Evaluating Patient Outcomes and Interpretation of Statistics in Cardiac Care

Date: September 18, 2024 Reference: Dillon et al. Naloxone and Patient Outcomes in Out-of-Hospital Cardiac Arrests in California. JAMA Network Open. August 20, 2024 Guest Skeptic: Dr. Chris Root is an emergency medicine and emergency medicine service (EMS) physician at the University of New Mexico, Albuquerque. Before attending medical school, he was a New York City Paramedic. Chris completed his emergency medicine residency and EMS fellowship at UNM. He currently practices emergency medicine in New Mexico in the ED, in the field with EMS and with the UNM Lifeguard Air Emergency Services. Case: You are working as a paramedic, and you respond to a cardiac arrest. On arrival, you find a 35-year-old male, pulseless and apneic with cardio-pulmonary resuscitation (CPR) in progress by a bystander. There is drug paraphernalia scattered around the room. You and your partner initiate high-quality CPR, place a supraglottic airway, establish intra-osseous (IO) access and administer epinephrine. Your partner asks if you want to administer naloxone as well. Background: We’ve discussed out-of-hospital cardiac arrest (OHCA) at least once or twice on the SGEM (see long list at end of blog). Today’s study looks at the role of naloxone in OHCA. Naloxone is a well-established medication used primarily for reversing opioid overdoses. As a competitive opioid antagonist, naloxone binds to opioid receptors in the central nervous system, effectively displacing opioids and reversing their effects, particularly respiratory depression. This makes naloxone an essential tool for emergency responders dealing with opioid-related incidents. Typically administered via intravenous (IV), intramuscular (IM), or intranasal (IN) routes, naloxone acts rapidly, often restoring normal breathing within minutes. Its safety profile is well-tolerated, with the primary adverse effects related to the abrupt reversal of opioid effects, such as acute withdrawal symptoms. Traditionally, naloxone has been used in cases of suspected opioid overdose where patients exhibit signs of severe respiratory depression or loss of consciousness (LOC). However, its role in broader emergency care contexts, such as OHCA, is evolving. Opioid-associated OHCA has become increasingly common due to the ongoing opioid crisis, with opioids contributing to a significant proportion of cardiac arrests [1-4]. In these scenarios, the pathophysiology involves opioid-induced respiratory depression leading to hypoxia, hypotension, and eventually cardiac arrest. Given this progression, naloxone's ability to counteract opioid effects offers a potential intervention point, even in cardiac arrest scenarios. Current guidelines from organizations like the American Heart Association (AHA) suggest considering naloxone in suspected opioid-associated OHCA cases [5]. However, the efficacy of naloxone in improving outcomes in such cardiac arrests remains a topic of ongoing research and debate. While naloxone is not traditionally viewed as a standard treatment in cardiac arrest care, its potential to address underlying opioid toxicity provides a rationale for its use in selected patients. This has led to variability in EMS protocols, with some agencies including naloxone in their cardiac arrest protocols while others do not specifically recommend it, highlighting a gap in definitive guidance [6]. As the landscape of OHCA continues to evolve, understanding the role of naloxone in these critical situations is vital for EMS providers. This discussion sets the stage for exploring naloxone's place in the management of cardiac arrest, particularly as new evidence emerges regarding its impact on outcomes such as return of spontaneous circulation (ROSC) and survival to hospital discharge. Clinical Question: Is naloxone administration in undifferentiated OHCA associated with survival to hospital discharge? Reference: Dillon et al. Naloxone and Patient Outcomes in Out-of-Hospital Cardiac Arrests in California. JAMA Network Open. August 20, 2024 Population: Adult patients (aged 18 or older) who received EMS treatment for nontraumatic OHCA in three Northern California counties between 2015 and 2023. Excluded: Patients under 18 and missing data regarding medication administration Exposure: Naloxone administration during resuscitation. Comparison: No naloxone administration during resuscitation. Outcome: Primary Outcome: Survival to hospital discharge Secondary Outcomes: Sustained ROSC (detectable pulse for at least 20 minutes or at the end of EMS care) Type of Study: Retrospective cohort study Authors’ Conclusions: “In this retrospective cohort of patients with OHCA, EMS-administered naloxone was associated with clinically significant improvements in ROSC and survival to hospital discharge.” Quality Checklist for Observational Study: Did the study address a clearly focused issue? Yes Did the