Objectives: This experimental study compared the effect of compression-to-ventilation (CV) ratios of 15:1, 15:2, and 30:2 on hemodynamics and resuscitation outcome in a canine model of a simulated, witnessed ventricular fibrillation (VF) cardiac arrest.
Methods: Thirty healthy dogs, irrespective of species (mean ± SD, 19.2 ± 2.2 kg), were used in this study. A VF arrest was induced. The dogs received cardiopulmonary resuscitation (CPR) and were divided into three groups based on the applied CV ratios of 15:1, 15:2, and 30:2. After 1 minute of untreated VF, 4 minutes of basic life support (BLS) was performed. At the end of the 4 minutes, the dogs were defibrillated with an automatic external defibrillator (AED) and advanced cardiac life support (ACLS) efforts were continued for 10 minutes or until restoration of spontaneous circulation (ROSC) was attained, whichever came first.
Results: None of the hemodynamic parameters, and arterial oxygen profiles was significantly different between the three groups during BLS- and ACLS-CPR. Eight dogs (80%) from each group achieved ROSC during BLS and ACLS. The survival rate was not different between the three groups. In the 15:1 and 30:2 groups, the number of compressions delivered over 1 minute were significantly greater than in the 15:2 group (73.1 ± 8.1 and 69.0 ± 6.9 to 56.3 ± 6.8; p < 0.01). The time for ventilation during which compressions were stopped at each minute was significantly lower in the 15:1 and 30:2 groups than in the 15:2 group (15.4 ± 3.9 and 17.1 ± 2.7 to 25.2 ± 2.6 sec/min; p < 0.01).
Conclusions: In a canine model of witnessed VF using a simulated scenario, CPR with three CV ratios, 15:1, 15:2, and 30:2, did not result in any differences in hemodynamics, arterial oxygen profiles, and resuscitation outcome among the three groups. CPR with a CV ratio of 15:1 provided comparable chest compressions and shorter pauses for ventilation between each cycle compared to a CV ratio of 30:2.
The optimal compression-to-ventilation (CV) ratio for cardiopulmonary resuscitation (CPR) still remains controversial. Although the CV ratio of 30:2 was recommended in the American Heart Association 2005 guidelines for CPR and emergency cardiovascular care, further validation of this guideline is needed.1 Interruption of chest compressions during CPR is detrimental to maintaining the coronary perfusion pressure (CPP), and the controversy over CV ratio is focused on how to minimize interruption of chest compressions during CPR. A higher CV ratio may deliver more compressions per minute and decrease interruptions for ventilation.2 However, a higher CV ratio or continuous compressions might also be associated with an increase in rescuer fatigue and/or a decrease in compression quality.3–5 Chest compression-only CPR might be as good as CPR with ventilation for a successful outcome,6 but the advantage of increased blood flow is offset by emerging arterial desaturation after 1.5 to 2 minutes of CPR, with progressively less oxygen being delivered to the body tissues.7
In a real situation involving resuscitating a victim with out-of-hospital cardiac arrest, usually a bystander, rather than a healthcare provider, performs basic life support (BLS), including CPR and first defibrillation with an automated external defibrillator (AED) during the first several minutes. CV ratio is a component of BLS only during the first several minutes of attempted resuscitation because healthcare providers perform continuous compressions without pauses for ventilation once an advanced airway is in place during advanced cardiac life support (ACLS). For this reason, the effect of various CV ratios should be tested during the first several minutes of cardiac arrest.
The aim of this experimental study was to compare the effects of CV ratios of 15:2, 30:2, and 15:1 on hemodynamics and resuscitation outcome in a canine model of simulated, witnessed, ventricular fibrillation (VF) cardiac arrest. We designed this study to investigate short-term survival as a primary outcome and to compare hemodynamic variables and arterial oxygen profiles.
This was an experimental protocol using dogs to assess various CV ratios. The study was reviewed and approved by the Animal Research Committee of the Wonju College of Medicine, Yonsei University.
