Assisted Ventilation During ‘Bystander’ CPR in a Swine Acute Myocardial Infarction Model Does Not Improve Outcome

Most cardiac arrests occur in prehospital settings.¹ Although bystander cardiopulmonary resuscitation (CPR) substantially improves survival from these out-of-hospital cardiac arrests,^2–5 potential rescuers are often reluctant to provide CPR.^6–8 The greatest barrier to performing bystander CPR seems to be concerns regarding mouth-to-mouth rescue breathing.^6–8 Moreover, single rescuer bystander CPR as recommended by the American Heart Association is a complex psychomotor task that is difficult to learn, remember, and perform.⁹ If bystander CPR could be safely and efficaciously performed without mouth-to-mouth rescue breathing, potential bystanders may be more willing and able to provide this life-saving intervention.⁸

Experimental data in animal models have established that assisted ventilation is not a necessary component of CPR in some circumstances.^10–15 During CPR for ventricular fibrillation (VF), the initial rate-limiting step for oxygen delivery to the myocardium is inadequate blood flow rather than inadequate ventilation. Arterial oxygen saturation can remain greater than 90% for several minutes of chest compressions without assisted ventilation.^11,15 In addition, passive ventilation during chest compressions and active gasping provide substantial ventilation for brief periods of CPR.^10–17 More importantly, in four swine fibrillatory cardiac arrest experiments, 24-hour survival was comparable after chest compressions with or without assisted ventilation.^10,12,13,15

Most animal models of cardiac arrest and CPR have ignored the importance of coronary artery abnormalities on CPR physiology. Nevertheless, laboratory investigations have established that coronary artery obstruction has profound effects on CPR physiology.^18–20 Because (1) greater than 5 minutes of CPR without assisted ventilation can lead to decreased arterial oxygenation^13,21 and (2) coronary artery obstruction further compromises the myocardium, we evaluated the need for assisted ventilation during 10 minutes of CPR in a swine cardiac arrest model with coronary artery obstruction. Our hypothesis was that simulated bystander CPR would improve 24-hour neurologically intact survival compared with no bystander CPR, but that assisted ventilation would provide no further benefit.

Methods

Preparation

Experimental protocols were approved by the Institutional Animal Care and Use Committee and followed the guidelines of the American Physiological Society. Experiments were performed on healthy domestic swine, weighing approximately 30 to 35 kg. After an overnight fast, the pigs were subjected to masked induction of anesthesia with isoflurane followed by oral endotracheal intubation. They were mechanically ventilated with a volume-limited time-cycled Harvard ventilator (model 661; Harvard Apparatus, Inc), on a mixture of room air and titrated isoflurane (approximately 1%). The tidal volume was initially set at 15 mL/kg and ventilator rate at 16 breaths per minute; ventilator settings were adjusted to maintain end-tidal carbon dioxide at 40±2 mm Hg.

After obtaining a surgical plane of anesthesia, introducer sheaths were placed in the right internal and external jugular veins, left external jugular vein, right carotid artery, and right femoral artery by cutdown technique. Continuous arterial pressure monitoring was performed via a 7F pigtail, micromanometer-tipped, solid state catheter (Millar Instruments) placed in the descending aorta near the diaphragm from the right femoral artery. A 5F coronary sinus catheter was placed through the right internal jugular vein. A 7F balloon-tipped flotation catheter was placed in the main pulmonary artery from the left external jugular vein and a 4F bipolar pacing catheter in the right ventricle.

A 5F Amplatz AL-1 angiographic catheter was advanced through a right common carotid artery 7F introducer sheath into the ascending aorta, and the left main coronary artery was cannulated. Angiography was performed to define the left coronary system. A 0.014-in, high-torque, floppy coronary guidewire was placed in the left anterior descending coronary artery with its tip extending to the distal aspect of the vessel. After removal of the Amplatz catheter, a steel cylinder was advanced over the guidewire with the use of an angioplasty catheter (Mansfield Scientific Inc) as a “pushing” catheter. The ostium of the left main coronary artery was negotiated by careful advancement of the angioplasty balloon (pushing the steel cylinder) as the angioplasty guide catheter was held still. The steel cylinder was 4.5 mm long, with an outer diameter of 2.0 mm and an internal diameter of 0.457 mm. The narrow lumen fit snugly over the coronary guidewire. Previous experience indicated that the outer diameter was appropriate for wedging into the mid left anterior descending coronary artery (LAD) of a 30- to 35-kg pig. The steel cylinders were wedged in the mid-LAD between the second and third diagonal arteries. After placement of the steel cylinder, coronary angiography was repeated in the first 10 animals, demonstrating complete coronary occlusion (Fig 1). Placement was documented fluoroscopically for subsequent animals. All animals developed ischemic changes on ECG after cylinder placement (Fig 2).

