- Email: [email protected]
Outcomes of patients resuscitated from cardiac arrest in the setting of drug overdose
Resuscitation 94 (2015) 23–27
Contents lists available at ScienceDirect
Resuscitation journal homepage: www.elsevier.com/locate/resuscitation
Clinical Paper
Outcomes of patients resuscitated from cardiac arrest in the setting of drug overdose Alexander Z. Katz ∗ , Anne V. Grossestreuer, David F. Gaieski, Benjamin S. Abella, Vinayak Kumar, Jeanmarie Perrone Department of Emergency Medicine, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA, United States
a r t i c l e
i n f o
Article history: Received 4 February 2015 Received in revised form 8 June 2015 Accepted 12 June 2015 Keywords: Cardiac arrest Drug overdose Overdose Targeted temperature management Prescription drugs
a b s t r a c t Objectives: To compare the attributes and clinical outcomes of patients with cardiac arrest in the setting of drug overdose (OD) to patients with cardiac arrest from non-drug related etiologies. Methods: We utilized a US inpatient cardiac arrest registry used to study targeted temperature management (TTM) to identify patients with cardiac arrest in the setting of drug overdose between 2005 and 2013. Data regarding the cardiac arrest, resuscitation interventions, use of post-arrest TTM, urine drug screen, survival, and neurologic outcome were examined. These results were compared to patients suffering cardiac arrest from other causes during the same time period using Wilcoxon rank-sum tests for continuous variables and chi-square tests on categorical variables. Results: Approximately 2.5% (64/2584) of cardiac arrests occurred in the setting of drug overdose. Patients in the OD cohort were younger, more likely to be male, and more likely to have an out-of-hospital cardiac arrest that was unwitnessed with no bystander CPR and from a non-shockable rhythm. However, the patients in the OD cohort had similar rates of survival and good neurologic outcomes (Cerebral Performance Category 1–2) compared to non-OD patients. A fraction of initially resuscitated patients in each group (8% in OD cohort vs. 15% in non-OD cohort, p = ns) did not receive post-arrest TTM due to prompt awakening following resuscitation. Conclusions: Patients resuscitated from cardiac arrest in the setting of drug OD have neurologic and survival outcomes comparable to non-OD patients despite lower rates of bystander CPR, shockable rhythms, and witnessed arrest. © 2015 Elsevier Ireland Ltd. All rights reserved.
1. Introduction There are approximately 424,000 EMS-assessed out of hospital cardiac arrests each year in the United States, with approximately 60% of these treated by EMS personnel.1 During cardiopulmonary resuscitation (CPR) one of the core tasks is to identify and treat reversible causes of arrest including drug overdose. Adverse cardiovascular events have been reported to occur in up to 9.3% of all drug overdoses and in up to 16.9% of hospital admissions for drug overdose.2 Recently, prescription drug overdose (OD) has become the leading cause of accidental death in the United States.3 In 2011, approximately 110 Americans died from drug OD each day, and prescription opioid analgesics accounted for over 16,900 (42%) of these deaths, representing more than a three-fold increase since
∗ Corresponding author at: University of Pennsylvania, Department of Emergency Medicine, Ground Ravdin, 3400 Spruce, Philadelphia, PA 19104, United States. E-mail address: [email protected] (A.Z. Katz). http://dx.doi.org/10.1016/j.resuscitation.2015.06.015 0300-9572/© 2015 Elsevier Ireland Ltd. All rights reserved.
1999.4 Heroin was responsible for an additional 4300 deaths in 2011 as well.5 In addition to the morbidity and mortality associated with prescription drug OD, they also constitute an enormous public health and financial burden on the healthcare system. Almost 500,000 emergency department visits in 2009 were due to the misuse or abuse of prescription opioid analgesics.4 Other illicit drugs including cocaine, heroin, amphetamines, and marijuana also continue to be commonly abused, and in 2012, 9.2% of all Americans over the age of 11 used an illicit drug in the prior month compared to 8.3% in 2002.6 This is a concerning trend that may further increase the incidence of lethal drug overdose. Although fatal drug OD have increased in the past decade, little is known about the outcomes of patients resuscitated from cardiac arrest in the setting of drug OD and how these outcomes compare to the outcomes of patients resuscitated from cardiac arrest secondary to other etiologies. There is a continuum of outcomes after cardiac arrest, beginning with return of spontaneous circulation (ROSC), followed by admission to the hospital, and subsequently including survival to hospital discharge and neurologic outcome at discharge.
24
A.Z. Katz et al. / Resuscitation 94 (2015) 23–27
There is little data about what percentage of cardiac arrests caused by drug overdoses have ROSC, are admitted to the hospital, survive to hospital discharge, or have a favorable neurological outcome at the time of discharge. Post-arrest targeted temperature management (TTM) is implemented in patients who remain comatose after ROSC to treat post-arrest anoxic encephalopathy and to optimize neurological outcomes. TTM has been demonstrated to improve both survival and neurological outcomes after cardiac arrest.7–9 While TTM has been in increasingly widespread clinical use over the past decade, the three major randomized studies focused on patients with arrests secondary to primary cardiac etiologies.7,8,10 This study examined characteristics of patients with cardiac arrest in the setting of drug overdose with the objective of comparing the attributes and outcomes of patients with cardiac arrest in the setting of drug OD to patients with cardiac arrest from non-OD causes.
Table 1 Cumulative data comparing demographic characteristics, characteristics about the arrests, and outcomes between cardiac arrests attributed to drug OD and cardiac arrests attributed to all other causes.
