Each year, more than 500,000 cardiac surgery procedures are performed in the United States, and, until recently, this number continued to grow annually. Cardiac surgical patients are routinely admitted to the intensive care unit (ICU) for monitoring of recovery from anesthesia and surgery, optimization of hemodynamics, weaning from ventilatory support, and monitoring for possible complications. The ICU has emerged as the dominant area where the complex transition from the operating room to sophisticated care occurs. With high volumes of cardiac surgery procedures, the postoperative care of these patients accounts for a significant percentage of ICU admissions at many institutions. Cardiothoracic surgery intensive care units (CT-ICUs) have evolved as a separate entity from the general surgical ICU as management for cardiac surgery patients has become streamlined and algorithm driven. Critical care is best managed when the service is designed for a homogeneous population with a circumscribed set of medical and surgical issues.
Traditionally, cardiac surgery patients remained in the ICU for a few days before discharge to the ward or step-down unit. Over the past decade, ICU management has changed in response to changing patient populations, new surgical and anesthetic techniques, and the penetration of managed care. Patients presenting for cardiac surgery are significantly older as the number of patients undergoing angioplasty and stenting procedures increases. Aggressive medical therapy and nonsurgical revascularization techniques also result in patients presenting for surgery at more advanced stages of disease and with substantially more comorbidities. Furthermore, given the current market of health maintenance organizations (HMOs) and other cost-containment strategies, there is an ongoing trend toward enhanced recovery after cardiac surgery (ERAS), promoting an earlier return to normal activities. These initiatives aim for reduction of complications, accelerated care (e.g., fast track), clinical care pathways, and earlier discharge from intensive care.
With the development of minimally invasive cardiac surgery, warm bypass, and off-pump bypass techniques, cardiac surgeons have altered the requirements for conventional postoperative recovery. The overall trend has been to move away from the high-dose, opioid-based anesthetic techniques of the past to newer forms of balanced anesthesia with shorter-acting induction agents (propofol and etomidate), volatile agents (isoflurane and sevoflurane), and reduced doses of opioids allow accelerated patient recovery from anesthesia. This trend has been accompanied by an adoption of multimodal postoperative analgesia including intravenous nonsteroidal analgesics and regional anesthesia (both nerve and fascial blocks). Consequently, intensivists need to develop strategies to manage this ever-changing patient population in an efficient and cost-effective manner while maintaining quality and minimizing morbidity and mortality. Effective postoperative management depends highly on each patient’s preoperative status, intraoperative events, and condition on ICU arrival. Although the majority of institutions utilize the surgical ICU or specialized CT-ICU for postoperative care, avoidance of ICU admission altogether may be the future for selected patient populations: some institutions utilize step-down units or short-stay intensive care for the weaning process and high-dependency care.
Transport and Initial Assessment
The transport of the freshly operated cardiac surgery patient from the operating room to the ICU is not without risks and should be as smooth as possible. Problems encountered during transport include acute changes in physiology (with hypovolemia being the most prevalent), sudden awakening, or serious bleeding. Once surgery is complete and the patient is stabilized in the operating room, the patient is transported to the ICU while still emerging from anesthesia. To improve comfort and safety, the patient is routinely monitored (electrocardiography [ECG], invasive blood pressure [BP] monitoring, and pulse oximetry) and maintained on sedation with short-acting agents such as propofol or dexmedetomidine supplemented with opioids when needed. This allows the patient to wake up gradually in the ICU while the intensivist team continues to monitor organ perfusion and postoperative complications.
During the transition of care from the anesthesiologist to the ICU team, the intensivist must become familiar with the patient’s medical history and intraoperative course. Ideally, the multidisciplinary hands-off process should follow a structured checklist and involve members from the anesthesia team, the surgical team and the ICU team. Preoperative and intraoperative events vary in magnitude and duration but typically result in a myocardium of reduced contractility and compliance, which affects the postoperative management and eventual outcome. Essential data include indication for surgery, preoperative cardiac function and major comorbidities, details of the surgical procedure (arterial and venous coronary bypass grafts, valves repaired or replaced, aortic surgery, etc.), duration of cardiopulmonary bypass and cross-clamp, difficulties weaning from bypass, and post-bypass cardiac function assessment. It is also important to know the patient’s individual response to fluid administration and vasoactive agents, the number and position of chest drains, requirement of cardiac pacing, and hemostasis before chest closure. Extra caution should be paid to the “re-do” procedures (repeat sternotomy or thoracotomy). They are technically more challenging and tend to result in greater blood loss and higher complication rates. Finally, the amount and type of fluid and blood products administered, the expected postoperative course, and specific postoperative guidelines (BP targets, time for weaning from ventilator support, initiation of anticoagulation, etc.) should be communicated to the ICU team.
