Los pacientes con un perímetro torácico mayor de 101 cm y las mujeres presentan un riesgo superior de lesiones graves debidas a la RCP. Palabras clave: Reanimación cardiopulmonar. Fracturas costales. Lesiones torácicas.

Introduction the best possible balance between the benefit of increased CC depth and the risk of causing harm to the pa- The survival rate of out-of-hospital cardiac arrest is tient6. Although maximum survival has been found to low1. The quality of chest compressions (CC), with an be associated with patients receiving CC between 40.3 adequate depth and frequency to allow the sternum and 55.3 mm7, the latest guidelines of the European to return to its position, is key to optimizing cardiac Resuscitation Council (ERC) support the recommenda- output during cardiopulmonary resuscitation (CPR)2. tions of the International Liaison Committee on Chest injuries secondary to CC are common and Resuscitation, which state that a compression depth of may be accompanied by visceral lesions3. The duration approximately 5 cm, but not more than 6 cm, is reaso- of CPR, female sex, age and depth of compression are nable for a medium sized adult8. the main known associated variables4,5. In the last deca- The objective of this study was to determine the in- de, resuscitation guidelines have attempted to establish cidence of serious rib cage damage (SRD) and severe

327 Azeli Y, et al. Emergencias 2019;31:327-334

visceral damage (SVD) in a prospective series of pa- mild ribcage damage (MRD). SVD was considered as tients who did not survive CPR with manual CC, and to any of the following lesions: hemopericardium, epicar- study the factors associated with them. dial contusion, dissection or hematoma of the thoracic aorta, pneumothorax, hemothorax, hepatic laceration, hepatic subcapsular hematoma, spleen injury, pulmo- Method nary hematoma. Any injury that could interfere with the cardiovascular or respiratory system or with ex- Design and stage sanguination of more than 800 ml14 was considered a life-threatening injury14. The Tarragona Clinical Pathological Register (ReCaPTa) is a prospective study of sudden cardiac Statistical analysis death carried out in the Camp de Tarragona Health Region (Catalonia, Spain), which has a population of The continuous variables are described with me- 511,622 inhabitants. This register is based on several dian and interquartile ranges and the categorical ones sources of information: data from all out-of-hospital with number of cases and percentages. The Student’s cardiac arrests attended by the Sistema d’Emergències T or Mann-Whitney’s U and the chi-square were used Mèdiques de Catalunya (SEM), data from autopsies to compare the subgroups according to severity. A performed by the Forensic Pathology Service, data ROC curve (Figures 1 and 2) was performed in order from survivors admitted to hospitals in the area, as to obtain a threshold value of the thoracic and abdo- well as the medical history of patients, through the minal perimeters that would allow predicting SRD primary care (PC) clinical history. The study was with the greatest possible sensitivity and specificity. approved by the ethical committee of the Joan XXIII Based on this value, the thoracic perimeter was cate- University Hospital (CI 65/2014) and its design has gorized as greater or less than 101 cm and the abdo- been previously published9. minal perimeter as greater or less than 100 cm. A bi- Patients treated by SEM who did not survive a ma- nary logistic regression model was constructed to nual CPR attempt between April 2014 and May 2016 assess the effect of various risk factors on SRD and and who died in situ were included in this paper. SVD. The odds ratio (OR) and confidence interval were Excluded were people under 18 years of age, trauma- initially calculated in a univariate analysis and subse- tic deaths, patients treated with a mechanical chest quently variables showing a p < 0.05 and variables compressor and those transferred to the hospital. All such as age and sex were included in a stepwise mul- patients were treated according to SEM protocols, tivariate analysis, regardless of their significance. The who followed the recommendations of the ERC resus- statistical package R and the software SPSS v23 were citation guidelines in force during the study period10. used.

Clinical and autopsy variables

Data on the SEM support were collected by the 1.0 team that cared for the patient and the variables included in the Utstein11 register were recorded. A team of PC researchers collected the medical history of the 0.8 patient. In Spain, according to the law, a forensic autopsy is required for all violent deaths where there is no known cause of death. This includes all natural and unexpected deaths occurring in the out-of-hospital se- 0.6 tting. All subjects were autopsied according to a speci- y fic protocol12 by forensic doctors from the Institute of

Forensic and Legal Medicine of Catalonia in Tarragona. Sensitivit 0.4 In addition to anthropometric variables such as wei- ght, height, and thoracic and abdominal perimeter, AUC: 0.591 (IC 95%: 0.454-0.728) the depth of the compression point was measured. For this purpose, the sternum was measured between 0.2 the sternal notch and the base of the xiphoid appen- dix and the compression point was located in the mi- ddle of the lower half of the sternum. The depth of 0.0 this point was measured in relation to the anterior 0.0 0.2 0.4 0.6 0.8 1.0 pectoral plane. Specificity SRD was defined as the presence of a sternal fractu- Figure 1. ROC curve for thoracic perimeter and severe thora- re or more than 6 unilateral rib fractures or more than cic bone injury. The chest circumference value of 101 has a 4 rib fractures if one of them was bilateral13. Patients sensitivity of 0.599 and a specificity of 0.652. AUC: area un- without lesions or without SRD criteria were considered der the curve; CI: confidence interval.

