For nearly three-quarters century, oxygen therapy has been recommended for the treatment of chest pain and acute myocardial injury (AMI). Barach was the first to report the positive effects of oxygen therapy on patients with myocardial infarction (AMI). 8, 9, Boland wrote in 1940 that there were two immediate goals in treating patients with acute coronary embolism.
They are support circulation and relief of pain. It is well-known that oxygen administration can have beneficial effects on cardiorespiratory function. This is especially true when there are significant levels of shock or pulmonary embolism. However, it is not well-known that oxygen administration can be beneficial in maintaining the cardiorespiratory function, especially in the case of severe shock or pulmonary edema.
Boland continues to explain that, although opiates can relieve pain from angina, it is not common for them to be effective in relieving pain. In cases where they are not sufficient, oxygen is an important “therapeutic addition.” Boothby and other members of the team who developed the BLB mask (Boothby Lovelace, Bulbulain), made similar observations about the role of 100% oxygen in the relief of chest pain.
Russek and his coworkers started to question the logic behind high oxygen concentrations for angina treatment in the absence of pulmonary edema. 12 The investigators examined 5 patients in the Master 2-step exercise test. Repeat testing showed electrocardiographic changes. They gave nitroglycerine before they began exercising. The researchers found that 100% oxygen did not prevent the RS-T segment or T wave changes in each of the five cases. The continuous administration of 100% oxygen for the duration of an exercise session did not result in the disappearance or reversal of electrocardiographic signs of myocardial oxia. In 4 cases, the changes were more apparent or more severe when 100% oxygen was administered. Furthermore, oxygen did not prevent or influence the duration of anginal pain in any of these cases. One tablet of nitroglycerin was administered before the test. This not only prevented the onset of anginal pain but also greatly affected the electrocardiographic response.
Despite the fact that this experiment was extremely well controlled and showed no beneficial effects of oxygen on chest pain, routine oxygen therapy for suspected myocardial damage continued for 50 years.
Recent research has focused on the role of oxygen therapy for AMI and published new recommendations. Ganz et al. demonstrated in their early work the importance of oxygen tension in coronary artery blood flow. However, high oxygen concentrations can also reduce coronary blood flow due to increased coronary vascular resistance. Myocardial oxygen consumption decreased by about 16% in these experiments. Similar results were also found in patients with congestive heart disease and those without. Hypoxemia may warrant oxygen administration. These studies show that oxygen administration may be necessary in the face of hypoxemia. 17 The table shows changes in coronary blood flow, myocardial oxygen intake, and coronary vascular resistance associated with oxygen delivery.
Impact of Hyperoxemia on Coronary Artery Circulation and Myocardial Oxygen Demand in Normals and Patients With Cardiac Disease
Rawles and Kenmure randomly assigned 200 patients with suspected AMI to either oxygen via face mask or air during their first 24 hours in hospital. This left 157 patients. There was no difference in mortality, arrhythmias incidence, analgesic use, or systolic intervals between the groups. The P aO 2 difference was larger in the oxygen group, and there was also a higher incidence of sinus tachycardia. These authors concluded that routine oxygen delivery to uncomplicated AMI was not necessary.
Beasley, Wijesinghe, and colleagues have concluded that routine oxygen use in AMI can lead to an increase in infarct size, increased mortality, and more frequent hospitalizations. Nearly 500 patients suffering from ST-elevation AMI will be enrolled in this trial. They will not receive hypoxia and either oxygen or air. Myocardial infarct severity will be the main outcome.
A Cochrane review of 2010 examined the use oxygen for AMI. The authors were unable to identify any randomized controlled trials in which one group received oxygen and the other did not. The studies involved 387 patients. There were 14 deaths and 3 times more likely for patients who received oxygen to die. Although these small numbers can’t prove causation, it clearly shows the need to reassess oxygen delivery in patients with AMI or normoxia. 22 The American Heart Association Emergency Cardiac Care Guidelines took these findings into consideration. According to current recommendations, emergency medical service (EMS) providers can deliver oxygen to patients with acute coronary disease. There is no evidence to support routine oxygen use in severe acute coronary syndrome. If there is dyspnea or hypoxemia, as determined by pulse oximetry, and signs of heart disease, oxygen should not be used routinely.