Thirty healthy dogs, irrespective of species, were used in this study. The dogs were initially sedated with 400 mg of intramuscular ketamine (Ketalar, Yuhan Corp., Seoul, Korea) and 0.3 mg/kg intramuscular xylazine (Rompun, Bayer Animal Health Corp., Ansan-Shi, Korea) and then further sedated with 200 mg of intravenous ketamine. The thoraces of all dogs were shaved by razor for application of the AED pads to the chest. After sedation, endotracheal intubation was done with a cuffed endotracheal tube. Electrocardiographic (ECG) monitoring was recorded continuously. Under aseptic conditions, the dogs were positioned supine and right femoral artery cannulation was done using the Seldinger method. Aortic blood pressures were recorded continuously, with a micromanometer-tipped catheter (5 Fr, Millar Instruments, Inc., Houston, TX). An introducer sheath (7.5 Fr, Arrow International Inc., Reading, PA) was placed in the right external jugular vein, and a micromanometer-tipped catheter was advanced to the right atrium to measure the right atrial pressure. An introducer sheath was placed in the right femoral vein for infusion of normal saline and intravenous medications. Left femoral artery cannulation was performed for arterial blood gas analysis. The right common carotid artery was surgically exposed, and an ultrasonic flow probe (T106, Transonic Systems Inc., Ithaca, NY) was placed around it to quantify blood flow. All dogs were treated with a heparin bolus (100 units/kg intravenously) once catheters were in place to prevent thrombosis formation. During the preparation, the dogs were ventilated with room air via a volume control ventilator (MDS Matrix 3000, Orchard Park, NY). The tidal volume was initially set at 12 mL/kg and the ventilator rate at 12 breaths per minute.
After baseline data were collected, a pacing catheter (5 Fr, bipolar lead, Arrow International Inc.) was positioned in the right ventricle. VF was then induced by delivering alternating electrical current at 60 Hz to the endocardium and confirmed by the ECG waveform and a decline in aortic pressure. Once VF was induced, the ventilator was disconnected from the endotracheal tube. After 1 minute of untreated VF, mimicking the activity of a bystander recognizing cardiac arrest and calling for help, 4 minutes of BLS was performed. Positive bag valve ventilations were delivered with a resuscitator bag (silicone resuscitator 870040, Laerdal Medical, Stavanger, Norway), with controlled peak airway pressure of 40 mm Hg. During BLS, each ventilation was given after performing a head-tilt chin-lift maneuver to mimic an airway maneuver by a bystander and make the rescuer use of time realistic. Approximately 300 to 400 mL of tidal volume for 1 second was delivered per breath.
The dogs received CPR and were randomly assigned to three groups: 1) the 15:1 group, provided with a manual rescue breath followed by 15 manual chest compressions; 2) the 15:2 group, provided with two manual rescue breaths followed by 15 manual compressions; or 3) the 30:2 group, provided with two manual rescue breaths followed by 30 manual compressions. Chest compression was performed by emergency medicine residents trained and experienced in BLS and ACLS with the metronome-guided rate of 100 compressions per minute. At the end of the 4-minute BLS-CPR, dogs were defibrillated with an AED (CU-ER1, CU Medical Systems Inc., Wonju Korea) delivering 80 J (≈4 J/kg) of energy. If VF persisted or asystole developed, 1 mg of epinephrine was administered intravenously, and all resuscitation efforts including provision of 100% oxygen were started following the 2005 ACLS pulseless arrest algorithm.11 Shock delivery was performed every 2 minutes during CPR, and 1 mg of epinephrine was administered intravenously every 3 minutes during CPR. ACLS-CPR efforts were continued for 10 minutes unless restoration of spontaneous circulation (ROSC) was attained. ROSC was defined as an unassisted pulse with a systolic arterial pressure of >50 mm Hg and a pulse pressure of >20 mm Hg lasting more than 1 minute. If ROSC was not achieved by 10 minutes, despite all efforts, the experiment was terminated. If ROSC was attained, dogs were observed for 2 hours of survival without any pharmacologic support and then euthanized by an intravenous injection of potassium chloride.