After removal of the angioplasty catheter and guidewire, a 7F pigtail catheter was placed into the left ventricle through the right carotid artery introducer sheath for infusion of microspheres. Fifteen minutes after placement of the mid-LAD cylinder, VF was induced, the pacing catheter was removed, and a 5F calibrated micromanometer-tipped catheter (Millar Instruments) was advanced through the introducer into the right atrium. All catheter placements were performed under fluoroscopic guidance.

Measurements

Right atrial and thoracic aortic pressure waveforms, ECG, and end-tidal carbon dioxide measurements were continuously monitored and recorded on a four-channel Gould ES 1000 recorder throughout the experiment until the 1- hour simulated intensive care unit (ICU) period ended. End-tidal carbon dioxide was measured with an infrared capnometer (model 47210A, Hewlett Packard) through a sensor attached to the ventilator circuit at the end of the endotracheal tube. Coronary perfusion pressure during CPR was calculated by subtracting right atrial relaxation (mid-diastolic) pressure from simultaneous aortic relaxation (mid-diastolic) pressure at a single point during three consecutive compression-relaxation cycles. Arterial blood gas specimens were obtained from the thoracic aorta, mixed venous specimens from the main pulmonary artery, and coronary sinus specimens through the coronary sinus catheter at baseline (before cardiac arrest) and during CPR (10 minutes after cardiac arrest). Oxygen saturation, PCO₂, PO₂, pH, and hemoglobin were measured with a blood gas analyzer (IL-1306 with model 482 co-oximeter, Instrumentation Laboratories). Cardiac output and regional blood flow to the left ventricle were determined by nonradioactive, colored microsphere technique at baseline (before cardiac arrest) and during CPR (in the interval from 8.5 to 11 minutes of VF), as previously described.^15,18–20

Experimental Protocol

After baseline data were collected, a steel cylinder was placed in the mid-LAD. After 15 minutes of observation, isoflurane was discontinued and VF was induced by applying 60-cycle alternating current to the endocardium through the pacing electrode (Fig 3). VF was confirmed by the typical ECG rhythm and precipitous decrease in arterial pressure. Mechanical ventilation was then discontinued. A 2-minute interval of untreated VF, mimicking a bystander recognizing the cardiac arrest and calling for help, was followed by a 10-minute basic life support period, consistent with typical paramedic response times.³ Animals were randomly assigned into one of three groups: (1) chest compressions plus assisted ventilation (CC+V); (2) chest compressions only (CC); and (3) no CPR during the 10-minute basic life support period (control group). The CC+V group had an endotracheal tube placed and received two bag-valve–endotracheal tube breaths followed by 15 manual chest compressions at the rate of 100/min, sequentially for 10 minutes. The “rescue breaths” were provided with a gas mixture of 17% oxygen and 4% carbon dioxide, consistent with expired air from a rescue breather.²² The CC group had endotracheal tubes removed and received 10 minutes of manual chest compressions at 100/min. Chest compression rates were metronome guided.

After the basic life support period (12 minutes after VF induction), all animals received advanced cardiac life support with the American Heart Association algorithms for VF, as if a paramedic group arrived. Defibrillation attempts started with 120 J on the first shock, followed by 200 J on subsequent attempts. CC animals were reintubated during the minute immediately preceding the first defibrillation attempt. If the three initial attempts at defibrillation were unsuccessful, CPR was restarted and epinephrine (1 mg) was administered intravenously. After epinephrine administration, CPR was continued for 30 seconds to allow circulation of the epinephrine before further attempts to defibrillate. CPR by this simulated “paramedic team” included ventilation with 100% oxygen on a volume-limited, time-cycled ventilator at a rate of 15 breaths/min and chest compressions manually at a rate of 100 compressions/min. Restoration of spontaneous circulation was defined as unassisted pulse with a systolic arterial pressure of at least 50 mm Hg and pulse pressure of at least 20 mm Hg lasting for at least 1 minute.

