Cardiac Arrest in the Operating Room: Resuscitation and Management for the Anesthesiologist: Part 1

Vivek K. Moitra, MD; Sharon Einav, MD; Karl-Christian Thies, MD; Mark E. Nunnally, MD; Andrea Gabrielli, MD; Gerald A. Maccioli, MD; Guy Weinberg, MD; Arna Banerjee, MD; Kurt Ruetzler, MD; Gregory Dobson, MD; Matthew D. McEvoy, MD; Michael F. O'Connor, MD, FCCM


Anesth Analg. 2018;126(3):876-888. 

In This Article

Precardiac Arrest Considerations

Surveys of anesthesiologists document lack of awareness of both basic and anesthesia-related knowledge of resuscitation and cardiac arrest.[23,24] One study documented delay in the cardioversion and defibrillation of patients with shockable rhythms in the perioperative setting.[25] To rescue a patient from crisis, caregivers must recognize the patient is in crisis and institute effective action.[26–28] Recognition that a patient is in crisis is more difficult in the periprocedural setting because the patient is sedated or under general anesthesia (precluding adequate monitoring of their mental status); their respirations are often controlled (preventing tachypnea or apnea); surgical positioning often frustrates assessment (lateral, prone, steep Trendelenburg); and large portions of their body are covered with drapes. Failure to rescue is an often-invoked "cause" of cardiac arrest and morbidity/mortality and is generally the product of hindsight bias shaping the evaluation of the care rendered.[29] While failure to rescue does occur, it almost certainly occurs less frequently than it is suggested. In many (likely most) instances, the underlying cause of crisis is so severe that the patient's demise is inevitable, even if maximal support is instituted in a timely fashion.[26,30]

Escalating Care

Escalation of care includes higher levels of monitoring and more advanced supportive measures. Decisions about higher levels of monitoring or evaluation require consideration of the patient's history, current clinical status, anesthetic, and procedure. Insertion of invasive monitors should not delay supportive care. Almost every unstable patient should be monitored with an arterial line. Central venous access is appropriate when monitoring central venous pressures or venous oxygen saturations help guide resuscitation, or when caregivers anticipate infusing vasoconstrictors over longer periods of time. Over the past decade, clinicians have increasingly performed point of care ultrasound in unstable patients to make quick diagnoses and manage a crisis.[31] The decision to escalate the level of monitoring is a clinical judgment that encompasses all relevant patient and surgical factors and is beyond the scope of these recommendations.

Clinical Progression to Shock

Anesthesiologists commonly administer titrated boluses of vasoactive drugs (ie, phenylephrine, ephedrine, vasopressin, norepinephrine, and epinephrine) to unstable patients. Often, small boluses of vasopressin (arginine vasopressin 0.5–2 units IV) may improve hemodynamics when escalating bolus doses of catecholamines have failed. The use of arginine vasopressin and its analogs in low-flow states, cardiac arrest, and hypotension refractory to catecholamines has been extensively documented.[32–35] A reasonable sequence of care for the unstable patient who is progressing toward shock is outlined in the first part of Table.

Left Ventricular Failure

Echocardiography and invasive monitors such as the pulmonary artery catheter guide the management of left ventricular (LV) failure. Hypovolemia can cause or contribute to shock in patients with poor LV function and should be remedied before institution of any pharmacologic therapies. Hypotensive euvolemic patients with LV shock are treated with inotropic agents and medications that reduce afterload.[36,37] In patients with known, significant diastolic dysfunction, therapy with lusitropic agents such as milrinone enhances ventricular relaxation to improve cardiac output. Increasingly, mechanical support with intraaortic balloon pumps, ventricular assist devices, and extracorporeal life support (also referred to as extracorporeal membrane oxygenation) is utilized in hospitalized patients who are believed to have good potential for recovery from severe LV shock, right ventricular (RV) shock, and cardiac arrest.[38,39] Figure 1 outlines 1 approach to manage a patient with LV shock.

Figure 1.