328 Azeli Y, et al. Emergencias 2019;31:327-334

1.0 1,140 OHCA treated by MES between April 2014 and May 2016

377 no CPR attempts 0.8 191 CPR interrupted by futility 13 CPR treated by primary care

559 attempted CPR treated by MES 0.6

y 144 admitted to hospital alive

Sensitivit 0.4 415 non-survivors

AUC: 0.565 (IC 95%: 0.423-0.707) 7 under 18 years of age 0.2 24 taken to hospital 70 treated with mechanical compressions 16 traumatic cardiac arrests 0.0

0.0 0.2 0.4 0.6 0.8 1.0 298 non-traumatic cardiac arrests treated with manual CPR Specificity

Figure 2. ROC curve for abdominal perimeter and severe tho- 110 autopsies racic bone injury. The abdominal perimeter value of 100 has a sensitivity of 0.543 and a specificity of 0.682. AUC: area un- 1 autopsy with der the curve; CI: confidence interval. incomplete data

Results 109 autopsies for final analysis

Basal Characteristics Figure 3. Patient Inclusion Flowchart. OHCA: out-of-hospital cardiac arrest; CPR: cardiopulmonary resuscitation; MES: medical emergency system. During the study period, SEM performed 559 CPR attempts. Nine patients were included in the final analysis (Figure 3). The mean age of the sample was hemopericardium in 40% of cases and hepatic hema- 63 years, interquartile range (IQR) 49-70 and 32.1% toma showed hemoperitoneum in 28.5% of cases. were women. The causes of death according to au- Four patients (3.7%) showed a life-threatening lesion, topsy were: cardiac (55%), drowning (12.8%), respi- all of them with hemopericardium, and two patients ratory (6.4%), intoxication (6.4%), hepatic-gastroin- also had liver laceration and one had significant he- testinal (5.5%), cerebral (3.7%) and other causes moperitoneum. None of the patients with hemoperi- (10%). cardium presented acute ischemic disease as a cause of death. Characteristics of the lesions Injury Risk Factors SRD was recorded by 63.3% and 14.7% by SVD. At least 77.1% presented a rib fracture and 43.1% presen- Table 3 shows the main injury risk factors. The ted sternum fracture. Table 1 describes the detail of the subgroup with SRD had a higher age than the MRD lesions. In the SRD subgroup, 68.1% showed at least group (63 [IQR 49-75] vs 59 [IQR 56-68] years, p = one sternal fracture, 78.3% more than 4 bilateral frac- 0.031). The total median CPR time was 35 min (IQR tures and 4.3% more than 6 unilateral fractures. In the 25-50). There were no significant differences be- MRD subgroup the median number of fractures was 0 tween subgroups regarding total, basic or advanced (IQR 0-2) and the percentage of bilateral rib fractures CPR times. Of the study population, 37.6% received was 10%; however, in the SRD subgroup the median witnessed CPR, with no differences found between number of rib fractures was 8 (IQR 6-11) and the per- groups. None of the 2 patients who received fibri- centage of bilateral rib fractures was 87%. The percen- nolysis during CPR had SVD. There were no signifi- tage of fractures according to their distribution in the cant differences between the subgroups regarding rib cage in each subgroup is shown in Figures 4, 5 and baseline antithrombotic therapy. The median body 6. The SRD subgroup was significantly associated with mass index (BMI) in the total population was 28 kg/ SVD compared to the MBI subgroup (21 vs 2.5%, p = m2 (IQR 25-31) and the median compression point 0.006). depth was 10 mm (IQR 7-13), with no differences The type of SVD and its distribution by sexes are between subgroups. Median chest and abdominal shown in Table 2. Epicardial contusions also showed perimeters were greater in the subgroup with SRD,