Intensive Care

All successfully resuscitated animals were supported for 1 hour in a simulated intensive care unit setting. Systolic blood pressure was sustained greater than 80 mm Hg with dopamine and/or volume administration, as clinically indicated. All pigs received 10 mL/kg of normal saline intravenously during the intensive care period. Ventricular arrhythmias were treated with lidocaine or electroshock therapy as necessary. Mechanical ventilation was provided with 100% oxygen and adjusted to obtain an end-tidal carbon dioxide of 40±2 mm Hg. Recurrent cardiac arrest was treated with standard CPR and advanced life support with the American Heart Association algorithms. At the end of 1 hour, all animals were weaned from pharmacological and ventilatory support. Throughout the intensive care period, isoflurane was administered, as necessary, to maintain adequate analgesia and anesthesia. Animals that survived the ICU period were transferred to observation cages for the next 24 hours. Survival and neurological status were evaluated at 24 hours after the initial cardiac arrest, as previously described.^10,12,15 Survivors were euthanatized by infusion of Beuthanasia.

Documentation of Infarct

The heart was excised from a 24-hour surviving animal immediately after it was euthanatized for documentation of infarct as previously described.¹⁹ The left and right main coronary arteries were simultaneously but separately cannulated with two 5F Amplatz catheters and were injected with 20 mL of Evans blue dye. Both atria were removed, and the remaining myocardium was sectioned from apex to base in 10-mm slices along a perpendicular plane to the long axis of the left ventricle. Each slice was then incubated in a 1% solution of triphenyltetrazolium chloride buffered to a pH of 7.4 at 37°C. After 15 minutes of incubation and after rinsing in sodium chloride, each slice was photographed from the basal side.

Data Analysis

Heart rates and blood pressures were collected from the graphic records at prearrest baseline and 15 minutes after resuscitation. Aortic and right atrial compression and relaxation pressures were collected from the graphic records during CPR. Average left ventricular myocardial blood flow was determined by adding epicardial and endocardial flows from the anterior, inferior, and lateral walls of the left ventricle and dividing by 6.

Myocardial oxygen delivery was calculated from average left ventricular myocardial blood flow×aortic oxygen content, and myocardial oxygen consumption was derived from average left ventricular myocardial blood flow×(aortic−coronary sinus oxygen content). Similarly, systemic oxygen delivery was calculated from cardiac output×aortic oxygen content, and systemic oxygen consumption was derived from cardiac output× (aortic−pulmonary artery oxygen content).

Continuous variables were evaluated by ANOVA or unpaired Student’s t tests. Continuous variables are described as mean±SE. Comparisons of discrete variables were evaluated by Fisher’s exact test.

Results

Forty-nine animals were studied. Six animals were excluded from further analysis because of complications during placement of the coronary artery cylinder, including 3 with proximal LAD cylinder placement, 2 with intractable VF during coronary artery instrumentation before cylinder placement, and 1 animal with anomalous coronary anatomy precluding cylinder placement. Mean weight of the remaining 43 animals was 33±1 kg.

Return of spontaneous circulation was attained in 8 of 14 CC animals, 8 of 15 CC+V animals, and 9 of 14 control animals. Only 3 animals were defibrillated by the first attempt: 2 CC animals that immediately attained return of spontaneous circulation and 1 CC+V animal that rapidly refibrillated and attained return of spontaneous circulation with the next defibrillation attempt. Five of 14 CC animals and 3 of 15 CC+V animals survived for 24 hours, whereas only 1 of 14 control animals survived for 24 hours (CC versus control, P=.07). Seven of the nine 24-hour survivors were neurologically normal by the swine cerebral performance scale (cerebral performance category 1). These seven normal 24-hour survivors were walking, drinking, eating, and acting as they had before the experiment. One CC pig and one CC+V pig survived 24 hours with moderate neurological impairment (cerebral performance category 3).