These scientific findings present a dilemma for both the EMS provider and the EMS physician. In a hurry and hurried setting, the EMS provider is responsible for diagnosing dyspnea. It might be more efficient to provide oxygen to all patients until they arrive at the hospital. Pulse oximetry is a method that can help identify hypoxemia. Hyperoxemia can be reduced by training EMS professionals to use low flow oxygen appliances instead of the common non-rebreathing mask.
Ronning and Guldvog compared the oxygen received by patients to either a nasal cannula or a control group that had no oxygen. The researchers followed patients for 7 months after discharge and determined if there was a trend towards increased mortality in patients with mild or moderate strokes. Patients with severe strokes and hypoxemia showed a significant improvement in their outcome. According to the authors, oxygen supplementation as a routine treatment is not beneficial for stroke victims.
Pancioli and his colleagues performed a retrospective chart review to evaluate the use of oxygen therapy in stroke patients. They found that oxygen therapy was indicated in less than half the patients. The indications were an S 2 >= 92%, breathing rate > 24 breaths/min and heart rate > 100 beats/min. Dyspnea, Cysis, Anemia (hemoglobin 9 g/dL), AMI, or chronic lung disease. Tachycardia and chronic lung disease were the most common indications for oxygen therapy. The study found that stroke patients are often treated with oxygen therapy without a prescription. It was suggested that decreasing oxygen usage could save money and help reduce costs. The study did not assess the outcomes.
Recent Chinese trials have shown that controlled oxygen therapy with an air-entrainment masque was less complicated than using a nasal cannula. A recent Chinese trial suggests that controlled oxygen therapy using an air-entrainment mask, as opposed to a nasal cannula, is associated with fewer complications.
This is partly due to animal studies that indicate that hyperoxia can lead to worse neurologic outcomes in cardiopulmonary resuscitation. These guidelines are partly based on animal studies that again suggest that hyperoxia can lead to worse neurologic outcomes.
This was further supported by a large clinical study of more than 6,000 cardiac arrest patients. The Emergency Medicine Shock Research Network investigators Kilgannon and Kilgannon found that hypoxia (33%), was three times more likely than normoxia (19%) or hyperoxia (18%) after cardiac arrest. However, hyperoxia was much more common. Hyperoxia exposure had a death rate of 1.8 (95%CI 1.5-2.2). These investigators also examined the mortality rates in ICU patients suffering from hyperoxia. They found a dose-dependent association between supranormal oxygen tension levels and in-hospital deaths. The relative death risk increased by 24% for every 100 mm Hg in P aO . They did not discover a threshold below or above which hyperoxia is more severe. 37
These data are convincing, but hypoxemia risks are still important. Further research is needed to determine the optimal F IO delivery for cardiopulmonary resuscitation and, perhaps, how to measure it.
While the presence of hypoxemia in the face of unknown lung disease should be treated with oxygen without delay in the pre-hospital environment, there are specific areas where oxygen therapy should be more carefully addressed. Perhaps the most complete guidelines for the use of oxygen in emergency situations is published by the British Thoracic Society.38 The following discussion will refer strictly to use in pre-hospital care of common diseases.
Chronic Obstructive Pulmonary Disease
The dangers of high flow oxygen in the presence of COPD were described over 50 years ago by Campbell in a landmark paper in the Lancet.39 He described 4 cases in patients with chronic bronchitis, emphysema, and bronchopneumonia. While the seminal observation was that high levels of inspired oxygen resulted in hypercapnia due to “respiratory depression,” it may be that the observation was correct but the attribution was incomplete. Work in the 1980s by Milic-Emili and colleagues suggests that alterations in ventilation-perfusion matching resulting from loss of adaptive regulatory mechanisms were responsible for hypercarbia.40–42 Regardless, over half a century ago, Campbell suggested that patients with acute on chronic respiratory disease should receive oxygen, “continuously by a method which permits the inspired concentration to be controlled within limits of ± 1% over the range of 24–35%.”