Upon the patient’s arrival in the ICU, the clinician should carefully note the position of the endotracheal tube and perform an inspection of all surgical sites, drains, and indwelling catheters. Important physical examinations focusing specifically on potential complications of cardiac surgery include pupillary response, pulmonary auscultation, cardiac auscultation, chest tube output, and urinary output. Patients may be hypothermic upon arrival, thus temperature should be rapidly assessed and warming initiated if necessary. A 12-lead ECG provides valuable information regarding baseline cardiac rhythm and potential ischemia. An admission chest radiograph (CXR) effectively checks for adequate positioning of all tubes and lines and provides assessment of lung volumes, pulmonary edema, and potential operative complications such as pneumothorax and hemothorax. CXR also provides a baseline cardiac silhouette for later comparison if cardiac tamponade were to develop. The extent of hemodynamic monitoring depends on the patient’s perioperative condition, intraoperative course, and anticipated complications after cardiac surgery. At a minimum, patients should have an arterial line, a central venous pressure (CVP) line, and a urinary catheter. The utility of pulmonary artery catheters (PAC) remains debatable. Since Connors and colleagues initially questioned the utility of PAC in critically ill patients, there have been numerous studies investigating the effect of PAC in various patient populations with different conclusions on patient outcome. Despite the lack of consensus, there is report of increased PAC use in cardiac surgeries from 2010 to 2014. A recent cohort analysis using propensity match demonstrates that PAC use was associated with decreased cardiopulmonary morbidity, decreased hospital length of stay, and increased risk of infection. There was no difference on the 30-day mortality between the PAC group and the control group. The increased utilization of intraoperative transesophageal echocardiography (TEE) has resulted in PAC being less frequently used during the surgery. However, in the immediate postoperative period, PAC data may guide therapy in patients with severe pulmonary hypertension, right heart failure, low cardiac output (CO), end-stage renal disease requiring renal-replacement therapy, or patients with severe diastolic dysfunction.
Routine laboratory analysis on admission includes arterial blood gases (ABG), electrolytes, complete blood count (CBC), and coagulation parameters. ABG and electrolyte data provide insight toward acid-base derangement, intravascular volume status, and optimization of ventilator support. The plasma levels of potassium and magnesium need to be kept above 4.0 and 2.0 mmol/L, respectively, to minimize the incidence of cardiac dysrhythmias. Potassium repletion may not be successful without restoration of magnesium stores. The ABG, hemoglobin/hematocrit, and electrolytes tests should be obtained every 4 to 8 hours or more frequently when clinically indicated, until when the patient’s condition has stabilized and the data can be obtained less frequently. As blood loss and marked inflammation associated with cardiovascular surgery may result in significant derangements in hemoglobin, platelets, and clotting factor activity, a CBC and standard coagulation tests are necessary to guide transfusion therapy. If significant coagulopathy is initially suspected on arrival in the ICU from high chest tube output, an activated clotting time may also be used as a rapid point-of-care test to check for adequate reversal of heparin given during surgery. The incorporation of thromboelastography (TEG) or rotational thromboelastometry (ROTEM) as a point-of-care monitor into a transfusion algorithm allows more specific diagnosis of coagulopathy and reduction of indiscriminate transfusion practices ( Fig. 31.1 ).
Immediate changes in respiratory function as a consequence of cardiac surgery and CPB include significant decreases in vital capacity, total lung capacity, and functional residual capacity, as well as significant atelectasis. There is also increased pulmonary edema correlating with the length of CPB. These disorders probably stem from a mix of surgical stress, complete deflation of the lungs during cardiopulmonary bypass, and incisional pain. In patients undergoing coronary artery bypass grafting (CABG), the average period of postoperative mechanical ventilation used to be between 1 and 2 days when high-dose, opioid-based anesthetics were used. With the modern balanced anesthetic technique, the “fast track” clinical care pathway targets early extubation, with the average mechanical ventilation time ranging from 6 hours or less to 24 hours postoperatively. Compared with late extubation (> 24 hours postoperatively), early endotracheal extubation after cardiac surgery has been shown to reduce both the ICU and hospital length of stay, to be more cost efficient, to improve left ventricular function, and to decrease cardiopulmonary morbidity.