Baseline weights, hemoglobin concentrations, and hemodynamic data before coronary artery cylinder placement did not differ significantly among the three groups (Table 1). Blood pressures obtained during CPR in the CC+V and CC groups did not differ except that the CC+V group had higher coronary perfusion pressures after 6 minutes of CPR than the CC group (19.8±3.4 versus 10.1±2.2 mm Hg, P<.05). There were no differences in aortic systolic pressures between the two groups during CPR, suggesting that the force of compressions was comparable in both groups (Table 2).

Hemodynamic data 15 minutes after resuscitation did not differ among the three groups. Difficulty with resuscitation and intensive care management was further estimated by comparing the three groups in terms of number of electrical shocks and epinephrine doses during resuscitation, need for dopamine or lidocaine during resuscitation or ICU management, and mean time until return of spontaneous circulation. No significant differences or trends were noted. The mean number of defibrillation attempts was 7.9±1.4 with CC, 7.7±1.1 with CC+V, and 9.4±1.4 with no “bystander” CPR.

Arterial and mixed venous Po₂ and oxygen saturation (So₂) generally did not differ at baseline, although the baseline mixed venous So₂ was slightly greater in the CC group than in the CC+V group (Table 3). Coronary sinus Po₂ and So₂ of the CC and CC+V groups also did not differ at baseline. During CPR, the arterial Po₂ and So₂ were higher in the CC+V group than the CC group. However, the mixed venous and coronary sinus Po₂ and So₂ during CPR did not differ.

Baseline arterial, mixed venous, and coronary sinus pH and Pco₂ did not differ (Table 4). During CPR, the arterial pH was higher and Pco₂ lower in the CC+V group than in the CC group. The mixed venous and coronary sinus pH and Pco₂ of the two experimental groups were not different.

Left ventricular myocardial regional blood flows were measured on 16 pigs (6 CC+V and 10 CC), but the data from one swine in each group were not evaluated because of proximal LAD cylinder placement in 1 (animal not included in any analyses) and a microsphere technical problem in another. There were no differences in left ventricular myocardial blood flows between the experimental groups at baseline or during CPR (Table 5). In addition, there were no differences in cardiac output, systemic oxygen delivery or consumption, or myocardial oxygen delivery or consumption between the two groups at baseline or during CPR. Furthermore, left and right kidney blood flows were quite similar at baseline and during CPR.

Thirty-nine of the 43 animals had active gasping, or “agonal,” respirations during CPR. Some had only a few agonal breaths; some had 5 to 7 deep gasps/min. Gasping was less frequent later in the CPR period. All CC and CC+V animals gasped, as well as 10 of the 14 controls. All nine 24-hour survivors gasped.

During CPR, animals that survived for 24 hours had markedly higher myocardial oxygen delivery (512±110 versus 179±42 mL/100 g per minute, P<.01) and consumption (470±105 versus 159±39 mL/100 g per minute, P<.01) than nonsurvivors. Similarly, animals with return of spontaneous circulation had markedly higher myocardial oxygen delivery (396±76 versus 110±24 mL/100 g per minute, P<.01) and consumption (363±71 versus 94±22 mL/100 g per minute, P<.01). In contradistinction, systemic oxygen delivery (6.37±1.06 versus 6.31±1.32 L/100 g per minute) and consumption (5.00±0.77 versus 5.22±1.33 L/100 g per minute) during CPR were not different between the survivors and nonsurvivors. Interestingly, all 7 animals with return of spontaneous circulation and microsphere data had myocardial oxygen delivery >200 mL/100 g per minute versus only 1 of 7 animals without return of spontaneous circulation, P<.01 (Fig 4).

The expected anterior myocardial infarction was demonstrated in the specimen stained with Evans blue dye and triphenyltetrazolium chloride (Fig 5).