Age (median, IQR) Male Location: out of hospital Witnessed Shockable Bystander CPR performed Duration (min) ROSC TTM (of those with ROSC) Excluded from TTM due to prompt awakening Survival Cerebral Performance Category CPC 1–2 at discharge
OD (n = 64)
Non-OD (n = 2520)
p
40 (28, 52) 46 (72) 58 (91) 20 (32) 5 (8) 9 (16) 28.5 (18, 49) 25 (39) 16 (64) 2 (8)
66 (54, 78) 1444 (57) 1737 (69) 1727 (72) 592 (25) 629 (34) 16 (8, 30) 1173 (47) 576 (49) 177 (15)
<0.001 0.02 <0.001 <0.001 0.003 0.004 0.0017 ns ns ns
10 (16) 8 (13)
469 (19) 409 (16)
ns ns
2. Methods 3. Results This study was approved by the Institutional Review Board of the University of Pennsylvania. This is a retrospective study using a US cardiac arrest registry, the Penn Alliance for Therapeutic Hypothermia (PATH), to identify patients with cardiac arrest between 2005 and 2013. This database includes twenty-eight member hospitals around the country, all of which contributed to the database, and was established to enable both quality assurance and research efforts surrounding post-arrest care. All patients presenting to participating hospitals with both out-of-hospital and in-hospital cardiac arrests as defined by loss of pulse requiring chest compressions are eligible for entry into the PATH database. A patient does not have to receive TTM to be entered into the database. Case entry is performed by a healthcare provider trained by the PATH database manager, and data integrity is ensured by a formal auditing process with feedback provided for correction. For our analysis, the PATH database was queried to identify all patients with cardiac arrest in the setting of drug OD. Attribution of drug OD as a suspected etiology for the cardiac arrest is designated by the provider involved in the patient’s care and extracted by the provider entering patients into the database. Data entry personnel at participant hospitals when entering “no” for the question of whether an arrest was cardiac in etiology are prompted to select an alternative etiology of arrest, of which “drug overdose” is an option. In some cases, Emergency Medical Services (EMS) reports and/or urine drug screen data were available to help confirm an OD, but in other cases the determination of arrest etiology was based solely on the information and history available to the clinicians involved as extracted by data entry personnel. For the OD patients who did not achieve ROSC, urine drug screen (UDS) data were not available. The cohort of patients who had a suspected drug overdose as the etiology of arrest was compared to that of patients whose cardiac arrest was deemed secondary to another cause. Demographic data including patient’s age and gender, as well as pre-, intra-, and post-arrest data for OD patients were analyzed and compared to the non-OD cardiac arrest patients during the same period. Outcome measures were survival and neurologic outcome after discharge which was categorized by the Cerebral Performance Category (CPC) score. CPC outcomes involve a scale that ranges from 1 to 5, representing worsening neurological deficits; in previous studies a favorable neurologic outcome has been defined as a CPC score of 1 or 2, and a poor neurologic outcome as a CPC score of 3, 4, or 5. Data analysis was performed using Wilcoxon rank-sum tests for non-parametric continuous variables and chi-square tests for categorical variables. Data were analyzed using Stata 13.0 (StataCorp, College Station, TX). Significant p values were defined as p < 0.05.
A small percentage (64/2584 = 2.5%) of the cardiac arrests from the PATH database were in the setting of overdose. When comparing the OD cohort and the non-OD cohort, the two groups differed in median age, sex, location of arrest, percentage with witnessed arrests, initial shockable rhythms, rate of bystander CPR, and duration of arrest (Table 1). Despite the fact that the OD cohort had lower rates of in-hospital arrests, witnessed arrests, shockable rhythms, and bystander CPR, these patients had similar rates of survival and good neurologic outcomes when compared to the non-OD cohort. These similar outcomes included rate of ROSC (39% vs. 47%, p = ns), rate of survival (16% vs. 19%, p = ns), and rate of recovery with CPC 1–2 (13% vs. 16%, p = ns). A fraction of patients in each group (OD: 2/25, 8%; non-OD: 177/1173, 15%; p = ns) did not receive TTM due to prompt awakening post-arrest. Due to the small number of patients that received TTM in the OD cohort (n = 16), we were underpowered to further analyze the effects of TTM in this group (Table 2). In order to further explore the nature of the patients who were classified as deaths in the setting of an OD, urine drug screen data were analyzed. Of the patients who achieved ROSC, 16/25 had a documented UDS. 15/16 (94%) were positive, with 11 (69%) containing opioids and 8 (50%) containing cocaine, with overlap between the two groups. Overall, 14/16 (88%) urine drug screens contained at least one opioid, benzodiazepine, or cocaine (Fig. 1). The other two OD cardiac arrests were in the setting of other ingestions, not clearly picked up on the UDS. 4. Discussion In this investigation we found that patients with cardiac arrest in the setting of drug OD had similar outcomes when compared to patients with cardiac arrest from non-drug OD etiologies. When comparing the OD cardiac arrest cohort to the non-OD cardiac Table 2 This table demonstrates the breakdown of survivors vs. non-survivors of (a) ODinduced and (b) non-OD-induced cardiac arrest patients who received TTM. Outcome
OD patients
Non-OD patients
n
n
Percent of total (%)
Outcomes of patients receiving TTM 11 69 Non-survivors 5 31 Survivors 4 25 CPC 1–2 Total
16
100
Percent of total (%)
341 235 200
59 41 35
576
100
A.Z. Katz et al. / Resuscitation 94 (2015) 23–27
25
Fig. 1. This figure depicts the breakdown of OD-induced cardiac patients who achieved ROSC, and the content of their urine drug screens. Opioids, benzodiazepines, and cocaine were frequently found in this patient population.
