REVIEW ARTICLE
Therapeutic Hypothermia in Children After Cardiac Arrest A Systematic Review and Meta-analysis Janice F. Bistritz, MS, CPNP, DNP, Lauren M. Horton, MS, CPNP, DNP-C, and Arlene Smaldone, PhD, CPNP-BC, CDE Background: Therapeutic hypothermia (TH) has been shown to be effective in resuscitation of some adults following cardiac arrest and infants with hypoxic ischemic encephalopathy, but has not been well studied in children. Objectives: The purpose of this systematic review/meta-analysis was to examine mortality, neurologic outcomes, and adverse events in children following use of TH. Results: A search of PubMed, the Cumulative Index to Nursing and Allied Health Literature, and the Institute for Scientific Information’s Web of Knowledge from 1946 to 2014 yielded 6 studies (3 retrospective and 3 prospective cohort studies) that met our inclusion criteria. Quantitative synthesis of mortality following TH (136 subjects) was 44% (95% confidence interval, 32-57) with 28% (95% confidence interval, 11-53) of survivors (42 subjects) demonstrating poor neurologic outcome. The most frequently reported adverse events were electrolyte imbalances and pneumonia. Conclusions: Evidence is insufficient to support the advantage of TH compared with normothermia in pediatric resuscitation. The adverse event profile appears to be different than that reported in adults. Further studies are needed before TH may be considered a standard protocol for children after cardiac arrest. Key Words: therapeutic hypothermia, cardiac arrest, children, systematic review, meta-analysis (Pediatr Emer Care 2015;31: 296–303)
P
ediatric cardiac arrest (CA) is an unusual yet devastating event, often associated with decreased survival and increased neurological morbidity. The economic costs of the sequelae are substantial, as is the loss of future productive life.1 Approximately 16,000 American children experience an out-ofhospital CA (OHCA) each year.1–3 Approximately 2% of children in pediatric intensive care units have an in-hospital CA (IHCA).2–4 The rates of survival to hospital discharge vary greatly between OHCA and IHCA. Out-of-hospital CA carries a poorer prognosis.5,6 Survival in children after OHCA is lower (2%-12%) compared with IHCA (27%-51%),1,7–9 with infants having the greatest mortality, whereas children and adolescents exhibit a higher survival rate compared with adults1,3,10 Poor neurologic outcomes among survivors of OHCA are very common (76%) compared with IHCA (24%-53%).4,7,8,11,12 To achieve favorable neurological recovery after successful resuscitation, it is imperative to stop the ischemia process. The cessation of cerebral blood flow from CA causes hypoxia, which leads to ischemic injury to the brain. Even when resuscitation efforts are successful, additional damage can result from reperfusion injury. This insult to the sensitive neurological structures From Columbia University School of Nursing, New York, NY. Disclosure: The authors declare no conflict of interest. Reprints: Janice F. Bistritz, MS, CPNP, DNP-C, Columbia University School of Nursing, 630 W 168th St, New York, NY 10032 (e‐mail: [email protected]). Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0749-5161
296
www.pec-online.com
leads to permanent damage. Many treatments have been attempted to reduce or reverse the hypoxic and reperfusion injury to the neurological system; however, none have demonstrated improved outcomes.13 For 40 years, researchers have studied the use of therapeutic hypothermia (TH) after CA to determine if its use would improve neurological outcomes.3 The mechanisms underlying the beneficial effects of hypothermia are complex and include reductions in the cerebral metabolism of glucose and oxygen consumption, thereby mitigating the destructive processes that occur after the insult. These destructive pathways include the accumulation of excitotoxic neurotransmitters, intracellular acidosis, neuroinflammation, apoptosis, free radical production, and seizure activity.14,15 In his systematic review, Bernard5 defined TH as the controlled lowering of core body temperature to approximately 34°C, the optimal balance between clinical effect and cardiovascular toxicity. This temperature provides for easier clinical management because shivering can occur between 34°C and 35.5°C and may require suppression by additional amounts of sedation or neuromuscular blockade.5 Lower core temperature was found to be counterproductive, as temperatures between 28°C and 31°C were associated with atrial fibrillation and ventricular fibrillation, respectively.5 To date, TH is the only treatment demonstrating laboratory and clinical efficacy in improving neurological outcomes in selected circumstances. Therapeutic hypothermia improves neurologic outcomes in some adults following OHCA,13,16–19 as well as in newborns who sustained perinatal asphyxia.19–22 Infrequent use of TH in the pediatric population was noted in an earlier systematic review5; in addition, in the 3 studies that examined TH in the treatment of pediatric survivors of submersion accidents, outcomes were inconsistent.9 At this time, there are no completed randomized trials utilizing TH in children after CA.23 The American Heart Association, in 2005, suggested that induced hypothermia be considered in children who remain comatose after cardiopulmonary resuscitation,24 and in 2010, the Committee for the International Consensus on Cardiopulmonary Resuscitation and Emergency Cardiovascular Care Science further recommended considering TH in the care of pediatric patients following resuscitation.25 The purpose of this systematic review/metaanalysis was to examine mortality, neurologic morbidity of survivors, and type and frequency of adverse events in pediatric patients treated with TH following resuscitation.