Data were digitized by a digital recording system (Powerlab, AD Instruments, Colorado Springs, CO). All the parameters (aortic and right atrial pressure and common carotid blood flow) were recorded continuously and analyzed at baseline, between Minutes 0 and 4 of CPR (1 to 5 min after VF cardiac arrest), between Minutes 4 and 14 of CPR (5 to 15 min after VF cardiac arrest), and 2 hours after ROSC if ROSC was achieved. CPP during CPR was calculated as the difference between aortic pressure and right atrial pressure in the relaxation phase using an electronic subtraction unit. Gradients of peak aortic pressures and CPPs from the first two compressions and the last two compressions were calculated by averaging the differences between mean pressures of the first two compressions and those of the last two compressions during BLS. End tidal CO2 (ETCO2), tidal volumes, and oxygen saturation were continuously measured (CO2SMO Plus, Novametrix Medical Systems, Wallingford, CT). Arterial blood gas specimens were collected from the left femoral artery at baseline (before cardiac arrest) and at 1, 5, 9, and 15 minutes after cardiac arrest. Arterial blood gas analyses including oxygen saturation, pCO2, pO2, pH, base excess, HCO3–, hemoglobin, and hematocrit were done with a blood gas analyzer (i-STAT1, Abbott Laboratories, Abbott Park, IL). Interruption time of compressions was measured as a time interval between the onset of aortic pressure drop from the last compression of the previous cycle and a very beginning of aortic pressure rise from the first compression of the cycle. Time for ventilation was measured as a time interval between an initial rise of inspiratory flow and the onset of expiratory upstroke of ETCO2 curve. Compressions delivered over 1 minute, interruption time of compressions over 1 minute, and time for ventilation at each cycle during BLS-CPR were calculated.
For analysis, hemodynamic data during BLS and ACLS were averaged from values measured during the entire period of BLS and ACLS. Data were summarized as mean ± standard error of the mean and coded into a computerized data processing software package (SPSS for Windows 12.0, SPSS Inc., Chicago, IL). One-way analysis of variance was used when appropriate to compare the three groups with regard to the aforementioned variables including hemodynamic data, arterial blood gas profiles, compressions delivered per minute, and pauses for ventilation. Post hoc comparison was performed with Tukey honestly significant difference post hoc test. Statistical differences of ROSC rate and 2-hour survival rate between the three groups were analyzed using the chi-square test. p-Values below 0.05 were defined as statistically significant.
The mean (± SD) weight of the dogs was 19.2 ± 2.2 kg with no significant difference between the three groups. There were no significant differences in aortic pressure, right atrial pressure, CPP, and common carotid blood flow between the three groups before VF induction.
Hemodynamic Parameters during BLS and ACLS-CPR
None of the hemodynamic parameters were significantly different among the three groups during BLS-and ACLS-CPR (Table 1). Aortic pressure at peak compression and common carotid blood flow during ACLS tended to be higher in the 30:2 group than the 15:1 and 15:2 groups, but this trend did not reach statistical significance. There were no differences in gradients of aortic pressures and CPPs from the first two compressions and the last two compressions of a CPR cycle among the three groups (Table 2). Aortic pressure at peak compression between the first two compressions and last two compressions during BLS-CPR tended to be higher in the 30:2 group than the 15:1 and 15:2 groups, but this trend did not reach statistical significance.
Table 1. Hemodynamic Parameters during BLS and ACLS CPR
Arterial blood oxygen saturation declined below 90% after 4 minutes of BLS-CPR and was maintained above 80% during BLS- and ACLS-CPR in all groups. There were no significant differences in arterial blood gas analysis at baseline, immediately after VF and first defibrillation, and 9 and 15 minutes after initiation of the experiment between the three groups (Table 3).
The success rate of defibrillation was 70% in the 15:1 group and 60% in the 15:2 and 30:2 groups at the first shock. There were no significant differences in the success rate of defibrillation during ACLS-CPR or hemodynamics during the compressions before defibrillation between the three groups.