Traditionally, ventilation with high tidal volume (10 mL/kg or more) was used to reduce atelectasis. However, current evidence suggests that lower tidal volume (6–8 mL/kg ideal body weight) is beneficial, with the goal of reducing barotrauma and volume trauma in the lungs that were already susceptible to injury by surgical stress. Lower tidal volume has been shown to provide higher success rate of ventilator-free at 6 hours after cardiac surgery as well as a lower incidence of reintubation. On the contrary, high tidal volume is a risk factor for organ failure and prolonged ICU stay after cardiac surgery. Positive end-expiratory pressure (PEEP) is generally started at 5 cm H 2 O and titrated up to adequate oxygenation. Whereas high versus low PEEP strategies are debatable throughout critical care practice, lower PEEP in postoperative cardiac surgery patients may facilitate improved hemodynamics as long as adequate alveolar recruitment and oxygenation is maintained. Since hyperoxia (PaO 2 > 120 mmHg) is detrimental to organ function with worse clinical outcome as a result of production of oxygen-free radical species and subsequent inflammatory responses, it is critical to reduce the fraction of inspired oxygen (F i O 2 ) as quickly as possible as long as adequate oxygenation is maintained (PaO 2 > 60 mmHg). The partial pressure of arterial carbon dioxide (PaCO 2 ) should be titrated to compensate for any concurrent metabolic acidosis, with the goal of a normalized pH. When criteria for extubation are met, patients can be transitioned to pressure support ventilation (PSV) mode. Early extubation after admission to the ICU is an essential component of fast-track protocols.
Routinely, weaning from ventilatory support can start once the patient starts to breathe over the rate set on the ventilator. Weaning should not be aggressive until the patient meets certain established criteria ( Table 31.1 ). These criteria relate to the patient’s neurologic, cardiac, respiratory, and renal status. They define an awake, normothermic, and hemodynamically stable patient who most likely will not need to return to the operating room for surgical bleeding. One method of weaning is to reduce the rate set on the ventilator gradually until the patient is consistently triggering breaths, after which the ventilator can be switched to a low level of pressure support (generally 5 cm of PEEP and 5 cm of support) for a formal spontaneous breathing trial (SBT). The patient who can maintain a reasonable gas exchange and is not tachypneic after 30 minutes on pressure support is ready to have the endotracheal tube removed. At some institutions, the patient’s pulmonary mechanics ( Box 31.1 ) are evaluated for further confirmation. An alternative weaning strategy is to attempt SBTs every 30 minutes as the patient continues to recover from anesthesia until a certain set criterion is satisfied (see Table 31.1 ). Both methods are safe and neither has any clear advantage over the other. Most patients undergoing cardiac surgery with normal pulmonary function preoperatively do not require a gradual wean from the ventilator. If the patient has remained stable during recovery from anesthesia and has no significant pulmonary disease, a rapid decrease in ventilator support to minimal levels can be safely instituted under close observation, and the patient can be evaluated for extubation. The method of removal from ventilatory assistance varies between institutions depending on the characteristics of the individual ICU.
|General criteria||Awake and follow commands; adequate analgesia|
|Hemodynamically stable a : MAP > 65 mmHg, CI > 2 L/min/m 2|
|Normal acid-base status: pH ≥ 7.32|
|Chest tube drainage < 50 mL/h|
|Urine output > 0.5 mL/kg/min|
|Core temperature > 36°C, no shivering|
|Airway criteria||Intact cough and gag reflex|
|Neuromuscular blockade fully reversed|
|Pulmonary secretions manageable|
|Chest radiograph within expectations|
|Gas exchange criteria||PaO 2 > 70 mmHg on 40% F i O 2 ; P a CO 2 < 50 mmHg; |
PEEP at 5 cm H 2 O
|No signs of respiratory distress: |
f/VT < 100 on PSV or T-piece;
|NIF ≤ -25 cmH 2 O|
|FVC > 10 mL/kg|
Respiratory rate < 25/min
Vital capacity > 12–15 mL/kg
Maximal negative inspiratory pressure ≤-25 cm H 2 O
Minute ventilation (V m ) < 10 L/min
Respiratory rate (f) / tidal volume (V T ) ratio < 105
pH ≥ 7.35
PaO 2 > 60 mmHg with F i O 2 < 40%
PaO 2 / F i O 2 ratio > 200
P A O 2 – PaO 2 < 350 mmHg
Long-Term Ventilation and Weaning
The occurrence of respiratory complications and the duration of endotracheal intubation have been shown to correlate with mortality in patients who have undergone cardiac operations. Because mechanical ventilation can have life-threatening complications, it should be discontinued at the earliest possible time. It is important to realize that prolonged intubation time may lead to respiratory tract mucociliary dysfunction, diminished clearing of secretions, atelectasis, delirium, and ventilator-associated pneumonia. The process of discontinuing mechanical ventilation, termed weaning, is one of the most challenging problems in intensive care, and it accounts for a considerable proportion of the workload of ICU staff.