arrest cohort, the OD patients had fewer in-hospital arrests, fewer witnessed arrests, were less likely to have an initial shockable rhythm, and were less likely to receive bystander CPR. Generalizations about these findings are limited by the small sample size of overdose arrests. A prior investigation utilizing the Cardiac Arrest Registry to Enhance Survival (CARES) demonstrated that out-of-hospital arrests, unwitnessed arrests, non-shockable rhythms, and lack of bystander CPR are associated with poorer outcomes in cardiac arrest patients.11 For CARES patients, survival after an unwitnessed arrest was 3.9% vs. 15.9% for a witnessed arrest; survival after presentation with a non-shockable rhythm was 4.2% vs. 27.1% if the initial rhythm was shockable; and if no bystander CPR was performed survival was 8.7% vs. 11.3% if bystander CPR was performed.11 While the population in our analysis is not identical to that of CARES, it is reasonable to assume these same prognostic factors hold true. Using this assumption, in our investigation the presence of higher rates of poor prognostic factors in the OD cardiac arrest cohort should have portended worse outcomes. However, this is not what we found. There are several possible explanations for the comparable survival rates and recovery with favorable neurologic outcomes in OD cardiac arrest cohort when compared to the non-OD cohort. First, the similarity in outcomes between our two cohorts and differences when compared to CARES patients could be a function of the age of the patient. The average age of the OD cardiac arrest patients in our investigation was significantly lower than that of non-OD cardiac arrest patients. Younger patients would be expected to have a better outcome after cardiac arrest. Furthermore, our findings may be a function of different pathological processes leading to the cardiac arrest in the OD cohort (e.g. hyperthermia/vasospasm in cocaine overdose and sudden respiratory arrest in opioid overdose) vs. the non-OD cohort (myocardial infarction and arrhythmia among many other causes). It may be that the cause of arrest from drug intoxication is often respiratory in nature and more readily reversible, making OD cardiac arrest inherently more treatable once CPR has begun. For example, in an opioid overdose when a patient suffers cardiac arrest from a hypoxic or hypercarbic insult in the setting of apnea or hypoventilation, naloxone or a definitive airway may promptly reverse the pathology causing the arrest, resulting in an increased likelihood of survival and a good neurologic outcome.
However, in a patient who has an arrest from a primary cardiovascular insult such as myocardial infarction or arrhythmia, the underlying process is not as easily treated and may cause ongoing problems including refractory cardiogenic shock or malignant arrhythmias. The underlying mechanisms leading to cell death and injury may also be very different in the setting of asphyxia vs. dysrhythmic arrest. While the UDS data are limited in our overdose cohort, it is significant to note that opioids were found in highest frequency followed by cocaine. A high proportion (14/16) of urine drug screens contained opioids, benzodiazepines, or cocaine, which is consistent with the pharmacotoxicology of these substances (see Table 3). The other two OD cardiac arrests with urine drug screens lacking these drugs were both attributed to alcohol, which may have acted as a primary CNS depressant in overdose. The mechanism of toxicity of these drugs is dose dependent and variable, so further analysis about outcomes in patients stratified by drug of abuse would be useful to describe. However, the frequency of co-ingestion in our patients limits this analysis. Urine drug screens were also used to further explore the patients who were deemed to have a cardiac arrest in the setting of drug OD. However, it is important to note that drug screens give a snapshot of a patient’s drug consumption within the last several days rather than immediately prior to his/her death, and thus serve as a surrogate marker to confirm recent exposure, but does not definitively confirm a specific drug ingestion as the cause of death. In addition, many drugs and toxins may not be detected on standard drug screens including oxycodone, a very commonly prescribed and abused opioid, and fentanyl. The rapid toxicology screens used at some of the hospitals contributing patients to PATH may be missing these opioids. We considered other methods including autopsy reports to confirm classification of each cardiac arrest as an OD or non-OD. However, this information was not available in the PATH database, so further analyses could not be performed. An additional topic that is relevant to cardiac arrest outcomes in OD patients relates to the utility and benefit of TTM. 16/25 (64%) of the OD patients who had ROSC received TTM as compared to 576/1173 (49%) of the non-OD patients. Because of the small number of OD patients who received TTM, there was insufficient power to make meaningful conclusions about the benefits of TTM as it specifically relates to this subset of patients. The outcomes of
26
A.Z. Katz et al. / Resuscitation 94 (2015) 23–27
Table 3 This table demonstrates the results of the urine drug screens for all the patients who had them documented in our analysis. Patient #
Cocaine
Opiates
Benzodiazepines
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
+ + +
+
+
+ + +
+
Total % of UDSs
8 50
+
THC
Amphetamines
Other
+
+
+ +
+ +
+ + +
+ + +
+
+ +
+ +
+ +
+ + +
+ +
+
11 69
7 44
5 31
+
our analysis suggest, however, that TTM is feasible in post-arrest patients where the arrest was caused by drug OD and that clinicians should consider TTM in this cohort of post-arrest patients. A recently published study reviewing city EMS data in Pittsburgh reached similar overall conclusions as our study: patients with cardiac arrests secondary to drug OD were younger, had more un-witnessed arrests, and had similar outcomes when compared to the cohort of patients with cardiac arrests from non-OD etiologies (19% survival to hospital discharge for OD arrests vs. 12% for non-OD arrests). However, they found a three-fold higher incidence of cardiac arrest secondary to drug OD than we report, estimating that 7.7% of their cardiac arrests were secondary to drug OD (vs. 2.5% in our analysis).12 This finding may be a result of different inclusion criteria. To classify a patient into the OD cohort, our analysis required a physician caring for the patient in the ED or the hospital to determine that the arrest was in the setting of an OD, while the Pittsburgh analysis classified patients into the OD cohort if naloxone was given in the field. Due to its relative risk benefit profile, naloxone is often given even without definitive confirmation of an overdose, so using naloxone as a marker for overdose can be problematic. In fact, 93% of the ODs in their cohort received naloxone; the PATH registry does not report data on pre-hospital naloxone. There may also be poor documentation of cardiac arrest in situations where primary drug overdose is suspected and naloxone administration is prioritized. This may lead to under-reporting of these patients in the PATH database.