Eight dogs (80%) of each group achieved ROSC during ACLS. Time to ROSC was not different between the three groups (6.8 ± 3.0 min in 15:1 group, 7.7 ± 3.3 min in 15:2 group, and 7.2 ± 4.4 min in 30:2 group; p > 0.05). Two-hour survival rate was 60% in the 15:1 group, 60% in the 15:2 group, and 70% in the 30:2 group. The survival rate was not different between the three groups (Table 4).
Table 4. Resuscitation Outcomes
15:1 (n = 10)
15:2 (n = 10)
30:2 (n = 10)
CV = compression-to-ventilation; ROSC = restoration of spontaneous circulation.
Data are reported as mean ± SD.
Time to ROSC (min)
6.8 ± 3.0
7.7 ± 3.3
7.2 ± 4.4
Two-hour survival (%)
Compression Rate, Interruption of Compressions, and Ventilation Duration during BLS-CPR
The number of compressions delivered over 1 minute was significantly greater in the 15:1 and 30:2 groups than in the 15:2 group (73.1 ± 8.1 and 69.0 ± 6.9 to 56.3 ± 6.8; p < 0.01). The time of interruption of compressions was shorter in the 15:1 and 30:2 groups than the 15:2 group (19.2 ± 3.7 and 20.0 ± 3.7 to 27.9 ± 3.7 sec/min; p < 0.01). The time for ventilation during which compressions were stopped at each minute was significantly lower in the 15:1 and 30:2 groups compared to the 15:2 group (15.4 ± 3.9 and 17.1 ± 2.7 to 25.2 ± 2.6 sec/min; p < 0.01). The time for ventilation at each cycle was lower in the 15:1 group compared to the 15:2 and 30:2 groups (3.3 ± 1.1 to 6.9 ± 1.5 and 7.6 ± 1.8 sec; p < 0.01; Figure 1).
Our investigation showed that hemodynamics, arterial blood gas profiles, and resuscitation outcomes were not significantly different in the three groups. This finding suggests that during the early phase of VF, the three CV ratios of 15:1, 15:2, or 30:2 are equivalent in terms of hemodynamics, arterial oxygenation, and short-term resuscitation outcome.
A CV ratio of 30:2 has recently been recommended in American Heart Association guidelines to increase the number of compressions, reduce the likelihood of hyperventilation, minimize interruption in chest compressions for ventilation, and simplify instruction for teaching and skill retention. Mathematical analysis indicates an optimum CV ratio of near 30:2 for standard performance and 60:2 for actual lay rescuer performance in the field.8 Based on theoretical analysis, Babbs and Kern9 propose that CV ratios in the range of 24:2 to 48:2 would be ideal to achieve maximal blood flow and oxygen delivery with interruption in chest compressions of 5 seconds per two breaths. In an animal model, a ratio of 30:2, compared to a ratio of 15:2, increased CPP, cerebral perfusion pressure, and carotid blood flow due to the greater number of compressions and fewer interruptions for ventilation.10 Based on these studies, a CV ratio of 30:2 has been chosen for standard CPR. However, the clinical efficacy of a CV ratio of 30:2 in real resuscitation situations has not yet been proven.