METHODS Search Strategy We searched the literature using PubMed, Cumulative Index to Nursing and Allied Health Literature, and Web of Knowledge to identify studies that examined the intervention of TH in the treatment of pediatric patients after CA. To reflect all research on this topic, we searched all studies published from 1946 until January 2014. Subject headings used in the search included therapeutic hypothermia, cardiac arrest, and child. These headings Pediatric Emergency Care • Volume 31, Number 4, April 2015
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Pediatric Emergency Care • Volume 31, Number 4, April 2015
were then exploded to obtain a more complete listing of all related resources. All studies reporting mortality and morbidity outcomes of TH in treatment of pediatric patients, from birth to age 21 years, who had CA were reviewed. The reference lists of all pertinent articles were hand searched to identify any additional studies not previously identified. Studies in which TH was utilized for reasons other than CA (eg, surgery, traumatic brain injury), where the outcomes of morbidity and mortality were not studied or reported, the subjects were older than 21 years, were published in a language other than English or Spanish, or were literature reviews, editorials, individual case studies, or letters to the editor were excluded.
Assessment of Study Quality Two reviewers (L.M.H., J.F.B.) independently appraised each study for quality using the Downs and Black26 criteria, a 27-item scale developed to assess the quality of nonrandomized intervention studies. The instrument has good test-retest and interrater reliability, high internal consistency, and good criterion validity. We modified the scale by deleting 9 items that were not relevant to observational studies. Each indicator was scored as 0 (criterion not met) or 1 (criterion met) with the exception of 1 item (indicator 5), which was scored as 0 (criterion not met), 1 (criterion partially met), or 2 (criterion met). After independent review of the studies, scores were assigned. If there was disagreement regarding the quality scores, the topic was discussed, and a consensus was reached. Studies that achieved a score of 15 or greater of the 18 criteria were considered to be of good quality.
Data Extraction and Synthesis Data extracted from each study included the general characteristics of the sample, location of CA, qualities of hypothermia, setting of data collection, percent mortality after cardiac event, neurologic outcome, and type and frequency of reported adverse events. Neurologic outcome was reported using the Pediatric Cerebral Performance Category (PCPC) or Glasgow Outcome Scale (GOS). The PCPC scale is a validated scale designed to assess functional morbidity and cognitive impairment after a critical injury.27 The scale contains 6 categories; each addresses a level of functional impairment, with lower scores indicating a higher level of function.28 Researchers frequently utilize this scale to determine the impact of the CA and compare populations4,7,10,29–35 across studies. A related measurement scale, the GOS,36,37 was originally devised to describe outcomes of severely head injured patients. This scale is a variation of the Glasgow Coma Scale, which is widely used in the Utstein protocol of reporting CA features.38 The GOS contains 5 categories, with higher scores indicating higher levels of function.39 Similar to the PCPC, the GOS has also been used by researchers to quantify the effect of subsequent neurologic injury after CA.40 Table 1 compares features of both scales. We defined poor neurologic outcome as PCPC of 4 or
Therapeutic Hypothermia After Cardiac Arrest
greater or GOS of 3 or less (severe disability, coma or vegetative state, or death).