5. Limitations Similar to all retrospective registries, the cardiac arrests entered into the PATH registry may not be entirely representative of all cardiac arrests as for example it does not include arrests that are pronounced in the field or in which resuscitation was never attempted at all, and may under-represent the number of cardiac arrests due to drug overdose. Data from the PATH registry was not independently validated but was entered by researchers at the participating hospitals using information provided by the clinicians involved in the patient’s care. Also, there is inherent variability in clinicians’ thresholds to terminate resuscitation, so this inevitably happens at different times for different patients and this variable termination may affect the OD or non-OD cohort selectively. Furthermore, as a retrospective study, we were unable to completely ascertain and control for certain biases. For example, we had no way to definitely confirm that patients in the OD cohort truly sustained a cardiac arrest from an OD etiology. We attempted to validate this
3 19
2 13
using UDS data, but there is no gold standard method to determine whether a cardiac arrest was directly caused or not caused by a drug OD. Also, it is theoretically possible that several of the positive UDS were from medications given for sedation after ROSC, although there is no direct evidence supporting this, and many of the UDS contained opiates such as methadone, codeine, and hydromorphone that are not commonly used for sedation. Ultimately it was a clinician’s decision as to whether a cardiac arrest was deemed an OD using all information available at the time of resuscitation. Confirmation was lacking in many cases. Additionally, we were unable to stratify the patients into age groups and other demographic categories and compare outcomes including ROSC, survival, and CPC scores between the two groups. The OD population median age is 26 years younger than the non-OD population which may be a contributing factor in their improved outcomes. Finally, because there were so many fewer OD cardiac arrests than non-OD arrests, our non-significant findings in terms of survival and neurologic outcome may be a result of insufficient power, although the percentages of favorable outcome are similar in both groups.
6. Conclusion Despite higher rates of OHCA and lower rates of bystander CPR, shockable rhythms, and witnessed arrest, patients with cardiac arrest in the setting of drug OD have neurologic and survival outcomes comparable to non-OD cardiac arrest patients. Therefore, while OD patients are a distinct subset of the cardiac arrest population, it is clear that resuscitation of these patients can have highly successful outcomes, even if the arrest is accompanied by traditionally poor prognostic features.
Conflict of interest statement Alex Katz – none; Anne Grossestreuer – none; David Gaieski – Honorarium: BARD, Database funding: Stryker; Benjamin Abella – Honoraria: Stryker Medical, CR Bard, Research grants: Stryker Medical, Equity ownership: Resuscor LLC; Vinayak Kumar – none; Jeanmarie Perrone – none.
Acknowledgements None.
A.Z. Katz et al. / Resuscitation 94 (2015) 23–27
References 1. Go AS, Mozaffarian D, Roger VL, et al. American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Heart disease and stroke statistics – 2014 update: a report from the American Heart Association. Circulation 2014;129:e28–92. 2. Manini AF, Nelson LS, Stimmel B, Vlahov D, Hoffman RS. Incidence of adverse cardiovascular events in adults following drug overdose. Acad Emerg Med 2012;19:843–9. 3. Ghitza UE, Straus MM, Tai B. Preventing deaths from rising opioid overdose in the US – the promise of naloxone antidote in community-based naloxone take-home programs. Subst Abuse Rehabil 2013;4:65–72. 4. CDC. Prescription painkiller overdoses in the US. CDC vital signs 2011. http:// www.cdc.gov/vitalsigns/pdf/2011-11-vitalsigns.pdf 5. National Center for Health Statistics/CDC. National Vital Statistics. Unpublished special tabulations (June 10, 2014). 6. CDC Health Publications and Information Products. Table 60; 2013. http://www. cdc.gov/nchs/data/hus/2013/060.pdf
27
7. HaCAS Group. Mild therapeutic hypothermia to improve the neurologic outcome after cardiac arrest. N Engl J Med 2002;346:549–56. 8. Bernard SA, Gray TW, Buist MD, et al. Treatment of comatose survivors of out-of-hospital cardiac arrest with induced hypothermia. N Engl J Med 2002;346:557–63. 9. Geocadin RG, Koenig MA, Jia X, Stevens RD, Peberdy MA. Management of brain injury after resuscitation from cardiac arrest. Neurol Clin 2008;26:ix. 10. Nielsen N. Targeted temperature management at 33 ◦ C versus 36 ◦ C after cardiac arrest. N Engl J Med 2013;369:2197–206. http://www.nejm.org/toc/nejm/369/ 23/ 11. McNally B, Robb R, Mehta M, et al. Out-of-hospital cardiac arrest surveillance – Cardiac Arrest Registry to Enhance Survival (CARES), United States, October 1, 2005–December 31, 2010. MMWR Surveill Summ 2011;60: 1–19. 12. Koller AC, Salcido DD, Callaway CW, Menegazzi JJ. Resuscitation characteristics and outcomes in suspected drug overdose-related out-of-hospital cardiac arrest. Resuscitation 2014;85:1375–9.