authors use an appropriate method to answer their question? Yes Was the cohort recruited in an acceptable way? Yes Was the exposure accurately measured to minimize bias? Yes Was the outcome accurately measured to minimize bias? Yes Have the authors identified all-important confounding factors? Unsure Was the follow-up of subjects complete enough? Yes How precise are the results? Precise Do you believe the results? No Can the results be applied to the local population? Unsure Do the results of this study fit with other available evidence? No Funding of the Study: One author reported a grant from the National Heart Lung and Blood Institute. One author reported grants from the Substance Abuse and Mental Health Services Administration outside of this work. Results: There were 8,195 people identified in this study with OHCAs. Of these patients, 1,165 received naloxone (14%) while 7,030 did not receive naloxone (86%). The median age was 65 years and 68% were male. Nine percent were drug related while 91% were not drug related. Key Result: Naloxone administration was associated with increased ROSC and increased survival to hospital discharge compared to those who did not receive naloxone. Primary Outcome: Survival to hospital discharge was 15.9% in the naloxone group vs 9.7% in the non-naloxone group. This gives an absolute risk difference (ARD) of 6.2% (95% CI; 2.3%-10.0%) P<0.001 Secondary Outcomes: Sustained ROSC was 34.5% in the naloxone group vs 22.9% in the non-naloxone group. This gives an ARD of 15.2% (95% CI; 9.9%-20.6%) P<0.001 Interestingly, they found higher ROSC with naloxone in both “drug-related” and “non-drug related” arrests. 1. Confounders and Neighbors: The exposed and unexposed groups were very different. Every p-value in Table 1 is significant, which is not what you typically see. The exposed group was more likely to be younger, to have an unwitnessed arrest, less likely to have ventricular fibrillation, etc. The authors generated propensity score models based on age, sex, initial cardiac rhythm, comorbid conditions, whether the OHCA was witnessed, and whether the cause of arrest was drug-related. Rosenbaum and Rubin defined propensity score matching in 1983 as “the probability of treatment assignment conditional on observed baseline covariates” [7].  It represents an attempt to balance two groups conditionally on the distribution of measured baseline covariates. Propensity score matching is a statistical attempt to decrease bias in an observational study. For those who want to take a deeper dive into this topic here are a few references [8-10]. They performed two types of regression analysis, first, an inverse probability weighted regression adjustment. For this analysis, they calculate the likelihood that an individual receives a given treatment (naloxone) and then, based on the characteristics above, a propensity score, and then they assign weights based on the inverse of the propensity score creating a “pseudo population” wherein the confounders are equally distributed amongst the treatment groups. Then they did nearest neighbor propensity score matching. They used the same propensity scores generated for the inverse probability weighted analysis and then paired every patient who received naloxone to a patient whose propensity score was most similar. These are robust statistical methods but they’re no substitute for a prospective randomized trial. 2. Unmeasured Confounders: Chart extraction can only capture so much. Whether or not an arrest was “drug-related” was based on the documentation in the EMS chart, yet 60% of patients who received naloxone had “non-drug related” 13% of patients in the no naloxone group also didn’t receive any epinephrine. There are always factors that determine why a clinician makes a certain decision that we’ll never be able to measure, which is why prospective RCTs are so important for treatment-oriented questions. 3. Patient-Oriented Outcome (POO): Discharge from the hospital was the primary outcome, which is an important patient-oriented outcome or POO, however, there is no data on patient status at the time of discharge. Especially with a disease process as potentially debilitating as cardiac arrest, it’s important to know not just that the patient left the hospital, but how they were when they left. Awake and talking or bed-bound with a trach and a PEG tube? An example of this would be the PARAMEDIC 2 trial that we covered on SGEM#238. 4. Calculating an NNT from Observational Data: This is something that epidemiologists and biostatisticians have been talking about for decades [11]. To remind everyone, the NNT estimates the average number of patients who need to be treated to positively impact one person with therapeutic benefit. As with any summary statistics, the NNT can imply a sense of certainty that is not justified. A major strength of the NNT is its simplicity, making complex research easier to understand.