Prehospital Administration of Oxygen for Chest Pain Patients Decreases Signficantly Following Implementation of the 2010 AHA Guidelines

Elliot Carhart, EdD, RRT, NRP & Joshua G. Salzman, MA, EMT-B


In 2010, the American Heart Association published an updated version of the guidelines on Resuscitation and Emergency Cardiovascular Care (ECC). These guidelines claimed insufficient evidence to support the routine administration of oxygen to patients with uncomplicated presentations of acute coronary syndromes (ACS).1 This was based in part on evidence showing that supplemental oxygen for such patients can increase coronary vascular resistance, decrease coronary blood flow,2 increase infarct size, and ultimately lead to increased mortality.3 As such, it was recommended that oxygen not be administered to patients experiencing uncomplicated presentations of ACS whose SpO2 is 94% or higher.1

The 2010 AHA guidelines held significant implications for emergency medical services (EMS) providers, who have historically taken part in the routine administration of supplemental oxygen despite a lack of evidence to support such interventions.4 While EMS is progressing towards evidence-based practice, there is no current literature documenting the adoption of this new guideline by the EMS community. The purpose of this study was to examine trends in oxygen administration for the time period since publication of the 2010 AHA guidelines to determine if EMS providers are making appropriate modifications to their practice. This will help determine if further educational interventions are necessary to promote adherence to the published standard of care for patients experiencing symptoms of ACS.


After receiving institutional review board (IRB) approval, we performed a retrospective analysis of data obtained from FisdapTM, a national, clinical skills tracking system for paramedic students. More than 900 EMS education programs in all 50 U.S. states use FisdapTM, including the majority of accredited paramedic programs. Detailed descriptions of the Fisdap data collection process have been previously described.5,6 In general, after registration and consent for use of their data for research purposes, paramedic students record clinical skills observed and performed for all patient encounters during the field internship portion of their academic program. The student’s paramedic preceptor for those field experiences verifies the student’s documentation prior to entry into Fisdap. The entered data is then then audited by the student’s instructor who subsequently flags the data as worthy of use for research purposes. Since data are collected in real time during the internship phase for each student, these records mirror the patient care data recorded on the actual patient care report by the supervising paramedic providing patient care onscene. This “carbon copy” patient care record allows for a unique window into paramedic clinical practice nationwide. The Fisdap dataset has been compared to the patient care database of a large, private ambulance provider and was shown to be similar in composition with respect to patient age and primary impression.7

Data were extracted from the database between June 2010 and December 2012. Prior to June 2010, paramedic students were unable to record vital sign data within FisdapTM. With this change, baseline data were available for the 6 months prior to publication date of the 2010 guideline. Records were included for this study if they met the following inclusion criteria: 1) student consent for research, 2) cardiac chest pain recorded as the chief complaint, and 3) SpO2 data available for review. Exclusion criteria included: 1) SpO2 < 94%, 2) cases for which oxygen was administered via positive pressure ventilation or continuous positive airway pressure, and 3) patients who could be considered clinically hemodyamically unstable and in need of O2 therapy. Hemodynamically unstable was defined as a systolic blood pressure < 100 or > 250 mm Hg or a heart rate < 60 or > 100 beats per minute.

For each patient encounter meeting inclusion criteria, the following variables were extracted from the database: 1) unique student identifier, 2) unique program identifier, 3) month and year of patient encounter, 4) first recorded SpO2 value, systolic blood pressure, and heart rate, 5) O2 administration (yes/no), 6) O2 administration route, and 7) O2 administration rate. The percent of patient encounters with oxygen administration was calculated for each year, and unadjusted logistic regression was used to determine if O2 administration rates changed significantly over the study period. Unadjusted logistic regression was also used to determine if there was a difference in the odds of receiving oxygen based on a patient’s SpO2 value.


A total of 16,820 patient encounters met the initial inclusion criteria. Following application of the exclusion criteria, 10,552 patient encounters by 2,447 paramedic students from 195 paramedic programs representing 49 states remained and were included for analysis (Figure 1).