Contents lists available at ScienceDirect
Resuscitation journal homepage: www.elsevier.com/locate/resuscitation
Clinical Paper
Outcomes of patients resuscitated from cardiac arrest in the setting of drug overdose Alexander Z. Katz ∗ , Anne V. Grossestreuer, David F. Gaieski, Benjamin S. Abella, Vinayak Kumar, Jeanmarie Perrone Department of Emergency Medicine, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA, United States
a r t i c l e
i n f o
Article history: Received 4 February 2015 Received in revised form 8 June 2015 Accepted 12 June 2015 Keywords: Cardiac arrest Drug overdose Overdose Targeted temperature management Prescription drugs
a b s t r a c t Objectives: To compare the attributes and clinical outcomes of patients with cardiac arrest in the setting of drug overdose (OD) to patients with cardiac arrest from non-drug related etiologies. Methods: We utilized a US inpatient cardiac arrest registry used to study targeted temperature management (TTM) to identify patients with cardiac arrest in the setting of drug overdose between 2005 and 2013. Data regarding the cardiac arrest, resuscitation interventions, use of post-arrest TTM, urine drug screen, survival, and neurologic outcome were examined. These results were compared to patients suffering cardiac arrest from other causes during the same time period using Wilcoxon rank-sum tests for continuous variables and chi-square tests on categorical variables. Results: Approximately 2.5% (64/2584) of cardiac arrests occurred in the setting of drug overdose. Patients in the OD cohort were younger, more likely to be male, and more likely to have an out-of-hospital cardiac arrest that was unwitnessed with no bystander CPR and from a non-shockable rhythm. However, the patients in the OD cohort had similar rates of survival and good neurologic outcomes (Cerebral Performance Category 1–2) compared to non-OD patients. A fraction of initially resuscitated patients in each group (8% in OD cohort vs. 15% in non-OD cohort, p = ns) did not receive post-arrest TTM due to prompt awakening following resuscitation. Conclusions: Patients resuscitated from cardiac arrest in the setting of drug OD have neurologic and survival outcomes comparable to non-OD patients despite lower rates of bystander CPR, shockable rhythms, and witnessed arrest. © 2015 Elsevier Ireland Ltd. All rights reserved.
1. Introduction There are approximately 424,000 EMS-assessed out of hospital cardiac arrests each year in the United States, with approximately 60% of these treated by EMS personnel.1 During cardiopulmonary resuscitation (CPR) one of the core tasks is to identify and treat reversible causes of arrest including drug overdose. Adverse cardiovascular events have been reported to occur in up to 9.3% of all drug overdoses and in up to 16.9% of hospital admissions for drug overdose.2 Recently, prescription drug overdose (OD) has become the leading cause of accidental death in the United States.3 In 2011, approximately 110 Americans died from drug OD each day, and prescription opioid analgesics accounted for over 16,900 (42%) of these deaths, representing more than a three-fold increase since
∗ Corresponding author at: University of Pennsylvania, Department of Emergency Medicine, Ground Ravdin, 3400 Spruce, Philadelphia, PA 19104, United States. E-mail address: [email protected] (A.Z. Katz). http://dx.doi.org/10.1016/j.resuscitation.2015.06.015 0300-9572/© 2015 Elsevier Ireland Ltd. All rights reserved.
1999.4 Heroin was responsible for an additional 4300 deaths in 2011 as well.5 In addition to the morbidity and mortality associated with prescription drug OD, they also constitute an enormous public health and financial burden on the healthcare system. Almost 500,000 emergency department visits in 2009 were due to the misuse or abuse of prescription opioid analgesics.4 Other illicit drugs including cocaine, heroin, amphetamines, and marijuana also continue to be commonly abused, and in 2012, 9.2% of all Americans over the age of 11 used an illicit drug in the prior month compared to 8.3% in 2002.6 This is a concerning trend that may further increase the incidence of lethal drug overdose. Although fatal drug OD have increased in the past decade, little is known about the outcomes of patients resuscitated from cardiac arrest in the setting of drug OD and how these outcomes compare to the outcomes of patients resuscitated from cardiac arrest secondary to other etiologies. There is a continuum of outcomes after cardiac arrest, beginning with return of spontaneous circulation (ROSC), followed by admission to the hospital, and subsequently including survival to hospital discharge and neurologic outcome at discharge.
24
A.Z. Katz et al. / Resuscitation 94 (2015) 23–27
There is little data about what percentage of cardiac arrests caused by drug overdoses have ROSC, are admitted to the hospital, survive to hospital discharge, or have a favorable neurological outcome at the time of discharge. Post-arrest targeted temperature management (TTM) is implemented in patients who remain comatose after ROSC to treat post-arrest anoxic encephalopathy and to optimize neurological outcomes. TTM has been demonstrated to improve both survival and neurological outcomes after cardiac arrest.7–9 While TTM has been in increasingly widespread clinical use over the past decade, the three major randomized studies focused on patients with arrests secondary to primary cardiac etiologies.7,8,10 This study examined characteristics of patients with cardiac arrest in the setting of drug overdose with the objective of comparing the attributes and outcomes of patients with cardiac arrest in the setting of drug OD to patients with cardiac arrest from non-OD causes.
Table 1 Cumulative data comparing demographic characteristics, characteristics about the arrests, and outcomes between cardiac arrests attributed to drug OD and cardiac arrests attributed to all other causes.