Discussion

In this realistic model of prehospital single-rescuer CPR with acute myocardial infarction, assisted ventilation did not improve outcome. These findings are consistent with the blood gas, blood flow, oxygen delivery, and oxygen consumption data. Although assisted ventilation during CPR resulted in slightly superior oxygenation of the arterial blood, acid-base status was not differentially affected. Perhaps more importantly, systemic and myocardial oxygen delivery and consumption were not different in the CC versus the CC+V groups. Consequently, initial resuscitation rates, survival, and neurological outcome were not different in the two groups.

Defibrillation is the definitive treatment of VF. However, CPR can be an important bridge for myocardial oxygen supply until defibrillation is available. In this investigation, 12 minutes of untreated VF (control group) or 2 minutes of untreated VF followed by 10 minutes of CPR for VF precluded initial return of spontaneous circulation despite defibrillation attempts in approximately half of the animals and ultimately precluded 24-hour survival in more than two thirds of the animals. Adequate myocardial oxygen delivery and consumption during this prolonged CPR period was apparently instrumental in preserving the potential of the myocardium for successful resuscitation and the potential of the animal for survival. Mean myocardial oxygen delivery and consumption were approximately threefold higher in animals with return of spontaneous circulation and 24-hour survival compared with those without. Impressively, all 7 animals with return of spontaneous circulation and microsphere data had myocardial oxygen delivery >200 mL/100 g per minute versus only 1 of 7 animals without return of spontaneous circulation, P<.01 (Fig 4).

If myocardial oxygen delivery and consumption are so important for survival, why isn’t pulmonary respiration, the exchange of oxygen and carbon dioxide, also important? It is. Idris and colleagues²³ demonstrated that hypoxia and hypercarbia independently adversely affect outcome from prolonged fibrillatory cardiac arrest. Not surprisingly, the excellent outcomes of the CC group in the present study occurred in the setting of adequate ventilation and oxygenation despite no assisted ventilation. The CC animals had a mean arterial pH and Pco₂ of 7.40 and 34 mm Hg, respectively, and mean Po₂ of 61 mm Hg after 8 minutes of chest compressions without assisted ventilation. Exhalation of small volumes of gas was heard and felt during chest compressions, and active gasping was frequently noted.

Consistent with these observations, we previously demonstrated that swine in fibrillatory cardiac arrest provided with chest compressions alone for 9.5 minutes maintained an arterial pH of 7.33 and arterial Pco₂ of 48 mm Hg.¹⁰ Moreover, in a canine VF model, arterial oxygen saturation was maintained at >90% and minute ventilation at 5.2 L/min during the fourth minute of chest compressions without assisted ventilation.¹¹ However, swine studies indicate that minute ventilation decreases after 3 to 10 minutes of CPR, possibly because of progressive atelectasis and thoracic deformity.^13,17,21 Provision of assisted ventilation during CPR can minimize this decrease in minute ventilation to some extent.¹³

Complementary to the findings in this study, several animal investigations have now established that (1) spontaneous gasping occurs frequently during CPR, (2) the gasping contributes substantially to minute ventilation, especially during “chest compressions only” CPR, and (3) gasping is associated with improved outcome.^12–16 Clark and coworkers²⁴ have similarly observed that agonal respirations occurred in 40% of 445 out-of-hospital cardiac arrests, and these gasping breaths were associated with increased survival.

In the current study, 24-hour survival tended to be better in the CC group compared with the control group. In two previous investigations, we established that outcome was better with either CC or CC+V compared with no CPR for 12 to 13 minutes.^10,15 Those findings are consistent with the clinical data that bystander CPR improves outcome from cardiac arrest.^2–5 The power of this study to demonstrate improved outcome with either bystander CPR technique was limited because of the lower survival rates in the two bystander CPR groups and the single 24-hour survivor in the control group. The additional severity of insult with the coronary artery obstruction and acute myocardial infarction resulted in a markedly lower survival rate compared with our previous investigations. In fact, our attempts to mimic bystander CPR with this combination of a realistic timeline and coronary artery obstruction resulted in resuscitation rates and survival rates similar to those in prehospital human studies. Unfortunately, these lower survival rates require higher numbers of animals for differences between groups to reach statistical significance. The single 24-hour survivor in the control group was the only control animal among 28 animals in our three studies with similar untreated VF intervals to survive for 24 hours neurologically intact. If this animal’s outcome was similar to the other 27, both bystander CPR techniques would have resulted in statistically improved outcomes compared with the control group.

