Author + information
- Received February 5, 2016
- Revision received July 15, 2016
- Accepted August 15, 2016
- Published online February 20, 2017.
- Muhammad Shahreyar, MDa,
- Geetanjali Dang, MDa,
- Mohammad Waqas Bashira,
- Gagan Kumar, MDb,
- Jawad Hussain, MDa,
- Shahryar Ahmad, MDa,
- Beneet Pandey, MSc,
- Atul Thakur, MDd,
- Sanjay Bhandari, MDa,
- Krishna Thandra, MDa,
- Jasbir Sra, MDc,
- Abdul J. Tajik, MDc and
- Arshad Jahangir, MDc,e,∗ ()
- aDivision of Hospital Medicine, Department of General Internal Medicine, Medical College of Wisconsin, Milwaukee, Wisconsin
- bPhoebe Putney Memorial Hospital, Albany, Georgia
- cAurora Cardiovascular Services, Aurora Sinai/Aurora St. Luke’s Medical Centers, Milwaukee, Wisconsin
- dSaint Mary’s Hospital, Department of General Internal Medicine, Cleveland Clinic, Cleveland, Ohio
- eCenter for Integrative Research on Cardiovascular Aging, Aurora Sinai/Aurora St. Luke's Medical Centers, Aurora Health Care, Milwaukee, Wisconsin
- ↵∗Address for correspondence:
Dr. Arshad Jahangir, Sheikh Khalifa bin Hamad Al Thani Center for Integrative Research on Cardiovascular Aging, and Aurora Cardiovascular Services, 2801 West Kinnickinnic River Parkway, Milwaukee, Wisconsin 53215.
Objectives This study sought to assess the impact of morbid obesity on outcomes in patients with in-hospital cardiac arrest (IHCA).
Background Obesity is associated with increased risk of out-of-hospital cardiac arrest; however, little is known about survival of morbidly obese patients with IHCA.
Methods Using the Nationwide Inpatient Sample database from 2001 to 2008, we identified adult patients undergoing resuscitation for IHCA, including those with morbid obesity (body mass index ≥40 kg/m2) by using International Classification of Diseases 9th edition codes and clinical outcomes. Outcomes including in-hospital mortality, length of stay, and discharge dispositions were identified. Logistic regression model was used to examine the independent association of morbid obesity with mortality.
Results Of 1,293,071 IHCA cases, 27,469 cases (2.1%) were morbidly obese. The overall mortality was significantly higher for the morbidly obese group than for the nonobese group experiencing in-hospital non-ventricular fibrillation (non-VF) (77% vs. 73%, respectively; p = 0.006) or VF (65% vs. 58%, respectively; p = 0.01) arrest particularly if cardiac arrest happened late (>7 days) after hospitalization. Discharge to home was significantly lower in the morbidly obese group (21% vs. 31%, respectively; p = 0.04). After we adjusted for baseline variables, morbid obesity remained an independent predictor of increased mortality. Other independent predictors of mortality were age and severe sepsis for non-VF and VF group and venous thromboembolism, cirrhosis, stroke, malignancy, and rheumatologic conditions for non-VF group.
Conclusions The overall mortality of morbidly obese patients after IHCA is worse than that for nonobese patients, especially if IHCA occurs after 7 days of hospitalization and survivors are more likely to be transferred to a skilled nursing facility.
