This qualitative review examines three retrospective cohort studies that evaluated endotracheal intubation (ETI) in nontraumatic adult out-of-hospital cardiac arrest (OOHCA) patients, as performed by paramedics in the prehospital setting. Studies were identified through a MEDLINE search, and three articles were discovered that compared intubation versus no intubation for these OOHCA patients with survival to hospital discharge as the measured outcome. Careful consideration and examination of these studies and their data supports the conclusion that ETI is associated with decreased survival to hospital discharge among adult nontraumatic OOHCA patients.
Introduction
Each year nearly 300,000 people with cardiovascular disease die from out-of-hospital cardiac arrest (OOHCA). For a quarter of a century, endotracheal intubation (ETI) has been the standard practice intervention for appropriate airway protection and subsequent ventilation during cardiopulmonary resuscitation (CPR) in the prehospital setting. Traditional CPR begins with securing a proper airway, breathing for the patient, and establishing circulatory support (commonly known as the “ABCs”). This mnemonic also establishes the initial priority and order of CPR interventions, although this priority has shifted in light of research demonstrating better survival in patients receiving chest compressions with minimal interruption (“cardiocerebral resuscitation” or CCR). The amount of time it takes to intubate a patient in the prehospital setting has been demonstrated to be significant, and thus the utility of ETI in OOHCA patients has been called into question. Today, intubating these patients remains a controversial topic, and the evolution of simpler advanced airway devices has fueled the fire for ongoing debate. The decrease in survival for OOHCA patients in trauma has been well defined, however studies exclusively comparing ETI with basic life support (BLS) airway management in nontraumatic OOHCA patients remain scarce. This paper reviews these few studies in a qualitative manner in order to discuss whether successful prehospital ETI affects survival in adult patients in nontraumatic OOHCA.
Methods
Search Strategy
A comprehensive MEDLINE search was used to identify a list of studies applicable to the focused clinical question. All potential articles were required to include ETI as the intervention and survival to hospital discharge as the appropriate outcome. The following MeSH criteria were searched in various combinations to generate a broad pool of potential articles: “prehospital,” “cardiac arrest,” “endotracheal intubation” or “intubation,” and “nontraumatic.” The search was limited to research written in the English language. Articles specifically examining ETI in OOHCA trauma patients were discarded, as well as those examining cardiac arrest intubations in the pediatric population (under 18 years of age). This search strategy identified three articles that met the criteria for this review.
Study Selection and Evaluation
Three retrospective cohort studies were identified. This study design was deemed appropriate for evaluating the clinical question, based on the accuracy of data collection and the possible ethical issues that a randomized-control trial (RCT) might pose in this setting. Patient care reports (PCRs) from Emergency Medical Services (EMS) records and/or hospital databases were examined by the three studies in order to synthesize data and evaluate survival outcomes. All three papers were published in 2010 and 2011, reflecting data that incorporated more recent CPR guidelines and standard practices. The ability to thoroughly analyze PCRs and hospital electronic medical records within the past few years allowed the three studies to be deemed a reliable method to answer the clinical question. A more comprehensive evaluation of each study follows.
Study Reviews
Studnek et al.
Summary. Acknowledging the growing research that shows CCR improves survival in OOHCAs, the authors sought to further investigate the advanced airway management of these patients. ETI may cause a significant interruption in chest compressions, and multiple unsuccessful attempts will decrease the amount of time that proper CPR is administered. Provider-performed unintentional hyperventilation (i.e. sufficiently raising intrathoracic pressure so that coronary perfusion is decreased) also remains a concern with advanced airway management. With this in mind, Studnek et al. retrospectively analyzed data from Mecklenburg County, NC, tracking OOHCAs that occurred between July 1, 2006 and December 31, 2008. Adult patients were excluded if they were involved in a drowning, electrocution, trauma, inter-facility transfer, or were pronounced dead on arrival (DOA). Patients were also removed if documentation did not provide the number of successful or failed ETI attempts during patient care. A total of 1,142 patients met the inclusion criteria. Of this group, 577 individuals (50.5%) were successfully intubated on the first attempt, 292 (25.5%) required multiple attempts, and 203 (17.8%) of the patients did not receive any ETI attempts.
Results. Of the 1,142 patients, 299 had prehospital return of spontaneous circulation (ROSC) and 118 (39.5%) survived to hospital discharge. Forty-eight (16.0%) had an unknown discharge status, and were thus conservatively classified as not surviving to discharge. It was determined that when comparing patients with one successful ETI attempt versus those with no attempts, the latter group of patients were 5.46 times more likely to survive to hospital discharge (95% CI = 3.36 to 8.90). This means that there is a 95% chance that when an ETI is not attempted, patients are between 3.36 to 8.90 times more likely to survive to hospital discharge than those who have received such an attempt. Despite this relatively wide confidence interval, the authors rightly concluded that there is “a negative association between prehospital endotracheal intubation attempts and survival from out-of-hospital cardiac arrest” (924). As survival is an important patient-oriented outcome, and since the confidence interval is statistically significant, these results were deemed substantial and the patient sample size adequate.
Discussion. It is important to note that management of OOHCA patients is one of the most difficult tasks in the prehospital setting. This complexity partly accounts for the substantial confounding variables that are present in all three of these retrospective cohort studies. Confounding factors include patient demographics (age, ethnicity), comorbidities (diabetes, infection), initial cardiac arrest rhythm, witnessed vs. unwitnessed arrests, presence of prearrival CPR, length of time from dispatch to arrival on scene, and time between arrest and 911 call, just to name a few. Studnek et al. attempted to account for some of these variables using regression analysis; however the confounding bias remains significant. This study demonstrated little selection bias, as all nontraumatic OOCHAs in the county were included in the study (with very specific and important exceptions). Since survival to hospital discharge was the clearly defined clinical outcome, and since it was contrasted a well-defined intervention – an ETI attempt, defined as “performing an intubation procedure by inserting the laryngoscope into the mouth, past the anterior teeth” – measurement bias was also minimized. However, ETI attempts were self-reported, thus there is indeed a possibility of misclassification of the intervention by the provider; this misclassification was most likely nondifferential, biasing the results towards the null, since there was no evidence or reasoning to suggest that different study groups would have an incentive to misreport their data. Measurement may also have been affected by the use of other advanced airway devices (e.g. King airway, Combitube) in non-intubated patients. This type of bias would be differential, as the effects of these techniques were not considered, and are unknown.
Several factors lend credibility to the study. The data was collected after 2005, which included the most recent American Heart Association (AHA) guidelines related to CPR and advanced cardiac life support (ACLS). Studnek et al. had the advantage of working in a county staffed with at least one paramedic on each ambulance, and with all firefighters having basic life support (BLS) and CPR certification (as did the other two studies). Also, if the discharge status or the medical record was unavailable, the patient was conservatively classified as not surviving to hospital discharge. This would bias the results towards the null, lending further credibility to the study’s conclusion. Generalizability, however, was scored lower due to the fact that most transports in Mecklenburg County occurred in the City of Charlotte. This should lead to shorter run times to scenes, and faster at-hospital arrivals. Thus it is difficult to assess the effectiveness of ETI in the rural setting.
Conclusion. Overall this study was fairly well-controlled as there was moderate accountability for confounding factors. The confidence interval was significant and the results were extremely patient-oriented and generalizable to the urban population. Score = 3 out of 4 (see Table 1).
Arslan Hanif et al.
Summary. The authors in this study sought to more clearly define the role of ETI in nontraumatic OOHCA adult patients; no one had compared ETI specifically against the use of the bag-valve mask (BVM), the most widely used airway management in basic life support. Arslan Hanif et al. retrospectively analyzed patient records from November 1, 1994 to June 30, 2008 from a 533-bed general hospital in southwestern Los Angeles County, searching for OOHCAs undergoing either airway management. Adult patients were excluded if they were involved in drowning, trauma, or drug overdose. DOA patients were not included since they were not transported to the hospital. A total of 1,294 OOHCAs met these inclusion criteria, and 55 (4.3%) of these patients survived to hospital discharge. A total of 1,027 (79.4%) patients were intubated, and 131 (10.1%) received BVM ventilation only. Five patients had incomplete medical records; three of these survived to discharge.
Results. It was determined that when comparing patients that were intubated to those only receiving BVM ventilations, the latter patients were 4.5 times more likely to survive to hospital discharge than those with ETI (95% CI = 2.3 to 8.9 – see Studnek et. al. results for explanation of terms), after adjusting for confounding factors. This analysis resulted in a p value less than 0.0001. This extremely small p value confirms the significance of the results, corresponding to a 0.01% chance that the null hypothesis (no difference between ETI and BVM in terms of survival) is true. Thus the authors concluded that “when compared to bag-valve-mask ventilation, endotracheal intubation was associated with decreased survival to hospital discharge among adult nontraumatic out-of-hospital cardiac arrest patients” (930). Since the results were statistically significant and survival is clearly a patient-oriented outcome, the sample size was deemed substantial and the findings significant.
Discussion. Arslan et al. also attempted to control for confounding variables. The authors adjusted for age, sex, race, notable past medical history (CHF, renal failure, MI, diabetes, etc.), and site of arrest (nursing facility, home), and rhythm upon arrival. Because past medical history specifically was controlled for, this study should score slightly better on adjusting for confounding bias, but the risk for bias remains significant due to collection of data since 1994. The AHA guidelines have changed substantially over the years, with more focus on compressions during CPR, and EMS practices have also advanced since the mid-1990s. Most of the OOHCAs in this study were pre-2005. Both of these factors would increase the amount of confounding bias present. This study had different exclusion criteria, although the two were comparable. All OOHCA patients were included (limiting selection bias), and since the non-intubated patients had a clearly defined intervention (BVM), the authors did a better job at controlling for measurement bias, specifically the differential type.
Other important considerations include the identification of patients with incomplete medical records. Statistical analysis demonstrated that inclusion of these patients would be insignificant to the findings, lending more credibility to the study. Generalizability, however, must be scored lower, as the study began in 1994 and was limited to an urban environment.
Conclusion. This study had moderate accountability for confounding variables, but did much better at controlling for measurement bias. The confidence interval and the p value were very significant, and the study was also patient-oriented with fair generalizability. Score = 4 out of 4.
Egly et al.
Summary. The authors noted the growing literature documenting adverse neurological outcomes and survival rates in trauma OOHCAs, and so decided to retrospectively study the outcomes of nontraumatic OOCHCAs in adult patients. Data was collected from a large tertiary care center in suburban Royal Oak, Michigan from January 1, 1995 to December 31, 2006. The authors noted that these cases would not have been significantly affected by the 2005 AHA guidelines. Patients were excluded if there was ROSC upon immediate defibrillation on scene, unknown age or intubation status, a traumatic origin of arrest, or if the patient was DOA. Egly et al. took the analysis one step further, dichotomizing patients with “shockable” rhythms (ventricular fibrillation and ventricular tachycardia, VF and VT respectively), and “non-shockable” rhythms (asystole and pulseless electrical activity), with survival to discharge still as the primary outcome. Ultimately, 1,515 total adult OOHCAs were examined, with 1,220 (86.2%) being intubated, and 93 (7.0%) surviving to hospital discharge.
Results. Of the 1,220 patients who were intubated, 75 (6.5%) survived to hospital discharge; of the 194 patients not intubated, 18 (10.0%) reached the same outcome. The authors concluded that “survival to hospital discharge was no different between the intubated and nonintubated groups (6.5% vs. 10.0%, p = 0.09)” (45). This corresponds to a 9% chance of rejecting the null hypothesis (no difference in survival between the two groups) when it is actually true, and thus these findings were deemed to be not significant (p values < 0.05 are often used an indicator of statistical significance). However, in patients with a VF/VT rhythm, patients who were intubated were less likely to survive to discharge (OR = 0.52, 95% CI = 0.27 to 0.998). Interestingly, patients who presented with a non-VF/VT rhythm were 2.68 times more likely to survive to hospital admission (OR = 2.68, 95% CI = 1.04 to 8.43), but they were no more likely to be discharged alive than their counterparts. Concluding that ETI correlated with a decreased survival rate to discharge for VF/VT patients, with a statistically significant confidence interval, these findings were deemed substantial and the patient size adequate for this conclusion.
Discussion. As in the previous two studies, the authors made a strong attempt to control for confounding variables. They specifically controlled for demographics, arrest location, bystander CPR, and arrest rhythm. Adjusted odds ratios were used, but the confounding bias that is inherent in retrospective cohort analyses remains significant. Including only patients that received ACLS with pre-2005 AHA guidelines worsens this bias. Again, similar exclusion criteria were used for choosing the patients. Egly et al. removed all selection bias by including all adult OOHCAs within the time frame in the study, and as with the previous two studies, measurement bias was minimized. Like Studnek et al., however, measurement bias may have been affected by the use of other advanced air devices in non-intubated patients (differential bias). The sample size was larger in this study, and the suburban environment may have increased the number of distant transports, lending credibility to the findings. Nonetheless, this observation is simply speculation. If true, it would increase the generalizability to the rural population, though overall generalizability is hindered by the use of pre-2005 AHA guidelines.
Conclusion. This study was moderately accountable for confounding variables, and other biases were adequately well-controlled. The p value demonstrated minimal correlation between ETI and decreased survival, except for in patients presenting with a VF/VT rhythm. Score = 3 out of 4.
Discussion
The results of this systematic review lead to the conclusion that ETI in adult nontraumatic OOHCA patients is negatively correlated with survival to hospital discharge. Some evidence suggests that this may be particularly valid for patients with a prehospital ventricular fibrillation or ventricular tachycardia rhythm. The three retrospective cohort studies were all conducted in the appropriate setting, with adequate sample size. Selection bias in all three was essentially zero, as each study used all the identifiable patients within similar exclusion criteria, and no research team was able to control for the confounding bias present in the study design (although valiant attempts were made). The major difference among the studies was the ability of Arslan Hanif et al. to control for added measurement bias by clearly dichotomizing the interventions: ETI vs. BVM. Use of a King Airway, Combitube, or other advanced airway device would have unknown effects on the data. For this reason, when summing the internal validity, generalizability, and relevance, Arslan Hanif et al. scored a 4, while Studnek et al. and Egly et al. scored a 3. All three studies, however, were well-controlled, had reasonable generalizability, and measured a very important patient outcome: survival.
Limitations
It is important to note that with only three studies in the literature, further research on the subject is needed. Randomized-controlled trials would be the better study design, and would be able to account for the many confounding biases that were present in these studies. This would give a better assessment of risk and benefit of ETI in this patient population. Also, all three authors admitted that they could not account for the lack of ETI in some patients. Why did some patients received ETI and others not? This is a curious finding and should be explored.
Conclusion
The results of this qualitative review lead to the conclusion that ETI in adult nontraumatic OOHCA patients is negatively correlated with survival to hospital discharge. Some evidence from the examined studies suggests that this may be particularly valid for patients with a prehospital ventricular fibrillation or ventricular tachycardia rhythm.
Studnek JR, Thestrup L, Vandeventer S, Ward SR, Staley K, Garvey L, Blackwell T. The association between prehospital endotracheal intubation attempts and survival to hospital discharge among out-of-hospital cardiac arrest patients. Acad Emerg Med. 2010;17(9):918-25.
Arslan Hanif M, Kaji AH, Niemann JT. Advanced airway management does not improve outcome of out-of-hospital cardiac arrest. Acad Emerg Med. 2010;17(9):926-31.
Egly J, Custodio D, Bishop N, Prescott M, Lucia V, Jackson RE, Swor RA. Assessing the Impact of Prehospital Intubation on Survival in Out-of-Hospital Cardiac Arrest. Prehosp Emerg Care. 2011;15(1):44-9.
(other sources and tables available upon request)
No comments:
Post a Comment