Data Analysis Studies reporting mortality and/or neurologic outcome following use of hypothermia during resuscitation were eligible for inclusion in meta-analysis. Data were extracted as frequency and sample size and standardized effect sizes computed. Neurologic outcome was determined for survivors of CA. A pooled effect was estimated for each outcome using a random-effects metaanalysis model. Heterogeneity was assessed using Cochran Q and I2 statistics. Data were analyzed using Comprehensive Metaanalysis statistical software (Biostat, Inc, Edgewood, NJ). Results are presented as forest plots.
RESULTS Figure 1 provides details of the literature search and ultimate selection of the included studies. Of 562 articles identified, 6 observational (3 retrospective29,35,40 and 3 prospective cohort30–32) studies were included in the systematic review. Based on their study design, these studies provide levels 3 to 4 evidence for prognostic studies,41 with level 1 evidence being of highest quality. Table 2 reports the results of the evaluation of the methodological quality of the included studies. The quality scores of the studies ranged from 13 to 18. The 3 retrospective studies29,35,40 and 2 of 3 of the prospective studies30,31 were found to be of good quality (≥15). Only the pilot study32 did not meet the standard for good quality. Of the 6 studies, only 1 adjusted analyses for confounding variables. Table 3 summarizes the characteristics of the included studies. A total of 355 subjects (hypothermia: n = 136, standard care: n = 219) were included in the studies, with ages ranging from younger than 1 year to 21 years. Sample sizes of the individual studies ranged between 6 subjects32 and 181 subjects.40 The majority of studies were conducted in the United States,30,31,40 with the remaining studies distributed among Canada, the United Kingdom, Taiwan, and Chile.29,32,35 One study was multisite29 and involved 5 university-affiliated tertiary care institutions. The 3 retrospective studies30,35,40 included a comparison group of normothermic resuscitation. Each study identified the location of CA as either out of or in hospital, with the majority (57.4% of the 136 subjects treated with TH) having sustained an OHCA. The methods of induction of TH varied among the studies and some studies used a combination of methods. The methods included use of ice packs,29,31,32 cooling blankets,29–32,35,40 extracorporeal membrane oxygenation (ECMO),29 and iced saline administered either by gastric lavage40 or intravenously.29–32 With the exception of ECMO, methods of inducing hypothermia are readily available in all tertiary care centers. The time to induction of TH was reported as being within 8 hours from time of CA,29,31,32,35 admission to the intensive care unit,40 or return of spontaneous circulation (ROSC).30
TABLE 1. Comparison of Scoring Criteria of PCPC and GOS PCPC 1 2 3 4 5 6
GOS
Category
Description
5 4
Normal Mild disability Moderate disability Severe disability Coma or vegetative state Brain death
Age-appropriate level of development or functioning Alert, conscious, possibility of mild neurologic deficit Conscious, independent activities of daily life, learning deficits Conscious, dependent for daily support Coma, cerebral unresponsiveness and some reflexive response Apnea and/or electroencephalographic silence
3 2 1
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.pec-online.com
297
Pediatric Emergency Care • Volume 31, Number 4, April 2015
Bistritz et al
Mortality and neurological performance scores were reported in all studies, although the end points varied from time of discharge30,31,35,40 to 6 months after CA.29,32 The PCPC score27 was used as a measurement of neurological outcome in the majority of studies.29–32,35 The process of instituting TH was the primary outcome in 2 of the studies: Topjian et al31 evaluated the feasibility of TH, and Fink et al40 assessed the frequency of adverse events as a result of TH. The addition of electroencephalogram monitoring during the hypothermia procedure and its use as a predictor of outcomes were other features addressed in the study of Kessler et al.30 Table 4 provides a synthesis of the findings of the studies included in this review. Mortality varied among studies. In the majority of studies,29–31,40 mortality for subjects treated with TH was 40% or greater, with 2 studies32,35 reporting mortality of less than 22%. Small sample size may have influenced this finding. In studies that compared mortality outcomes for subjects treated with TH with those receiving standard care or normothermia,29,35,40 a statistically significant advantage for the treatment group was found in only 1 study.35 In the study conducted by Doherty and colleagues,29 unadjusted mortality was higher for those who had received TH; however, when results were adjusted for confounders, there were no statistically significant differences between groups. Poor neurological performance, reported as a PCPC of 4 or greater, was greater than 33% in 3 of the studies.29–31 Lin and colleagues35 reported poor neurological outcome as a PCPC of 3 to 5, different than the parameters set by the other studies. Although poor outcome was found in only 18% of the TH group, it is difficult to compare the findings to the other studies. Fink and colleagues40 assessed poor neurologic outcome using the GOS scoring system as 3 or less and found no significant differences between the TH group and the standard care group. Bustos and colleagues32 reported the most favorable neurologic outcome for survivors of TH with no subject scoring in the unfavorable range. Adverse events were reported in each study; 1 study30 restricted its reporting of adverse events to seizures. Table 5 presents the frequency of occurrence of adverse events in children who
FIGURE 1. Literature search and selection of studies.