Prior to release of the new guidelines (2010), 71.9% (95% CI 69.8-74.0%) of patients with SpO2 ≥ 94% received supplemental O2. Rates of O2 administration were significantly lower in 2011 (64%; 95% CI 62.7-65.3%) and in 2012 (53.1%; 95% CI 51.5–54.7%). Figure 2 depicts the steady decline in oxygen administration by quarter each year contained in the study period. The most common method of O2 administration was nasal cannula (78.7%; 95% CI 77.6–79.6%), with 20.8% (95% CI 19.8–21.8%) of patients receiving oxygen via non-rebreather mask. The odds of a hemodynamically stable chest pain patient with SpO2 ≥ 94% receiving supplemental oxygen in 2011 were 1.4 times lower compared to patients in 2010 (95% CI 1.3–1.6). Similarly, the odds of patients in 2012 receiving supplemental oxygen were 2.3 times lower compared to patients in 2010 (95% CI 2.0–2.6).

Table 1 shows the percent of patients receiving O2 at each level of SpO2 value ≥ 94%. Results of the unadjusted logistic regression show the odds of receiving supplemental oxygen decreased by 4% for each 1% increase in SpO2 beyond the 94% threshold (OR = 0.96; 95% CI 0.94–0.98). This result can be attributed to the large drop in O2 administration between patients with an SpO2 of 94 or 95% (71.9% and 66.4%, respectively) compared to those with an SpO2 ≥ 96% (range = 61.8-58.1%). There is no significant difference in O2 administration rates for patient with SpO2 ≥ 96% (OR = 1.0; 95% CI 0.97–1.03).


Our study showed a progressive decline in oxygen administration in the 2 years following the publication of the 2010 American Heart Association guidelines. This suggests that many EMS medical directors and agencies are using these guidelines as a basis for treatment protocol and guideline development. Additionally, it appears providers have taken these protocol and guideline modifications and adapted their clinical practice. However, there still appears to be a large number of patients receiving oxygen when it may not be indicated, and the rate of change between 2010 and 2012 (18.8% absolute change) is noteworthy.

The potentially negative implications for patients who continue to receive oxygen therapy when their SpO2 levels are ≥ 94% include increased coronary vascular resistance, decreased coronary blood flow,2 increased infarct size, and potentially increased mortality.3 Unfortunately, these physiologic metrics cannot be monitored directly by EMS providers in the field. Patients rarely display decreased vascular resistance and coronary blood flow or increased infarct size in a clinically recognizable way during the relatively short period of time it takes for most EMS agencies to transport to the hospital. Previous research has shown EMS medical directors are more likely to change their practice in response to evidence of potential harm as opposed to the absence of evidence demonstrating benefit.8 It seems this guideline change lands squarely in the latter of the two situations. Adherence to this guideline change may be improved by providing additional focused education on the rationale behind these changes, along with emphasis of the potentially negative effects of inappropriate treatment and potential cost savings of avoiding over-treatment. A more active monitoring of this guideline change by individual EMS agencies, or on a national level, could also influence the rate of change in a positive direction.

Although it appears EMS agencies and medical directors are adopting this recommendation, it is important to recognize adoption of AHA guidelines is not absolute nationwide. We do not know how many EMS agencies included in our study made an intentional decision to not include this change in their patient care protocols or guidelines. We also know the initial rollout for instructors began in late 2010 and printed course materials were released in the 1st and 2nd quarters of 2011. Figure 2 shows an initial drop between the 3rd Quarter 2010 and the 1st Quarter 2011, then a leveling off until 4th Quarter of 2011. We suspect 4th Quarter 2011 is a reasonable time frame for the guideline change to be delivered in the standardized ACLS curriculum, as well as sufficient time for agencies to incorporate these changes into updated agency protocols.

As noted above, there is no mandate for EMS providers to follow the AHA guidelines. For those who choose to follow these guidelines, there is room for interpretation with regards to defining complicated and uncomplicated presentations of ACS. These terms are not explicitly defined, which could lead to differing beliefs regarding the appropriateness of oxygen administration for ambiguous clinical presentations. While lacking an explicit definition, these guidelines did include examples of complicated clinical presentations. These examples (e.g. hypoxemia & heart failure) suggest a distinction based on oxygen delivery to the tissue (DO2), which is a product of cardiac output and arterial oxygen content (CaO2). In this regard, patientswith ACS experiencing a compromise in any facet of DO2 (e.g. SpO2 < 94%) would be considered to have a complicated presentation and should receive supplemental oxygen. We attempted to exclude patients whose chief complaint and additional vital signs were consistent with potentially complicated patient presentations. Nonetheless, we do not know how individual medical directors have defined “uncomplicated” in their local treatment guidelines. Further, it is reasonable to suspect that the decision to administer oxygen may have been influenced by variables not captured in our data set (e.g. pain severity or ECG changes).