Prolonged interruption of chest compressions between each compression cycle is detrimental to hemodynamics during CPR. Berg et al.11 demonstrated that interrupting chest compressions for rescue breathing could adversely affect myocardial hemodynamics during CPR for VF. The 4-second pause for ventilation following each compression cycle has been shown to decrease aortic pressure during relaxation, resulting in reduced CPP and myocardial blood flow. A prospective ECG study found that an interruption of chest compressions for more than 5 seconds decreases the probability of successful defibrillation with ROSC in patients starting with a high-to-medium chance of success.12 Ideally, a pause of less than 5 seconds is recommended for two rescue breaths following each chest compression cycle, but in practice, the interruption of chest compressions for rescue breathing by lay rescuers requires about 16 seconds.13,14 A CV ratio of 15:1 is the same as a CV ratio of 30:2 in terms of providing compressions and ventilations. Our results confirm that CV ratios of 15:1 and 30:2 deliver similar compressions and time for ventilation per minute. The time for ventilation at each CPR cycle with a CV ratio of 15:1 was less than 5 seconds, but those with CV ratios of 30:2 and 15:2 were longer than 6 seconds. Differences in aortic pressure at peak compression tended to be lower in the 15:1 group than the 30:2 or 15:2 group, which was reflected in less deterioration of hemodynamics in the 15:1 group due to the shorter pause between CPR cycles. Yannopoulos et al.15 demonstrated in an animal experiment that a reduction in the ventilation frequency, by increasing the CV ratio from 15:2 to 15:1, improved hemodynamics and vital organ perfusion pressures without compromising oxygenation and acid–base balance by reducing intratracheal pressure. Therefore, a CV ratio of 15:1 has the advantage of shortening the time for ventilation at each CPR cycle over a CV ratio of 30:2, with comparable hemodynamics and arterial oxygen profiles. In addition, by providing only a single ventilation between compression sets, the rescuer can start the next cycle of CPR without any need to wait for exhalation after giving a breath. In this regard, providing a single ventilation between compression sets of any number might have practical advantages over any CV ratio with two ventilations.
In our study, we tested the effect of various CV ratios for CPR during the first 5 minutes of VF. When an AED is present, the clinically relevant time intervals were a mean of 3.5 minutes from collapse to AED attachment and a further 0.9 minutes to the delivery of the first defibrillation shock.16 We formulated a scenario of resuscitation with BLS and an AED shock performed by a bystander, followed by full support of ACLS by healthcare providers. The scenario consisted of 1 minute without CPR to check the response of the victim and activate the emergency medical system; 4 minutes of BLS-CPR and the first defibrillation shock with an AED, mimicking immediate response of a bystander; and 10 minutes of ACLS-CPR, mimicking treatment by healthcare providers. Once healthcare providers arrive at the scene and the advanced airway is in place, a CV ratio is meaningless, because pauses for ventilation are no longer necessary. Our scenario to test the effect of various CV ratios represented a typical emergency medical response including the early BLS-CPR, defibrillation within 5 minutes by a bystander, and ACLS-CPR by healthcare providers for a witnessed, out-of-hospital, cardiac arrest.
Pauses for ventilation were not identical to those in a real resuscitation situation, although our experimental scenario did simulate a witnessed cardiac arrest with bystander CPR. The difficulty and time required to give the breaths and the effectiveness of the breaths might be quite different in non-intubated humans than in intubated dogs. Therefore, the data from our study cannot be extrapolated to layperson CPR. Second, our scenario simulated an ideal response to the victim with out-of-hospital cardiac arrest witnessed by bystanders. Not all emergency medical systems experience this kind of emergency response to victims with out-of-hospital cardiac arrest. Not all bystanders can perform CPR or defibrillate the victim with an AED. Therefore, our study results are limited to the situation where a bystander is familiar with CPR, an AED is available, and the emergency medical response is adequate. Third, the interval of untreated VF was only 1 minute in this study. After such a short period, there is minimal loss of vascular tone. Thus, perfusion achieved with any combination of compression and ventilations may be optimal. This situation might minimize the opportunity to see differences between CV ratios because of a “ceiling effect.” Fourth, compression-only CPR was not tested in our experiment. Several investigations have revealed that compression-only CPR resulted in similar or superior resuscitation outcomes, compared to CPR with various CV ratios.17,18 In our cardiac arrest model with a short downtime, compression-only CPR might be as effective as the other modes of ventilation. Finally, because we could not measure actual depth of compression, the quality of CPR might affect the results, although all emergency residents have been certified as a CPR provider.
In a canine model of witnessed VF using a simulated scenario, CPR with three CV ratios, 15:1, 15:2, and 30:2, did not result in any differences in hemodynamics, arterial oxygen profiles, and resuscitation outcomes among the three groups. CPR with a CV ratio of 15:1 provided comparable chest compressions and shorter pauses for ventilation between each cycle compared to a CV ratio of 30:2.
This study was supported by a grant (02-PJ3-PG6-EV01-001) from Ministry of Health and Welfare, Republic of Korea.