TABLE 2. Quality Appraisals of the Included Studies Quality of Reporting
External Validity
Prospective Cohort Studies Author
Year
I
II
III
IV
V
VI
VII
VIII
IX
X
XI
XII
XIII
XIV
XV
XVI
XVII
XVIII
Total
Kessler et al.30 Topjian et al31 Bustos et al32
2011 2011 2012
1 1 1
1 1 1
1 1 1
1 1 1
1 1 1
1 1 1
1 1 1
0 1 1
1 1 1
1 0 0
1 1 0
0 0 0
1 1 0
1 1 1
1 1 1
1 1 1
1 1 1
0 0 0
15 15 13
Retrospective Cohort Studies Authors 29
Doherty et al Fink et al40 Lin et al35
Year
I
II
III
IV
V
VI
VII
VIII
IX
X
XI
XII
XIII
XIV
XV
XVI
XVII
XVIII
Total
2009 2010 2013
1 1 1
1 1 1
1 1 1
1 1 1
2 2 2
1 1 1
1 1 1
1 1 1
1 1 1
1 1 1
1 1 1
1 0 1
0 1 0
1 1 1
1 1 1
1 1 1
1 1 1
1 0 1
18 17 18
I indicates objective described; II, outcomes described; III, sample characteristics described; IV, intervention described; V, confounders described; VI, findings described; VII, variability of data described; VIII, adverse events described; IX, characteristics of subjects lost to follow-up described; X, probability values for outcomes described; XI, subjects were representative of the population from which they were recruited; XII, subjects who were prepared to participate representative of the population from which they were recruited; XIII, were staff, places, and facilities where subjects were treated representative of the population from which they were recruited; XIV, if any study results were based on “data dredging” and whether this was made clear; XV, statistical tests were appropriate; XVI, main outcome measures used were valid and reliable; XVII, study subjects were recruited over the same period; XVIII, results were adjusted for confounders; Each criterion scored as 0, not met; 1, criterion met with exception of criterion V scored as 0, not met; 1, partially met; 2, criterion met.
298
www.pec-online.com
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Pediatric Emergency Care • Volume 31, Number 4, April 2015
Therapeutic Hypothermia After Cardiac Arrest
TABLE 3. Characteristics of the Included Studies
Author, Year
Sample Size Age
Characteristics of the Sample
Hypothermia Procedure
n = 79 Majority IH (95%) Induction: ≤6 h Doherty et al,29 2009 Aged >40 wk to 12 h after ROSC, ▪ ECMO PICU admission ▪ Cooling blanket after resuscitation ▪ Ice packs ▪ Iced IV fluids n = 181 Majority IH (54.9%) Induction: 5-8 h Fink et al,40 2010 Aged 1 week to 21 y Study inclusion criteria: Duration: 16-48 h Admitted to PICU with Type: ROSC after CA ▪ Cooling blanket ▪ Ice lavage with IV iced saline n = 35 Majority OH (83%) Induction: 7.3 ± 0.2 h Kessler et al,30 2011 0.18-16.6 y Study inclusion criteria: Duration 24 h admitted to PICU with Type: ROSC after CA ▪ Cooling blanket 31 n = 12 Majority OH (83%) Induction: 8 h with Type: ROSC after CA ▪ Cooling blanket ▪ Ice packs n=6 Majority OH (83%) Induction:
Therapeutic Hypothermia in Children After Cardiac Arrest A Systematic Review and Meta-analysis Janice F. Bistritz, MS, CPNP, DNP, Lauren M. Horton, MS, CPNP, DNP-C, and Arlene Smaldone, PhD, CPNP-BC, CDE Background: Therapeutic hypothermia (TH) has been shown to be effective in resuscitation of some adults following cardiac arrest and infants with hypoxic ischemic encephalopathy, but has not been well studied in children. Objectives: The purpose of this systematic review/meta-analysis was to examine mortality, neurologic outcomes, and adverse events in children following use of TH. Results: A search of PubMed, the Cumulative Index to Nursing and Allied Health Literature, and the Institute for Scientific Information’s Web of Knowledge from 1946 to 2014 yielded 6 studies (3 retrospective and 3 prospective cohort studies) that met our inclusion criteria. Quantitative synthesis of mortality following TH (136 subjects) was 44% (95% confidence interval, 32-57) with 28% (95% confidence interval, 11-53) of survivors (42 subjects) demonstrating poor neurologic outcome. The most frequently reported adverse events were electrolyte imbalances and pneumonia. Conclusions: Evidence is insufficient to support the advantage of TH compared with normothermia in pediatric resuscitation. The adverse event profile appears to be different than that reported in adults. Further studies are needed before TH may be considered a standard protocol for children after cardiac arrest. Key Words: therapeutic hypothermia, cardiac arrest, children, systematic review, meta-analysis (Pediatr Emer Care 2015;31: 296–303)
