Objective. Data regarding pediatric in-hospital cardiopulmonary resuscitation (CPR) have been limited because of retrospective study designs, small sample sizes, and inconsistent definitions of cardiac arrest and CPR. The purpose of this study was to prospectively describe and evaluate pediatric in-hospital CPR with the international consensus-derived epidemiologic definitions from the Utstein guidelines.
Methods. All 129 in-hospital CPRs during 12 months at a 122-bed university children’s hospital in Sao Paulo, Brazil, were described and evaluated using Utstein reporting guidelines. These guidelines include standardized descriptions of hospital variables, patient variables, arrest/event variables, and outcome variables. CPR was defined as chest compressions and assisted ventilation provided because of cardiac arrest or because of severe bradycardia with poor perfusion. Outcome variables included sustained return of spontaneous circulation, 24-hour survival, 30-day survival, 1-year survival, and neurologic status of survivors by the Pediatric Cerebral Performance Category Scale.
Results. Of the 6024 children admitted to the hospital, 176 (3%) had an episode that met the criteria for provision of CPR and 129 (2%) received CPR, 86 for clinical cardiac arrest and 43 for bradycardia with poor perfusion. Most of the children (71%) had preexisting chronic diseases. The most common precipitating causes were respiratory failure (61%) and shock (29%). The initial cardiac rhythm was asystole in 71 children (55%), pulseless electrical activity in 12 (9%), ventricular fibrillation in 1, and bradycardia with pulses and poor perfusion in 43 (33%). Eighty-three children (64%) attained sustained return of spontaneous circulation (>20 minutes), 43 (33%) were alive at 24 hours, 24 (19%) were alive at 30 days, and 19 (15%) were alive at 1 year. Although many factors correlated with 24-hour survival, multivariate logistic regression analysis revealed independent association of 24-hour survival with respiratory failure as the precipitating cause (odds ratio [OR]: 4.92; 95% confidence interval [CI]: 1.73–14.0), bradycardia with pulses as the initial event (OR: 2.68; 95% CI: 1.01–7.1), and shorter duration of CPR (OR: 0.92; 95% CI: 0.89–0.96 for each elapsed minute). Similarly, 30-day survival was independently associated with respiratory failure as the precipitating cause and shorter duration of CPR. Thirty-day survival decreased by 5% with each elapsed minute of CPR. Nineteen (91%) of the 21 survivors to hospital discharge and 16 (83%) of the 19 1-year survivors had no demonstrable long-term change in neurologic function from their pre-CPR status.
Conclusions. During this study, CPR was uncommon but not rare. Respiratory failure was the most common precipitating cause, followed by shock. Preexisting chronic diseases were prevalent among these children. Asystole was the most common initial cardiac rhythm, and bradycardia with pulses and poor perfusion was the second most common. Ventricular fibrillation was rare, but children with acute cardiac diseases, such as cardiac surgery and acute cardiomyopathies, were not admitted to this children’s hospital. CPR was effective: nearly two thirds of these children were initially successfully resuscitated, and one third were alive at 24 hours compared with imminent death without CPR and advanced life support. Nevertheless, survival progressively decreased over time, generally as a result of the underlying disease process. One-year survival was 15%. Importantly, most of these survivors had no demonstrable change in gross neurologic function from their pre-CPR status.