Most cardiac surgical patients are expeditiously weaned from mechanical ventilation, but up to 6% of all patients undergoing CABG surgery require mechanical ventilation for more than 1 day and approximately 2% remain on the ventilator for more than 2 weeks. Identification of preoperative risk factors limiting the ability to wean from mechanical ventilation has been difficult, but common factors include advanced age, marked neurologic deficits, acute renal failure with volume overload, limited perioperative cardiac function, unstable hemodynamics, and sepsis. Other intraoperative and postoperative predictors of failure to wean include prolonged CPB, preexisting pulmonary disease (e.g., chronic obstructive pulmonary disease [COPD]), diaphragmatic paralysis, malnutrition, and high oxygen requirements.
The difference in oxygen consumption between spontaneous and total mechanical ventilation can be substantial. Routine preoperative pulmonary function tests have failed to serve as predictors for prolonged mechanical ventilation. Risk stratification with established scoring systems has led to mixed results when predicting the length of endotracheal intubation after cardiac surgery. Methods used to wean from mechanical ventilation include synchronized intermittent mandatory ventilation (SIMV), PSV, and T-piece trials with spontaneous breathing. Until the early 1990s, all weaning methods were considered equally effective, and the intensivist’s judgment was regarded as the critical determinant. This has changed with the results of randomized controlled trials that revealed that the period of weaning is up to three times longer with SIMV compared with SBTs. SBTs may be as short as 30 minutes and appear to be as effective when performed once a day as several times a day.
Protocol-driven weaning from mechanical ventilation or automated weaning protocols may facilitate the challenging task of weaning. Scheduled assessment of respiratory mechanics may also predict successful weaning. Among patients who cannot be weaned, disconnection from the ventilator is followed almost immediately by an increase in respiratory rate and a fall in tidal volume. The respiratory rate/tidal volume (f/V T ) ratio, also known as the rapid shallow breathing index, has been used with some success as a predictor of failure to wean. An f/V T ratio greater than 105 resulted in 95% of patients failing to wean, whereas an f/V T ratio less than 105 resulted in a weaning success of 80%. In a randomized trial, a two-stage approach to weaning—systematic measurement of predictors, including f/V T , followed by a single daily trial of spontaneous breathing—was compared with conventional management. Although the patients assigned to the two-stage approach were generally sicker than those assigned to conventional weaning, they were weaned twice as rapidly. This approach was not only cost effective but also lowered the incidence of complications when compared with conventional management.
In addition to the increase in respiratory effort, an unsuccessful attempt at spontaneous breathing causes considerable cardiovascular stress. Patients can have substantial increases in right and left ventricular afterload, with increases of 39% and 27% in pulmonary and systemic arterial pressures, respectively, most probably because the negative swings in intrathoracic pressure are more extreme. While most patients demonstrate improved hemodynamics with weaning of positive pressure ventilation secondary to improved preload and cessation of sedation, such afterload increases may be very poorly tolerated in those with severe heart failure. One of the most common reasons for failure to wean is pulmonary edema, which worsens gas exchange and increases the work of breathing. The correct diagnosis is made with a careful clinical examination, assessment of net fluid balance, and review of a chest radiograph. The etiology of the pulmonary edema will guide therapy. Careful attention to fluid balance and aggressive diuresis or use of ultrafiltration as indicated for patients with good cardiac function are essential components of therapy. Patients with poor cardiac function and pulmonary edema need afterload reduction and/or inodilator therapy in combination with diuresis in preparation for successful extubation.
Long-term management of failure to wean is facilitated by performing a tracheostomy, which allows reduced sedation, better pulmonary toilet, and utilization of intermittent ventilation, but carries the risk of higher rates of mediastinitis. A comparison of open versus bedside percutaneous dilatational tracheostomy in cardiothoracic surgical patients revealed no significant clinical differences but the potential for significant cost savings with the latter.