P
ediatric cardiac arrest (CA) is an unusual yet devastating event, often associated with decreased survival and increased neurological morbidity. The economic costs of the sequelae are substantial, as is the loss of future productive life.1 Approximately 16,000 American children experience an out-ofhospital CA (OHCA) each year.1–3 Approximately 2% of children in pediatric intensive care units have an in-hospital CA (IHCA).2–4 The rates of survival to hospital discharge vary greatly between OHCA and IHCA. Out-of-hospital CA carries a poorer prognosis.5,6 Survival in children after OHCA is lower (2%-12%) compared with IHCA (27%-51%),1,7–9 with infants having the greatest mortality, whereas children and adolescents exhibit a higher survival rate compared with adults1,3,10 Poor neurologic outcomes among survivors of OHCA are very common (76%) compared with IHCA (24%-53%).4,7,8,11,12 To achieve favorable neurological recovery after successful resuscitation, it is imperative to stop the ischemia process. The cessation of cerebral blood flow from CA causes hypoxia, which leads to ischemic injury to the brain. Even when resuscitation efforts are successful, additional damage can result from reperfusion injury. This insult to the sensitive neurological structures From Columbia University School of Nursing, New York, NY. Disclosure: The authors declare no conflict of interest. Reprints: Janice F. Bistritz, MS, CPNP, DNP-C, Columbia University School of Nursing, 630 W 168th St, New York, NY 10032 (e‐mail: [email protected]). Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved. ISSN: 0749-5161
296
www.pec-online.com
leads to permanent damage. Many treatments have been attempted to reduce or reverse the hypoxic and reperfusion injury to the neurological system; however, none have demonstrated improved outcomes.13 For 40 years, researchers have studied the use of therapeutic hypothermia (TH) after CA to determine if its use would improve neurological outcomes.3 The mechanisms underlying the beneficial effects of hypothermia are complex and include reductions in the cerebral metabolism of glucose and oxygen consumption, thereby mitigating the destructive processes that occur after the insult. These destructive pathways include the accumulation of excitotoxic neurotransmitters, intracellular acidosis, neuroinflammation, apoptosis, free radical production, and seizure activity.14,15 In his systematic review, Bernard5 defined TH as the controlled lowering of core body temperature to approximately 34°C, the optimal balance between clinical effect and cardiovascular toxicity. This temperature provides for easier clinical management because shivering can occur between 34°C and 35.5°C and may require suppression by additional amounts of sedation or neuromuscular blockade.5 Lower core temperature was found to be counterproductive, as temperatures between 28°C and 31°C were associated with atrial fibrillation and ventricular fibrillation, respectively.5 To date, TH is the only treatment demonstrating laboratory and clinical efficacy in improving neurological outcomes in selected circumstances. Therapeutic hypothermia improves neurologic outcomes in some adults following OHCA,13,16–19 as well as in newborns who sustained perinatal asphyxia.19–22 Infrequent use of TH in the pediatric population was noted in an earlier systematic review5; in addition, in the 3 studies that examined TH in the treatment of pediatric survivors of submersion accidents, outcomes were inconsistent.9 At this time, there are no completed randomized trials utilizing TH in children after CA.23 The American Heart Association, in 2005, suggested that induced hypothermia be considered in children who remain comatose after cardiopulmonary resuscitation,24 and in 2010, the Committee for the International Consensus on Cardiopulmonary Resuscitation and Emergency Cardiovascular Care Science further recommended considering TH in the care of pediatric patients following resuscitation.25 The purpose of this systematic review/metaanalysis was to examine mortality, neurologic morbidity of survivors, and type and frequency of adverse events in pediatric patients treated with TH following resuscitation.