Treatment algorithm of LV failure with cardiogenic shock. CVP indicates central venous pressure; ECMO, extracorporeal membrane oxygenation; Hct, hematocrit; IABP, intraaortic balloon pump; LV, left ventricular; LVAD, left ventricular assist device; PEEP, positive end-expiratory pressure; PPV, pulse pressure variation; pRBC, packed red blood cell; PTX, pneumothorax; R/O, rule out; Scvo2, central venous oxygen saturation; SPV, systolic pressure variation; Svo2, mixed venous oxygen saturation; SVR, systemic vascular resistance in dyne·second−1·cm–5·m–2; TEE, transesophageal echocardiography; ZEEP, zero end-expiratory pressure.

RV Failure

Similar to LV shock, RV shock is best guided by a combination of invasive monitors such as the pulmonary artery catheter and/or echocardiography. In most instances, an acute rise in pulmonary vascular resistance (often in the setting of a chronic cause of pulmonary hypertension) causes and sustains RV shock.[40] A combination of inotropes, systemic arterial vasoconstrictors, and pulmonary artery vasodilators such as nitric oxide manage RV shock. In contrast to the management of LV shock, the use of systemic arterial vasoconstrictors for RV dysfunction may improve end-organ perfusion and cardiac output (Figure 2).[40,41] Administering vasopressin to enhance blood pressure may decrease the pulmonary vascular resistance-to-systemic vascular resistance ratio because vasopressin's constrictive effects spare the pulmonary vasculature compared to norepinephrine and phenylephrine.[42,43] Over the past several years, mechanical support devices, including ventricular assist devices and extracorporeal membrane oxygenation, have been utilized with increasing frequency in patients with RV shock.[38,39]

Figure 2.

Treatment algorithm of RV failure with cardiogenic shock. CVP indicates central venous pressure; ECMO, extracorporeal membrane oxygenation; Hct, hematocrit; IABP, intraaortic balloon pump; iNO, inhaled nitric oxide; PEEP, positive endexpiratory pressure; pRBC, packed red blood cell; PVR, pulmonary vascular resistance; R/O, rule out; RV, right ventricular; RVAD, right ventricular assist device; Scvo2, central venous oxygen saturation; Spo2, pulse oximeter oxygen saturation; Svo2, mixed venous oxygen saturation; TEE, transesophageal echocardiography; ZEEP, zero endexpiratory pressure.

Hypovolemia and Systolic and Pulse Pressure Variation

Hypovolemia can cause perioperative hypotension, circulatory crisis, and shock. Over the past decade, pulse pressure variation (PPV) and systolic pressure variation (SPV) have replaced central venous pressure monitoring as bedside indicators of volume responsiveness in hypotensive patients. While these measurements are most reliable in intubated, mechanically ventilated patients who are synchronous with appropriate ventilator settings (>8 mL/kg), there is a growing literature that suggests that SPV and PPV can be measured in spontaneously breathing patients with only slightly diminished reliability.[44,45] If the PPV or SPV exceeds the threshold value of 12%–15%, fluid administration or increased preload will likely increase stroke volume.[46] The plethysmographic signal from a pulse oximeter may suggest fluid responsiveness.[47] Importantly, the presence of RV shock or any of the causes of obstructive shock (auto-PEEP, cardiac tamponade, tension pneumothorax, pulmonary hypertensive crisis, and abdominal compartment syndrome) will produce elevated SPV and PPV that do "not" predict volume responsiveness.[48,49] Excessive tidal volumes (>10 mL/kg), increased residual volume and lung compliance (emphysema), and decreased chest wall compliance (third-degree chest burn, obesity, prone position) increase PPV and SPV, and criteria for volume responsiveness should be adjusted in these conditions.[50] Assessing heart–lung interactions via PPV or stroke volume variation in the setting of cardiac arrhythmias such as atrial fibrillation or frequent premature ventricular contractions is not reliable.[50]

Hypotension and PPV/SPV values of <10% suggest that hypotension and shock will not improve with fluid resuscitation. While the passive leg raise (a quick, reversible, and often easy-to-perform maneuver that raises the patient's legs to assess changes in blood pressure and hemodynamics) also predicts volume responsiveness, it is not especially practical in the operating room.[51–54] Although ultrasound assessment of inferior vena cava diameter variation with respiration may predict volume responsiveness, it is also not practical during a wide variety of operations (abdominal, cardiac, thoracic) or patient positions (lateral, prone, sitting).[55,56] Evaluation of the SVC diameter is possible with either transesophageal echocardiography or transthoracic echocardiography and is more practical in many operative settings. Esophageal Doppler assessment of aortic blood velocity is also predictive of volume responsiveness, but once again, requires instrumenting the esophagus and expertise in interpreting the data.[57] Practically speaking, a patient who is acutely and severely hypotensive should be volume resuscitated (with blood products if hemorrhage or undetected surgical bleeding is likely) as monitoring is escalated, and volume responsiveness is assessed via changes in blood pressure and heart rate.