Age (median, IQR) Male Location: out of hospital Witnessed Shockable Bystander CPR performed Duration (min) ROSC TTM (of those with ROSC) Excluded from TTM due to prompt awakening Survival Cerebral Performance Category CPC 1–2 at discharge
OD (n = 64)
Non-OD (n = 2520)
p
40 (28, 52) 46 (72) 58 (91) 20 (32) 5 (8) 9 (16) 28.5 (18, 49) 25 (39) 16 (64) 2 (8)
66 (54, 78) 1444 (57) 1737 (69) 1727 (72) 592 (25) 629 (34) 16 (8, 30) 1173 (47) 576 (49) 177 (15)
<0.001 0.02 <0.001 <0.001 0.003 0.004 0.0017 ns ns ns
10 (16) 8 (13)
469 (19) 409 (16)
ns ns
2. Methods 3. Results This study was approved by the Institutional Review Board of the University of Pennsylvania. This is a retrospective study using a US cardiac arrest registry, the Penn Alliance for Therapeutic Hypothermia (PATH), to identify patients with cardiac arrest between 2005 and 2013. This database includes twenty-eight member hospitals around the country, all of which contributed to the database, and was established to enable both quality assurance and research efforts surrounding post-arrest care. All patients presenting to participating hospitals with both out-of-hospital and in-hospital cardiac arrests as defined by loss of pulse requiring chest compressions are eligible for entry into the PATH database. A patient does not have to receive TTM to be entered into the database. Case entry is performed by a healthcare provider trained by the PATH database manager, and data integrity is ensured by a formal auditing process with feedback provided for correction. For our analysis, the PATH database was queried to identify all patients with cardiac arrest in the setting of drug OD. Attribution of drug OD as a suspected etiology for the cardiac arrest is designated by the provider involved in the patient’s care and extracted by the provider entering patients into the database. Data entry personnel at participant hospitals when entering “no” for the question of whether an arrest was cardiac in etiology are prompted to select an alternative etiology of arrest, of which “drug overdose” is an option. In some cases, Emergency Medical Services (EMS) reports and/or urine drug screen data were available to help confirm an OD, but in other cases the determination of arrest etiology was based solely on the information and history available to the clinicians involved as extracted by data entry personnel. For the OD patients who did not achieve ROSC, urine drug screen (UDS) data were not available. The cohort of patients who had a suspected drug overdose as the etiology of arrest was compared to that of patients whose cardiac arrest was deemed secondary to another cause. Demographic data including patient’s age and gender, as well as pre-, intra-, and post-arrest data for OD patients were analyzed and compared to the non-OD cardiac arrest patients during the same period. Outcome measures were survival and neurologic outcome after discharge which was categorized by the Cerebral Performance Category (CPC) score. CPC outcomes involve a scale that ranges from 1 to 5, representing worsening neurological deficits; in previous studies a favorable neurologic outcome has been defined as a CPC score of 1 or 2, and a poor neurologic outcome as a CPC score of 3, 4, or 5. Data analysis was performed using Wilcoxon rank-sum tests for non-parametric continuous variables and chi-square tests for categorical variables. Data were analyzed using Stata 13.0 (StataCorp, College Station, TX). Significant p values were defined as p < 0.05.
A small percentage (64/2584 = 2.5%) of the cardiac arrests from the PATH database were in the setting of overdose. When comparing the OD cohort and the non-OD cohort, the two groups differed in median age, sex, location of arrest, percentage with witnessed arrests, initial shockable rhythms, rate of bystander CPR, and duration of arrest (Table 1). Despite the fact that the OD cohort had lower rates of in-hospital arrests, witnessed arrests, shockable rhythms, and bystander CPR, these patients had similar rates of survival and good neurologic outcomes when compared to the non-OD cohort. These similar outcomes included rate of ROSC (39% vs. 47%, p = ns), rate of survival (16% vs. 19%, p = ns), and rate of recovery with CPC 1–2 (13% vs. 16%, p = ns). A fraction of patients in each group (OD: 2/25, 8%; non-OD: 177/1173, 15%; p = ns) did not receive TTM due to prompt awakening post-arrest. Due to the small number of patients that received TTM in the OD cohort (n = 16), we were underpowered to further analyze the effects of TTM in this group (Table 2). In order to further explore the nature of the patients who were classified as deaths in the setting of an OD, urine drug screen data were analyzed. Of the patients who achieved ROSC, 16/25 had a documented UDS. 15/16 (94%) were positive, with 11 (69%) containing opioids and 8 (50%) containing cocaine, with overlap between the two groups. Overall, 14/16 (88%) urine drug screens contained at least one opioid, benzodiazepine, or cocaine (Fig. 1). The other two OD cardiac arrests were in the setting of other ingestions, not clearly picked up on the UDS. 4. Discussion In this investigation we found that patients with cardiac arrest in the setting of drug OD had similar outcomes when compared to patients with cardiac arrest from non-drug OD etiologies. When comparing the OD cardiac arrest cohort to the non-OD cardiac Table 2 This table demonstrates the breakdown of survivors vs. non-survivors of (a) ODinduced and (b) non-OD-induced cardiac arrest patients who received TTM. Outcome
OD patients
Non-OD patients
n
n
Percent of total (%)
Outcomes of patients receiving TTM 11 69 Non-survivors 5 31 Survivors 4 25 CPC 1–2 Total
16
100
Percent of total (%)
341 235 200
59 41 35
576
100
A.Z. Katz et al. / Resuscitation 94 (2015) 23–27
25
Fig. 1. This figure depicts the breakdown of OD-induced cardiac patients who achieved ROSC, and the content of their urine drug screens. Opioids, benzodiazepines, and cocaine were frequently found in this patient population.