METHODS Search Strategy We searched the literature using PubMed, Cumulative Index to Nursing and Allied Health Literature, and Web of Knowledge to identify studies that examined the intervention of TH in the treatment of pediatric patients after CA. To reflect all research on this topic, we searched all studies published from 1946 until January 2014. Subject headings used in the search included therapeutic hypothermia, cardiac arrest, and child. These headings Pediatric Emergency Care • Volume 31, Number 4, April 2015
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Pediatric Emergency Care • Volume 31, Number 4, April 2015
were then exploded to obtain a more complete listing of all related resources. All studies reporting mortality and morbidity outcomes of TH in treatment of pediatric patients, from birth to age 21 years, who had CA were reviewed. The reference lists of all pertinent articles were hand searched to identify any additional studies not previously identified. Studies in which TH was utilized for reasons other than CA (eg, surgery, traumatic brain injury), where the outcomes of morbidity and mortality were not studied or reported, the subjects were older than 21 years, were published in a language other than English or Spanish, or were literature reviews, editorials, individual case studies, or letters to the editor were excluded.
Assessment of Study Quality Two reviewers (L.M.H., J.F.B.) independently appraised each study for quality using the Downs and Black26 criteria, a 27-item scale developed to assess the quality of nonrandomized intervention studies. The instrument has good test-retest and interrater reliability, high internal consistency, and good criterion validity. We modified the scale by deleting 9 items that were not relevant to observational studies. Each indicator was scored as 0 (criterion not met) or 1 (criterion met) with the exception of 1 item (indicator 5), which was scored as 0 (criterion not met), 1 (criterion partially met), or 2 (criterion met). After independent review of the studies, scores were assigned. If there was disagreement regarding the quality scores, the topic was discussed, and a consensus was reached. Studies that achieved a score of 15 or greater of the 18 criteria were considered to be of good quality.
Data Extraction and Synthesis Data extracted from each study included the general characteristics of the sample, location of CA, qualities of hypothermia, setting of data collection, percent mortality after cardiac event, neurologic outcome, and type and frequency of reported adverse events. Neurologic outcome was reported using the Pediatric Cerebral Performance Category (PCPC) or Glasgow Outcome Scale (GOS). The PCPC scale is a validated scale designed to assess functional morbidity and cognitive impairment after a critical injury.27 The scale contains 6 categories; each addresses a level of functional impairment, with lower scores indicating a higher level of function.28 Researchers frequently utilize this scale to determine the impact of the CA and compare populations4,7,10,29–35 across studies. A related measurement scale, the GOS,36,37 was originally devised to describe outcomes of severely head injured patients. This scale is a variation of the Glasgow Coma Scale, which is widely used in the Utstein protocol of reporting CA features.38 The GOS contains 5 categories, with higher scores indicating higher levels of function.39 Similar to the PCPC, the GOS has also been used by researchers to quantify the effect of subsequent neurologic injury after CA.40 Table 1 compares features of both scales. We defined poor neurologic outcome as PCPC of 4 or
Therapeutic Hypothermia After Cardiac Arrest
greater or GOS of 3 or less (severe disability, coma or vegetative state, or death).
