CLINICAL INVESTIGATION
Cardiopulmonary Resuscitation in the Hospitalized Patient: Impact of System-Based Variables on Outcomes in Cardiac Arrest Phani Kantamineni, MD, Vamsi Emani, MD, Ankur Saini, MD, Hardeep Rai, MD and Abhijit Duggal, MD, MPH
Abstract: Background: A better understanding of the factors affecting the outcome of inpatient cardiopulmonary resuscitation (CPR) is crucial in making key clinical decisions. We aim to study the impact of various patient-related and hospital-related variables in a community-based teaching setup that could affect the prognosis of in-hospital cardiac arrests. Methods: We analyzed the data on all patients who experienced cardiac arrest while hospitalized at a community teaching hospital in Youngstown, Ohio. A multivariable logistic regression was performed to identify patient- and system-based variables associated with mortality in inpatient cardiac arrest. Results: A total of 417 in-hospital cardiopulmonary arrests were recorded during the study period. We analyzed 299 events in our final sample. One hundred sixty-four patients (54.8%) achieved return of spontaneous circulation and 137 (48.5%) survived the cardiopulmonary arrest for at least 24 hours. The duration of CPR, age, initial rhythm, witnessed events and sex were strongly associated with mortality in our univariate analysis. After adjustment for age, location and whether the code was witnessed, the timing of the week, initial rhythm, the duration of CPR and the sex of the patient retained prognostic significance in predicting the mortality. Conclusions: In our study, we report a 17.4% survival to hospital discharge after an in-hospital cardiopulmonary arrest and subsequent CPR, similar to rates reported in larger multicenter studies. Prolonged duration of CPR (.10 minutes) and male sex were found to be associated with worse outcomes. We report the impact of system-based variables such as physician and nursing staffing during different days of the week, on survival in these patients. Key Indexing Terms: Cardiopulmonary resuscitation; Inpatient cardiac arrest; Mortality. [Am J Med Sci 2014;348(5):377–381.]
T
he introduction of basic life support and advanced cardiac life support (ACLS) as guidelines for a standardized response to all cardiac arrests have had a dramatic impact on the immediate survival of patients. These gains, however, have not been as remarkable in hospitalized patients who undergo a cardiac arrest.1 This lack of improvement in outcome measures has in part been explained by the impact of an aging and sicker inpatient population offsetting the advances in medicine over the past 2 decades.1 The in-hospital incidence of cardiac arrest is estimated to be between 1 and 5 events per 1000 hospital admissions,2 and despite the introduction of advanced directives as part of standard patient care, most of these patients From the Department of Internal Medicine (PK, VE, AS), ValleyCare Health System/NEOMED, Northside Medical Center, Youngstown, Ohio; and Medical Intensive Care Unit (HR, AD), Respiratory Institute, Cleveland Clinic, Cleveland, Ohio. Submitted September 25, 2013; accepted in revised form March 24, 2014. The authors have no financial or other conflicts of interest to disclose. Correspondence: Abhijit Duggal, MD, MPH, Medical Intensive Care Unit, Respiratory Institute, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195 (E-mail: [email protected]).
The American Journal of the Medical Sciences
end up receiving cardiopulmonary resuscitation (CPR) as part of their treatment plan.3 It is important to understand the likelihood of success of CPR in patients admitted to the hospital, as they are much sicker and perhaps have different factors associated with cardiopulmonary arrest compared with the general population.4 Recent data have identified variables such as the initial rhythm on record, early initiation of CPR, early return of spontaneous circulation (ROSC) and a younger age to be associated with better outcomes.5–7 But there are only sparse data on the impact of the complex interplay between the patient-related variables (age, sex, etc), the hospital-based variables (place and timing of arrest in a hospital) and the system-based variables (size of the hospital, academic affiliation), on the overall outcomes of inpatients undergoing CPR.4,5,8,9 A better understanding of these factors should translate into a more consistent response to inpatient cardiac arrest in different hospital settings. Also it might lead to a more comprehensive and appropriate dialog between physicians and the patients when they discuss the factors that determine a patients’ wish to proceed with life-sustaining measures. In our study, we attempt to identify the factors that might affect outcomes in hospitalized patients suffering a cardiopulmonary arrest, in a community-based teaching setup. We also compare these results with outcomes reported from larger databases, and tertiary academic centers, to see if there are any differences in the reported outcomes.
MATERIALS AND METHODS We performed a retrospective chart review from January 2005 to December 2007, for all patients on whom a “code blue” CPR protocol was activated at Northside Medical Center, Youngstown, Ohio. All the nurses and physicians working at our hospital are required to be basic life support and ACLS certified. An express (code blue) team responds to all code blue calls. On weekdays, this team comprises a senior medical resident, a critical care nurse, an anesthesia and a critical care attending. Over the weekends, the team is modified and comprises a senior medical resident, a critical care nurse and a certified nurse anesthetist. We excluded patients less than 18 years and instances of inappropriate “false code blues” (situations where CPR protocol was activated, but after initial evaluation was not implemented; as the patients were hemodynamically intact). We only evaluated the first episode of CPR in patients who underwent multiple instances of cardiopulmonary arrest in a 24-hour period. Definitions Definitions of cardiac arrest, witnessed arrest, CPR, chest compressions, date of the arrest, location of the arrest, first monitored rhythm, shockable/nonshockable rhythm, ROSC and survival to hospital discharge were used based on the Utstein guidelines for standardized reporting in CPR.10 There is a wide
Volume 348, Number 5, November 2014
377
Kantamineni et al
variation in the reporting of ROSC, especially in hospitalized patients as a number of them can suffer multiple cardiopulmonary arrests over a short period of time. We defined the ROSC as the presence of a palpable rhythm after 20 minutes of successful CPR. We describe the survival at 24 hours after a successful CPR and also survival at discharge. We decided a priori that to better understand the impact of in-hospital cardiopulmonary arrest in terms of system-based variables, mortality at 24 hours after ROSC was a better time point to describe. For our univariate and multivariable analyses, we use survival at 24 hours as the outcome of interest. This measure is validated by the Utstein guidelines, as a measurable variable.10 We also describe survival to hospital discharge based on these guidelines.10 We used the cutoff time for CPR as #10 minutes or .10 minutes, as it has been recognized as a prognostic factor in previous studies.11 We decided to use 07:00 to 19:00 hours, and 19:01 to 06:59 hours as cutoff points for time, as medical staff and residents change their shifts at this time. Weekends were defined as from 19:01 hours on Friday until 06:59 hours on Monday, which closely mirrors existing literature.9 Univariate Analysis Descriptive statistics are reported as frequency (percentages) for categorical variables. Continuous variables are reported as either means (standard deviation) or median (interquartile ranges) based on the distribution of the individual variables. We used a Student t test, or Wilcoxon’s rank sum test as appropriate for continuous variables, and x2 test or Fisher’s exact test for discrete variables. Confidence intervals (CI) and P values reflect a 2-tailed a level of 0.05. Statistical analyses were conducted using SAS version 9.3 (SAS Institute, Cary, NC). Multivariable Analysis Patients with missing data were excluded from the regression model. We included all the clinically significant prognostic variables with a statistically significant result (P 5 0.20) on the univariate analysis. We used the “limiting sample size” rule to determine the number of prognostic variables associated with mortality on the univariate analysis and used these in a multiple logistic regression model. Before including these factors in our multivariable model, we tested for multicolinearity by looking at the tolerance and variance inflation among the variables of interest. Selected variables were introduced in a forward stepwise manner to the logistic regression model. We compared models using the Akaike information criterion and the 22 likelihood ratio x2 test as appropriate. The goodness-offit of the logistic regression model was assessed with the Hosmer-Lemeshow test. We evaluated the C-statistic to look at the area under the curve for the receiver operating characteristic curve to finalize the most effective model for determining variables that are associated with mortality in this patient population.