329 Azeli Y, et al. Emergencias 2019;31:327-334

Table 1. LInjuries secondary to cardiopulmonary resuscitation in the total population and according to the presence of serious rib damage or visceral damage Total MRD SRD No SVD SVD N = 109 N = 40 N = 69 p N = 93 N = 16 p n (%) n (%) n (%) n (%) n (%) Thoracic Bone Injuries Sternal fracture 47 (43.1) 0 47 (68.1) < 0.001 37 (39.8) 10 (62.5) 0.09 Costal fracture 84 (77.1) 17 (42.5) 67 (97.1) < 0.001 70 (75.3) 14 (87.5) 0.282 Flail chest 3 (2.8) 0 3 (4.3) 0.181 2 (2.2) 1 (6.3) 0.355 Bilateral rib fracture 64 (58.7) 4 (10.0) 60 (87.0) < 0.001 51 (54.8) 13 (81.3) 0.047 Thoracic vertebral fracture 0 0 0 0 0 Rib fractures [median (IQR)] 5 (1-9) 0 (0-2) 8 (6-11) < 0.001 4 (1-9) 9 (7-13) 0.005 Other injuries Skin lesion 27 (24.8) 7 (17.5) 20 (29.0) 0.181 19 (20.4) 8 (50.0) 0.011 Retroesternal hematoma 16 (14.7) 3 (7.5) 13 (18.8) 0.107 8 (8.6) 8 (50.0) < 0.001 Anterior mediastinal hematoma 18 (16.5) 2 (5.0) 16 (23.2) 0.014 10 (10.8) 8 (50.0) < 0.001 Airway Injury 11 (10.5) 3 (7.5) 8 (11.6) 0.495 7 (7.5) 4 (25.0) 0.032 Severe visceral injury 16 (14.7) 1 (2.5) 15 (21.7) 0.006 SRD- serious ribcage damage; MRD- mild ribcage damage; SVD-severe visceral damage; IQR-interquartile range. although without statistical significance (101 [IQR male patients (56.3 vs 28%, p = 0.025) and with a 95-107] vs 99 [IQR 93-103] cm, p = 0.192, and 102 history of osteoporosis (12.5 vs 2.1%, p = 0.038). [IQR 92-107] vs 98 [IQR 89-108], p = 0.108). In the multivariate analysis, the chest circumferen- However, the SRD subgroup was associated with a ce > 101 cm with an OR of 2.45 (95% CI 1.03-5.84) higher proportion of patients with a chest circumfe- was the sole independent risk factor for SRD, and the rence greater than 101 cm (56.5 vs 30%, p = 0.016) female sex was the sole risk factor for SVD, with an and with an abdominal circumference greater than OR of 5.02 (95% CI 1.18-21.25). Tables 4 to 6 show 100 cm (62.3 vs 37.5%, p = 0.017). The SVD sub- the incidence of lesions with other Utstein group was associated with a higher proportion of fe- variables11.

Left side Right side

Rib Sternal Paravertebral Posterolateral Anterolateral Paraesternal Paraesternal AnterolateralPosterolateralParavertebral number fraction

1 000 2.3 00 00 2 004.7 2.3 2.3 4.7 00 3 002.3 004.7 7 2.3 0 4 007 2.3 0 11.6 2.3 00 5 009.3 002.3 9.3 00 6 009.3 004.7 00 7 0000 0 2.3 00 8 000 2.3 00 9 000000 10 000000 11 00 00 12 0 0 Figure 4. Distribution of lesions in the subgroup of mild ribcage damage (%). Graphic representation based on a study by Kashiwagui et al.33.

Left side Right side

Ribde Sternal Paravertebral Posterolateral Anterolateral Paraesternal Paraesternal AnterolateralPosterolateralParavertebral number fraction

1 0016.7 15.2 26.7 16.7 00 2 0 1.5 45.5 16.7 24.2 39.4 1.5 0 3 0 3 60.6 16.7 6.1 34.8 51.5 3 0 4 0 1.5 59.1 15.2 45.5 30.3 53 1.5 0 5 0 1.5 57.6 13.6 4.5 19.7 48.5 1.5 0 6 0045.5 3 9.1 37.9 00 7 0022.7 0016.7 00 8 003 7.6 00 9 001.5 1.5 00 10 000000 11 00 00 12 0 0 Figure 5. Distribution of lesions in the subgroup of serious ribcage damage (%). Graphic representation based on a study by Kashiwagui et al.33.

330 Azeli Y, et al. Emergencias 2019;31:327-334

Left side Right side

Rib Sternal Paravertebral Posterolateral Anterolateral Paraesternal Paraesternal AnterolateralPosterolateralParavertebral number fraction

1 0018.8 31.2 37.5 12.5 00 2 0050 37.5 43.8 43.8 00 3 0056.2 31.2 12.5 43.8 50 00 4 0043.8 25 50 31.2 43.8 00 5 0050 18.8 0 25 31.2 00 6 0037.5 6.2 18.8 25 00 7 0018.8 006.2 00 8 006.2 6.2 00 9 000 6.2 00 10 000000 11 00 00 12 0 0 Figure 6. Distribution of bone lesions in the subgroup of severe visceral damage (%). Graphical representation based on a study by Kashiwagui et al.33.