Adequacy of tissue perfusion and CO can routinely be assessed with the clinical evaluation of heart rate (HR), heart rhythm, BP, skin perfusion, capillary refill, and urine output. A patient with good skin perfusion and normal values for mean arterial blood pressure (MAP), HR, and urine output probably has a normal CO and sufficient tissue perfusion. However, these parameters remain insensitive for dysoxia (abnormal tissue oxygen utilization) and are considered to be poor surrogates of markers for oxygen delivery at the tissue level, because tissue oxygenation is determined by the net balance between cellular oxygen supply and oxygen demand. Indirect measures of tissue perfusion and oxygenation include the calculation of oxygen delivery or the measurement of mixed venous oxygen saturation (SvO 2 ). SvO 2 can be readily measured either intermittently from blood gas analysis or continuously with a fiberoptic PA catheter. The importance of monitoring arterial lactate levels in critically ill patients has been advocated, and concentrations of greater than 2 mmol/L are generally considered a biochemical marker of inadequate oxygenation. However, one must keep in mind that administration of beta-adrenergic inotropes may create a type-B lactic acidosis through increased glycolytic activity in skeletal muscles, which can confound the use of rising lactate to detect hypoperfusion. Other methods to assess tissue perfusion include near-infrared spectroscopy, gastrointestinal tonometry, and direct monitoring of tissue oxygenation with miniaturized implantable Clark electrodes. A recent study compared using normal capillary refill time (CRT) with serum lactate level as a resuscitation target in patients with septic shock. There was 8.5% of risk reduction of 28-day mortality in the CRT group compared with the lactate group; however, there was no significant difference between the two groups. In clinical practice, the combination of both direct monitoring of tissue perfusion and indirect biomarkers may provide best resuscitation strategy in critically ill patients.
One of the most dynamic physiologic variables during the first hour of postoperative ICU care is the MAP, which changes rapidly as a result of dynamic alterations in preload, afterload, and ventricular function. Fluid shifts and venodilation results in decreased preload, while arteriolar vasodilation from widespread inflammation and sedatives results in decreased afterload. Myocardial stunning from ischemia, cardiopulmonary bypass, or deep hypothermic circulatory arrest may result in both poor ventricular compliance and contractility. Arrhythmogenic disturbances may also have significant effects on MAP. Bradycardia can result from conduction system edema, surgical damage, or drug effects, whereas atrial fibrillation or other losses of atrioventricular synchrony impede filling of stiff, postoperative ventricles. Although hypotension has no concrete definition, the general consensus is that a MAP goal of 60 to 80 mmHg is ideal and that a systolic BP of less than 90 mmHg or a MAP of less than 60 mmHg denotes hypotension. However, the patient’s baseline BP must be considered because that determines the range of autoregulation for end organs such as the brain or kidney. In older adult patients and those with preexisting cerebrovascular or renal disease, a higher MAP may be required to perfuse adequately tissue beds that are used to chronic hypertension. Diastolic BP is a major determinant of myocardial blood flow, therefore attention must be paid to diastolic BP if myocardial ischemia is evident. Occasionally, the cardiac surgeon may request a lower BP target if they deem the risk of bleeding to be particularly high. Such cases demand a careful balance of ensuring adequate organ perfusion while also minimizing bleeding risk.
Hypovolemia is the most common cause of hypotension in postoperative patients. Peripheral rewarming causes vasodilation, requiring expansion of circulating blood volume for treatment. Ongoing bleeding requires intravascular volume replacement. Diagnosis of hypovolemia relies on clinical observation (vital signs, chest tube output, urinary output), chest radiograph, and point of care ultrasound (POCUS).
Although measurements of CVP and pulmonary capillary occlusion pressure (PCOP) have been commonly used to determine the need for fluid responsiveness, increasing data suggest that both the absolute values and their trends can be misleading, with poor negative and positive predictive values. When used in appropriate clinical situations, more accurate information is gained from dynamic indicators such as pulse pressure variation (PPV), stroke volume variation (SVV), and inferior vena cava (IVC) distensibility/collapsibility. Because each of these methods attempts to estimate the response of the cardiovascular system to a fluid changes, the best confirmation of fluid responsiveness may be an improvement in stroke volume (and thus CO) measured by a PA catheter, echocardiography, or carotid velocity-time-integral (VTI) immediately after the application of a small amount of crystalloid or a passive leg raise test. Patients who do not respond to small amounts of fluid are unlikely to respond to more volume.
The best resuscitation fluid to use in cardiac surgery patients depends upon the clinical scenario. Patients in hemorrhagic shock from rapid blood loss require rapid transfusion of blood products, initially given at 1:1:1 ratios of packed red blood cells, plasma, and platelets to reduce coagulopathy. Frequent measurements of hemoglobin, platelets, prothrombin time (PT), partial thromboplastin time (PTT), and fibrinogen will help guide initial therapy, while the use of TEG or ROTEM may provide further insight to the function of the coagulation cascade as a whole and provide accurate assessment of further product needs.