RESULTS A total of 417 in-hospital cardiopulmonary arrests were recorded during the study period. We excluded 118 of these cardiopulmonary arrests (7 patients with age ,18), 77 were false codes and 34 were repeat codes (repeat cardiopulmonary arrest and CPR within 24 hours) (Figure 1). We analyzed 299 events in our final sample, among which 171 (57.2%) were men and 128 (42.8%) were women. Of the patients included in our final analysis, 137 (48.5%) survived the cardiopulmonary arrest for at least 24 hours. Among them, 62.04% (n 5 85) expired during the same hospital stay
378
FIGURE 1. Flowchart detailing the patients excluded from our analysis, along with specific reason for exclusion.
and 37.96% (n 5 52) survived to discharge; thus, our overall survival at the time of discharge was 17.5% (Table 1). From the patients who survived to hospital discharge, 61.5% had significant physical and/or neurological deficits, which required their transfer to a nursing home. Only 20 patients (6.7%) were able to go home after an inpatient cardiopulmonary arrest. Also of note was the fact that almost 80% of the initial rhythms were nonshockable. In our univariate analysis, the duration of CPR, initial rhythm and the day of the week were strongly associated with mortality (Table 2). Although mortality was higher in men compared with women, and in inpatients with an unwitnessed initial event, it did not reach statistical significance. Similarly, patients who suffered the initial cardiac arrest on non–intensive care unit (non-ICU) floors also had a higher mortality, but it did not reach statistical significance (Table 2). Of the 137 patients who survived, 111 patients (48.2%) had a witnessed cardiac arrest. Mortality was significantly higher among those patients who had a nonshockable rhythm compared with patients with a shockable rhythm (59.5% versus 33.8%). Duration of CPR was also associated with the outcome. Approximately 60.2% survived when the duration of CPR was #10 minutes compared with only 39% when it was .10 minutes. There was no difference in mortality associated with the time (day versus night) of the cardiopulmonary arrest (52.7% versus 55.3%). In our multivariable analysis, after adjustment for age, location (ICU versus non-ICU floors) and whether the event was witnessed or not, we found that the timing (weekend versus weekday) and the initial rhythm at the time of CPR were still associated with the overall mortality in our patient population (Table 3). Also, patients who had a cardiac arrest during the weekend had a 2.9 times more odds of dying than patients who had a cardiac arrest during a weekday (95% CI: 1.589–5.347, P 5 0.0005). After adjusting for all these factors, the duration of CPR (,10 minutes compared with .10 minutes) and the sex of the patient also retained prognostic significance when evaluating the mortality of these patients. Men had a higher mortality and were 2 times more likely to die than women (95% CI: 1.233–3.448, P 5 0.0058) (Table 3). Volume 348, Number 5, November 2014
Impact of System-Based Variables on Outcomes in Cardiac Arrest
TABLE 1. Patient characteristics and observed variables for patients analyzed in our study Variables Patients, n (%) (n 5 299) Age, yr 20 to ,25 25 to ,35 35 to ,45 45 to ,55 55 to ,65 65 to ,75 75 to ,85 85 and older Sex Male Race White Black Others Arrest witnessed Yes No Time of day, hr 07:00–19:00 19:01–06:59 Time of week Weekday Weekend Initial cardiac rhythm Shockable rhythm Nonshockable rhythm Hospital site Intensive care unit Medical-surgical wards Duration of CPR, min ,10 .10 Discharge Home Nursing home Return of spontaneous circulationa 24-hr survival Survival at hospital discharge
2 3 16 30 39 60 109 40
(0.6) (1) (5.3) (10) (13) (20) (36.4) (13.3)
171 (57.2) 219 (73.2) 66 (22.1) 14 (4.7) 230 (76.9) 69 (23.1) 131 (43.8) 168 (56.2) 231 (77.2) 68 (22.8) 62 (20.7) 237 (79.3) 121 (40.7) 178 (59.3) 93 (31.1) 206 (68.9) 20 32 164 137 52
(6.7) (10.7) (54.8) (45.8) (17.4)
a Return of spontaneous circulation defined as the presence of a circulating rhythm after 20 minutes of a successful cardiopulmonary resuscitation.
DISCUSSION The rate of survival to discharge, following an inpatient CPR has been historically reported at 7% to 26%.12,13 We report an almost 18% survival to hospital discharge in our single center retrospective study. These numbers are in accordance with a number of multicenter studies previously published in the literature.4,8,14 Patients experiencing a cardiopulmonary arrest in a hospital are much sicker and the rate of nonshockable rhythms (pulseless electrical activity or asystole) is much higher compared with out-of-hospital arrests.15 Respiratory failure or respiratory distress associated with hypoxemia is a precursor to cardiac arrest in the majority of these patients. Meaney et al6 reported that Ó 2014 Lippincott Williams & Wilkins
patients with shockable rhythms were less likely to have preexisting respiratory failure, pneumonia, hypotension or renal failure compared with patients who have pulseless electrical activity or asystole on presentation. Nonshockable rhythms have been associated with poorer outcomes in literature, and this finding is again reflected in our study. Analyses of previous studies,16,17 exploring outcomes in hospitalized patients, have shown that patient-related factors such as the presence of comorbidities, functional status and severity of disease at the initial presentation are all associated with poor outcomes. The ACLS protocol does not address these inherent differences in the population and treats all patients as similar at the time of a cardiac arrest. This approach is appropriate for initial resuscitation of all patients and perhaps is very effective in the outpatient setting where untrained individuals are expected to follow a simple protocol-based format, but this might not be the best approach to inpatient cardiac arrests. Based on the findings of our study, we feel that it is very challenging to address this heterogeneity in patients sustaining an in-hospital cardiac arrest, and that physicians need to look at each episode individually to decide the most appropriate intervention during a cardiac arrest. Much more focus needs to be placed on the complex interplay of system-based variables such as staffing of the institution during nights and weekends, the presence of in-house medical personnel and the training of personnel performing the CPR. These factors are perhaps much more amenable to modification and perhaps hospital systems need to optimize these resources to maximize successful outcomes in inpatient cardiac arrest. The optimal duration of CPR has always been a challenge for clinicians. Although older studies have shown longer resuscitation times to be associated with worse outcomes,18,19 a recent multicenter retrospective study suggested that a systematic increase in the duration of resuscitation might improve survival in hospitalized patients undergoing a cardiac arrest.11 The authors compared the hospitals with the shortest median resuscitation attempts with the hospitals that had the longest attempts. But these findings need to be carefully interpreted, as some of their outcomes are perhaps a reflection of differences in the kinds of patients being admitted to different hospitals. Patients with a high functional capacity and a lower rate of comorbidities are more likely to get a longer period of resuscitation compared with chronically ill patients and perhaps also have a higher likelihood of survival because of their overall health status, independent of the duration of resuscitation. Our study shows poorer outcomes when resuscitation measures were prolonged for more than 10 minutes, although a third of the events studied had a resuscitation period of more than 20 minutes. We feel that these results emphasize that defining a fixed period of resuscitation in patients undergoing inpatient cardiac arrests oversimplifies a very complex issue and perhaps the optimal duration of CPR is dependent on individual patients and to a certain degree on the clinical judgment of the treating physician. Outcomes in inpatient cardiac arrests have been associated with the time of the day in multiple studies.9,20 Hospitalized patients undergoing a cardiac arrest at night have lower rates of survival than similar patients who had a cardiac arrest during the day.1 Our study did not show this difference; we did, however, note a higher mortality associated with cardiac arrests occurring during the weekend. A study by Girotra et al8 noted no difference in survival among patients undergoing CPR between the days of a week. These findings again might be a reflection of a system-based effect, and these outcomes might plausibly reflect the difference in staffing models in different studies. There might also be an effect of the presence or absence of in-house residents and house staff on these outcomes. At our
379
Kantamineni et al
TABLE 2. A univariate analysis of variables associated with patients who had a return of spontaneous circulation (ROSC) and survived over the first 24 hours after in-hospital cardiopulmonary arrest, as compared to the patients who died within the first 24 hours Alive at 24 hr after ROSC Dead within 24 hr after ROSC Variables (n 5 137) (n 5 162) P Age, yr Male, n (%) Race, n (%) White Others Time of day, n (%) Day Night Time of week, n (%) Weekday Weekend Cardiac arrest witnessed by health care provider, n (%) Witnessed Unwitnessed Location, n (%) Intensive care unit Medical-surgical wards Initial cardiac rhythm recorded, n (%) Ventricular fibrillation Ventricular tachycardia Pulseless electrical activity Asystole Duration of cardiopulmonary resuscitation, n (%) #10 min .10 min
institution, we have 24-hour staffing with residents and interns during the weekdays and an inadequacy of both attending staff and house staff during weekends. The debate of age affecting the outcomes of CPR has never been conclusively settled. Advanced age has been previously reported to be associated with poorer outcomes in some studies.5,7,21 But, these results have never been replicated in the analysis of large multicenter databases.1,13 In our multivariable analysis, advanced age was not associated with poor outcomes.
TABLE 3. Adjusted mortality following multivariable analysis Odds ratio estimates Effect CPR ,10 min Time of the week Age Initial rhythm Location of arrest Witnessing of events Sex
Point estimate
95% Wald confidence interval
2.948 2.915 0.996 3.510 1.101 1.357
1.699–5.113 1.589–5.347 0.979–1.013 1.847–6.671 0.711–1.705 0.695–2.650
2.066
1.233–3.448
The results are described in odds ratio estimates with 95% confidence interval for all variables associated with mortality due to inpatient cardiopulmonary arrest.
380
71.5 6 13.7 71 (41.5)
69 6 15.8 100 (58.5)
0.14 0.08
102 (46.5) 35 (43.7)
117 (53.4) 45 (56.2)
0.66
62 (47.3) 75 (44.6)
69 (52.7) 93 (55.3)
0.64
97 (42) 40 (59)
134 (58) 28 (41)
0.01
111 (48.2) 26 (37.7)
119 (51.8) 43 (62.3)
0.13
62 (51.2) 75 (42.1)
59 (48.7) 103 (57.9)
0.12
22 19 49 47
(64.7) (67.8) (41.5) (39.5)
56 (60.2) 81 (39.3)
12 9 69 72
(35.3) (32.1) (58.4) (60.5)
0.004
37 (39.7) 125 (60.7)
0.001
Clinicians need to consider baseline comorbidities and severity of associated diseases along with age in their discussions regarding advanced directives and CPR in older individuals. Gender has been reported as a significant prognostic factor in critical illness. Although female patients have better survival after hospital discharge for sepsis and Acute Respiratory Distress Syndrome,22 these differences have not been shown consistently in previous studies evaluating outcomes in patients with cardiac arrest.3,23,24 However, a recent meta-analysis by Ebell et al13 reported male gender to be associated with a decreased likelihood of immediate survival. Our study corroborates these findings, as men had a poorer survival to discharge following an inpatient CPR. Gender-based outcomes in comorbidities like sepsis, coronary artery disease, hypertension and so on may partly be confounding the results of this finding.25 Contrary to similar studies published in the past,1,5 our study did not show any differences in outcome of inpatient CPR based on whether an event was witnessed. A staggered staffing model with residents, hospitalists and intensivists, and a higher rate of telemetry monitoring for inpatients at our institution coupled with standardized training for all hospital staff most likely resulted in shorter response times and earlier recognition of cardiopulmonary arrest, contributing to these results. This conclusion could be limited to our center alone and caution should be exercised when reviewing these results, as traditionally witnessed events have been shown to have better results than those that were not witnessed. There are a number of limitations to our study. It is a retrospective, single-centered observational study and the Volume 348, Number 5, November 2014
Impact of System-Based Variables on Outcomes in Cardiac Arrest
impact of residual confounding affecting our results needs to be acknowledged. We also recognize that there might be inaccuracies in documentation. The inability to assess the competency of the resuscitative measures is an indigenous limitation of such study designs. We also acknowledge that this study was performed before the implementation of the new ACLS guidelines. We did not include neurologic outcomes of patients who were discharged, and evaluation of short-term and long-term cognitive deficits might be a better outcome measurement tool than survival alone. Another surrogate marker for overall outcomes is discharge to home compared with other health care facilities like nursing homes. This, however, involves a multidisciplinary approach involved in the care of patients after resuscitation. We also do not report on the use of therapeutic hypothermia in our patients, as the implementation of this therapeutic modality is still very inconsistent in inpatient cardiac arrest. In a recent study by Mikkelsen et al,26 a therapeutic hypothermia protocol as a part of post–cardiac arrest care was implemented only in 2% of inpatient cardiac arrest patients.
CONCLUSION Understanding the factors that might affect the outcome of in-hospital CPR is important for physicians. Clinicians need to look at the patient- and system-based variables separately when they are evaluating outcomes associated with cardiopulmonary arrests. Improvement of system-based variables such as staffing patterns, staggered staffing to cover nights and weekends and a higher ratio of monitored beds in hospitals might help us improve the ROSC in a number of patients undergoing an in-hospital cardiac arrest. But this might not change the survival to hospital discharge. Survival to hospital discharge after an inpatient CPR at our teaching hospital was found to be similar to larger multicenter studies. As described in previous studies,18,19 a longer duration of resuscitative efforts (.10 minutes) does not translate into better results when compared with a shorter duration of time (#10 minutes). Decisions around prolonged resuscitative measures should be based on individual patient characteristics such as the severity of illness, comorbidities and baseline functional status. Inappropriate prolonged CPR may result in an increased burden on the health care resources. The impact of such futile interventions can be potentially taxing to patient families, and physicians should address these issues in the case of higher risk patients much earlier in the course of their hospitalization. REFERENCES 1. Brindley PG, Markland DM, Mayers I, et al. Predictors of survival following in-hospital adult cardiopulmonary resuscitation. CMAJ 2002; 167:343–8. 2. Hodgetts TJ, Kenward G, Vlackonikolis L, et al. Incidence, location and reasons for avoidable in-hospital cardiac arrest in a district general hospital. Resuscitation 2002;54:115–23. 3. Murphy DJ. Do-not-resuscitate orders. Time for reappraisal in longterm-care institutions. JAMA 1988;260:2098–101. 4. Ehlenbach WJ, Barnato A, Randall C. Epidemiologic study of inhospital cardiopulmonary resuscitation in the elderly. N Engl J Med 2009;361:22–31. 5. Larkin GL, Copes WS, Nathanson BH, et al. Pre-resuscitation factors associated with mortality in 49,130 cases of in-hospital cardiac arrest. Resuscitation 2010;81:302–11.
Ó 2014 Lippincott Williams & Wilkins
6. Meaney PA, Nadkarni VM, Kern KB, et al. Rhythms and outcomes of adult in-hospital cardiac arrest. Crit Care Med 2010;38:101–8. 7. Meaney PA, Nadkarni VM, Cook EF, et al. Higher survival rates among younger patients after pediatric intensive care unit cardiac arrests. Pediatrics 2006;118:2424–33. 8. Girotra S, Brahmajee K, Nallamothu BK, et al. Trends in survival after in-hospital cardiac arrest. N Engl J Med 2012;367:1912–20. 9. Peberdy MA, Ornato JP, Larkin GL, et al. Survival from in-hospital cardiac arrest during nights and weekends. JAMA 2008;299:785–92. 10. Cummins RO, Chamberlain D, Hazinski MF, et al. Recommended guidelines for reviewing, reporting, and conducting research on inhospital resuscitation: the in-hospital “Utstein style.” American Heart Association. Circulation 1997;95:2213–39. 11. Goldberger ZD, Chan PS, Berg RA, et al. Duration of resuscitation efforts and survival after in-hospital cardiac arrest: an observational study. Lancet 2012;380:1473–81. 12. Tresch D, Heudebert G, Kutty K, et al. Cardiopulmonary resuscitation in elderly patients hospitalized in the 1990s: a favorable outcome. J Am Geriatr Soc 1994;42:137–41. 13. Ebell MH, Becker LA, Barry HC, et al. Survival after in-hospital cardiopulmonary resuscitation: a meta-analysis. J Gen Intern Med 1998;13:805–16. 14. Peberdy MA, Kaye W, Ornato JP, et al. Cardiopulmonary resuscitation of adults in the hospital: a report of 14720 cardiac arrests from the National Registry of Cardiopulmonary Resuscitation. Resuscitation. 2003;58(3):297–308. 15. Nadkarni VM, Larkin GL, Peberdy MA, et al. First documented rhythm and clinical outcome from in-hospital cardiac arrest among children and adults. JAMA 2006;295:50–7. 16. Ebell MH. Prearrest predictors of survival following in-hospital cardiopulmonary resuscitation: a meta-analysis. J Fam Pract 1992;34:551–8. 17. Cohn EB, Lefevre F, Yarnold PR, et al. Predicting survival from inhospital CPR: meta-analysis and validation of a prediction model. J Gen Intern Med 1993;8:347–53. 18. Ballew KA, Philbrick JT, Caven DE, et al. Predictors of survival following in-hospital cardiopulmonary resuscitation: a moving target. Arch Intern Med 1994;154:2426–32. 19. Bedell SE, Delbanco TL, Cook EF, et al. Survival after cardiopulmonary resuscitation in the hospital. N Engl J Med 1983;309:569–76. 20. Dumot JA, Daniel J, et al. Outcome of adult cardiopulmonary resuscitations at a tertiary referral center including results of “limited” resuscitations. Arch Intern Med 2001;161:1751–8. 21. Rakic D, Rumboldt Z, Carevic V, et al. In-hospital cardiac arrest and resuscitation outcomes: rationale for sudden cardiac death approach. Croat Med J 2005;46. p. 907. 22. Valentin A, Jordan B, Lang T, et al. Gender-related differences in intensive care: a multiple-center cohort study of therapeutic interventions and outcome in critically ill patients. Crit Care Med 2003;31:1901–7. 23. Albert CM, McGovern BA, Newell JB, et al. Sex differences in cardiac arrest survivors. Circulation 1996;93:1170–6. 24. Taffet GE, Teasdale TA, Luchi RJ. In-hospital cardiopulmonary resuscitation. JAMA 1988;260:2069–72. 25. Fowler R, Sabur N, Ling P, et al. Sex-and age-based differences in the delivery and outcomes of critical care. CMAJ 2007;177:1513–9. 26. Mikkelsen ME, Christie JD, Abella BS, et al. Use of therapeutic hypothermia after in-hospital cardiac arrest. Crit Care Med 2013;41: 1385–95.
381
Cardiopulmonary Resuscitation in the Hospitalized Patient: Impact of System-Based Variables on Outcomes in Cardiac Arrest Phani Kantamineni, MD, Vamsi Emani, MD, Ankur Saini, MD, Hardeep Rai, MD and Abhijit Duggal, MD, MPH
Abstract: Background: A better understanding of the factors affecting the outcome of inpatient cardiopulmonary resuscitation (CPR) is crucial in making key clinical decisions. We aim to study the impact of various patient-related and hospital-related variables in a community-based teaching setup that could affect the prognosis of in-hospital cardiac arrests. Methods: We analyzed the data on all patients who experienced cardiac arrest while hospitalized at a community teaching hospital in Youngstown, Ohio. A multivariable logistic regression was performed to identify patient- and system-based variables associated with mortality in inpatient cardiac arrest. Results: A total of 417 in-hospital cardiopulmonary arrests were recorded during the study period. We analyzed 299 events in our final sample. One hundred sixty-four patients (54.8%) achieved return of spontaneous circulation and 137 (48.5%) survived the cardiopulmonary arrest for at least 24 hours. The duration of CPR, age, initial rhythm, witnessed events and sex were strongly associated with mortality in our univariate analysis. After adjustment for age, location and whether the code was witnessed, the timing of the week, initial rhythm, the duration of CPR and the sex of the patient retained prognostic significance in predicting the mortality. Conclusions: In our study, we report a 17.4% survival to hospital discharge after an in-hospital cardiopulmonary arrest and subsequent CPR, similar to rates reported in larger multicenter studies. Prolonged duration of CPR (.10 minutes) and male sex were found to be associated with worse outcomes. We report the impact of system-based variables such as physician and nursing staffing during different days of the week, on survival in these patients. Key Indexing Terms: Cardiopulmonary resuscitation; Inpatient cardiac arrest; Mortality. [Am J Med Sci 2014;348(5):377–381.]
T
he introduction of basic life support and advanced cardiac life support (ACLS) as guidelines for a standardized response to all cardiac arrests have had a dramatic impact on the immediate survival of patients. These gains, however, have not been as remarkable in hospitalized patients who undergo a cardiac arrest.1 This lack of improvement in outcome measures has in part been explained by the impact of an aging and sicker inpatient population offsetting the advances in medicine over the past 2 decades.1 The in-hospital incidence of cardiac arrest is estimated to be between 1 and 5 events per 1000 hospital admissions,2 and despite the introduction of advanced directives as part of standard patient care, most of these patients From the Department of Internal Medicine (PK, VE, AS), ValleyCare Health System/NEOMED, Northside Medical Center, Youngstown, Ohio; and Medical Intensive Care Unit (HR, AD), Respiratory Institute, Cleveland Clinic, Cleveland, Ohio. Submitted September 25, 2013; accepted in revised form March 24, 2014. The authors have no financial or other conflicts of interest to disclose. Correspondence: Abhijit Duggal, MD, MPH, Medical Intensive Care Unit, Respiratory Institute, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195 (E-mail: [email protected]).
The American Journal of the Medical Sciences
end up receiving cardiopulmonary resuscitation (CPR) as part of their treatment plan.3 It is important to understand the likelihood of success of CPR in patients admitted to the hospital, as they are much sicker and perhaps have different factors associated with cardiopulmonary arrest compared with the general population.4 Recent data have identified variables such as the initial rhythm on record, early initiation of CPR, early return of spontaneous circulation (ROSC) and a younger age to be associated with better outcomes.5–7 But there are only sparse data on the impact of the complex interplay between the patient-related variables (age, sex, etc), the hospital-based variables (place and timing of arrest in a hospital) and the system-based variables (size of the hospital, academic affiliation), on the overall outcomes of inpatients undergoing CPR.4,5,8,9 A better understanding of these factors should translate into a more consistent response to inpatient cardiac arrest in different hospital settings. Also it might lead to a more comprehensive and appropriate dialog between physicians and the patients when they discuss the factors that determine a patients’ wish to proceed with life-sustaining measures. In our study, we attempt to identify the factors that might affect outcomes in hospitalized patients suffering a cardiopulmonary arrest, in a community-based teaching setup. We also compare these results with outcomes reported from larger databases, and tertiary academic centers, to see if there are any differences in the reported outcomes.
MATERIALS AND METHODS We performed a retrospective chart review from January 2005 to December 2007, for all patients on whom a “code blue” CPR protocol was activated at Northside Medical Center, Youngstown, Ohio. All the nurses and physicians working at our hospital are required to be basic life support and ACLS certified. An express (code blue) team responds to all code blue calls. On weekdays, this team comprises a senior medical resident, a critical care nurse, an anesthesia and a critical care attending. Over the weekends, the team is modified and comprises a senior medical resident, a critical care nurse and a certified nurse anesthetist. We excluded patients less than 18 years and instances of inappropriate “false code blues” (situations where CPR protocol was activated, but after initial evaluation was not implemented; as the patients were hemodynamically intact). We only evaluated the first episode of CPR in patients who underwent multiple instances of cardiopulmonary arrest in a 24-hour period. Definitions Definitions of cardiac arrest, witnessed arrest, CPR, chest compressions, date of the arrest, location of the arrest, first monitored rhythm, shockable/nonshockable rhythm, ROSC and survival to hospital discharge were used based on the Utstein guidelines for standardized reporting in CPR.10 There is a wide
Volume 348, Number 5, November 2014
377
Kantamineni et al
variation in the reporting of ROSC, especially in hospitalized patients as a number of them can suffer multiple cardiopulmonary arrests over a short period of time. We defined the ROSC as the presence of a palpable rhythm after 20 minutes of successful CPR. We describe the survival at 24 hours after a successful CPR and also survival at discharge. We decided a priori that to better understand the impact of in-hospital cardiopulmonary arrest in terms of system-based variables, mortality at 24 hours after ROSC was a better time point to describe. For our univariate and multivariable analyses, we use survival at 24 hours as the outcome of interest. This measure is validated by the Utstein guidelines, as a measurable variable.10 We also describe survival to hospital discharge based on these guidelines.10 We used the cutoff time for CPR as #10 minutes or .10 minutes, as it has been recognized as a prognostic factor in previous studies.11 We decided to use 07:00 to 19:00 hours, and 19:01 to 06:59 hours as cutoff points for time, as medical staff and residents change their shifts at this time. Weekends were defined as from 19:01 hours on Friday until 06:59 hours on Monday, which closely mirrors existing literature.9 Univariate Analysis Descriptive statistics are reported as frequency (percentages) for categorical variables. Continuous variables are reported as either means (standard deviation) or median (interquartile ranges) based on the distribution of the individual variables. We used a Student t test, or Wilcoxon’s rank sum test as appropriate for continuous variables, and x2 test or Fisher’s exact test for discrete variables. Confidence intervals (CI) and P values reflect a 2-tailed a level of 0.05. Statistical analyses were conducted using SAS version 9.3 (SAS Institute, Cary, NC). Multivariable Analysis Patients with missing data were excluded from the regression model. We included all the clinically significant prognostic variables with a statistically significant result (P 5 0.20) on the univariate analysis. We used the “limiting sample size” rule to determine the number of prognostic variables associated with mortality on the univariate analysis and used these in a multiple logistic regression model. Before including these factors in our multivariable model, we tested for multicolinearity by looking at the tolerance and variance inflation among the variables of interest. Selected variables were introduced in a forward stepwise manner to the logistic regression model. We compared models using the Akaike information criterion and the 22 likelihood ratio x2 test as appropriate. The goodness-offit of the logistic regression model was assessed with the Hosmer-Lemeshow test. We evaluated the C-statistic to look at the area under the curve for the receiver operating characteristic curve to finalize the most effective model for determining variables that are associated with mortality in this patient population.
RESULTS A total of 417 in-hospital cardiopulmonary arrests were recorded during the study period. We excluded 118 of these cardiopulmonary arrests (7 patients with age ,18), 77 were false codes and 34 were repeat codes (repeat cardiopulmonary arrest and CPR within 24 hours) (Figure 1). We analyzed 299 events in our final sample, among which 171 (57.2%) were men and 128 (42.8%) were women. Of the patients included in our final analysis, 137 (48.5%) survived the cardiopulmonary arrest for at least 24 hours. Among them, 62.04% (n 5 85) expired during the same hospital stay
378
FIGURE 1. Flowchart detailing the patients excluded from our analysis, along with specific reason for exclusion.
and 37.96% (n 5 52) survived to discharge; thus, our overall survival at the time of discharge was 17.5% (Table 1). From the patients who survived to hospital discharge, 61.5% had significant physical and/or neurological deficits, which required their transfer to a nursing home. Only 20 patients (6.7%) were able to go home after an inpatient cardiopulmonary arrest. Also of note was the fact that almost 80% of the initial rhythms were nonshockable. In our univariate analysis, the duration of CPR, initial rhythm and the day of the week were strongly associated with mortality (Table 2). Although mortality was higher in men compared with women, and in inpatients with an unwitnessed initial event, it did not reach statistical significance. Similarly, patients who suffered the initial cardiac arrest on non–intensive care unit (non-ICU) floors also had a higher mortality, but it did not reach statistical significance (Table 2). Of the 137 patients who survived, 111 patients (48.2%) had a witnessed cardiac arrest. Mortality was significantly higher among those patients who had a nonshockable rhythm compared with patients with a shockable rhythm (59.5% versus 33.8%). Duration of CPR was also associated with the outcome. Approximately 60.2% survived when the duration of CPR was #10 minutes compared with only 39% when it was .10 minutes. There was no difference in mortality associated with the time (day versus night) of the cardiopulmonary arrest (52.7% versus 55.3%). In our multivariable analysis, after adjustment for age, location (ICU versus non-ICU floors) and whether the event was witnessed or not, we found that the timing (weekend versus weekday) and the initial rhythm at the time of CPR were still associated with the overall mortality in our patient population (Table 3). Also, patients who had a cardiac arrest during the weekend had a 2.9 times more odds of dying than patients who had a cardiac arrest during a weekday (95% CI: 1.589–5.347, P 5 0.0005). After adjusting for all these factors, the duration of CPR (,10 minutes compared with .10 minutes) and the sex of the patient also retained prognostic significance when evaluating the mortality of these patients. Men had a higher mortality and were 2 times more likely to die than women (95% CI: 1.233–3.448, P 5 0.0058) (Table 3). Volume 348, Number 5, November 2014
Impact of System-Based Variables on Outcomes in Cardiac Arrest
TABLE 1. Patient characteristics and observed variables for patients analyzed in our study Variables Patients, n (%) (n 5 299) Age, yr 20 to ,25 25 to ,35 35 to ,45 45 to ,55 55 to ,65 65 to ,75 75 to ,85 85 and older Sex Male Race White Black Others Arrest witnessed Yes No Time of day, hr 07:00–19:00 19:01–06:59 Time of week Weekday Weekend Initial cardiac rhythm Shockable rhythm Nonshockable rhythm Hospital site Intensive care unit Medical-surgical wards Duration of CPR, min ,10 .10 Discharge Home Nursing home Return of spontaneous circulationa 24-hr survival Survival at hospital discharge
2 3 16 30 39 60 109 40
(0.6) (1) (5.3) (10) (13) (20) (36.4) (13.3)
171 (57.2) 219 (73.2) 66 (22.1) 14 (4.7) 230 (76.9) 69 (23.1) 131 (43.8) 168 (56.2) 231 (77.2) 68 (22.8) 62 (20.7) 237 (79.3) 121 (40.7) 178 (59.3) 93 (31.1) 206 (68.9) 20 32 164 137 52
(6.7) (10.7) (54.8) (45.8) (17.4)
a Return of spontaneous circulation defined as the presence of a circulating rhythm after 20 minutes of a successful cardiopulmonary resuscitation.
DISCUSSION The rate of survival to discharge, following an inpatient CPR has been historically reported at 7% to 26%.12,13 We report an almost 18% survival to hospital discharge in our single center retrospective study. These numbers are in accordance with a number of multicenter studies previously published in the literature.4,8,14 Patients experiencing a cardiopulmonary arrest in a hospital are much sicker and the rate of nonshockable rhythms (pulseless electrical activity or asystole) is much higher compared with out-of-hospital arrests.15 Respiratory failure or respiratory distress associated with hypoxemia is a precursor to cardiac arrest in the majority of these patients. Meaney et al6 reported that Ó 2014 Lippincott Williams & Wilkins
patients with shockable rhythms were less likely to have preexisting respiratory failure, pneumonia, hypotension or renal failure compared with patients who have pulseless electrical activity or asystole on presentation. Nonshockable rhythms have been associated with poorer outcomes in literature, and this finding is again reflected in our study. Analyses of previous studies,16,17 exploring outcomes in hospitalized patients, have shown that patient-related factors such as the presence of comorbidities, functional status and severity of disease at the initial presentation are all associated with poor outcomes. The ACLS protocol does not address these inherent differences in the population and treats all patients as similar at the time of a cardiac arrest. This approach is appropriate for initial resuscitation of all patients and perhaps is very effective in the outpatient setting where untrained individuals are expected to follow a simple protocol-based format, but this might not be the best approach to inpatient cardiac arrests. Based on the findings of our study, we feel that it is very challenging to address this heterogeneity in patients sustaining an in-hospital cardiac arrest, and that physicians need to look at each episode individually to decide the most appropriate intervention during a cardiac arrest. Much more focus needs to be placed on the complex interplay of system-based variables such as staffing of the institution during nights and weekends, the presence of in-house medical personnel and the training of personnel performing the CPR. These factors are perhaps much more amenable to modification and perhaps hospital systems need to optimize these resources to maximize successful outcomes in inpatient cardiac arrest. The optimal duration of CPR has always been a challenge for clinicians. Although older studies have shown longer resuscitation times to be associated with worse outcomes,18,19 a recent multicenter retrospective study suggested that a systematic increase in the duration of resuscitation might improve survival in hospitalized patients undergoing a cardiac arrest.11 The authors compared the hospitals with the shortest median resuscitation attempts with the hospitals that had the longest attempts. But these findings need to be carefully interpreted, as some of their outcomes are perhaps a reflection of differences in the kinds of patients being admitted to different hospitals. Patients with a high functional capacity and a lower rate of comorbidities are more likely to get a longer period of resuscitation compared with chronically ill patients and perhaps also have a higher likelihood of survival because of their overall health status, independent of the duration of resuscitation. Our study shows poorer outcomes when resuscitation measures were prolonged for more than 10 minutes, although a third of the events studied had a resuscitation period of more than 20 minutes. We feel that these results emphasize that defining a fixed period of resuscitation in patients undergoing inpatient cardiac arrests oversimplifies a very complex issue and perhaps the optimal duration of CPR is dependent on individual patients and to a certain degree on the clinical judgment of the treating physician. Outcomes in inpatient cardiac arrests have been associated with the time of the day in multiple studies.9,20 Hospitalized patients undergoing a cardiac arrest at night have lower rates of survival than similar patients who had a cardiac arrest during the day.1 Our study did not show this difference; we did, however, note a higher mortality associated with cardiac arrests occurring during the weekend. A study by Girotra et al8 noted no difference in survival among patients undergoing CPR between the days of a week. These findings again might be a reflection of a system-based effect, and these outcomes might plausibly reflect the difference in staffing models in different studies. There might also be an effect of the presence or absence of in-house residents and house staff on these outcomes. At our
379
Kantamineni et al
TABLE 2. A univariate analysis of variables associated with patients who had a return of spontaneous circulation (ROSC) and survived over the first 24 hours after in-hospital cardiopulmonary arrest, as compared to the patients who died within the first 24 hours Alive at 24 hr after ROSC Dead within 24 hr after ROSC Variables (n 5 137) (n 5 162) P Age, yr Male, n (%) Race, n (%) White Others Time of day, n (%) Day Night Time of week, n (%) Weekday Weekend Cardiac arrest witnessed by health care provider, n (%) Witnessed Unwitnessed Location, n (%) Intensive care unit Medical-surgical wards Initial cardiac rhythm recorded, n (%) Ventricular fibrillation Ventricular tachycardia Pulseless electrical activity Asystole Duration of cardiopulmonary resuscitation, n (%) #10 min .10 min
institution, we have 24-hour staffing with residents and interns during the weekdays and an inadequacy of both attending staff and house staff during weekends. The debate of age affecting the outcomes of CPR has never been conclusively settled. Advanced age has been previously reported to be associated with poorer outcomes in some studies.5,7,21 But, these results have never been replicated in the analysis of large multicenter databases.1,13 In our multivariable analysis, advanced age was not associated with poor outcomes.
TABLE 3. Adjusted mortality following multivariable analysis Odds ratio estimates Effect CPR ,10 min Time of the week Age Initial rhythm Location of arrest Witnessing of events Sex
Point estimate
95% Wald confidence interval
2.948 2.915 0.996 3.510 1.101 1.357
1.699–5.113 1.589–5.347 0.979–1.013 1.847–6.671 0.711–1.705 0.695–2.650
2.066
1.233–3.448
The results are described in odds ratio estimates with 95% confidence interval for all variables associated with mortality due to inpatient cardiopulmonary arrest.
380
71.5 6 13.7 71 (41.5)
69 6 15.8 100 (58.5)
0.14 0.08
102 (46.5) 35 (43.7)
117 (53.4) 45 (56.2)
0.66
62 (47.3) 75 (44.6)
69 (52.7) 93 (55.3)
0.64
97 (42) 40 (59)
134 (58) 28 (41)
0.01
111 (48.2) 26 (37.7)
119 (51.8) 43 (62.3)
0.13
62 (51.2) 75 (42.1)
59 (48.7) 103 (57.9)
0.12
22 19 49 47
(64.7) (67.8) (41.5) (39.5)
56 (60.2) 81 (39.3)
12 9 69 72
(35.3) (32.1) (58.4) (60.5)
0.004
37 (39.7) 125 (60.7)
0.001
Clinicians need to consider baseline comorbidities and severity of associated diseases along with age in their discussions regarding advanced directives and CPR in older individuals. Gender has been reported as a significant prognostic factor in critical illness. Although female patients have better survival after hospital discharge for sepsis and Acute Respiratory Distress Syndrome,22 these differences have not been shown consistently in previous studies evaluating outcomes in patients with cardiac arrest.3,23,24 However, a recent meta-analysis by Ebell et al13 reported male gender to be associated with a decreased likelihood of immediate survival. Our study corroborates these findings, as men had a poorer survival to discharge following an inpatient CPR. Gender-based outcomes in comorbidities like sepsis, coronary artery disease, hypertension and so on may partly be confounding the results of this finding.25 Contrary to similar studies published in the past,1,5 our study did not show any differences in outcome of inpatient CPR based on whether an event was witnessed. A staggered staffing model with residents, hospitalists and intensivists, and a higher rate of telemetry monitoring for inpatients at our institution coupled with standardized training for all hospital staff most likely resulted in shorter response times and earlier recognition of cardiopulmonary arrest, contributing to these results. This conclusion could be limited to our center alone and caution should be exercised when reviewing these results, as traditionally witnessed events have been shown to have better results than those that were not witnessed. There are a number of limitations to our study. It is a retrospective, single-centered observational study and the Volume 348, Number 5, November 2014
Impact of System-Based Variables on Outcomes in Cardiac Arrest
impact of residual confounding affecting our results needs to be acknowledged. We also recognize that there might be inaccuracies in documentation. The inability to assess the competency of the resuscitative measures is an indigenous limitation of such study designs. We also acknowledge that this study was performed before the implementation of the new ACLS guidelines. We did not include neurologic outcomes of patients who were discharged, and evaluation of short-term and long-term cognitive deficits might be a better outcome measurement tool than survival alone. Another surrogate marker for overall outcomes is discharge to home compared with other health care facilities like nursing homes. This, however, involves a multidisciplinary approach involved in the care of patients after resuscitation. We also do not report on the use of therapeutic hypothermia in our patients, as the implementation of this therapeutic modality is still very inconsistent in inpatient cardiac arrest. In a recent study by Mikkelsen et al,26 a therapeutic hypothermia protocol as a part of post–cardiac arrest care was implemented only in 2% of inpatient cardiac arrest patients.