Discussion sing the lesions used, which diagnosed sternal fracture in only 4% of cases. Our series reveals that two-thirds of subjects who do Age is a known risk factor for injuries secondary to not survive a CPR attempt with manual chest compres- CPR4. With age, bone mineralization and porosity in- sions have either SRD or SVD. In addition, SRD is rela- crease, while bone cortical thinning occurs. The result is ted to thorax perimeter, abdominal perimeter, and age, a more elastic tissue and more vulnerable to fractures. and SVD is associated with females. Women have been associated with a higher inciden- Autopsies are considered the standard of reference ce of bone fractures, especially of the sternum. The for the study of post-CPR lesions15. In other similar lower thickness of the breastbone compared to men, prospective studies, the incidence of costal fractures va- the higher relative age of women when suffering a car- ries between 64.6 and 73.7%, and that of sternal frac- diac arrest and a higher rate of degenerative bone di- tures between 54.2 and 66.3%16,17. The incidence of rib seases justify their greater vulnerability. Furthermore, fractures in our series was 77.1%, and the incidence of women have been described as an independent risk sternal fractures was 43.1%. These differences may be factor for visceral lesion19, although not all studies coin- attributable to the different study populations or to cide17,20. Recently, the major traumatic injuries secon- small differences between the protocols used. dary to CPR of survivors of the Parisian Registry have In our study, 3.7% of patients suffered a life-threate- been analysed and were associated with the female sex, ning injury. In the literature, the percentage of patients age and the occurrence of CRA at home18. exhibiting life-threatening injuries secondary to CPR As in other studies, we have found no relationship ranged from 0 to 13.7%13,16-18. Hemopericardium occu- between witnessed CPR and the incidence of lesions se- rred in 3.7% and epicardial injury in 4.6% of cases. In condary to CPR18. another similar study, epicardial lesion was found in Our study series has shown a non-significant trend 8.4% of cases after manual CPR16. This difference could in the duration of CPR as a risk factor for injury. Only be attributed to our lower incidence of sternal the subgroup with SRD received longer advanced CPR fractures17. time (24 vs 29 min, p = 0.086). In the literature this re- Our data differ from those of Koster et al., who lationship is also unclear21 and studies are needed on found SRD in 41.3% of a series of patients who recei- the safety of prolonged CPR. It is possible that the inci- ved manual CPR13, a low percentage compared to our dence of lesions is bimodal, producing a first peak at series, perhaps explained both by the high percentage the beginning22, due to the initial stiffness of the tho- of survivors and by the different methods for diagno- rax, and a second peak as time passes, which could mask the effect of the first one. Until now, the study of the relationship between Table 2. Severe visceral damage and distribution by sex body size and the incidence of fractures had been ba- Patients Men Women sed mainly on BMI. As in other studies5,23,24, we have N = 16 N = 7 N = 9 found no differences in the subgroups studied in rela- n (%) n (%) n (%) tion to BMI, possibly due to the small number of pa- Hemopericardium 4 (3.7) 2 (50) 2 (50) tients with a BMI > 30. Only the subgroup with SVD Aortic hematoma 3 (2.7) 1 (33.3) 2 (66.6) presents a higher BMI than the subgroup without SVD, 1 (50) 1 (50) Hemoperitoneum 2 (1.8) but without reaching statistical significance. Epicardial contusion 5 (4.6) 1 (20) 4 (80) Hemothorax 3 (2.7) 0 (0) 3 (100) There is limited data to help establish the risk of in- Pulmonary hematoma 2 (1.8) 1 (50) 1 (50) jury based on the size of patients. The biomechanical Liver laceration 3 (2.7) 1 (33.3) 2 (66.6) properties of the thorax have a great interindividual va- Liver hematoma 7 (6.4) 4 (57.1) 3 (42.9) riability25, not only explained by the composition and Spleen injury 0 0 0 properties of the tissues, but also by the geometric cha-