arrest cohort, the OD patients had fewer in-hospital arrests, fewer witnessed arrests, were less likely to have an initial shockable rhythm, and were less likely to receive bystander CPR. Generalizations about these findings are limited by the small sample size of overdose arrests. A prior investigation utilizing the Cardiac Arrest Registry to Enhance Survival (CARES) demonstrated that out-of-hospital arrests, unwitnessed arrests, non-shockable rhythms, and lack of bystander CPR are associated with poorer outcomes in cardiac arrest patients.11 For CARES patients, survival after an unwitnessed arrest was 3.9% vs. 15.9% for a witnessed arrest; survival after presentation with a non-shockable rhythm was 4.2% vs. 27.1% if the initial rhythm was shockable; and if no bystander CPR was performed survival was 8.7% vs. 11.3% if bystander CPR was performed.11 While the population in our analysis is not identical to that of CARES, it is reasonable to assume these same prognostic factors hold true. Using this assumption, in our investigation the presence of higher rates of poor prognostic factors in the OD cardiac arrest cohort should have portended worse outcomes. However, this is not what we found. There are several possible explanations for the comparable survival rates and recovery with favorable neurologic outcomes in OD cardiac arrest cohort when compared to the non-OD cohort. First, the similarity in outcomes between our two cohorts and differences when compared to CARES patients could be a function of the age of the patient. The average age of the OD cardiac arrest patients in our investigation was significantly lower than that of non-OD cardiac arrest patients. Younger patients would be expected to have a better outcome after cardiac arrest. Furthermore, our findings may be a function of different pathological processes leading to the cardiac arrest in the OD cohort (e.g. hyperthermia/vasospasm in cocaine overdose and sudden respiratory arrest in opioid overdose) vs. the non-OD cohort (myocardial infarction and arrhythmia among many other causes). It may be that the cause of arrest from drug intoxication is often respiratory in nature and more readily reversible, making OD cardiac arrest inherently more treatable once CPR has begun. For example, in an opioid overdose when a patient suffers cardiac arrest from a hypoxic or hypercarbic insult in the setting of apnea or hypoventilation, naloxone or a definitive airway may promptly reverse the pathology causing the arrest, resulting in an increased likelihood of survival and a good neurologic outcome.
However, in a patient who has an arrest from a primary cardiovascular insult such as myocardial infarction or arrhythmia, the underlying process is not as easily treated and may cause ongoing problems including refractory cardiogenic shock or malignant arrhythmias. The underlying mechanisms leading to cell death and injury may also be very different in the setting of asphyxia vs. dysrhythmic arrest. While the UDS data are limited in our overdose cohort, it is significant to note that opioids were found in highest frequency followed by cocaine. A high proportion (14/16) of urine drug screens contained opioids, benzodiazepines, or cocaine, which is consistent with the pharmacotoxicology of these substances (see Table 3). The other two OD cardiac arrests with urine drug screens lacking these drugs were both attributed to alcohol, which may have acted as a primary CNS depressant in overdose. The mechanism of toxicity of these drugs is dose dependent and variable, so further analysis about outcomes in patients stratified by drug of abuse would be useful to describe. However, the frequency of co-ingestion in our patients limits this analysis. Urine drug screens were also used to further explore the patients who were deemed to have a cardiac arrest in the setting of drug OD. However, it is important to note that drug screens give a snapshot of a patient’s drug consumption within the last several days rather than immediately prior to his/her death, and thus serve as a surrogate marker to confirm recent exposure, but does not definitively confirm a specific drug ingestion as the cause of death. In addition, many drugs and toxins may not be detected on standard drug screens including oxycodone, a very commonly prescribed and abused opioid, and fentanyl. The rapid toxicology screens used at some of the hospitals contributing patients to PATH may be missing these opioids. We considered other methods including autopsy reports to confirm classification of each cardiac arrest as an OD or non-OD. However, this information was not available in the PATH database, so further analyses could not be performed. An additional topic that is relevant to cardiac arrest outcomes in OD patients relates to the utility and benefit of TTM. 16/25 (64%) of the OD patients who had ROSC received TTM as compared to 576/1173 (49%) of the non-OD patients. Because of the small number of OD patients who received TTM, there was insufficient power to make meaningful conclusions about the benefits of TTM as it specifically relates to this subset of patients. The outcomes of
26
A.Z. Katz et al. / Resuscitation 94 (2015) 23–27
Table 3 This table demonstrates the results of the urine drug screens for all the patients who had them documented in our analysis. Patient #
Cocaine
Opiates
Benzodiazepines
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
+ + +
+
+
+ + +
+
Total % of UDSs
8 50
+
THC
Amphetamines
Other
+
+
+ +
+ +
+ + +
+ + +
+
+ +
+ +
+ +
+ + +
+ +
+
11 69
7 44
5 31
+
our analysis suggest, however, that TTM is feasible in post-arrest patients where the arrest was caused by drug OD and that clinicians should consider TTM in this cohort of post-arrest patients. A recently published study reviewing city EMS data in Pittsburgh reached similar overall conclusions as our study: patients with cardiac arrests secondary to drug OD were younger, had more un-witnessed arrests, and had similar outcomes when compared to the cohort of patients with cardiac arrests from non-OD etiologies (19% survival to hospital discharge for OD arrests vs. 12% for non-OD arrests). However, they found a three-fold higher incidence of cardiac arrest secondary to drug OD than we report, estimating that 7.7% of their cardiac arrests were secondary to drug OD (vs. 2.5% in our analysis).12 This finding may be a result of different inclusion criteria. To classify a patient into the OD cohort, our analysis required a physician caring for the patient in the ED or the hospital to determine that the arrest was in the setting of an OD, while the Pittsburgh analysis classified patients into the OD cohort if naloxone was given in the field. Due to its relative risk benefit profile, naloxone is often given even without definitive confirmation of an overdose, so using naloxone as a marker for overdose can be problematic. In fact, 93% of the ODs in their cohort received naloxone; the PATH registry does not report data on pre-hospital naloxone. There may also be poor documentation of cardiac arrest in situations where primary drug overdose is suspected and naloxone administration is prioritized. This may lead to under-reporting of these patients in the PATH database.