Data Analysis Studies reporting mortality and/or neurologic outcome following use of hypothermia during resuscitation were eligible for inclusion in meta-analysis. Data were extracted as frequency and sample size and standardized effect sizes computed. Neurologic outcome was determined for survivors of CA. A pooled effect was estimated for each outcome using a random-effects metaanalysis model. Heterogeneity was assessed using Cochran Q and I2 statistics. Data were analyzed using Comprehensive Metaanalysis statistical software (Biostat, Inc, Edgewood, NJ). Results are presented as forest plots.
RESULTS Figure 1 provides details of the literature search and ultimate selection of the included studies. Of 562 articles identified, 6 observational (3 retrospective29,35,40 and 3 prospective cohort30–32) studies were included in the systematic review. Based on their study design, these studies provide levels 3 to 4 evidence for prognostic studies,41 with level 1 evidence being of highest quality. Table 2 reports the results of the evaluation of the methodological quality of the included studies. The quality scores of the studies ranged from 13 to 18. The 3 retrospective studies29,35,40 and 2 of 3 of the prospective studies30,31 were found to be of good quality (≥15). Only the pilot study32 did not meet the standard for good quality. Of the 6 studies, only 1 adjusted analyses for confounding variables. Table 3 summarizes the characteristics of the included studies. A total of 355 subjects (hypothermia: n = 136, standard care: n = 219) were included in the studies, with ages ranging from younger than 1 year to 21 years. Sample sizes of the individual studies ranged between 6 subjects32 and 181 subjects.40 The majority of studies were conducted in the United States,30,31,40 with the remaining studies distributed among Canada, the United Kingdom, Taiwan, and Chile.29,32,35 One study was multisite29 and involved 5 university-affiliated tertiary care institutions. The 3 retrospective studies30,35,40 included a comparison group of normothermic resuscitation. Each study identified the location of CA as either out of or in hospital, with the majority (57.4% of the 136 subjects treated with TH) having sustained an OHCA. The methods of induction of TH varied among the studies and some studies used a combination of methods. The methods included use of ice packs,29,31,32 cooling blankets,29–32,35,40 extracorporeal membrane oxygenation (ECMO),29 and iced saline administered either by gastric lavage40 or intravenously.29–32 With the exception of ECMO, methods of inducing hypothermia are readily available in all tertiary care centers. The time to induction of TH was reported as being within 8 hours from time of CA,29,31,32,35 admission to the intensive care unit,40 or return of spontaneous circulation (ROSC).30
TABLE 1. Comparison of Scoring Criteria of PCPC and GOS PCPC 1 2 3 4 5 6
GOS
Category
Description
5 4
Normal Mild disability Moderate disability Severe disability Coma or vegetative state Brain death
Age-appropriate level of development or functioning Alert, conscious, possibility of mild neurologic deficit Conscious, independent activities of daily life, learning deficits Conscious, dependent for daily support Coma, cerebral unresponsiveness and some reflexive response Apnea and/or electroencephalographic silence
3 2 1
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
www.pec-online.com
297
Pediatric Emergency Care • Volume 31, Number 4, April 2015
Bistritz et al
Mortality and neurological performance scores were reported in all studies, although the end points varied from time of discharge30,31,35,40 to 6 months after CA.29,32 The PCPC score27 was used as a measurement of neurological outcome in the majority of studies.29–32,35 The process of instituting TH was the primary outcome in 2 of the studies: Topjian et al31 evaluated the feasibility of TH, and Fink et al40 assessed the frequency of adverse events as a result of TH. The addition of electroencephalogram monitoring during the hypothermia procedure and its use as a predictor of outcomes were other features addressed in the study of Kessler et al.30 Table 4 provides a synthesis of the findings of the studies included in this review. Mortality varied among studies. In the majority of studies,29–31,40 mortality for subjects treated with TH was 40% or greater, with 2 studies32,35 reporting mortality of less than 22%. Small sample size may have influenced this finding. In studies that compared mortality outcomes for subjects treated with TH with those receiving standard care or normothermia,29,35,40 a statistically significant advantage for the treatment group was found in only 1 study.35 In the study conducted by Doherty and colleagues,29 unadjusted mortality was higher for those who had received TH; however, when results were adjusted for confounders, there were no statistically significant differences between groups. Poor neurological performance, reported as a PCPC of 4 or greater, was greater than 33% in 3 of the studies.29–31 Lin and colleagues35 reported poor neurological outcome as a PCPC of 3 to 5, different than the parameters set by the other studies. Although poor outcome was found in only 18% of the TH group, it is difficult to compare the findings to the other studies. Fink and colleagues40 assessed poor neurologic outcome using the GOS scoring system as 3 or less and found no significant differences between the TH group and the standard care group. Bustos and colleagues32 reported the most favorable neurologic outcome for survivors of TH with no subject scoring in the unfavorable range. Adverse events were reported in each study; 1 study30 restricted its reporting of adverse events to seizures. Table 5 presents the frequency of occurrence of adverse events in children who
FIGURE 1. Literature search and selection of studies.