Data regarding pediatric in-hospital cardiopulmonary resuscitation (CPR) have been limited and difficult to interpret because nearly all of the reported studies are retrospective chart reviews of small numbers of children.1–5 Moreover, definitions of cardiac arrest and CPR interventions have been inconsistent.1–4 In the early 1990s, international conferences of resuscitation experts in Utstein, Norway, resulted in the development of consensus guidelines for uniform reporting of data from out-of-hospital cardiac arrests and in-hospital resuscitation, the so-called Utstein style.6,7 Similarly, guidelines for uniform reporting of pediatric advanced life support, the pediatric Utstein style, emanated from an international task force in 1994.3 However, the recent exhaustive international scientific review and statement entitled “Guidelines 2000 for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care: International Consensus on Science” could not identify any prospective pediatric cardiac arrest or CPR studies that used these Utstein-style reporting guidelines.1
The main purpose of this study was to prospectively describe and evaluate in-hospital CPR in a large tertiary care university children’s hospital, using the Utstein reporting guidelines. CPR was defined as closed-chest compressions and assisted ventilation intended to restore spontaneous, effective ventilation and circulation. On the basis of the previous retrospective inpatient studies, we hypothesized that duration of CPR >15 minutes, use of >2 doses of epinephrine, preexisting chronic disease, and initial electrocardiographic rhythm during resuscitation would be important predictors of outcome.2,4,8–24 Moreover, with consistent prospective documentation of the initial cardiac rhythm, we predicted a higher incidence of ventricular fibrillation (VF) or pulseless ventricular tachycardia (VT) than previously reported in other pediatric inpatient studies, because of recent outpatient pediatric studies and experimental laboratory investigations indicating a higher incidence even when the underlying cause of the cardiac arrest is asphyxia.25–27
The Children’s Institute of Sao Paulo University College of Medicine is a 122-bed tertiary care children’s hospital in the largest city of South America. More than 6000 patients are admitted to the hospital each year, and >650 are admitted to the intensive care unit (ICU). There is no cardiac surgery at this hospital and no inborn neonatal care. Intensive care for premature infants and infants with common neonatal problems is provided in the neonatal ICU at the University’s obstetric hospital. Pediatric cardiac surgery and acute pediatric nonsurgical cardiac care for patients with known heart disease (congestive heart failure, arrhythmias, acute myocarditis) are provided at the Heart Institute within the greater university complex. Trauma patients, including all pediatric trauma patients, are directed to the Surgical Emergency Department in the same complex rather than the Children’s Institute. Except for cardiac surgery and trauma surgery, the full array of pediatric surgical care was provided in the Children’s Institute.
This prospective investigation was approved by the Commission of Ethics in Research of the Children’s Institute. Because of the unexpected and sudden nature of these events and because this was an observational study, informed consent for inclusion was waived by the commission. All patients whose CPR was initiated in the hospital from July 1997 to June 1998 were included and followed for 1 year. Patients whose CPR was initiated in the prehospital setting were excluded from analysis. Consent for telephone follow-up after hospital discharge was obtained from the parents.
CPR was provided by pediatric residents, pediatric surgical residents, nurses, and faculty. Pediatric intensive care and pediatric emergency medicine faculty were on-call in-hospital 24 hours per day and available as part of the cardiac resuscitation team. The pediatric critical care medicine physicians and nurses had completed training in the American Heart Association Pediatric Advanced Life Support Course. There was an extensive educational effort for medical and nursing personnel to ensure capture of all eligible patients, including easy availability of the telephone extensions and pagers of in-hospital research team members 24 hours per day. In addition, the research team checked each unit for potential missed CPR events twice each day. CPR was performed by the professionals who were responsible for patient care, without interference from the observing research team.
CPR was defined as chest compressions and assisted ventilation provided because of cardiac arrest or because of bradycardia with poor perfusion. Cardiac arrest was defined as the cessation of cardiac mechanical activity, determined by the absence of a palpable central pulse, unresponsiveness, and apnea. Bradycardia with poor perfusion was defined as a heart rate <60 beats per minute and grossly inadequate perfusion despite support with oxygen and effective ventilation.
The data collection form was translated and adapted from the in-hospital Utstein-style guidelines because the pediatric Utstein-style guidelines document focused on prehospital cardiac arrests.7 Adaptations and inclusions per the pediatric Utstein style were added in collaboration with the American Heart Association National Registry of CPR Science Advisory Board (eg, the inclusion of bradycardia with poor perfusion and the use of pediatric neurologic outcome measures).3 In general, event-related data were obtained during and immediately after the CPR event. The members of the research team interviewed the physicians and nurses who provided CPR as soon as possible after the resuscitation and reviewed the medical records.
The Utstein style addresses 4 sets of variables: hospital variables (described above), patient variables, arrest/event variables, and outcome variables. Patient-related variables include age, gender, race, site of occurrence, presence of a witness and monitoring, previously instituted advanced life support interventions, previous provision of CPR, diagnosis at hospital admission, and presence of preexisting chronic disease. The arrest/event variables are immediate precipitating cause, initial cardiac rhythm, therapeutic interventions and respective times, the time interval from cardiac arrest/event until return to spontaneous circulation (ROSC), duration of the ROSC, and interval between the arrest/event and the end of the CPR.