5. Limitations Similar to all retrospective registries, the cardiac arrests entered into the PATH registry may not be entirely representative of all cardiac arrests as for example it does not include arrests that are pronounced in the field or in which resuscitation was never attempted at all, and may under-represent the number of cardiac arrests due to drug overdose. Data from the PATH registry was not independently validated but was entered by researchers at the participating hospitals using information provided by the clinicians involved in the patient’s care. Also, there is inherent variability in clinicians’ thresholds to terminate resuscitation, so this inevitably happens at different times for different patients and this variable termination may affect the OD or non-OD cohort selectively. Furthermore, as a retrospective study, we were unable to completely ascertain and control for certain biases. For example, we had no way to definitely confirm that patients in the OD cohort truly sustained a cardiac arrest from an OD etiology. We attempted to validate this
3 19
2 13
using UDS data, but there is no gold standard method to determine whether a cardiac arrest was directly caused or not caused by a drug OD. Also, it is theoretically possible that several of the positive UDS were from medications given for sedation after ROSC, although there is no direct evidence supporting this, and many of the UDS contained opiates such as methadone, codeine, and hydromorphone that are not commonly used for sedation. Ultimately it was a clinician’s decision as to whether a cardiac arrest was deemed an OD using all information available at the time of resuscitation. Confirmation was lacking in many cases. Additionally, we were unable to stratify the patients into age groups and other demographic categories and compare outcomes including ROSC, survival, and CPC scores between the two groups. The OD population median age is 26 years younger than the non-OD population which may be a contributing factor in their improved outcomes. Finally, because there were so many fewer OD cardiac arrests than non-OD arrests, our non-significant findings in terms of survival and neurologic outcome may be a result of insufficient power, although the percentages of favorable outcome are similar in both groups.
6. Conclusion Despite higher rates of OHCA and lower rates of bystander CPR, shockable rhythms, and witnessed arrest, patients with cardiac arrest in the setting of drug OD have neurologic and survival outcomes comparable to non-OD cardiac arrest patients. Therefore, while OD patients are a distinct subset of the cardiac arrest population, it is clear that resuscitation of these patients can have highly successful outcomes, even if the arrest is accompanied by traditionally poor prognostic features.
Conflict of interest statement Alex Katz – none; Anne Grossestreuer – none; David Gaieski – Honorarium: BARD, Database funding: Stryker; Benjamin Abella – Honoraria: Stryker Medical, CR Bard, Research grants: Stryker Medical, Equity ownership: Resuscor LLC; Vinayak Kumar – none; Jeanmarie Perrone – none.
Acknowledgements None.
A.Z. Katz et al. / Resuscitation 94 (2015) 23–27
References 1. Go AS, Mozaffarian D, Roger VL, et al. American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Heart disease and stroke statistics – 2014 update: a report from the American Heart Association. Circulation 2014;129:e28–92. 2. Manini AF, Nelson LS, Stimmel B, Vlahov D, Hoffman RS. Incidence of adverse cardiovascular events in adults following drug overdose. Acad Emerg Med 2012;19:843–9. 3. Ghitza UE, Straus MM, Tai B. Preventing deaths from rising opioid overdose in the US – the promise of naloxone antidote in community-based naloxone take-home programs. Subst Abuse Rehabil 2013;4:65–72. 4. CDC. Prescription painkiller overdoses in the US. CDC vital signs 2011. http:// www.cdc.gov/vitalsigns/pdf/2011-11-vitalsigns.pdf 5. National Center for Health Statistics/CDC. National Vital Statistics. Unpublished special tabulations (June 10, 2014). 6. CDC Health Publications and Information Products. Table 60; 2013. http://www. cdc.gov/nchs/data/hus/2013/060.pdf
27
7. HaCAS Group. Mild therapeutic hypothermia to improve the neurologic outcome after cardiac arrest. N Engl J Med 2002;346:549–56. 8. Bernard SA, Gray TW, Buist MD, et al. Treatment of comatose survivors of out-of-hospital cardiac arrest with induced hypothermia. N Engl J Med 2002;346:557–63. 9. Geocadin RG, Koenig MA, Jia X, Stevens RD, Peberdy MA. Management of brain injury after resuscitation from cardiac arrest. Neurol Clin 2008;26:ix. 10. Nielsen N. Targeted temperature management at 33 ◦ C versus 36 ◦ C after cardiac arrest. N Engl J Med 2013;369:2197–206. http://www.nejm.org/toc/nejm/369/ 23/ 11. McNally B, Robb R, Mehta M, et al. Out-of-hospital cardiac arrest surveillance – Cardiac Arrest Registry to Enhance Survival (CARES), United States, October 1, 2005–December 31, 2010. MMWR Surveill Summ 2011;60: 1–19. 12. Koller AC, Salcido DD, Callaway CW, Menegazzi JJ. Resuscitation characteristics and outcomes in suspected drug overdose-related out-of-hospital cardiac arrest. Resuscitation 2014;85:1375–9.