Patients who are deemed volume responsive but have normal blood counts and coagulation parameters and are without significant amounts of active bleeding are probably better served by resuscitation with crystalloids to avoid the risk of transfusion reactions, infections, and immunomodulation associated with blood products. As to the choice of crystalloid, recent evidence suggests that the use of lactated Ringer and Plasmalyte as opposed to normal saline results in the avoidance of a hyperchloremic metabolic acidosis and improved renal outcomes. The use of colloid-containing fluids in resuscitation continues to be a topic of debate. Although the use of colloids may result in the need for less fluid administration to achieve clinically significant results, they are much more expensive, have not shown consistent benefits in large clinical trials of critically ill patients, and have the theoretical risk of distributing oncotically active substances throughout the body where they act as a nidus for the development of tissue edema. If colloid resuscitation is chosen in the cardiac surgery population, the use of albumin is preferred over hetastarch because of associations with increased renal injury from the latter.
Alternative diagnostic and therapeutic options need to be considered if preload resuscitation is not successful in improving MAP. Vasodilation with associated, often severe, hypotension is a frequent complication occurring in up to 8% of patients after CPB and cardiac surgery. Although the underlying mechanisms of this vasodilatory shock remain elusive, low ejection fraction and angiotensin-converting enzyme inhibitor use have been identified as predisposing factors, and anesthetic drugs, peripheral rewarming, anemia, and variable degrees of a systemic inflammatory response to the extracorporeal circuit may also contribute. By directly measuring CO and CVP, PA catheters allow the calculation of SVR and a rapid diagnosis of arterial vasodilation as the cause of low MAP. Before aggressive measures to increase SVR are taken, it is important to ensure that a patient is volume replete and has an adequate cardiac contractility. Vasopressors may provide an important temporizing measure to maintain coronary perfusion pressure while these processes are being corrected but should then be reevaluated. Common first choices of agents include norepinephrine and vasopressin. The two have been directly compared as a first-line vasopressor in both the postcardiac surgical and septic populations. Although mortality and primary composite outcomes were similar between the agents, vasopressin resulted in less need for renal replacement therapy and decreased rates of atrial fibrillation. When high-dose norepinephrine (> 0.1 μg/kg/min) is required, arginine vasopressin has been proven useful in increasing BP and reducing norepinephrine requirements, especially in patients undergoing placement of a left ventricular assist device (LVAD). Phenylephrine is another option if a pure vasoconstrictor is needed, while epinephrine and dopamine are useful agents to provide both inotropy and SVR support. In the case of severe, refractory vasoplegia, methylene blue and hydroxocobalamin may be useful to reduce arteriolar nitric oxide levels. When resorting to these agents, be aware that methylene blue can cause hemolysis in patients with G6PD deficiency and serotonin syndrome in patients who take serotonergic medications. Hydroxocobalamin likewise interferes with laboratory tests that reply on colorimetric analysis and may cause blood-leak alarms in dialysis machines. Regardless of the choice of agent, when vasoconstrictors are used for a prolonged period, it is essential to avoid hypovolemia because severe peripheral hypoperfusion with gangrene may result.
Calcium increases BP without decreasing CO and has a negligible effect on HR. However, routine use of calcium to treat hypotension and low CO after cardiac surgery is controversial. Bolus administration of calcium may impair internal mammary artery (IMA) flow and potentially triggers vasospasm. Furthermore, it is undesirable for ionized calcium levels to be abnormally high, therefore repeated administration should be guided by blood levels. Low blood levels may result from citrate accumulation with rapid blood product transfusion and can reduce vascular responsiveness to catecholamines. Patients with a significant hepatic failure are especially vulnerable to hypocalcemia from product transfusion because of an inability to metabolize citrate rapidly. If ongoing blood product transfusion is needed, calcium levels should be checked frequently to maintain normal ionized calcium level.