- cardiopulmonary resuscitation
- in-hospital cardiac arrest
- morbid obesity
- Nationwide Inpatient Sample
- ventricular fibrillation
Obesity is highly prevalent in the U.S. population and has been associated with increased prevalence of cardiovascular risk factors (1,2) and higher mortality in the general population (3). On the basis of body mass index (BMI), patients are subclassified as being underweight (BMI: <18.5 kg/m2), in a normal range (BMI: 18.5 to 24.99 kg/m2), overweight (BMI: 25 to 29.99 kg/m2), obese class I (BMI: 30 to 34.99 kg/m2), obese class II (BMI: 35 to 39.99 kg/m2) or obese class III (BMI: ≥40 kg/m2) (4). Nearly 35% of the adult population in the United States is obese, and 6.4% are morbidly obese, defined as having BMI ≥40 kg/m2 (5). Although the association between morbid obesity and increased risk of sudden cardiac death is recognized (2,6), little is known about outcomes after cardiac arrest in morbidly obese patients. Theoretically, obese patients are more prone to difficulties in airway establishment and chest compressions in case of sudden cardiac arrest, and the problem is likely to worsen as the patient’s weight increases. Several studies have defined an obesity paradox in patients with cardiovascular disease, where obese patients with established cardiovascular conditions including heart failure (7), atrial fibrillation (8), coronary artery disease (9), and hypertension (10) were found to have a better prognosis than nonobese or nonoverweight patients. However, the relationship between survival and BMI is U-shaped, with both morbidly obese and lean counterparts showing worse outcomes at extremes of weight (1). Such a relationship between morbid obesity and survival outcomes for in-hospital cardiac arrest (IHCA) patients is not fully defined, but important to determine because of the presumed difficulty with morbid obesity-related body habitus on cardiopulmonary resuscitation. Therefore, we suspected that the overall outcomes, including mortality, length of hospital stay, and non-home discharge disposition would be significantly worse in morbidly obese patients than in nonobese patients after IHCA requiring cardiopulmonary resuscitation. A nationally representative database developed by the Agency of Healthcare Research and Quality was used to address this question to ensure sufficient sample size that enhanced generalization of results (11). The primary aim of the study was to determine the mortality differences between morbidly obese and nonobese patients with IHCA early (within 48 h) or late (1 week) after hospitalization due to ventricular fibrillation (VF) or non-VF arrest. Secondary outcomes of interest were length of stay, and discharge disposition.
We used the Nationwide Inpatient Sample (NIS), the largest all-payer inpatient care database publicly available in the United States. This administrative dataset was created by the Agency for Healthcare Research and Quality as part of the Healthcare Cost and Utilization Project and contains data for 5 to 8 million hospital stays from approximately 1,000 hospitals sampled, to approximate a 20% stratified sample of U.S. community hospitals. NIS includes appropriate weights to allow the production of national estimates. Each hospitalization is treated as an individual entry in the database and includes International Classification of Diseases-9th edition-Clinical Modification (ICD-9-CM) codes for the principal diagnosis and up to 14 secondary diagnoses and 15 procedures associated with that stay. NIS includes information for all hospitalizations at participating hospitals, regardless of payer, including private insurance and the uninsured. Details about the structure of Healthcare Cost and Utilization Project NIS database are available online. We used data from 2001 to 2008 for this study. Because we used publicly available data with no patient identifiers, this study was exempt from formal review by the Institutional Review Board of our institution.
We used ICD-9-CM procedure codes 99.60 and 99.63 to identify adult patients 18 years of age or older who had cardiac arrest and underwent cardiopulmonary resuscitation (CPR). These codes represent in-hospital CPR. Because patients with non-VF arrest have significantly different and worse outcomes than those with VF arrest (12,13), they were analyzed separately. We identified those with ventricular flutter using ICD-9-CM code 427.41 and 427.42. To address the main question regarding the impact of morbid obesity that can compromise cardiopulmonary resuscitation compared to that in nonobese group, we focused on those with morbid obesity (BMI: ≥40 kg/m2; ICD-9-CM code 278.01) and excluded intermediate groups (overweight or obese; BMI: 25 to 39 kg/m2) with ICD-9-CM codes 278.00 and 278.02. The study question was addressed by carefully matching the morbidly obese group with the nonobese group for various demographic and comorbid conditions. Missing values with regard to mortality were excluded.
Definition of variables
We used NIS variables to identify patient characteristics of age, sex, race, and insurance status. Information about race is omitted from 20% to 25% of records in any given year. Those with missing race information were grouped together into the category “unknown.” We identified individual comorbidities using ICD-9-CM codes. Acute medical conditions associated with cardiac arrest were identified using standard ICD-9-CM codes, and we used those conditions for adjustment in matching. We identified presence of severe sepsis, venous thromboembolism, pneumonia, pneumothorax, and cardiac tamponade. Because severe sepsis is one of the common causes of cardiac arrest, we used ICD-9-CM codes according to the scheme validated by Martin et al. (14). We used number of organs failing as the surrogate for severity of acute disease. Prolonged mechanical ventilation (≥96 h) was identified using ICD-9-CM code 96.72.