TABLE 2. Quality Appraisals of the Included Studies Quality of Reporting
External Validity
Prospective Cohort Studies Author
Year
I
II
III
IV
V
VI
VII
VIII
IX
X
XI
XII
XIII
XIV
XV
XVI
XVII
XVIII
Total
Kessler et al.30 Topjian et al31 Bustos et al32
2011 2011 2012
1 1 1
1 1 1
1 1 1
1 1 1
1 1 1
1 1 1
1 1 1
0 1 1
1 1 1
1 0 0
1 1 0
0 0 0
1 1 0
1 1 1
1 1 1
1 1 1
1 1 1
0 0 0
15 15 13
Retrospective Cohort Studies Authors 29
Doherty et al Fink et al40 Lin et al35
Year
I
II
III
IV
V
VI
VII
VIII
IX
X
XI
XII
XIII
XIV
XV
XVI
XVII
XVIII
Total
2009 2010 2013
1 1 1
1 1 1
1 1 1
1 1 1
2 2 2
1 1 1
1 1 1
1 1 1
1 1 1
1 1 1
1 1 1
1 0 1
0 1 0
1 1 1
1 1 1
1 1 1
1 1 1
1 0 1
18 17 18
I indicates objective described; II, outcomes described; III, sample characteristics described; IV, intervention described; V, confounders described; VI, findings described; VII, variability of data described; VIII, adverse events described; IX, characteristics of subjects lost to follow-up described; X, probability values for outcomes described; XI, subjects were representative of the population from which they were recruited; XII, subjects who were prepared to participate representative of the population from which they were recruited; XIII, were staff, places, and facilities where subjects were treated representative of the population from which they were recruited; XIV, if any study results were based on “data dredging” and whether this was made clear; XV, statistical tests were appropriate; XVI, main outcome measures used were valid and reliable; XVII, study subjects were recruited over the same period; XVIII, results were adjusted for confounders; Each criterion scored as 0, not met; 1, criterion met with exception of criterion V scored as 0, not met; 1, partially met; 2, criterion met.
298
www.pec-online.com
© 2015 Wolters Kluwer Health, Inc. All rights reserved.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Pediatric Emergency Care • Volume 31, Number 4, April 2015
Therapeutic Hypothermia After Cardiac Arrest
TABLE 3. Characteristics of the Included Studies
Author, Year
Sample Size Age
Characteristics of the Sample
Hypothermia Procedure
n = 79 Majority IH (95%) Induction: ≤6 h Doherty et al,29 2009 Aged >40 wk to 12 h after ROSC, ▪ ECMO PICU admission ▪ Cooling blanket after resuscitation ▪ Ice packs ▪ Iced IV fluids n = 181 Majority IH (54.9%) Induction: 5-8 h Fink et al,40 2010 Aged 1 week to 21 y Study inclusion criteria: Duration: 16-48 h Admitted to PICU with Type: ROSC after CA ▪ Cooling blanket ▪ Ice lavage with IV iced saline n = 35 Majority OH (83%) Induction: 7.3 ± 0.2 h Kessler et al,30 2011 0.18-16.6 y Study inclusion criteria: Duration 24 h admitted to PICU with Type: ROSC after CA ▪ Cooling blanket 31 n = 12 Majority OH (83%) Induction: 8 h with Type: ROSC after CA ▪ Cooling blanket ▪ Ice packs n=6 Majority OH (83%) Induction:
