Year in review 2011: Critical Care - Out-of-hospital cardiac arrest and trauma

In 2011, numerous studies were published in Critical Care focusing on out-of-hospital cardiac arrest, cardiopulmonary resuscitation, trauma, and some related airway, respiratory, and response time factors. In this review, we summarize several of these studies, including those that brought forth advances in therapies for the post-resuscitative period. These advances involved hypothesis-generating concepts in therapeutic hypothermia as well as the impact of early percutaneous coronary artery interventions and the potential utility of extracorporeal life support after cardiac arrest. There were also articles pertaining to the importance of timing in prehospital airway management, the outcome impact of hyperoxia, and the timing of end-tidal carbon dioxide measurements to predict futility in cardiac arrest resuscitation. In other articles, additional perspectives were provided on the classic correlations between emergency medical service response intervals and outcomes.


Introduction
During the year 2011, many innovative and thoughtprovoking resuscitation studies were published in Critical Care. A distinct portion of these articles focused on outof-hospital cardiac arrest (OHCA) and cardiopulmonary resuscitation (CPR), ranging from emergency medical service (EMS) response intervals and airway management in both trauma and cardiac arrest cases to the early use of percutaneous coronary artery intervention (PCI) and extracorporeal life support (ECLS), and several of the articles pertained to respiratory factors in CPR. Many of the studies discussed the timing of interventions, including the eff ect of the time to reach target temperature in therapeutic hypothermia.

Time intervals to reach target core temperature and outcome following induced hypothermia after resuscitation from cardiac arrest
Despite recent improvements in enhancing successful resuscitation from OHCA, overall long-term outcomes remain poor in most venues [1,2]. One well-touted intervention that has been shown to have benefi t in recent years is the use of post-resuscitative mild therapeutic hypothermia (MTH), showing both improved neurologic outcomes and long-term survival [3][4][5][6]. Consensus guidelines for post-arrest management now advocate its use in survivors of OHCA, particularly those who present with ventricular dysrhythmias [7]. Nevertheless, the ideal timeline for inducing the core temperature goal in MTH remains elusive [6,8].
Th is past year in Critical Care, Haugk and colleagues [9] sought to address this question in a single-center retrospective cohort study. Th e investigators evaluated 2,536 OHCA patients, of whom 23% (588) underwent MTH with a target core temperature of less than 34°C. Th e authors divided subjects into three groups: those arriving at the goal temperature (a) in less than 120 minutes, (b) in 120 to 220 minutes, and (c) more than 220 minutes. Th ey defi ned the achievement of a good neuro logic outcome as having a Pittsburgh Cerebral Perfor mance Category (CPC) score of 1 or 2 (normal or only mild impairment) at 6-month follow-up [9].
Th e authors found that patients reaching the goal tempera ture faster had signifi cantly worse 6-month neuro logic outcomes (odds ratio (OR) 1.86, 95% confidence interval (CI) 1.03 to 3.38) but that the time taken to achieve the goal temperature had no association with 6-month all-cause survival [9]. At fi rst glance, these fi ndings call into question guidelines advocating rapid cooling [8,10].
However, while it is indeed possible that faster cooling was directly causal in the poorer neurologic outcomes, it is more likely that the adverse outcomes were simply a refl ection of more pronounced cerebral and systemic insults in these patients. In this line of thinking, those severe insults resulted in dysfunction of cerebral and systemic thermoregulation and thus a more rapid cooling.
Th is study was clearly a provocative and hypothesisgenerating addition to the critical care literature. While one might argue that their fi ndings prompt the need for prospective controlled evaluations of alternative timelines for inducing hypothermia, the study also reinforced the universal need for scientists and clinicians to be cautious with the face-value interpretations of results. Impact of adding percutaneous coronary artery intervention as part of a routine post-resuscitation care bundle following cardiac arrest Cardiac arrest may result from numerous precipitating factors involving both cardiac and non-cardiac etiologies. However, the most common cause of OHCA is ongoing myocardial ischemia leading to a lethal cardiac dysrhythmia such as ventricular fi brillation (VF) [11]. Even when resuscitative eff orts lead to a successful return of spontaneous circulation (ROSC), OHCA patients still have a poor prognosis in many cases.