Ventilation During Severe Shock or Cardiac Arrest

Over the past 2 decades, multiple clinical studies have demonstrated either no harm or a mortality or outcome benefit when patients with respiratory failure or acute respiratory distress syndrome are ventilated with lower tidal volumes and permissive hypercapnia; a strategy during which carbon dioxide (CO2) levels rise and pH falls as long as the oxygen saturation stays above 90%.[58–63]

Hyperventilation is deleterious in both shock and cardiac arrest. Studies of ventilation during shock repeatedly demonstrate that the duration of increased intrathoracic pressure is proportional to the ventilation rate, tidal volume, inspiratory time, and delayed chest decompression and is inversely proportional to coronary and cerebral artery perfusion.[59,64–66] Ventilation at 20 breaths·minute−1 during cardiopulmonary resuscitation (CPR) is associated with significantly lower survival than ventilation at 10 breaths·minute−1. BLS guidelines continue to emphasize avoiding hyperventilation during CPR and recommend higher compression-to-ventilation ratio (eg, 30:2) for victims of all ages (except newborns).[1] Even with an endotracheal tube, the respiratory rate should be 10 breaths·minute−1 or less, with an inspiratory time of 1 second, and the tidal volume limited to "chest rise" (approximately 500 mL in a 70-kg adult).[3] An algorithm for coordinating airway management with CPR is shown in Figure 3. Newer devices that provide a combination of automatic CPR and an airway in-line negative inspiratory valve (allowing increased venous return during chest decompression) may be associated with an increased rate of return of spontaneous circulation (ROSC), but no increase in survival to hospital discharge.[67–70]

Figure 3.

Intubation during CPR. CPR indicates cardiopulmonary resuscitation; ETT, endotracheal tube; LMA®, laryngeal mask airway; ROSC, return of spontaneous circulation.

Because positive pressure ventilation decreases venous return and hypoventilation seems to cause no harm, it is reasonable to ventilate patients in shock with the lowest ventilator settings compatible with a saturation of 90% or greater.


Auto-PEEP, also known as intrinsic PEEP or gas trapping, is a well-described but often difficult to recognize cause of circulatory collapse and PEA.[71] Auto-PEEP occurs almost exclusively in patients with obstructive lung disease, especially asthma and chronic obstructive pulmonary disease (emphysema). In patients with obstructive lung disease, mechanical ventilation that does not allow sufficient time for complete exhalation produces a gradual accumulation of air (volume) and pressure (end-expiratory pressure) in the alveoli. This pressure is transmitted to the pulmonary capillaries, and then to the great vessels in the thorax, where it decreases both venous return and cardiac output. Clinical reports demonstrate that as auto-PEEP increases, venous return decreases.[72,73]

The presence of auto-PEEP can be inferred whenever the expiratory flow waveform does not return to the zero baseline in between breaths. In the absence of a flow waveform display, auto-PEEP can be diagnosed by disconnecting the endotracheal tube from the ventilator for 10–20 seconds, and observing a "step-up" gain of invasive or noninvasive systolic blood pressure. Dramatic improvement in response to this maneuver should prompt maximal therapy for obstruction lung disease/bronchospasm, and mechanical ventilation with both small tidal volumes (<6 mL/kg), a low respiratory rate (<10/min), and a short inspiratory time (which will produce a paradoxical and acceptable increase in the peak inspiratory pressures). Given that auto-PEEP is an important cause of an unacceptable circulation, it should be quickly ruled out in any unstable patient. The Lazarus phenomenon, a seemingly miraculous recovery and ROSC after the discontinuation of resuscitative efforts, can diagnose circulatory collapse from auto-PEEP during resuscitation.[74]