The outcome variables include sustained ROSC (>20 minutes), 24-hour survival, 7-day survival, 30-day survival, hospital discharge, and survival after 6 months and 1 year. The follow-up after 6 months and 1 year was performed by medical record review and/or by telephone contacts with the patient’s family. Neurologic outcome was determined using the pediatric cerebral performance category (PCPC) scales3,9 as follows: category 1, normal, age-appropriate neurodevelopmental functioning; category 2, mild disability; category 3, moderate disability; category 4, severe disability; category 5, coma/vegetative state; and category 6, death. The pre-CPR categorization was based on historical data and chart review. Categorization at the time of discharge was determined by the discharge examination. Categorization at 6 months and 1 year post-CPR was generally determined by a telephone interview.
For patients with multiple arrest events, only the initial index arrest/event and resuscitation were described and analyzed. The data were entered in the Microsoft Access software, version 7.0 (Microsoft, Redmond, WA). Continuous variables were described as median and mean ± standard deviation. Comparisons of discrete variables were evaluated with χ2 analyses. P ≤ .05 was considered significant. All statistical analyses were performed with the use of the Epi-Info software, version 6.04 (Centers for Disease Control and Prevention, Atlanta, GA).
The χ2 test was used to calculate the relative risks for death at 24 hours and 30 days after CPR. Prognostic factors that were associated with survival (P < .10 by χ2 test) were submitted to multivariate logistic regression analysis, using SAS 6.08 software (SAS Institute Inc, Cary, NC). Odds ratios (OR) for survival and the 95% confidence intervals (CI) were determined for prognostic factors that were independently associated with survival.
During this 12-month investigation, 6024 patients were admitted to the Children’s Institute in Sao Paulo, Brazil. Figure 1 is a visual overview of the CPR performance and outcome, the pediatric Utstein-style template modified for in-hospital pediatric CPR by Suominem et al.10 A total of 176 (3%) had at least 1 episode of cardiac arrest or bradycardia with poor perfusion. Chest compressions and assisted ventilation were provided for 129 patients (2%) with 142 arrests/events. An additional 47 patients were declared dead without attempted resuscitation. These 47 DNAR patients were in the terminal phase of chronic, incurable, and debilitating diseases. Of the 677 patients admitted to the pediatric ICU (PICU), 94 (14%) received chest compressions and assisted ventilation.
The CPR attempts were generally successful initially, but survival progressively decreased over time (Fig 1). Sustained ROSC was attained in 64% of the patients. One third of the patients survived for at least 24 hours; however, the survival rate decreased to 15% by 1 year.
The characteristics of the study population and the correlation of these factors with 24-hour survival are described in Table 1. The patients’ ages varied from 3 days to 21 years, with the mean age of 3.6 ± 4.1 years and median of 1 year. Most of these patients (71%) had chronic diseases on admission (Table 2); 58% were boys. These patients were clearly recognized as high risk: 96% were witnessed by a medical professional at the time of the event, 76% had continuous electrocardiographic (ECG) monitoring at that time, and 73% were in the ICU, most receiving aggressive life-support interventions (Table 1). CPR was promptly provided for the 124 patients whose arrests/events were witnessed by a medical professional.
The most common precipitating causes of the events were respiratory failure (61%) and circulatory shock (29%) (Fig 2). In particular, 18% were in septic shock. The associated initial ECG rhythm was documented in 127 of the 129 children (Table 3): 76% by medical professional observation of continuous ECG monitoring immediately preceding pulselessness and the rest by quick look defibrillator paddles or rapid attachment to a continuous ECG monitor in the first minutes of CPR. Asystole was the most common initial rhythm (55%). Bradycardia with palpable pulse and hypoperfusion, unresponsive to oxygenation and ventilation, was noted in 33%. VF was the initial ECG rhythm in only 1 child, and pulseless VT was the initial ECG rhythm in none.
Most of the patients (83%) received epinephrine during resuscitation. The number of doses of epinephrine administered varied from 1 to 10. The mean was 3.28 ± 1.90, and the median was 3. Of the 107 patients who received epinephrine, 57% received >3 doses. The mean duration of CPR was 32 ± 35 minutes; the median was 23 minutes.
Many variables individually correlated with 24-hour survival (Tables 1 and 3). Respiratory failure as the immediate precipitating cause of the arrest/event and bradycardia with pulses as the initial event were each positively associated with 24-hour survival. Admission diagnosis of sepsis, shock as the immediate precipitating cause, initial ECG rhythm of asystole or pulseless electrical activity, administration of 2 or more doses of epinephrine, administration of bicarbonate, and duration of the resuscitation >15 minutes all were negatively associated with 24-hour survival. Controlling for confounding variables, multivariate logistic regression analysis revealed independent association of 24-hour survival with respiratory failure as the precipitating cause of the arrest (OR: 4.92; 95% CI: 1.73–14.0), bradycardia with pulses as the initial event (OR: 2.68; 95% CI: 1.01–7.1), and duration of the resuscitation (OR: 0.92; 95% CI: 0.89–0.96 for each elapsed minute). This is represented graphically as a 24-hour survival probability curve in Fig 3.
The factors that were positively associated with 30-day survival were respiratory failure as immediate precipitating cause of the arrest/event and bradycardia with pulses as the initial event. The variables that were negatively associated with 30-day survival were asystole as initial ECG rhythm, administration of 3 doses or more of epinephrine, administration of bicarbonate, and resuscitation duration of >15 minutes. Multivariate logistic regression analysis revealed independent positive association of 30-day survival with respiratory failure as the precipitating cause of the arrest (OR: 3.31; 95% CI: 1.01–10.81) and negative association with longer duration of the resuscitation (OR: 0.95; 95% CI: 0.91–0.98 for each elapsed minute). This is represented graphically as a 30-day survival probability curve in Fig 4.
The neurologic status before and after CPR, age, precipitating cause, and associated chronic disease of the 21 children who were discharged alive and the additional child who survived for 1 year without hospital discharge are listed in Table 4. Seven of these 22 surviving children had preexisting cerebral performance deficits. Three patients were lost to follow-up after hospital discharge; 1 before the 6-month assessment and 2 between the 6-month and 1-year assessments. The 1 who was lost to follow-up before the 6-month assessment was last evaluated 3 months post-CPR with moderate neurologic disability. None of the 3 had an underlying life-threatening disease, and 2 were neurologically normal at 6 months post-CPR, so all 3 were assumed to be alive at 1 year. The neurologic status assigned to each was the last available. Nineteen (91%) of the 21 discharged from the hospital and 16 (83%) of the 19 1-year survivors had no demonstrable change from pre-CPR status in the cerebral performance category at the last evaluation. Moreover, 13 (81%) of the 16 1-year survivors who were assessed at 1 year had no demonstrable change in the cerebral performance category. One patient with severe neurologic sequelae at 1 year had 3 separate cardiac arrests and remained hospitalized for the entire 12 months after the initial arrest; the other with severe neurologic sequelae had preexisting aplastic anemia.
Only 2 of 37 patients with shock as the precipitating cause survived for 1 year, and both were severely neurologically impaired at 1 year (Table 4). One was severely neurologically impaired before CPR (PCPC 4) and did not demonstrably deteriorate. They were also the only 2 1-year survivors who received >2 doses of epinephrine and the only 2 1-year survivors who received >30 minutes of CPR.
This is the first prospective in-hospital study of pediatric CPR using the Utstein guidelines for uniform data reporting.10 CPR, defined as assisted ventilation and closed-chest compressions, was uncommon but not rare. Two percent of children who were admitted to the hospital were provided with CPR, as were 14% of patients who were admitted to the ICU. Respiratory failure was the most common immediate precipitating cause of the event, followed by circulatory shock. These patients were clearly recognized as high risk: 96% were witnessed by a medical professional at the time of the event, 76% had continuous ECG monitoring at that time, and 73% were in the ICU, most receiving aggressive life-support interventions. Compared with imminent death, CPR was effective: sustained ROSC >20 minutes was attained in 64% of initial CPR attempts; 33% of these patients survived for at least 24 hours, and 15% were alive 1 year later. Independent predictors of 24-hour survival were respiratory failure as the precipitating cause, the presence of a palpable pulse when CPR commenced, and shorter duration of CPR. Most of these patients had preexisting chronic comorbid diseases; however, outcome was not affected by the presence of these comorbid chronic diseases. Neurologic status after successful CPR was ultimately the same as the prearrest neurologic status in 91% of those discharged from the hospital and 83% of the 1-year survivors. Contrary to our hypothesis, the initial ECG rhythm during cardiac arrest was rarely VF or pulseless VT (<1%).