Some aspects of this study protocol would tend to bias the data in favor of the assisted ventilation group compared with the situation during prehospital single-rescuer CPR. Both groups received excellent CPR. It is unlikely that excellent compressions and mouth-to-mouth ventilation would be provided by a single rescuer in the field. Blood flow obviously decreases rapidly during pauses for ventilation. Transitions from ventilation to compressions and vice versa are likely to be much more difficult for a single rescuer than for our experienced, multi-individual research team. Moreover, the CC+V animals benefited from mechanical ventilation and optimal airway management with an endotracheal tube. Mouth-to-mouth rescue breathing is not as controlled, effective, or safe.

Important limitations of this study also include lack of blinding, limited power, and applicability to human cardiac arrest victims. By its very nature, this study could not be blinded. Standardized resuscitation and postresuscitation protocols were strictly adhered to in order to minimize treatment bias. Comparability of aortic systolic pressures and myocardial and renal blood flows during CPR indicate that the chest compressions were similar for both groups.

The power of this study was limited by the small number of animals. However, the evidence that assisted ventilation did not improve outcome was stronger than the simple lack of a value of P<.05. The physiological data also demonstrated lack of important benefits with assisted ventilation, and there was no tendency toward improved outcome with CC+V. In fact, the absolute number and percentage of 24-hour survivors were greater in the CC group than the CC+V group (in all likelihood the result of random variation). More convincing is the consistency of these data with the 24-hour outcome data from four previously published swine CPR studies comparing chest compressions with and without assisted ventilation.^10,12,13,15 In those studies, 38 of 47 (81%) CC animals and 36 of 45 (80%) CC+V animals attained 24-hour neurologically intact survival. The different experimental protocols have included brief untreated VF intervals with 100% survivals,^10,12,13 5-minute untreated VF intervals with 50% survivals,¹⁵ and the present study with 20% to 35% survivals. The remarkable consistency of these findings despite different experimental designs is an important counterweight to the argument of inadequate power in each individual experiment.

We attempted to more accurately mimic human CPR physiology with the addition of coronary artery obstruction and acute myocardial infarction in this model. Nevertheless, upper airway anatomy of pigs differs from that of human beings. In addition, neurological responses, such as active maintenance of the airway and ventilatory drive, may differ. Finally, human sudden cardiac arrest victims do not always experience acute coronary artery obstruction and myocardial infarction at the time of the sudden cardiac arrest.

Safar and others^25-30 clearly demonstrated in the 1950s that mouth-to-mouth rescue breathing is superior to various chest compression and back compression techniques for ventilating paralyzed, anesthetized adults and children. In particular, upper airway obstruction precluded any ventilation in many of the subjects. Clinical experience in the operating room and other settings confirm that upper airway obstruction can preclude effective ventilation. However, limited clinical observations in humans have demonstrated that excellent minute ventilation can be attained during CPR with the active compression-decompression device (plunger) despite no assisted ventilation or establishment of an airway.³¹ In addition, as noted above, many cardiac arrest victims exhibit active gasping. The roles of gasping and airway tone in cardiac arrest victims appear to be important, yet they are poorly understood phenomena deserving further investigation.

A prospective study of 3053 prehospital cardiac arrests suggests that our findings are applicable to humans.^4,5 Belgian physicians who were present on ambulances evaluated the quality and efficiency of prehospital bystander CPR. Long-term survival was comparable among those treated with good-quality chest compressions alone (17 of 116, or 15%) and those treated with good-quality chest compressions plus mouth-to-mouth ventilation (71 of 443, or 16%). The outcomes were superior with either of these techniques compared with those receiving no CPR (123 of 2055 survival, or 6%, P<.001).

In conclusion, this investigation with a realistic acute myocardial infarction model of bystander CPR further establishes that assisted ventilation is not a necessary component of CPR for swine with fibrillatory cardiac arrests. Limited data suggest this may also be true in humans. Further elucidation of this issue in humans has important clinical and public policy implications.

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

Footnotes