Positive-pressure ventilation may produce relative hypovolemia, by eliminating the normal venous pressure gradient for filling the right atrium. Higher driving pressure and PEEP will both impede the flow of blood into the chest, especially in the presence of hypovolemia and consequently low filling pressures. This effect may be magnified in patients with COPD because of dynamic hyperinflation, where air trapping prevents full exhalation before the next breath is delivered by the ventilator. The presence of intrinsic positive end-expiratory pressure (auto-PEEP) is diagnostic of this condition. Adjustment of ventilator settings to increase expiratory time will help alleviate this problem. The development of a tension pneumothorax as the result of chest tube obstruction or occult intraoperative lung injury not drained by a chest tube (usually the right side) can also acutely compromise right heart filling. Assurance that breath sounds are present and symmetric on arrival in the ICU, as well as scrutiny of the initial postoperative chest radiograph, will identify this potential complication and guide the therapy (e.g., chest tube placement). Thoracic ultrasound is also very sensitive for the diagnosis of pneumothorax in the hands of a trained operator. It may allow diagnosis in an unstable patient faster than waiting for a chest radiograph.
Some patients present with postoperative hypertension instead of hypotension in the ICU. Systemic BP needs to be controlled to a MAP range of 70 to 80 mmHg because an excessive MAP may augment bleeding and create excessive afterload, subsequently myocardial contractility and compliance are compromised and myocardial oxygen demand increases. Postoperative hypertension may be caused by hypoxia, hypercarbia, pain, inadequate sedation, or shivering. Before therapy with vasodilating agents is initiated, these causes need to be ruled out. Nitroprusside and nitroglycerin are routinely used and alternative medications include nicardipine, labetalol, hydralazine, and esmolol. The vasodilation provided by propofol may also be helpful to modify BP rapidly in an intubated patient. Nitroprusside is widely used because it acts rapidly and can be easily titrated to effect and in response to sudden changes in preload and afterload. However, its use has been associated with the need for compensatory volume replacement, reflex tachycardia, cyanide toxicity, and methemoglobinemia. The hallmark of intravenous nicardipine is arterial specificity, which allows precise titration of BP without affecting the intravascular volume. This property is significant in perioperative hypertension, because arterial vasoconstriction with varying degrees of intravascular hypovolemia is a central characteristic of those patients. Nicardipine results in decreased mean arterial pressure and systemic vascular resistance, as well as an increase in CO, but it does not affect filling pressures.
The routine use of continuous or intermittent measurements of calculated hemodynamic indices is not necessary for low-risk patients undergoing elective CABG. When PA catheters are used in high-risk or complex cardiac surgical patients, the target cardiac index (CI, equal to CO per body surface area) is greater than 2.0 L/min/m 2 . Optimization of the CI is aimed at maximizing the pumping capacity of the cardiovascular system. A decreasing CI should be addressed before signs of overt hypoperfusion and tissue hypoxia develop, even if the baseline CI was low. Clinical measures of right and left ventricular performance, such as preload or volume status, afterload (SVR), HR, and myocardial contractility, are essential to guide therapy toward a stable hemodynamic status. Blood, crystalloids, or colloids may be infused to increase preload. CI can also be improved by optimizing afterload with a vasodilator such as nitroprusside. With an optimized preload and afterload, CI can sometimes be increased by atrial or atrioventricular (AV)-sequential pacing using epicardial pacing wires at a higher rate over the intrinsic HR. If preload and afterload are optimized, stroke volume (SV) would normally remain the same and a higher HR would increase the CI. If the target CI of greater than 2.0 L/min/m 2 is not achieved with optimization of preload, afterload, and HR, inotropic agents are usually required to improve cardiac contractility. There do not appear to be significant differences in patient outcome among the commonly used inotropes, but cardiac surgeons and intensivists often have personal preferences based on familiarity with the agent and clinical experience. One critical distinction to make in a hypotensive patient is between the relative intravascular hypovolemia from vasodilation during warming and the hypovolemia secondary to bleeding. Myocardial contractility may further diminish during the immediate postoperative period , mandating continued inotropic support or augmentation of support with mechanical assistance (e.g., intra-aortic balloon pump [IABP]).
An array of inotropic agents exists, including epinephrine, milrinone, dobutamine, and dopamine. Intimate knowledge of each drug’s inotropic, vasodilatory or vasoconstrictive, and chronotropic profile is warranted to achieve the desired effect. Combinations of inotropic and vasoconstrictive agents are selected when significantly reduced contractility is associated with severe hypotension. This is typical for phosphodiesterase inhibitors such as milrinone that often provide the wanted inotropic effect but not without significant vasodilation. Therefore, norepinephrine or vasopressin may be required in combination with milrinone to address the typical postoperative vasodilation that is related to the milrinone. Combining milrinone and epinephrine allows utilizing their different inotropic actions together with the vasoconstrictive effect of epinephrine for the treatment of hypotension. If the need for milrinone is anticipated in the operating room, it is ideally started during CPB to avoid a loading dose with its inherent hypotension. In addition to its inotropic effect, milrinone also provides lusitropy, which is beneficial in patients with severe diastolic dysfunction. Dopamine may be selected as an alternative choice agent either to increase CO or to allow weaning of epinephrine, but it might result in excessive tachycardia and increase the risk of atrial fibrillation. Dobutamine is generally associated with a positive chronotropic response and may be a good choice to increase CO if pulmonary artery pressures are elevated and baseline HR is low.