Previous studies of in-hospital pediatric cardiac arrests are difficult to compare and interpret because of inconsistent definitions of cardiac arrest, CPR, event times and intervals, specific interventions, and outcomes.1–5 For example, the best outcomes have been reported in series that have defined cardiac arrest as the code team was called13,14 and series that included pure respiratory arrests that were responsive to assisted ventilation alone.12,18,19 These and other studies defined CPR as ventilatory assistance, chest compressions (without criteria for providing chest compressions), or administration of emergency drugs. Such operator-dependent variables do not necessarily reflect the patients’ clinical status or the need for such interventions. Only 1 published study of in-hospital pediatric CPR, by Suominem et al, 10 included the Utstein-style epidemiologic definitions. Those data from a large children’s hospital in Finland are somewhat limited because they were abstracted from retrospective chart reviews. Nevertheless, the findings from that study are generally similar to our data and much easier to compare and contrast with our data than the rest of the literature.
The Finnish study revealed that CPR was provided for 118 (52%) of the 227 patients who experienced a cardiac arrest or bradycardia with poor perfusion during a 5-year period.10 The incidence of CPR for these events was 1% of all hospital admissions and 6% of ICU admissions. Most of the CPR (64%) occurred in the ICU. Initially successful resuscitation (sustained ROSC) was attained in 63%, and the 1-year survival was 18%. Multivariate analysis indicated that only shorter duration of external compressions was independently associated with survival. These findings were generally similar to the data in our study despite differences in the patient populations. The Finnish study was at the Hospital for Children and Adolescents in Helsinki, where all pediatric open-heart surgery in the country is performed. In contrast, there is no pediatric cardiac surgery at the Children’s Institute in Sao Paulo. Consequently, the most common precipitating cause of cardiac arrest in the Finnish study was cardiovascular (71%), whereas respiratory failure (61%) was the most common in our study. Moreover, the initial rhythm was VF/VT in 11% of the Finnish patients compared with <1% of ours.
It is often stated that CPR in children results in dismal outcome, suggesting that CPR is ineffective. These 2 pediatric studies indicate that in-hospital CPR and advanced life support can be remarkably effective. Nearly two thirds of these patients were initially successfully resuscitated (ie, attained sustained ROSC). Survival progressively decreased with time, attributable in large part to the underlying disease processes. Most of these arrests/events occurred in the PICU as a result of progressive life-threatening illnesses that had not responded to treatment. The 1-year survival rates of 15% and 18% are generally superior to outcomes from out-of-hospital pediatric CPR and substantially superior to the certain 0% survival rate if CPR and advanced life support were not provided.
In our review of the literature, only 2 other series had similar definitions of in-hospital pediatric CPR and 1-year follow-up.9,10 These were the Finnish study noted above and a retrospective review of CPR at Arkansas Children’s Hospital from 1986 to 1993. Despite different patient populations on 3 different continents, the outcomes were remarkably similar (Table 5). Because of somewhat similar data collection, we also included outcomes from a retrospective study of children who received >2 minutes of CPR in a specialized pediatric cardiac ICU.11 Most of the patients were postcardiac surgery (26 of 32); follow-up was provided 6 months after CPR. It is interesting that 4 of these children received mechanical cardiopulmonary support during CPR as a rescue therapy.