Despite the unknowns discussed above, our findings do suggest a transition is underway. The inverse relationship between SpO2 values between 94% and 100% and oxygen administration suggests that the EMS community is transitioning treatment for these patients fromthe historical “can’t hurt, might help” approach. The trend for less supplemental O2 at higher SpO2 levels is also important to recognize. This may reflect a higher comfort level of individual providers with particular SpO2 values (e.g. 100%) rather than outright accepting or rejecting the 94% SpO2 threshold.


In addition to the inherent limitations of retrospective analyses, specifically missing SpO2 values and errors due to self-reported data, there are further limitations to this study. First, we do not know if the mere presence of a paramedic student during a patient encounter influences patient care decisions. In the case of oxygen administration, this may mean paramedics treat patients more conservatively than they would otherwise. Without a prospective study designed to evaluate this phenomenon, we are unable to assess if this factor influenced our study results. Second, patients with secondary complaints of shortness of breath, or who received treatments for dyspnea, were not excluded from our analysis. Including these patients may inflate the number receiving oxygen appropriately. However, we believe our physiologic exclusion criteria helped restrict our findings to the stable chest pain patient who likely did not require oxygen administration. Third, we used the first recorded SpO2 value as part of our inclusion criteria, but we are not able to be certain that this value was obtained prior to oxygen administration or if it changed during the course of transport. Fourth, although we believe our inclusion and exclusion criteria maximized our ability to include the uncomplicated patients only, there is the possibility our dataset contains patients who could be considered complicated. Finally, there may be other factors that contribute to the reduction in oxygen administration over time that were not evaluated (e.g. gender, age) in our unadjusted univariate regression models.


This is the first examination of prehospital administration of supplemental O2 following release of the 2010 updated AHA guidelines. The prehospital administration of supplemental O2 decreased significantly following release of the 2010 updated guidelines; however, our data revealed that 50% of patients not meeting criteria for administration still received supplemental O2. There are several potential explanations for this finding, including variation in the interpretation of the AHA guidelines, lack of provider compliance, and the limitations of our statistical analysis. Regardless, additional educational activities may be needed to reinforce the pathophysiology of supplemental O2 in the uncomplicated cardiac chest pain patient with a SpO2 ≥ 94%. This may assist in promoting the cultural acceptance of this change. Further investigation should also attempt to determine the degree to which the perceptions of individual EMS providers and medical directors affect the implementation and compliance with treatment recommendations from large national and international bodies. Finally, adherence by individual providers and agencies to this guideline change through continuous quality improvement may be warranted.


1. O’Connor RE, Brady W, Brooks SC, Diercks D, Egan J, Ghaemmaghami, C, Menon V, O’Neil BJ, Travers AH, & Yannopoulos D. Part 10: Acute coronary syndromes: 2010 American Heart Association guidelines for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation 2010;122(18):S787–S817.

2. McNulty PH, King N, Scott S, Hartman G, McCann J, Kozak M, Chambers CE, Demers LM, Sinoway LI. Effects of supplemental oxygen administration on coronary blood flow in patients undergoing cardiac catheterization. American Journal of Physiology Heart & Circulatory Physiology 2005; 288(3): H1057–62.

3. Wijesinghe M, Perrin K, Ranchord A, Simmonds M, Weatherall M, Beasley R. Routine use of oxygen in the treatment of myocardial infarction: Systematic review. Heart 2009;95(3):198–202.

4. Branson RD, Johannigman JA. Prehospital oxygen therapy. Respiratory Care 2013;58(1): 86–97.

5. Wang HE, Seitz SR, Hostler D, Yealy DM. Defining the learning curve for paramedic student endotracheal intubation. Prehospital Emergency Care 2005;9(2):156–62.

6. Salzman JG, Page D, Kaye K, Stetham N. Paramedic student adherence to the National Standard Curriculum recommendations. Prehospital Emergency Care 2007;11(4): 448–52.

7. Bourn S. A Comparison of the Clinical Environment for the Paramedic Intern and the Practicing Paramedic [Abstract]. Presented at the National EMS Educators Association Symposium, St. Louis, MO, September 11–14h, 2008.

8. Youngquist ST, Gausche-Hill M, Squire BT, Koenig WJ. Barriers to adoption of evidence-based prehospital airway management practices in California. Prehospital Emergency Care 2010;14(4):505–9.