331 Azeli Y, et al. Emergencias 2019;31:327-334

Table 3. Risk factors in the total population and according to the presence of chest damage or severe visceral damage Total MRD SRD No SVD SVD N = 109 N = 40 N = 69 p N = 93 N = 16 p n (%) n (%) n (%) n (%) n (%) Epidemiological data Age [median (IQR)] 63 (49-70) 59 (56-68) 63 (49-75) 0.031 62 (48-69) 63 (46-81) 0.697 Women 35 (32.1) 10 (25.0) 25 (36.2) 0.226 26 (28.0) 9 (56.3) 0.025 CRA features Basic CPR time [min (IQR)] 10 (3-15) 7 (2-13) 10 (5-16) 0.303 10 (3-15) 8 (1-14) 0.457 Advanced CPR Time [min (IQR)] 25 (16-40) 29 (16-44) 25 (19-38) 0.815 24 (16-40) 29 (25-45) 0.086 Total CPR time [min (IQR)] 35 (25-50) 37 (21-50) 38 (25-52) 0.712 37 (25-49) 45 (27-53) 0.206 Witness CPR 41 (37.6) 12 (30) 29 (42.0) 0.319 32 (34.4) 9 (56.3) 0.301 Fibrinolysis in CPR 2 (1.8) 1 (2.5) 1 (1.4) 0.695 2 (2.1) 0 0.554 Medical history Osteoporosis 4 (3.6) 1 (2.5) 3 (4.3) 0.551 2 (2.1) 2 (12.5) 0.038 Anti-aggregate treatment 11 (10.1) 7 (17.5) 4 (5.7) 0.069 10 (10.7) 1(6.2) 0.579 Anticoagulant treatment 5 (4.6) 4 (10.0) 1 (1.4) 0.052 5 (5.4) 0 0.339 Anthropometric variables BMI [kg/m2 (IQR)] 28 (25-31) 29 (24-31) 28 (25-31) 0.784 28 (24-31) 31 (26-34) 0.163 DCP [mm (IQR)] 10 (7-13) 9 (7-13) 11 (7-14) 0.398 10 (7-14) 10 (6-12) 0.241 Abdominal circumference [cm (IQR)] 101 (91-108) 98 (89-108) 102 (92-107) 0.108 100 (90-107) 104 (95-107) 0.628 Abdominal circumference > 100 cm 58 (53.2) 15 (37.5) 43 (62.3) 0.017 48 (51.6) 10 (62.5) 0.445 Chest circumference [cm (IQR)] 100 (94-107) 99 (93-106) 101 (95-107) 0.192 100 (94-107) 102 (96-113) 0.628 Chest circumference > 101 cm 51 (46.7) 12 (30.0) 39 (56.5) 0.016 42 (45.2) 9 (56.3) 0.558 BMI - body mass index; DCP- depth of compression point; CPR - cardiopulmonary resuscitation; IQR - interquartile range. racteristics of the thorax. Age and obesity horizontalize adipose tissue are associated with improved survival the ribs, increasing the length of the costal arch26,27, rate after in-hospital cardiac arrest29. which leads to increased chest stiffness and greater vul- The tendency of CPR maneuvers to be prolonged nerability to bone fractures after trauma28. This would over time has increased concern about the adverse hae- explain the higher proportion of patients with elevated modynamic effect of serious chest injuries secondary to abdominal and thoracic perimeters who developed SRD CPR. Initial results from theoretical models have made it in our study. possible to initiate their specific study30. Research into Our results could guide resuscitation teams to maxi- strategies to customize the compression depth and mize the quality of CPR performed, especially in pa- compression point31 according to each patient’s needs tients with characteristics consistent with variables asso- could improve the quality of CPR in the future. ciated with serious injury, such as chest circumference > Our study has several limitations. On the one hand, 101 cm and female sex. the depth of the CCs given to patients is unknown, as Performing CPR on obese subjects is difficult and a there was no way to monitor the quality of the CCs. In real challenge for rescuers. It is plausible that the effort addition, the categorization of severe lesions has been required to obtain effective CC in this type of patient performed as in a previous study to facilitate compari- may increase the risk of fractures. It has recently been son, but their clinical usefulness is not yet well establi- described in adults that smaller transverse areas of the shed12. A possible inclusion bias cannot be ruled out as chest and shallower depth of subcutaneous thoracic 188 patients who did not have an autopsy during the

Table 4. Factors associated with the onset of severe bone damage Univariate analysis Multivariate analysis OR (95% CI) p OR (I5% CI) p Age 1.031 (1.002-1.060) 0.034 Sex 1.705 (0.716-4.060) 0.228 Total CPR time 1.009 (0.983-1.035) 0.517 Chest circumference 1.030 (0.994-1.067) 0.104 Chest circumference > 101 cm 2.786 (1.195-6.492) 0.018 2.456 (1.032-5.844) 0.042 Body Mass Index 0.993 (0.918-1.074) 0.866 Abdominal perimeter 1.025 (0.994-1.057) 0.110 Abdomen 100 2.530 (1.131-5.659) 0.024 Depth of the compression point 1.046 (0.958-1.142) 0.314 Fibrinolysis 0.574 (0.035-9.428) 0.697 Osteoporosis 2.000 (0.197-20.289) 0.558 Pre-aggregation treatment 0.301 (0.079-1.151) 0.079 Previous anticoagulant treatment 0.140 (0.015-1.331) 0.087 CI: 95% confidence interval; CPR: cardiopulmonary resuscitation; OR: odds ratio.

332 Azeli Y, et al. Emergencias 2019;31:327-334

Table 5. Factors related to the appearance of visceral damage Univariate analysis Multivariate analysis OR (IC 95%) p OR (IC 95%) p Age 1.009 (0.971-1.048) 0.645 Sex 3.313 (1.118-9.821) 0.031 5.016 (1.184-21.251) 0.029 Total CPR 1.013 (0.983-1.044) 0.396 Thoracic perimeter > 100 cm 1.378 (0.471-4.030) 0.559 Body Mass Index 1.065 (0.961-1.181) 0.230 Abdominal circumference > 100 cm 1.528 (0.513-4.552) 0.447 Depth of the compression point 0.909 (0.791-1.045) 0.180 Osteoporosis 7.125 (0.877-57.894) 0.066 Previous anti-aggregating treatment 0.544 (0.062-4.793) 0.584 SRD 10.833 (1.373-85.485) 0.024 CPR: cardiopulmonary resuscitation; SRD: severe ribcage damage. study period were not studied. Finally, this study does Article not commissioned by the Editorial Board and peer-reviewed externally not include survivors, which could have helped to bet- Acknowledgements: The ReCaPTa study has received a grant from the ter characterize different types of chest, provide data Consell Català de Ressuscitació. for the study of chest biomechanics during CPR, and We would like to thank the research team involved in the ReCaPTa understand the possible clinical relevance of lesions that study for their work and collaboration as authors: Xavier Escalada, Isaac occur during resuscitation. Lucas Guarque, from the Sistema d’Emergències Mèdiques de Catalun- ya; Inés Landin, Pilar Torralba, Cristina Amaya and Carlos Laguna, from As a conclusion, we can say that SRD secondary to the Instituto de Medicina Legal y Forense de Cataluña; Vanesa López manual CPR are frequent and associated with SVD. Madrid, from the Emergency Department of the Sant Joan Hospital in Reus; Jordi Bladé, from the Institut Català de Salut; Alex Gui-llem, from Patients with a thoracic perimeter greater than 101 cm the CAP in Cambrils (SAGESSA), and Laura Fernandez Sender, from the and women are at greater risk of suffering a severe in- Subacute Unit of the Tecla Network. We would also like to thank the Swedish national registry of out-of-hos- jury. Thoracic perimeter appears to be a new variable pital cardiac arrests for their collaboration, and Angels Mora and Xavier associated with chest lesions secondary to CPR with Jiménez Fábrega, of the Sistema d’Emergències Mèdiques de Catalunya, possible clinical implications. for their collaboration. Finally, we would like to thank the technical team, whose work helps us to make this record: Ruth Alvarez, Adnan Basic and Eva Hernando. Conflicting interests: The authors declare no conflicts of interest in relation to this article. Contribution of authors: All authors have confirmed their authorship Addendum in the author’s responsibilities document, publication agreement and transfer of rights to EMERGENCIAS. Additional data are available online at: http://dx.doi.org/ Financing: This study has received a grant for the best research project in the field of resuscitation sciences in 2014 awarded by the Consell Català de Ressuscitació. References Ethical Responsibilities: All authors confirm in the author’s responsibilities document, publication agreement and transfer of rights to EMER- 1 Rosell-Ortiz F, Escalada-Roig X, Fernández del Valle P, Sánchez-Santos GENCIAS that confidentiality and respect for patients’ rights have been L, Navalpotro-Pascual JM, Echarri-Sucunza A, et al. Out-of-hospital maintained, as well as international ethical considerations. cardiac arrest (OHCA) attended by mobile emergency teams with a

Table 6. Causes of death and Utstein variables and its relationship with the severity of lesions secondary to cardiopulmonary resuscitation Total MRD SRD No SVD SVD N = 109 N = 40 N = 69 p N = 93 N = 16 p n (%) n (%) n (%) n (%) n (%) Causes of death Cardiac 60 (55) 28 (70.0) 32 (46.4) 52 (55.9) 8 (50.0) Respiratory 7 (6.4) 2 (5.0) 5 (7.2) 5 (5.4) 2 (12.5) Drowning 14 (12.8) 1 (2.5) 13 (18.8) 13 (14) 1 (6.2) Intoxication 7 (6.4) 4 (10.0) 3 (4.3) 0.083 7 (7.5) 0 0.250 Liver-gastrointestinal 6 (5.5) 2 (5.0) 4 (5.8) 5 (5.4) 1 (6.2) Cerebral 4 (3.7) 1 (2.5) 3 (4.3) 4 (4.3) 0 (0) Other noncardiac 11 (10.0) 2 (5.0) 9 (13.0) 7 (7.5) 4 (25) Location Domicile 56 (54.9) 22 (64.7) 34 (50.0) 45 (52.3) 11 (68.8) Public place 39 (38.2) 10 (29.4) 29 (42.6) 0.367 37 (43.0) 2 (12.5) 0.020 Health Center 7 (6.9) 2 (5.9) 5 (7.4) 4 (4.7) 3 (18.8) Initial Rhythm Asystole 77 (78.6) 24 (70.6) 53 (82.8) 0.160 66 (80.5) 11 (68.8) 0.295 Ventricular Fibrillation 17 (17.3) 7 (20.6) 10 (15.6) 0.537 13 (15.9) 4 (25.0) 0.377 Pulseless electrical activity 3 (3.1) 2 (5.9) 1 (1.6) 0.237 2 (2.4) 1 (6.3) 0.418 SRD- serious ribcage damage; MRD- mild ribcage damage; SVD- severe visceral damage.

333 Azeli Y, et al. Emergencias 2019;31:327-334

physician on board. Results of the Spanish OHCA Registry ful resuscitation. Resuscitation. 2014;85:1708-12. (OSHCAR). Resuscitation. 2017;113:90-5. 17 Ihnát Rudinská L, Hejna P, Ihnát P, Tomášková H, Smatanová M, 2 Zuercher M, Hilwig RW, Ranger-Moore J, Nysaether J, Nadkarni VM, Dvořáček I. Intra-thoracic injuries associated with cardiopulmonary Berg MD, et al. Leaning during chest compressions impairs cardiac resuscitation – Frequent and serious. Resuscitation. 2016;103:66-70. output and left ventricular myocardial blood flow in piglet cardiac 18 Champigneulle B, Haruel PA, Pirracchio R, Dumas F, Geri G, Arnaout arrest. Crit Care Med. 2010;38:1141-6. M, et al. Major traumatic complications after out-of-hospital cardiac 3 Kralj E, Podbregar M, Kejžar N, Balažic J. Frequency and number of arrest: insights from the Parisian registry. Resuscitation. resuscitation related rib and sternum fractures are higher than gene- 2018;128:70-5. rally considered. Resuscitation. 2015;93:136-41. 19 Koga Y, Fujita M, Yagi T, Nakahara T, Miyauchi T, Kaneda K, et al. 4 Ram P, Menezes RG, Sirinvaravong N, Luis SA, Hussain SA, Madadin Effects of mechanical chest compression device with a load-distribu- M, et al. Breaking your heart—A review on CPR-related injuries. Am J ting band on post-resuscitation injuries identified by post-mortem Emerg Med. 2018;36:838-42. computed tomography. Resuscitation. 2015;96:226-31. 5 Hellevuo H, Sainio M, Nevalainen R, Huhtala H, Olkkola KT, 20 Cha KC, Kim YW, Kim HI, Kim OH, Cha YS, Kim H, et al. Tenhunen J, et al. Deeper chest compression – More complications Parenchymal lung injuries related to standard cardiopulmonary res- for cardiac arrest patients? Resuscitation. 2013;84:760-5. uscitation. Am J Emerg Med. 2017;35:117-21. 6 Møller Nielsen A, Rasmussen LS. Damage and depth of chest com- 21 Hoke RS, Chamberlain D. Skeletal chest injuries secondary to car- pressions. Resuscitation. 2013;84:713-4. diopulmonary resuscitation. Resuscitation. 2004;63:327-38. 7 Stiell IG, Brown SP, Nichol G, Cheskes S, Vaillancourt C, Callaway 22 Baubin M, Rabl W, Pfeiffer KP, Benzer A, Gilly H. Chest injuries after CW, et al. What is the optimal chest compression depth during active compression-decompression cardiopulmonary resuscitation out-of-hospital cardiac arrest resuscitation of adult patients? (ACD-CPR) in cadavers. Resuscitation. 1999;43:9-15. Circulation. 2014;130:1962-70. 23 Yamaguchi R, Makino Y, Chiba F, Torimitsu S, Yajima D, Inokuchi G, 8 Perkins GD, Handley AJ, Koster RW, Castrén M, Smyth MA, et al. Frequency and influencing factors of cardiopulmonary resusci- Olasveengen T, et al. European Resuscitation Council guidelines for tation-related injuries during implementation of the American Heart resuscitation 2015: Section 2. Adult basic life support and automa- Association 2010 guidelines: a retrospective study based on autopsy ted external defibrillation. Resuscitation. 2015;95:81-99. and postmortem computed tomography. Int J Legal Med. 9 Azeli Y, Barbería E, Jiménez-Herrera M, Bonet G, Valero-Mora E, 2017;131:1655-63. Lopez-Gomariz A, et al. The ReCaPTa study - A prospective out of 24 Setälä P, Hellevuo H, Huhtala H, Kämäräinen A, Tirkkonen J, Hoppu hospital cardiac arrest registry including multiple sources of survei- S. Risk factors for cardiopulmonary resuscitation-related injuries sus- llance for the study of sudden cardiac death in the Mediterranean tained during out-of-hospital cardiac arrests. Acta Anaesthesiol area. Scand J Trauma Resusc Emerg Med. 2016;24:127. Scand. 2018;62:1290-6. 10 Soar J, Nolan JP, Böttiger BW, Perkins GD, Lott C, Carli P, et al. 25 Beesems SG, Hardig BM, Nilsson A, Koster RW. Force and depth of European Resuscitation Council guidelines for resuscitation 2015: mechanical chest compressions and their relation to chest height Section 3. Adult advanced life support. Resuscitation. and gender in an out-of-hospital setting. Resuscitation. 2015;95:100-47. 2015;91:67-72. 11 Perkins GD, Jacobs IG, Nadkarni VM, Berg RA, Bhanji F, Biarent D, et 26 Holcombe SA, Wang SC, Grotberg JB. The effect of age and demo- al. Cardiac arrest and cardiopulmonary resuscitation outcome re- graphics on rib shape. J Anat. 2017;231:229-47. ports: update of the Utstein resuscitation registry templates for 27 Shi X, Cao L, Reed MP, Rupp JD, Hoff CN, Hu J. A statistical human out-of-hospital cardiac arrest. Resuscitation. 2015;96:328-40. rib cage geometry model accounting for variations by age, sex, sta- 12 Azeli Y, Barbería E, Landín I, Torralba P, Amaya C, Laguna C. Lesiones ture and body mass index. J Biomech. 2014;47:2277-85. por reanimación cardiopulmonar en autopsias forenses: protocolo 28 Wang S, Poster CS, Lange AW, Brede C, Lange D. Structural and del Registro Clínico-Patológico de Tarragona (ReCaPTa). Rev Esp material changes in the aging thorax and their role in crash protec- Med Leg. 2016;42:120-5. tion for older occupants. Stapp Car Crash J. 2005;49:231-49. 13 Koster RW, Beenen LF, van der Boom EB, Spijkerboer AM, Tepaske R, 29 Liu JZ, Ye S, Cheng T, Han TY, Li Q, Li RX, et al. The effects of thora- van der Wal AC, et al. Safety of mechanical chest compression devi- cic cage dimension and chest subcutaneous adipose tissue on outco- ces AutoPulse and LUCAS in cardiac arrest: a randomized clinical trial mes of adults with in-hospital cardiac arrest: A retrospective study. for non-inferiority. Eur Heart J. 2017;38:3006-13. Resuscitation. 2019;141:151-7. 14 Krischer JP, Fine EG, Davis J, Nagel EL. Complications of cardiac res- 30 Azeli Y, Lorente-Olazabal JV, Monge-García MI, Bardají A. Adverse uscitation. Chest. 1987;92:287-91. haemodynamic effects of serious thoracic injuries during CPR: a 15 Smekal D, Hansen T, Sandler H, Rubertsson S. Comparison of compu- study based on the Campbell diagram. Resuscitation. 2019 (En ted tomography and autopsy in detection of injuries after unsuccessful prensa). cardiopulmonary resuscitation. Resuscitation. 2013;84:357-60. 31 Azeli, Y, Suazo, M, Puigjaner, D, Amejide, A, Herrero, J, Fortuny G. 16 Smekal D, Lindgren E, Sandler H, Johansson J, Rubertsson S. CPR- Assessing the safety of a displaced chest compression to the left and related injuries after manual or mechanical chest compressions with caudally. First results using a detailed thoracic finite element model. the LUCASTM device: A multicentre study of victims after unsuccess- Resuscitation. 2017;118S:e5-42.

334