The precipitating cause of the arrest/event is an important determinant of outcome. For example, the outcome from shock that progressed to cardiac arrest or bradycardia with poor perfusion was especially dismal with an 8% 24-hour survival rate (relative risk of death by 24 hours: 1.63), a 5% (2 of 37) hospital discharge rate, and a 5% 1-year survival rate in our study. Both of these 1-year survivors were severely neurologically impaired. Most of these patients had septic shock with progressive deterioration despite aggressive medical management. Similarly, Torres et al9 noted that none of their 44 pediatric cardiac arrest victims with sepsis syndrome survived for 1 year. Moreover, in 4 other studies, only 1 of 23 children with sepsis that progressed to cardiac arrest survived.12,16,18,21 In addition, a multicenter study of cardiac arrests in PICUs revealed a 7% (3 of 44 patients) survival-to-hospital-discharge rate among these children when the diagnosis in the first 24 hours of PICU admission was an infectious disease.23
In contrast, when the precipitating disease process was more readily reversible, as in our patients with respiratory failure and our patients who still had pulses when CPR was started, the outcomes were superior. The children with respiratory failure as the precipitating cause had a 24-hour survival rate of 47% (relative risk of death: 0.6). The children who presented with severe bradycardia and poor perfusion rather than absence of pulses (asystole, pulseless electrical activity, or VF) had a 24-hour survival rate of 51% (relative risk of death: 0.58). These superior outcomes are consistent with other published results from special resuscitation circumstances that are readily reversible, such as the use of mechanical cardiopulmonary support (eg, extracorporeal membrane oxygenator) in the setting of postoperative cardiac surgery with myocardial failure.11,28,29
The low prevalence of VF was unexpected. Two relatively recent studies demonstrated VF as the initial rhythm in 19% to 24% of out-of-hospital pediatric cardiac arrest victims, including VF during asphyxial arrests.25,26 Moreover, in a piglet model of out-of-hospital asphyxial cardiac arrest, the incidence of VF was 28% at some time during the arrest.27 However, none of the piglets exhibited VF when cardiac arrest, defined as loss of aortic pulsations, initially occurred (approximately 10 minutes after clamping the endotracheal tube). The piglet VF occurred during the subsequent 10-minute period of untreated cardiac arrest, a common time interval for out-of-hospital cardiac arrest. Why was the incidence of VF so low in the present study? In contrast to pediatric out-of-hospital arrests, the arrests/events in this in-hospital study were almost always observed and, therefore, treated more promptly (ie, perhaps before secondary VF may have otherwise occurred). In addition, VF in out-of-hospital cardiac arrests is more common among older children. For example, VF was noted in 3% of children who were 0 to 8 years old with out-of-hospital cardiac arrests versus 17% of victims who were 8 to 30 years old.30 Only 13% of our patients in the present study were 8 to 21 years old. Finally, VF was the initial rhythm in 11% and occurred in 20% at some time during the arrest in the Finnish Utstein-style in-hospital pediatric CPR study.10 The Finnish patient population included children with acute cardiac disease and cardiothoracic surgery, whereas such patients were not included in this study because they were admitted to the Heart Institute in Sao Paulo rather than the Children’s Institute.
Animal models clearly demonstrate that successful resuscitation from cardiac arrest depends on the duration of the cardiac arrest before CPR is provided (no flow), duration of CPR (low flow), the ability to establish adequate myocardial blood flow with CPR (quality of CPR), and early defibrillation in the case of VF.4 For asphyxial cardiac arrests, duration of untreated asphyxia and early reestablishment of adequate oxygenation and ventilation are also important.31,32 The survival rates in this study are superior to those from out-of-hospital arrests, in part because the events were witnessed and effective CPR and advanced life support were provided promptly.1,2,4 Although we were unable to evaluate rigorously the quality of CPR, presumably the prompt administration of CPR with adequate oxygenation, ventilation, and myocardial perfusion minimized the periods of no flow and low flow, thereby resulting in the superior outcome.
An important ethical question is, “When should CPR efforts for children be terminated?” Although it is clear that successful outcomes are inversely related to the duration of CPR, the maximum duration of resuscitation with potential for good outcomes is not clear. Several studies have reported no survivors after resuscitation efforts that lasted >10, 15, 20, or 30 minutes.8,12,20,33 In contrast, neurologically intact survival has been described after prolonged CPR (>60 minutes), especially when there was severe prearrest hypothermia (eg, ice water submersion),34 severe toxic drug exposure,35 or use of a mechanical cardiopulmonary support system.11,28 Nevertheless, in most settings, prolonged periods of low flow (>30 minutes) are unlikely to result in survival and may increase the relative risk of neurologically impaired survival.1,36–38 Although 7% (3 of 44) of the patients in our study survived 24 hours after >30 minutes of CPR, only 2 were 1-year survivors and both were severely neurologically impaired. The wisdom of providing CPR for >30 minutes must be questioned. These concerns should be tempered in the settings of 1) severe prearrest hypothermia, 2) severe toxic drug exposure, or 3) other potentially reversible processes when mechanical cardiopulmonary support (eg, extracorporeal membrane oxygenator) is immediately available.
Another factor commonly considered regarding termination of resuscitation efforts is the number of doses of epinephrine administered.1,8 In several in-hospital and out-of-hospital studies, none of the children who received >2 doses of epinephrine survived to hospital discharge.8,18–20,22,33 Therefore, some have recommended terminating resuscitation if there is no ROSC after 2 doses of epinephrine because additional efforts would be futile. In the present study, 16% (10 of 61) of the children who were treated with >2 doses of epinephrine survived 24 hours, and 3% (2 of 61) were discharged from the hospital and alive at 1 year. Although 24-hour survival was inversely related to the number of epinephrine doses, the multivariate analysis indicated that the association with decreased survival was independently attributable to longer duration of CPR, absence of respiratory failure as the precipitating cause, and absence of a pulse when CPR commenced rather than number of epinephrine doses. In addition, several other studies have documented long-term survival after >2 doses of epinephrine.24,29,35–37,39 Unfortunately, as in the present study, some of those survivors were severely neurologically impaired.24,36,37,39
The ultimate goal of CPR, like any other life-saving effort, is neurologically intact survival. In the present study, the final neurologic status was the same as the prearrest status in 91% of those discharged from the hospital and 83% of the 1-year survivors (Table 4). Similarly, other pediatric in-hospital studies that addressed the neurologic status at 6 months or 1 year after CPR found that 8 of 11,11 2 of 3,20 6 of 8,24 and 7 of 99 patients were either normal or unchanged from baseline status. These findings are reassuring and markedly superior to neurologic outcomes reported in survivors of out-of-hospital pediatric CPR.1,2,4 More subtle neurologic changes, such as learning disabilities, were not assessed with the PCPC scales. Morris et al40 demonstrated subtle neurologic deficits in survivors of cardiac arrest; however, they did not determine the presence or absence of neurologic deficits in the prearrest state. In our study, most of the children (71%) had preexisting chronic disorders, and 7 of 22 long-term survivors had pre-CPR gross neurologic dysfunction. Likewise, in the 2 other in-hospital pediatric studies that systematically addressed pre-CPR neurologic status and 6-month or 1-year post-CPR status, 3 of 8 6-month survivors in 1 study24 and 5 of 9 1-year survivors in another9 were grossly abnormal before CPR.
Although the Utstein reporting style used in this study was developed by international resuscitation council consensus, there are inherent problems with practical operational definitions. Cardiac arrest was defined as the cessation of the detectable cardiac mechanical activity, determined by the absence of a palpable central pulse, unresponsiveness, and apnea. However, it is recognized that detection of a pulse is difficult, particularly in infants. Detection of pulse was not an issue for the 115 of the 129 ECG monitored and witnessed events treated with chest compressions because of documented asystole, VF, or severe bradycardia with poor perfusion and palpable pulses (Table 3). It is possible that some of the other 14 patients (12 with pulseless electrical activity and 2 with unknown rhythms) had pulses and did not need chest compressions. In light of their lack of other signs of circulation (apnea, no movement, no cough, no response to stimulation) and poor 24-hour survival (only 3 of 14 vs 40 of 115 among the others), it is probable that these 14 patients needed chest compressions. Another potential problem is that the true initial arrest ECG rhythms cannot be known for patients who were not continuously ECG monitored immediately before the onset of pulselessness. Nevertheless, 98 children had continuous ECG monitoring before CPR was provided; their initial ECGs at the time of their events were witnessed by medical professionals. Another 29 had ECG rhythm determination within 1 or 2 minutes of the initiation of chest compressions. It is possible but unlikely that these patients converted from VF to another rhythm during this time.