Our primary outcome of interest was in-hospital mortality. Secondary outcomes included hospital length of stay, and discharge disposition. We grouped discharge disposition as home, home with health care, transfers to other healthcare facilities (under which NIS includes skilled nursing facilities, intermediate care, inpatient rehabilitation, psychiatric hospitals, inpatient hospice, and others), transfers to other acute care hospitals, and other disposition (which includes against medical advice, unknown, and missing). Because cardiac arrest can occur anytime during the hospital admission and the outcomes may vary during different times, we examined the outcomes with respect to timing of the CPR with the admission day. We thus studied outcomes in 3 time categories: those who had CPR within 48 h of admission, those who underwent CPR between days 3 and 7 after admission, and those who underwent CPR after day 7 of admission.
We performed all statistical analyses using Stata IC version 11.0 software (Stata Corp., College Station, Texas). For our descriptive analyses, we used Pearson’s chi-square test and Student t test to compare categorical and continuous variables as appropriate for their distribution. Because length of stay was not normally distributed, we used Wilcoxon rank sum test for comparison and report results as medians and interquartile ranges. Statistical significance for all tests was defined as a p value of <0.05.
We analyzed VF and non-VF arrest separately. We used greedy nearest-neighbor matching with replacement to create matched cases (those with morbid obesity) and controls (not morbidly obese). We restricted the matches using caliper matching within 0.01. We checked the matched groups for balance and common support. The groups were matched for age, sex, race, insurance, hospital bed size, teaching hospital, year, and presence of comorbid conditions, including severe sepsis, pneumonia, venous thromboembolism, pneumothorax, cardiac tamponade, chronic lung disease, obstructive sleep apnea, hypertension, coronary artery disease, congestive heart failure, atrial fibrillation, stroke, peripheral artery disease, diabetes mellitus, hyperlipidemia, smoking, alcoholism, collagen vascular diseases, rheumatic disease, human immunodeficiency virus infection, paraplegia/hemiplegia, solid organ cancer, metastatic cancer, and lymphoma/leukemia and severity of disease by using number of organ failures.
We then used separate multivariate logistic regression models to elucidate the independent association of morbid obesity with in-hospital mortality in patients receiving CPR for VF and non-VF arrest. The model was adjusted for age, sex, insurance, Charlson-Deyo’s comorbidity index, number of organ failures, and year. Sensitivity analyses were performed using individual comorbidities in place of Charlson-Deyo’s comorbidity index and by excluding the variables race and insurance. We performed similar regression analyses in the 3 time category subgroups (Figures 1B and 2B⇓⇓).
Impact of morbid obesity on mortality, length of hospital stay, and discharge disposition in non-VF arrest patients
There were 552,083 patients who underwent CPR for non-VF arrest between 2001 and 2008. Of these, 13,662 patients (2.5%) were morbidly obese. We matched 13,633 morbidly obese patients with nonobese people in demographic and clinical characteristics as described above. Both of the groups were well matched in age, sex, insurance status, hospital characteristics, and comorbidity burden. The severity of disease was also similar in the 2 groups (Table 1).
The overall mortality was significantly higher in morbidly obese group than in the nonobese group (76.5% vs. 73.2%, respectively; p = 0.006). After we adjusted for demographic and clinical characteristics, the odds of mortality were 1.21 times higher for morbidly obese group (95% confidence interval [CI]: 1.07 to 1.37), and this difference was mostly accounted by the CPR outcomes that were performed later (>7 days) during the hospital stay (Figures 1A and 1B). The mortality in morbidly obese patients undergoing in-hospital CPR for non-VF arrests was not significantly higher than that for their counterparts if the CPR happened to be in the first 48 h or ≤7 days (Figure 1). There was also a significantly higher rate of discharge to skilled nursing facilities (56.3% vs. 50.5%, respectively; p = 0.04). Discharge to home was significantly lower in the morbidly obese (21.4% vs. 30.5%, respectively; p = 0.04) in the group with CPR for non-VF arrest after seven days of hospitalization (Figure 3).
The median length of hospital stay (4 days vs. 5 days, respectively; p = 0.06) and total hospital charges ($33,797 vs. $36,436, respectively; p = 0.07) were not significantly different between the 2 groups (Table 2).
Impact of morbid obesity on mortality, length of hospital stay, and discharge disposition in VF arrest patients
There were 115,319 patients who underwent in-hospital CPR for VF arrest between 2001 and 2008. Of these, 2,681 patients (2.3%) were morbidly obese. We matched 2,676 morbidly obese patients with 2,516 patients who were not morbidly obese in demographics and clinical characteristics like age, sex, race, insurance type, hospital characteristics, comorbidity burden, and severity of disease (Table 3).
The mortality was significantly higher in the morbidly obese group than in the group that was not morbidly obese (65% vs. 58%, respectively; p = 0.01). After we adjusted for demographic and clinical characteristics, the odds of mortality were 1.45 times higher for the morbidly obese group (95% CI: 1.12 to 1.87). Similar to non-VF arrests, the difference was accounted for mostly by the CPR performed later (>7 days) during the hospital stay. The mortality in morbidly obese persons undergoing in-hospital CPR for VF arrests was not significantly higher than the counterparts who were not morbidly obese if the CPR occurred in the first 7 days (Figure 2). There were no other significant differences between the 2 groups in disposition of survivors to home or to a facility.
The median lengths of hospital stay (4 days for both) and total hospital charges ($44,737 vs. $48,339, respectively; p = 0.65) were not significantly different between the 2 groups (Table 4).
Predictors of mortality in the non-VF and VF arrest groups
On the multivariate logistic regression model of the overall cohort, higher odds of mortality were found in the morbidly obese group than in the nonobese group undergoing CPR for non-VF (adjusted odds ratio [OR]: 1.22; 95% CI: 1.08 to 1.39) or VF arrest (adjusted OR: 1.42; 95% CI: 1.09 to 1.85). Within the non-VF arrest group, patients with venous thromboembolism, severe sepsis, cirrhosis, stroke, hematological malignancy, and rheumatologic conditions had a significantly higher mortality (Table 5). In the VF-arrest group, patients with diabetes mellitus and severe sepsis had significantly higher mortality (Table 6). Similarly, increasing age was associated with a significant increase in mortality in both the VF (adjusted OR: 1.02; 95% CI: 1.01 to 1.04) and non-VF arrest groups (adjusted OR: 1.01; 95% CI: 1.01 to 1.02). Males were associated with significantly lower mortality in both of the groups. Other significant predictors of mortality are summarized in Tables 5 and 6.
The main finding of this study is that morbid obesity is associated with increased mortality after IHCA due to either non-VF or VF arrest if it occurs late during hospitalization, and among survivors, discharge to home is significantly lower. Advanced age and severe sepsis are independent predictors of increased mortality for both non-VF and VF arrest patients, whereas venous thromboembolism, cirrhosis, stroke, hematological malignancy, and rheumatologic conditions are independent predictors of mortality in non-VF group. Male sex is associated with significantly lower mortality in both groups.
The overall prevalence of obesity is increasing, and within the obese group, the relative increase in morbidly obese individuals has been the greatest (15–18). Morbid obesity is a risk factor for adverse cardiovascular outcome (2,19), and morbidly obese patients have poorer prognosis than nonobese patients after acute coronary events, percutaneous interventions, and coronary artery bypass graft surgery (20–23). The outcome for patients after IHCA is better than out-of-hospital cardiac arrest (19). However, the impact of morbid obesity on survival and discharge disposition after IHCA is not clear. This is an important question because the efficacy of chest compressions and airway protection may be compromised in morbidly obese individuals because of their body habitus; therefore, the simple goal for this study was to determine the impact of morbid obesity (BMI: ≥40 kg/m2) on resuscitation outcomes in patients who had an IHCA compared to that in a nonobese group. The study was not designed to study the effect of various BMI groups on resuscitation outcomes and, therefore, intermediate groups, that is, overweight or obese patients (BMI: 25 to 39 kg/m2) were excluded from the analysis. The study question was addressed by carefully matching the morbidly obese group with the nonobese group for various demographic and comorbid conditions from a large cohort of patients in the NIS database, identified to have IHCA due to non-VF or VF arrest within 48 h, 3 to 7 days, and >7 days after hospitalization. Both VF and non-VF arrests were associated with higher risk of death in the morbidly obese group. The timing of IHCA had important implications, as higher mortality was seen with IHCA later during hospitalization, and in the morbidly obese group the survival was lower than that of the nonobese group if IHCA occurred more than 1 week after hospitalization. Compared to VF arrest, mortality was higher in the non-VF arrest group in both the morbidly obese and nonobese groups. This is consistent with what has been previously reported among out-of-hospital cardiac arrest patients (12,13). Among the survivors of cardiac arrest, the hospital discharge disposition was also different for the VF than for the non-VF groups, with a greater number of VF arrest patients dismissed to home than non-VF arrest patients in both morbidly obese and obese groups (40% to 45% vs. 29% to 33%, respectively) (Tables 2 and 4). In the non-VF group, more morbidly obese patients were discharged to facilities (Table 2), indicating worse functional status and outcomes, which is also reflected by a trend toward greater length of hospital stay and median hospital charges than in the nonobese group.
These results that highlight differences in outcomes between morbidly obese and nonobese group add to the limited information available on outcomes in morbidly obese patients after IHCA. A similar impact on survival in obese children with BMI >95th percentile was previously reported by Srinivasan et al. (24). Morbid obesity was associated with difficulty with maintaining airway, venous access, increased frequency of epinephrine injections, longer CPR duration, worse event survival and survival to hospital discharge in obese children. In a separate study, outcomes in adult patients with out-of-hospital cardiac arrest and return of spontaneous circulation were reported based on BMI without any significant difference in survival between patients with normal BMI and those who were overweight and in which obesity was defined as BMI of ≥30 kg/m2, although overweight group had better neurological outcomes. No information for those with morbid obesity (BMI: ≥40 kg/m2) was presented (25). The impact of severe obesity (defined as BMI >35 kg/m2) on poor survival from cardiac arrest due to shockable but not unshockable rhythm was reported in another large study (26). Our findings in a large cohort of patients hospitalized regardless of the reason for admission, demonstrating increased odds of mortality in morbidly obese group after IHCA compared to nonobese group is consistent with these prior reports. The mortality was higher in those with IHCA later in hospitalization, suggesting delayed cardiac arrest is associated with worse outcomes and the presence of morbid obesity is associated with higher mortality. In-hospital care of morbidly obese patients is more challenging because of their body habitus in the setting of acute emergency. There is a potential of facing difficult airway access and management with delays in intubation and/or administration of vasoactive medications in morbidly obese patients that may affect resuscitation outcomes. Chest compression depth may also affect circulatory support in the morbidly obese compared to that in nonobese during resuscitation. However, preliminary reports of compression depth and the duration of interrupted CPR do not seem to suggest that these are of inferior quality in very obese patients (27). Higher thoracic impedance associated with increase in BMI may also reduce defibrillation success (26,28,29). Dargin et al. (30) studied the effect of BMI on intubation success rate and reported that obesity predicted difficulty in intubation and higher post-intubation complications including a higher risk of pneumonia, acute respiratory distress syndrome, and other in-hospital complications in obese and morbidly obese patients (30). The reason for increased mortality after IHCA late (>7 days) into hospitalization in morbidly obese patients in our study, however, is not clear as this information could not be ascertained from the NIS, which is a limitation inherent to the nature of this database used for this hypothesis-generating study. It might be due to worsening of disease as hospitalization continues or increased risk of arrest or other complications while recovering from acute hospitalization.
We could not find any difference in severity of illness at hospitalization between the nonobese and obese groups who had VF arrest as judged by number of organ failures, but advanced age, severe sepsis and diabetes were predictive of poor survival. In the non-VF group, the presence of venous thromboembolism, severe sepsis, cirrhosis, stroke, hematological malignancy and rheumatologic conditions had poor survival. Information about the duration of stay within the intensive care setting, progression of disease or whether arrest occurred more often out of an intensive care unit late after hospitalization could not be obtained. Because resuscitation is likely more successful if the arrest occurs in an intensive care unit (ICU) or a setting more experienced with cardiac resuscitation (cardiology versus orthopedic ward), it is possible that greater mortality late in hospitalization is due to its occurrence more frequently in a non-ICU environment. This is an interesting hypothesis that could not be tested with the NIS database that only provides the overall length of stay in the hospital and not the ICU length of stay. This and other questions, including causes of cardiac arrest and mortality, can be better addressed using an institutional database with precise information about the IHCA.