More recently, a number of trials have evaluated the role of pre-hospital oxygen use in patients with COPD. Durrington et al evaluated pre-hospital oxygen delivery in Norfolk and Norwich in the United Kingdom, where transport times are commonly longer than 30 min.43 They hypothesized that with prolonged transport times the impact of excess oxygen delivery might be exaggerated. They retrospectively reviewed patients admitted with exacerbations of COPD over 2 time frames. The first 108 patients were reviewed, and, after a period of training, a second group was reviewed. The intervention between the 2 time frames was introduction of an air-entrainment oxygen mask set at an FIO2 of 0.28 for initial management. Patients in the first group with FIO2 > 0.28 had a greater incidence of acidosis, hypercarbia, and hyperoxia (Fig. 1). These patients were also more likely to have a complicated hospital course, more likely to receive aminophylline, noninvasive ventilation, and invasive ventilation, and had a higher mortality rate.
Changes in PaCO2 and PaO2 during pre-hospital oxygen delivery at FIO2 ≤ 0.28, initial FIO2 ≥ 0.28 changed to ≤ 0.28, and FIO2 ≥ 0.28. (Based on data from Reference 43.)
Cameron and others evaluated the outcome of patients with COPD exacerbation admitted to a New Zealand regional emergency department over 30 months.44 There were 680 patients presenting with COPD exacerbation, and 254 emergency admissions, representing 180 patients, were available for analysis. They found that hyperoxemia occurred in nearly a quarter of all admissions, and that, compared to normoxemia, the number of adverse outcomes were significantly greater (odds ratio 9.1, 95% CI 4.08–20.6). They also demonstrated that hypoxemia was associated with adverse outcomes (odds ratio 2.16, 95% CI 1.11–4.20). Clinically speaking, the hyperoxia group presented with an SpO2 > 96%, and the hypoxemia group with an SpO2 of < 88%. These data argue strongly for the use of titrated oxygen delivery in the patient with COPD.
Austin and colleagues compared high flow oxygen versus titrated oxygen treatment in a group of patients with COPD exacerbation in Tasmania.45 They evaluated 214 patients with a diagnosis of COPD exacerbation, of whom 117 received high flow oxygen and 97 received titrated oxygen therapy. In this group the mortality was 9% (11/117) in the high flow oxygen group, and 2% (2/97) in the titrated oxygen group. Patients receiving titrated oxygen were far less likely to have acidosis or hypercapnia. Other authors have made similar findings in this group of patients.46
The British Thoracic Society has been a vocal advocate of changing practice to reduce excess oxygen exposure in these patients.6,47 The key to success is translation of this information into practice by EMS providers. This will include the addition of specialized equipment to provide high flow, low FIO2 to patients at risk.48 The British Thoracic Society suggests that all EMS units carry a non-rebreathing mask for high flow oxygen delivery, a nasal cannula or simple mask for low flow oxygen delivery, tracheostomy masks for patients with tracheostomy or laryngectomy, and a 28% air entrainment mask. Another important issue is educating pre-hospital providers to operate up-draft nebulizers with room air if metered-dose inhalers are unavailable. During nebulizer treatments, oxygen can be provided at 2 L/min via nasal cannula. Another interesting aspect is the concept of providing patients with cards identifying themselves as having COPD and requesting use of a low FIO2 air entrainment mask during ambulance transport. This has been suggested by the British Thoracic Society. A possible example is shown in Figure 2.
Proposed example of an oxygen alert card that can be provided to patients with chronic respiratory disease at risk for hypercarbia. This includes patients with COPD as well as those with neuromuscular disease and obesity hypoventilation syndrome.
Oxygen therapy in patients with COPD is the classic story of Goldilocks and the Three Bears. Oxygen delivery cannot be too much or too little, but just right. The complexity of making the just right approach a reality requires substantial education, training, and cooperation from patients, hospitals, and EMS providers.49
Congestive Heart Failure and Pulmonary Edema
Congestive heart failure and pulmonary edema represent a common etiology behind the pre-hospital complaint of “shortness of breath.”50 While hypoxemia is a clear symptom of pulmonary edema, administration of oxygen has limited value in a fluid-filled lung. The use of pre-hospital CPAP and noninvasive ventilation has increased to treat congestive heart failure and pulmonary edema.51–57 These studies have demonstrated reduced need for intubation, shorter hospital stays, and reduced costs with CPAP and or noninvasive ventilation, compared to oxygen with similar medical therapies.
With respect to the use of oxygen in congestive heart failure and pulmonary edema, the paper by Bledsoe et al addresses this directly.51 They evaluated 340 patients with respiratory distress in the pre-hospital setting. Nearly half of the patients presented with symptoms of congestive heart failure/acute pulmonary edema. These authors found that CPAP at an FIO2 of 0.28–0.3 was effective in treating patients with congestive heart failure and pulmonary edema. Only 6% of patients required pre-hospital intubation, and 70% of patients had improved upon arrival at the emergency department.
Other Pulmonary Diseases
The use of pre-hospital oxygen therapy in patients with pneumonia and asthma has been reported.58–60 Patients with pneumonia are likely to have a component of chronic lung disease, and, as such, high flow oxygen delivery can result in elevated PaCO2.58 Majumdar et al have shown that in patients with pneumonia a low SpO2 (< 92%) is associated with greater mortality and morbidity at 30 days. This study evaluated 2,923 patients treated in Canadian emergency departments.59 They found that using a threshold for admission of an SpO2 < 90% resulted in a number of subjects requiring readmission. These authors believe that raising the threshold for admission to 92% would be safer. This may provide a threshold for titrated oxygen therapy in pre-hospital care.
In patients with asthma, data suggest that oxygen should be provided only in the face of hypoxemia. Additionally, the work by Perrin et al60 suggests that high inspired oxygen leads to hypercapnia in a fashion similar to that seen during treatment of COPD.
Traumatic illness and injury represent a wide variety of pathologies, commonly classified globally as blunt or penetrating. The spectrum of traumatic injury includes orthopedic fractures, pulmonary contusion, pneumothorax, airway obstruction, blood loss, traumatic brain injury, and solid organ injuries. While each can contribute to hypoxemia in a number of ways, the typical trauma patient in pre-hospital care is approached in a similar fashion based on the guidelines of Pre-hospital Trauma Life Support.61
The requirement for oxygen in the treatment of trauma patients is not well described. Data from the Pre-hospital Trauma Life Support (PHTLS) manual suggest that oxygen delivery should be provided based on the patient’s breathing frequency (Fig. 3). This methodology, along with the recommendations for oxygen delivery devices, results in the common default to a non-rebreathing mask at 15 L/min. The tables in Figure 3, are from different chapters in the Pre-hospital Trauma Life Support (PHTLS) manual. Together, however, they tend to encourage the use of high FIO2 in both spontaneously breathing patients and those requiring ventilatory support. With short transport times and multiple tasks required of the EMS provider, the use of the non-rebreathing mask simplifies care and reduces the number of required oxygen delivery devices in stock. Anecdotally, we have seen trauma patients with gunshot wounds to extremities arrive in the emergency department talking on their cell telephone held under the non-rebreather. The presumption is that high FIO2 is not toxic in these short time frames and that hypoxemia is far more dangerous.
Recommendations from the Pre-hospital Trauma Life Support (PHTLS) manual for FIO2 delivery, based on breathing frequency and FIO2 capabilities of devices. The linking of this data results in a recommendation of an FIO2 of > 0.85 in most cases. If the patient is spontaneously breathing this requires use of a non-rebreathing mask. NA = not applicable.
The United States Special Operations Command provides some guidance on the use of oxygen for traumatic injury on the battlefield.62 These guidelines suggest that oxygen should be provided when oximetry identifies hypoxemia (SpO2 < 92%). When oximetry is unavailable, oxygen is considered in casualties with loss of consciousness, traumatic brain injury, hemorrhagic shock, or casualty at altitude.
These recommendations carry conventional wisdom. Following traumatic brain injury, a single incidence of hypoxemia results in a worse neurologic outcome. This effect is associative; it is not clear if it is cause and effect.63–65 The use of oxygen in hemorrhagic shock is based on providing adequate oxygen delivery following massive blood loss. Kirkman and colleagues66 have shown that, following blast injury in a porcine model of controlled/uncontrolled hemorrhage, maintenance of SpO2 at 95% is associated with prolonged survival times, compared to animals breathing room air. They also noted that animals breathing air had progressive decreases in base deficit, corresponding to inadequate oxygen delivery. Two thirds of the animals supplied oxygen survived, compared to no survivors in the room air group. Unfortunately, these authors did not define the amount of oxygen required to reverse hypoxemia. As oxygen is a precious, finite resource on the battlefield, knowing how much oxygen to deploy is critical.
Stockinger and McSwain evaluated oxygen use in pre-hospital trauma patients in New Orleans over a 33 month time frame. During this time, 5,090 trauma patients were brought to the trauma center. Of these, 57% of patients did not receive pre-hospital oxygen, while 43% did. Patients who received oxygen in the pre-hospital setting had a significantly greater mortality rate (2.3% vs 1.1%). There was no difference in the outcomes among patients with blunt or penetrating injury. These data are intriguing. The authors concluded that oxygen does not improve outcome in trauma patients. It, however, might simply demonstrate that sicker patients receive oxygen in the pre-hospital environment. The number of subjects and range of injury severity is large, resulting in a very low mortality rate.67
How much oxygen trauma patients require is yet to be defined. Clearly, 15 L/min by non-rebreathing mask is often excessive. We have shown that in mechanically ventilated casualties an average oxygen flow of 3 L/min would suffice in a majority of patients.68 We have also presented preliminary data on the incidence of hypoxemia in patients following trauma and the utilization of oxygen in suspected traumatic brain injury.69,70 These data suggest that hypoxemia is not uncommon in pre-hospital transport of trauma patients (approximately 30%). However, a number of pre-hospital causes of hypoxemia (airway obstruction, pneumo/hemothorax) are not amenable to oxygen treatment. Further work to determine the flow of oxygen to reverse hypoxemia is warranted.
Data suggest that oxygen should be given only to patients who demonstrate hypoxemia and that oxygen should be titrated to a normal SpO2. As in other conditions, hyperoxemia may have untoward effects in trauma patients, with and without traumatic brain injury.71
A number of other pre-hospital conditions warrant the use of oxygen therapy in pre-hospital care or have been singled out for oxygen use. A small list is provided below.
While pregnancy causes a restrictive pulmonary defect, and pregnant patients may present with a variety of critical illnesses in pre-hospital care, there is no evidence that pregnancy requires oxygen therapy more than any other health condition.72–74 Oxygen therapy in pregnancy should be titrated to underlying pathophysiology and guided by oximetry or blood gas analysis.
Breathlessness is seen in patients with hypoxemia, but also in patients at the end of life, those with anxiety, and other illness not associated with hypoxemia. There is no evidence that oxygen delivery alleviates breathlessness in non-hypoxemic patients.75
Sickle Cell Crisis
Oxygen is often prescribed for the treatment of sickle cell disease, particularly in crisis. However there is a dearth of evidence supporting the effectiveness of oxygen in all patients. Oxygen therapy has not been shown to affect the duration of a pain crisis or to be useful in patients with acute chest syndrome with normoxemia. Oxygen should be administered only if hypoxemia is present.76–78
Carbon Monoxide Poisoning
The delivery of high concentrations of oxygen to patients with carbon monoxide poisoning is one condition where hyperoxemia is desirable. The half-life of carboxyhemoglobin is 4–5 hours breathing room air. Breathing 100% oxygen reduces this to 40 min.79,80
Oxygen is life-saving in the face of hypoxemia. The relatively common belief that oxygen is non-toxic over short exposures, logistical challenges, and training issues in pre-hospital care have resulted in routine excess oxygen use: excess not only in dose but in indications for conditions. At present, oxygen should be titrated to alleviate hypoxemia and prevent hyperoxemia, with treatment of carbon monoxide poisoning being the exception.