One root cause of the cardiac arrest and subsequent post-resuscitation morbidity and mortality is ongoing myocardial ischemia, most often a result of reversible coronary artery occlusion. While previous studies have shown no benefi t to systemic thrombolysis, there does appear to be some advantage to a more direct coronary intervention, regardless of initial electrocardiogram (ECG) fi ndings [12,13]. Furthermore, studies have indicated the outcome advantages of PCI after OHCA either as an individual intervention [14] or performed in combination with MTH [14][15][16].
In an attempt to further confi rm the bundled eff ects of PCI and MTH on outcomes after OHCA, Gräsner and colleagues [17] retrospectively analyzed the German Resuscitation Registry. Th e clinical course and outcomes of 584 patients who had achieved ROSC and hospital admission -out of 2,973 cases of OHCA -were analyzed. Primary outcomes examined included (a) 24-hour survival and (b) a CPC score of 1 or 2 at the time of hospital discharge. Of the 584 admitted patients, 81 (14%) received PCI alone, 73 (12.5%) received both PCI and MTH, and 106 (18%) received MTH alone. Of the 154 (26%) patients receiving PCI, 95% were considered to have had an arrest of presumed cardiac etiology, 66% of which were associated with initial ECG rhythms of VF or ventricular tachycardia (VT  [17]. Needless to say, the retrospective nature of this article leads to questions about selection bias, particularly a 'survival bias' (for example, patients surviving long enough to receive PCI obviously had a better chance of receiving the intervention and thus the correlation with survival). Nevertheless, despite those potential limitations, the data still seem to reconfi rm the likely advantage of PCI, including the possible synergy of the bundled approach of combining MTH with PCI following OHCA [18]. Moreover, this article helps to support the potential benefi t of PCI for all patients with OHCA of presumed cardiac etiology regardless of initial ECG fi ndings [13].

Oxygenation levels and end-tidal carbon dioxide concentrations during cardiopulmonary resuscitation eff orts
Published literature suggests that changes in oxygen saturation during resuscitation from cardiac arrest may aff ect outcome [19,20]. Whereas animal studies have shown an association between hyperoxia and increased mortality, results of human studies are less clear [21]. A 2010 study by the Emergency Medicine Shock Research Network (EMShockNet) research group showed an independent and stepwise relationship between hyperoxia and in-hospital mortality after resuscitation from cardiac arrest [20]. In an attempt to validate that work, Bellomo and colleagues [22] identifi ed all of the blood gas analyses obtained during the fi rst 24 hours of intensive care unit (ICU) admission for all survivors of cardiac arrest over a decade (2000 to 2009) by using a large Australian-New Zealand database. Th e authors published their fi ndings in Critical Care in 2011 [22].
In this study, relative 'hypoxia' was defi ned as an arterial partial pressure of oxygen (PaO 2 ) of less than 60 mm Hg or a PaO 2 /fraction of inspired oxygen (PaO 2 / FiO 2 ) ratio of less than 300 mm Hg whereas 'hyperoxia' was defi ned as a PaO 2 of more than 300 mm Hg. All other patients were classifi ed as 'normoxic' . Primary outcomes evaluated included in-hospital mortality and time to death in hours. A secondary outcome was successful discharge to home. Among the 12,806 patients who met inclusion criteria, 5.4% were subsequently excluded because of missing data or ICU readmission and 58% (6,968) died in the hospital.
In terms of outcomes, there was a signifi cant survival benefi t for those with normoxia versus hypoxia or hyperoxia (P <0.0001). Of these survivors, 65% were ultimately discharged to home. However, once injury severity was added to a regression model, the eff ect on in-hospital mortality was markedly decreased (OR 1.2, 95% CI 1.1 to 1.4 for hypoxia and OR 1.2, 95% CI 1.0 to 1.5 for hyperoxia), and hyperoxia no longer showed a statistically signifi cant association with discharge to home. Furthermore, the results demonstrated that an increased overall survival rate (65%), as compared with that of the previous EMShockNet study (44%), was possibly explained by a very high rate of MTH (median temperature of 34.9°C) in this cohort.
Whereas previous research has suggested an association between hyperoxia and poorer outcomes, this retrospective analysis by Bellomo and colleagues [22] demonstrated much weaker associations [20]. Recognizing the more concerning potential for hypoxia by restricting inspired FiO 2 , the investigators recom mended that clinicians take caution in ordering deliberate decreases in FiO 2 to avoid hyperoxia and that related protocols be revisited.
Another respiratory parameter that has been widely used in the management of CPR has been the quantitative monitoring of end-tidal carbon dioxide (ETCO 2 ). ETCO 2 has gained extensive support for monitoring advanced airway patency as well as the quality of CPR. It also has been used to predict the likelihood of successful resuscitation and thus used to help invoke termination of resuscitative eff orts [10,[23][24][25][26]. However, Grmec and colleagues [27] also demonstrated a signifi cant diff erence in initial ETCO 2 between asphyxial cardiac arrest and arrest secondary to VF/VT. In this case, the initial ETCO 2 was not a valid predictor of ROSC [27].
Accordingly, in a 2011 Critical Care publication, Lah and colleagues [28] reported a prospective evaluation of 325 OHCA patients between 2006 and 2009 at a single center in Slovenia. After the exclusion of 211 patients for various reasons, the remaining patients were divided into two groups: (a) 51 patients presenting with asphyxial arrest and an initial cardiac rhythm of asystole or pulseless electrical activity (PEA) and (b) 63 patients with cardiac arrest of presumed cardiac etiology and an initial rhythm of VF or pulseless VT. ETCO 2 measurements were made immediately following endotracheal intubation (ETI) and every minute thereafter. Interest ingly, initial ETCO 2 values were signifi cantly higher in the asphyxial group (6.74 ± 4.22 kPa versus 4.51 ± 2.47 kPa), but this diff erence was no longer signifi cant after 3 minutes of CPR. In addition, patients achieving successful ROSC had an initial ETCO 2 of more than 1.33 kPa, a fi nding consistent with established literature describing the futility of extended CPR in patients with lower initial levels [24][25][26].
One of the key fi ndings of this study, however, was that the ETCO 2 level was the most predictive of ROSC after the fi fth minute of CPR (asphyxial group: 6.09 ± 2.63 kPa versus 4.47 ± 3.35 kPa, P = 0.006; VF/VT group: 5.63 ± 2.01 kPa versus 4.26 ± 1.86 kPa, P = 0.015). Th us, this important work indicates that the association between ETCO 2 and ROSC should be established after 5 minutes of CPR performed in accordance with published guidelines [29].

Extracorporeal life support for out-of-hospital cardiac arrest
In recent years, ECLS has been proposed as an eff ective therapy not only for in-hospital cardiac arrest but also for OHCA [30][31][32][33][34][35][36]. However, outcomes with this novel therapy have been mixed. Le Guen and colleagues [37] added to this growing body of literature in a 2011 Critical Care article describing their ECLS experience with 51 OHCA patients whose arrest was witnessed but ROSC was not achieved after 30 minutes of resuscitative eff orts and the institution's mobile cardiothoracic surgery team was available. In cases of refractory cardiac arrest, an automated compression device was used. ECLS was established by using femoral-femoral cannulation, and the mean time to ECLS was 120 minutes from initiation of CPR. All patients also received 24 hours of MTH.
Of the 51 patients meeting inclusion criteria, 9 (18%) failed successful cannulation and only 17 (42%) were alive at 24 hours. Two (4% of the total 51 candidates) survived to day 28 (95% CI 1% to 13%). Both surviving patients had a no-fl ow time of less than 5 minutes but a low-fl ow time of more than 100 minutes. One was successfully weaned at day 5, but the other required 36 days. Both patients had a Glasgow Outcomes Scale score of 4 or 5 (relatively favorable outcomes) at 28 days and 6 months [37].
Although this study by Le Guen and colleagues [37] showed lower survival rates than those of previous studies, both survivors had favorable neurologic outcomes and 88% of non-survivors expired within 48 hours, contradicting the fear that ECLS may unnecessarily prolong life in futile cases. Also, the long (2-hour) mean initiation time may have contributed to these poorer outcomes. Th erefore, this study supports the potential benefi ts of ECLS for OHCA and suggests important steps for improvements in ECLS protocols such as earlier implementation [37].

Additional perspectives on emergency medical service response intervals and outcomes following cardiopulmonary arrest
Shorter elapsed time intervals from collapse to initiation of CPR have been established as important contributors to favorable outcomes following OHCA [38][39][40]. However, most of these studies have focused on OHCA of presumed cardiac etiology. Koike and colleagues [41] used a large nationwide database to examine undiff er entiated OHCA and described their fi ndings in a 2011 Critical Care article. By means of logistic regression techniques, 109,350 cases in which the collapse was witnessed by bystanders were analyzed. Only half of the cases were of presumed cardiac etiology. Bystanderperformed CPR was initiated in 45% of cases, and an automated external defi brillator (AED) was used in 0.8%. Th e mean collapse to EMS CPR interval (CECI) was 14.5 minutes.
Th e authors demonstrated a strong correlation between CECI and 1-month survival (R 2 = 0.98) and 1-month CPC scores of 1 or 2 (R 2 = 0.95). Consistent with existing literature, 1-month survival and good neurologic outcomes were associated with a cardiac etiology, bystander CPR, and public AED use [42][43][44]. In essence, this study confi rmed existing information, but with a much larger and diverse data set, and strongly reinforces the critical importance of public access to AEDs as well as community training in bystander CPR [45].
Another perspective on response intervals was provided this past year in Critical Care by Neukamm and colleagues [46], who used the German Resuscitation Registry, a national OHCA database established by the German Society for Anesthesiology and Intensive Care to evaluate 'response time reliability' (RTR) on patient survival to hospital admission [46,47]. RTR was defi ned as the percentage of cases in which the fi rst responding EMS crew arrived on the scene within 8 minutes of the call for help [46].
Th e group examined 2,330 cases over several years in seven selected EMS systems servicing a variety of population densities in Germany. All systems performed resuscitation based on 2005 guidelines [29]. Of the cases examined, 46.7% achieved ROSC in the fi eld, 42.8% survived to hospital admission, and 15.4% survived to hospital discharge. As expected, there was a signifi cant association between higher RTR and survival to hospital admission (24.4/100,000 population/year for RTR of more than 70% versus 15.6/100,000 population/year for RTR of less than 70%, OR 1.57, 99% CI 1.274 to 1.935), although no signifi cant diff erences were found in rates of ROSC. However, all seven EMS systems achieved higherthan-expected rates of ROSC based on resus ci tation after cardiac arrest (RACA) scores [47,48]. Th is fi nding was possibly due to the high utilization of therapeutic hypothermia in these seven centers (46.2% versus 13.7% across the entire region).
Th e frequency of bystander-initiated CPR was low overall and showed great variability across systems, ranging from 1.3% to 28.6% (P <0.001). Th e two largest cities had the fastest response times (67.9% and 64.2% of resuscitations were initiated within 8 minutes), whereas two other systems achieved RTR in only 56% and 53% of cases (P <0.001). Th e less reliable systems had the lowest frequencies of CPR performance as well. In essence, these fi ndings indicated specifi c areas for improvement in terms not only of achieving RTR targets but of frequency of bystander CPR. In addition, it should be noted that 8 minutes is an arbitrary target, and further delineation of even shorter RTRs may be worthwhile.

Evolving insights on prehospital airway management
Airway management in OHCA has been the topic of much debate in recent years. ETI has long been considered the primary advanced airway maneuver, but their benefi t over extraglottic airways (EGAs) in the OHCA setting has become uncertain for many medical decision makers [49]. In addition, the optimal time in which an advanced airway should be placed has not been defi ned. Th erefore, in a 2011 Critical Care article, Kajino and colleagues [50] compared outcomes with EGA and ETI use by EMS personnel for the treatment of witnessed non-traumatic OHCA in Osaka, Japan, and also examined the timeline of placement of the airway and its relationship to outcomes.
Of 5,377 adult patients meeting inclusion criteria, there were no signifi cant diff erences in 1-month survival chances between EGA (9.8%) and ETI (10.7%) and no signifi cant diff erence in 1-month neurologic outcomes between the EGA (3.6%) and ETI (3.6%). However, there was a statistically signifi cant diff erence in ROSC between EGA placement (10.1%) and ETI (16.6%). In context though, only advanced practitioners with additional training are allowed by protocol to give catecholamines and place endotracheal tubes. Th erefore, the addition of catecholamines may explain this improved ROSC despite the similar long-term outcomes. Furthermore, the use of these advanced practitioners had an independent association with favorable neurologic outcomes (OR 1.86, 95% CI 1.04 to 3.34).
Still, one of the most interesting fi ndings of this study was the eff ect of time to establishment of an advanced airway and outcome. Th e investigators found a statistically signifi cant decrease in favorable neurologic outcome with each passing minute (adjusted OR 0.91, 95% CI 0.88 to 0.95). So these data seem to indicate that, if an advanced airway is to be placed, early placement may very well be benefi cial. However, others may caution that the early airway placement would be acceptable but only assuming that there are no signifi cant interruptions to the performance of high-quality chest compressions [50,51].
In the realm of trauma care, while advanced airway management is typically part of the protocol in severe injury, the advantages of prehospital ETI for more moderate trauma are less clear [52,53]. In a 2011 Critical Care article, Hussman and colleagues [54] examined the German Association for Trauma Surgery database from 2005 to 2008 to evaluate the impact of prehospital ETI on moderately injured patients. Th e authors performed a matched-pairs analysis of 600 patients from Austria or Germany who were undergoing prehospital ETI, were more than 16 years old, had a Glasgow Coma Scale score of 13 to 15 and maximum injury severity per body region (Abbreviated Injury Score) of not more than 3, and who received no packed red blood cells in the emergency department. Patients were matched to 600 non-intubated controls.
Th e authors found no signifi cant diff erence in survival to hospital discharge between groups. However, mortality in both groups was quite low: 1% (ETI) and 0.5% (control). Both groups also had low rates of emergency operative intervention. However, the ETI patients had signifi cantly higher levels of volume replacement; decreases in hemoglobin, prothrombin time, and platelet count; and increased levels of hypothermia. Th ey also had signifi cantly lower blood pressures and higher heart rates at hospital admission and received more prehospital invasive procedures, including chest tubes and CPR. Furthermore, the ETI group had a signifi cantly higher rate of multi-organ failure and sepsis. Finally, intubated patients had signifi cantly longer ICU and total hospital stays and a 26.1% higher hospital cost.
Th ese fi ndings may refl ect simply selection bias and inadequate case controls. Also, the sympatholytic properties of sedatives and paralytics associated with intubation may explain the variation in vital signs [55]. In addition, ETI patients received signifi cantly more volume replacement, which may have contributed to hypothermia, clotting factor dilution, and decreases in hemoglobin and platelet counts [56]. Furthermore, the detrimental eff ects of positive pressure ventilation in the face of relative hypovolemia could also aff ect blood pressure and pulse. Likewise, the incidence of CPR was higher in the intubated cohort, again refl ecting a 'sicker' population.
At fi rst glance, therefore, this study may give pause to using ETI in moderately injured patients, but one might argue that the results refl ect that a sicker population had received ETI and that it may even have been benefi cial considering the relative equality in outcomes. Th e jury is still out, but this article gives us more concepts to consider in this ongoing debate.

Conclusions
In articles published in Critical Care in 2011, out-ofhospital cardiac arrest and trauma care were central topics of research. Authors added to the growing evidence in support of PCI for survivors of OHCA, and time to goal temperature for MTH was found to be a complicated issue. Concerns over hyperoxia in cardiac arrest survivors were again raised, although the impact of this fi nding may not be as clear as previously thought. Th e ability of ETCO 2 measurement to predict return of pulse was further defi ned in terms of timelines for making decisions to terminate resuscitative eff orts. Th e articles also provide some evidence for ECLS as a potential intervention for OHCA. Authors further highlighted the importance of faster responses by prehospital care providers, bystander CPR, and early AED use in the management of OHCA and provided new perspectives such as response time reliability. Finally, it appears that the earliest placement of an advanced airway is important and that extraglottic airways may be a useful alternative to ETI in OHCA. However, in the moderately injured patient, providers may need to be more conservative in their decision to place an endotracheal tube. Nevertheless, depending upon the patient, ETI may still have important utility in trauma care.