Prospective assessment of pediatric in-hospital CPR using the Utstein reporting style demonstrated that 1) CPR was uncommon but not rare; 2) respiratory failure was the most common precipitating cause of the event, followed by circulatory shock; 3) the initial ECG abnormality was most commonly asystole, and bradycardia was the next most common; 4) VF was rare; 4) preexisting chronic diseases were prevalent; 5) most 1-year survivors had no change from their pre-CPR functional neurologic status; 6) many factors correlated with survival, but the only independent predictors of 24-hour survival were respiratory failure as the precipitating cause, the presence of a palpable pulse when CPR commenced, and shorter duration of CPR; and 7) the relative risk of death at 30 days increased by 5% for each elapsed minute of CPR. We encourage others to evaluate and report their pediatric CPR experiences using this Utstein style of uniform data reporting, especially at institutions with trauma and cardiac surgery patients. An improved understanding of the epidemiology of in-hospital pediatric CPR is a necessary step in our quests to identify risk factors for sudden cardiac arrest and improve treatment and outcomes.
- Received June 5, 2001.
- Accepted August 31, 2001.
- Reprint requests to (R.A.B.) Pediatrics/3302, 1501 N Campbell Ave/Box 245073, Tucson, AZ 85724-5073. E-mail:
- ↵American Heart Association in Collaboration with the International Liaison Committee on Resuscitation. Guidelines 2000 for cardiopulmonary resuscitation and emergency cardiovascular care, pediatric advanced life support and pediatric basic life support. Circulation.2000;102(suppl I) :I253– I342
- ↵Zaritsky A, Nadkarni V, Hazinski MF, et al. Recommended guidelines for uniform reporting of pediatric advanced life support. The Pediatric Utstein Style: a statement for healthcare professionals from a task force of the American Academy of Pediatrics, the American Heart Association, and the European Resuscitation Council [review]. Pediatrics.1995;96 :765– 779
- ↵Berg RA, Goetting MG. Pediatric sudden death. In: Paradis NA, Halperin HR, Nowak RM, eds. Cardiac Arrest: The Science and Practice of Resuscitation Medicine. Baltimore, MD: Williams & Wilkins;1996:671– 694
- ↵Carpenter TC, Stenmark KR. High-dose epinephrine is not superior to standard-dose epinephrine in pediatric in-hospital cardiopulmonary arrest. Pediatrics.1997;99 :403– 408
- ↵Cummins RO, Chamberlain D, Abramson NS, et al. Recommended guidelines for uniform reporting of data from out-of-hospital cardiac arrest: the Utstein style. Circulation.1991;84 :960– 975
- ↵Cummins RO, Chamberlain D, Hazinski MF, et al. Recommended guidelines for reviewing, reporting, and conducting research on in-hospital resuscitation: the in-hospital “Utstein style.” Circulation.1997;95 :2213– 2239
- ↵Innes PA, Summers CA, Boyd IM, Molyneux EM. Audit of paediatric cardiopulmonary resuscitation. Arch Dis Child.1993;68 :487– 491
- ↵Friesen RM, Duncan P, Tweed WA, Bristow G. Appraisal of pediatric cardiopulmonary resuscitation. Can Med Assoc J.1982;126 :1055– 1058
- ↵del Nido P, Dalton HJ, Thompson AE, Siewers RD. Extracorporeal membrane oxygenator rescue in children during cardiac arrest after cardiac surgery. Circulation.1992;86 :300– 304
- ↵Berg RA, Hilwig R, Kern KB, Ewy GA. Bystander chest compressions and assisted ventilation independently improve outcome from piglet asphyxial pulseless “cardiac arrest.” Circulation.2000;101 :1743– 1748
- ↵Southall DP, Kilpatrick SM. Imipramine poisoning: survival of a child after prolonged cardiac massage. BMJ.1974;4 :508
- ↵Christiansen DW, Jansen P, Perkin RM. Outcome and acute care hospital costs after warm water near drowning in children. Pediatrics.1997;99 :715– 721
- ↵Nichter MA, Everett PB. Childhood near-drowning: is cardiopulmonary resuscitation always indicated? Crit Care Med.1989;89 :993– 995
- ↵Dieckmann RA, Vardis R. High-dose epinephrine in pediatric out-of-hospital cardiopulmonary arrest. Pediatrics.1995;95 :901– 913
- Copyright © 2002 by the American Academy of Pediatrics