Our data should be interpreted cautiously and in the right clinical context, as we must acknowledge some important limitations. Despite careful use of a well-characterized database along with good matching between morbidly obese and nonobese with non-VF or VF arrest, the information is from an administrative database with risk of miscoding leading to underestimation or overestimation of diagnoses or procedures. The large sample size of our study; however, may have compensated, to some degree, for some of these limitations. Information about specific causes of death or circumstances of IHCA, duration of intensive care stay, progression of the disease severity during the course of hospitalization, duration of CPR, time to return of spontaneous circulation, medications used during CPR and the frequency of their administration was not available. Therefore, whether differences in outcome early vs. late during hospitalization between the morbidly obese and nonobese group were due to clinical or resuscitation-related factors could not be determined. This information is important and should be obtained from large institutional-based medical record systems where information that is more precise could help identify factors that influence mortality.
Despite its limitations, this study using a large national database demonstrates a small but significant risk for higher mortality in those with morbid obesity with both VF and non-VF arrest, especially late during hospitalization. Further studies to identify clinical and resuscitation-related factors associated with higher mortality are needed to better prepare for in-hospital management of this expanding population cohort. Morbid obesity is associated with increased risk of sudden cardiac death; however, little is known about the impact of morbid obesity on survival of patients who experience IHCA. Using a nationwide database of hospitalized patients and matching of morbidly obese with nonobese patients for baseline characteristics and comorbidities and adjusting for potential confounding factors, this study provides unique information on higher mortality after ventricular fibrillation and non-VF arrest, especially if it occurs late after hospitalization.
COMPETENCY IN MEDICAL KNOWLEDGE: A knowledge gap exists about the impact of morbid obesity on survival after an IHCA event and the study provides unique information on a small but significant risk for higher mortality in those with morbid obesity when compared to nonobese patients with both ventricular fibrillation and non-VF arrest, especially late during hospitalization.
TRANSLATIONAL OUTLOOK: Further studies of patient characteristics, comorbidities, causes of IHCA, and duration of CPR which could potentially impact outcome in patients having a cardiac arrest while hospitalized are warranted to better prepare for in-hospital management of this catastrophic event in morbidly obese and nonobese patients.
The authors acknowledge Susan Nord and Jennifer Pfaff, Aurora Cardiovascular Services, for editorial preparation, and Brian Miller and Brian Schurrer, Aurora Research Institute, for their help in preparing figures.
Dr. Jahangir is supported by U.S. National Institutes of Health National Heart, Lung, and Blood Institute and Aurora Health Care intramural funding.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors' institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- body mass index
- in-hospital cardiac arrest
- Nationwide Inpatient Sample
- ventricular fibrillation
- Received February 5, 2016.
- Revision received July 15, 2016.
- Accepted August 15, 2016.
- 2017 American College of Cardiology Foundation
- Lavie C.J.,
- Milani R.V.,
- Ventura H.O.
- Poirier P.,
- Giles T.D.,
- Bray G.A.,
- et al.
- ↵World Health Organization. Global Database on Body Mass Index: Body Mass Index Classifications. 2006. Available at: http://apps.who.int/bmi/index.jsp?introPage=intro_3.html. Accessed August 10, 2015.
- Lavie C.J.,
- De Schutter A.,
- Patel D.A.,
- Romero-Corral A.,
- Artham S.M.,
- Milani R.V.
- ↵National Inpatient Sample (NIS) Database Documentation. Healthcare Cost and Utilization Project (HCUP). National Inpatient Sample. Available at: http://www.hcup-us.ahrq.gov/db/nation/nis/nisdbdocumentation.jsp. Accessed June 28, 2015.
- Go A.S.,
- Mozaffarian D.,
- Roger V.L.,
- et al.
- Das S.R.,
- Alexander K.P.,
- Chen A.Y.,
- et al.
- Lavie C.J.,
- Milani R.V.,
- Ventura H.O.
- Payvar S.,
- Kim S.,
- Rao S.V.,
- et al.
- Buschur M.E.,
- Smith D.,
- Share D.,
- et al.
- Srinivasan V.,
- Nadkarni V.M.,
- Helfaer M.A.,
- et al.
- Testori C.,
- Sterz F.,
- Losert H.,
- et al.
- Jain R.,
- Nallamothu B.K.,
- Chan P.S.,
- American Heart Association National Registry of Cardiopulmonary Resuscitation (NRCPR) investigators
- Edelson D.P.,
- Abella B.S.,
- Kim S.,
- VadenHoek T.L.,
- Becker L.B.