CONCLUSION Understanding the factors that might affect the outcome of in-hospital CPR is important for physicians. Clinicians need to look at the patient- and system-based variables separately when they are evaluating outcomes associated with cardiopulmonary arrests. Improvement of system-based variables such as staffing patterns, staggered staffing to cover nights and weekends and a higher ratio of monitored beds in hospitals might help us improve the ROSC in a number of patients undergoing an in-hospital cardiac arrest. But this might not change the survival to hospital discharge. Survival to hospital discharge after an inpatient CPR at our teaching hospital was found to be similar to larger multicenter studies. As described in previous studies,18,19 a longer duration of resuscitative efforts (.10 minutes) does not translate into better results when compared with a shorter duration of time (#10 minutes). Decisions around prolonged resuscitative measures should be based on individual patient characteristics such as the severity of illness, comorbidities and baseline functional status. Inappropriate prolonged CPR may result in an increased burden on the health care resources. The impact of such futile interventions can be potentially taxing to patient families, and physicians should address these issues in the case of higher risk patients much earlier in the course of their hospitalization. REFERENCES 1. Brindley PG, Markland DM, Mayers I, et al. Predictors of survival following in-hospital adult cardiopulmonary resuscitation. CMAJ 2002; 167:343–8. 2. Hodgetts TJ, Kenward G, Vlackonikolis L, et al. Incidence, location and reasons for avoidable in-hospital cardiac arrest in a district general hospital. Resuscitation 2002;54:115–23. 3. Murphy DJ. Do-not-resuscitate orders. Time for reappraisal in longterm-care institutions. JAMA 1988;260:2098–101. 4. Ehlenbach WJ, Barnato A, Randall C. Epidemiologic study of inhospital cardiopulmonary resuscitation in the elderly. N Engl J Med 2009;361:22–31. 5. Larkin GL, Copes WS, Nathanson BH, et al. Pre-resuscitation factors associated with mortality in 49,130 cases of in-hospital cardiac arrest. Resuscitation 2010;81:302–11.
Ó 2014 Lippincott Williams & Wilkins
6. Meaney PA, Nadkarni VM, Kern KB, et al. Rhythms and outcomes of adult in-hospital cardiac arrest. Crit Care Med 2010;38:101–8. 7. Meaney PA, Nadkarni VM, Cook EF, et al. Higher survival rates among younger patients after pediatric intensive care unit cardiac arrests. Pediatrics 2006;118:2424–33. 8. Girotra S, Brahmajee K, Nallamothu BK, et al. Trends in survival after in-hospital cardiac arrest. N Engl J Med 2012;367:1912–20. 9. Peberdy MA, Ornato JP, Larkin GL, et al. Survival from in-hospital cardiac arrest during nights and weekends. JAMA 2008;299:785–92. 10. Cummins RO, Chamberlain D, Hazinski MF, et al. Recommended guidelines for reviewing, reporting, and conducting research on inhospital resuscitation: the in-hospital “Utstein style.” American Heart Association. Circulation 1997;95:2213–39. 11. Goldberger ZD, Chan PS, Berg RA, et al. Duration of resuscitation efforts and survival after in-hospital cardiac arrest: an observational study. Lancet 2012;380:1473–81. 12. Tresch D, Heudebert G, Kutty K, et al. Cardiopulmonary resuscitation in elderly patients hospitalized in the 1990s: a favorable outcome. J Am Geriatr Soc 1994;42:137–41. 13. Ebell MH, Becker LA, Barry HC, et al. Survival after in-hospital cardiopulmonary resuscitation: a meta-analysis. J Gen Intern Med 1998;13:805–16. 14. Peberdy MA, Kaye W, Ornato JP, et al. Cardiopulmonary resuscitation of adults in the hospital: a report of 14720 cardiac arrests from the National Registry of Cardiopulmonary Resuscitation. Resuscitation. 2003;58(3):297–308. 15. Nadkarni VM, Larkin GL, Peberdy MA, et al. First documented rhythm and clinical outcome from in-hospital cardiac arrest among children and adults. JAMA 2006;295:50–7. 16. Ebell MH. Prearrest predictors of survival following in-hospital cardiopulmonary resuscitation: a meta-analysis. J Fam Pract 1992;34:551–8. 17. Cohn EB, Lefevre F, Yarnold PR, et al. Predicting survival from inhospital CPR: meta-analysis and validation of a prediction model. J Gen Intern Med 1993;8:347–53. 18. Ballew KA, Philbrick JT, Caven DE, et al. Predictors of survival following in-hospital cardiopulmonary resuscitation: a moving target. Arch Intern Med 1994;154:2426–32. 19. Bedell SE, Delbanco TL, Cook EF, et al. Survival after cardiopulmonary resuscitation in the hospital. N Engl J Med 1983;309:569–76. 20. Dumot JA, Daniel J, et al. Outcome of adult cardiopulmonary resuscitations at a tertiary referral center including results of “limited” resuscitations. Arch Intern Med 2001;161:1751–8. 21. Rakic D, Rumboldt Z, Carevic V, et al. In-hospital cardiac arrest and resuscitation outcomes: rationale for sudden cardiac death approach. Croat Med J 2005;46. p. 907. 22. Valentin A, Jordan B, Lang T, et al. Gender-related differences in intensive care: a multiple-center cohort study of therapeutic interventions and outcome in critically ill patients. Crit Care Med 2003;31:1901–7. 23. Albert CM, McGovern BA, Newell JB, et al. Sex differences in cardiac arrest survivors. Circulation 1996;93:1170–6. 24. Taffet GE, Teasdale TA, Luchi RJ. In-hospital cardiopulmonary resuscitation. JAMA 1988;260:2069–72. 25. Fowler R, Sabur N, Ling P, et al. Sex-and age-based differences in the delivery and outcomes of critical care. CMAJ 2007;177:1513–9. 26. Mikkelsen ME, Christie JD, Abella BS, et al. Use of therapeutic hypothermia after in-hospital cardiac arrest. Crit Care Med 2013;41: 1385–95.
381