Modes of Mechanical Support
Mechanical circulatory support (MCS) is either initiated preemptively in the operating room or provided in the ICU when low CO persists despite adequate preload, optimal systemic vascular resistance, and maximal inotropic support. Forms of short-term MCS include IABPs, single-catheter microaxial pumps (Impella, [Abiomed Inc., Danvers, MA]), multicatheter extracorporeal pumps (TandemHeart [Cardiac Assist Inc., Pittsburgh, PA], CentriMag [Thoratec, Pleasanton, CA]), and venoarterial extracorporeal life support (VA ECLS). Long-term, “durable” devices include implantable LVADs and the total artificial heart (SynCardia [SynCardia Systems, Inc, Tucson, AZ]). A detailed description of these devices and their management is outside the scope of this text, but a brief summary is presented.
The decision to initiate MCS must be individualized for each patient. Persistent low CO and hypotension in conjunction with an elevated SVR, fluid nonresponsiveness, inadequate response to inotropes, and signs of tissue hypoperfusion should generally prompt consideration of MCS options. A potentially helpful tool is the cardiac power index (CPI), which normalizes cardiac power output to body surface area (BSA) with the following formula: CPI = (MAP × CI)/451. CPI has been found to be correlated with mortality in cardiogenic shock in multiple studies. Retrospective data suggest that values below 0.34 watts/m 2 at the time of device implantation are associated with increased 90-day mortality. CPI below that level should thus suggest rapid initiation of mechanical support.
Once the decision to undertake MCS is made, the specific device is chosen in consultation with surgical and interventional cardiology teams. If right heart function is deemed adequate, commonly used left-sided devices are the IABP and Impella. An IABP increases coronary perfusion pressure by inflation during diastole and also reduces afterload by deflation during systole. The resulting improvement in myocardial oxygen supply and decrease in demand has the potential to increase CO significantly. However, the device requires sinus rhythm without marked tachycardia for proper inflation timing at aortic valve closure, as well as a patent aortic valve (no more than mild aortic regurgitation present). It also provides very little direct hemodynamic support. Impellas, on the other hand, pump blood directly from the left ventricle (LV) into the ascending aorta and do so regardless of HR or rhythm. Current evidence does not support a mortality benefit of one device versus the other in undifferentiated cardiogenic shock. If needed, higher flow rates may be achieved through the use of extracorporeal centrifugal devices that utilize separate inflow and outflow cannulas such as the CentriMag and TandemHeart. These devices may also be configured to provide right ventricular support (CentriMag) or offer full biventricular support (TandemHeart).
If biventricular failure is present, temporary MCS is best provided by VA ECLS. Venous and arterial cannulas may be placed centrally in the operating room or peripherally at the bedside. By removing blood from the right atria, pumping it through an oxygenator, and reinjecting it into the arterial system (the aorta in central cannulation; usually the femoral artery in peripheral cannulation), VA ECLS allows temporary, near-total unloading of a failing cardiopulmonary system. Like all other continuous flow support devices, adequate flows are highly preload dependent and afterload sensitive. Common complications include bleeding, thrombosis, and limb ischemia. VA ECLS may also lead to LV distension, with corresponding LV thrombus formation and wall ischemia, if there is lack of LV unloading and native contractility remains poor. Options for LV unloading include Impella, IABP, atrial septostomy, or direct surgical vent via apex or left atrium. Peripheral cannulation has the further complication of differential hypoxia, where poorly oxygenated blood from the lungs is ejected out of a recovering ventricle and competes with well-oxygenated blood from the femoral cannula, potentially causing hypoxic conditions in the right side of the aortic arch. Detection is facilitated by monitoring pulse oximetry and arterial blood gas from the right upper extremity.
Weaning from temporary support devices in the post-cardiac surgery setting is facilitated by transesophageal echocardiography (TEE). Device flow can be gradually decreased while paying careful attention to ventricular function and hemodynamic values. Be aware that most devices have minimum recommended flow settings to prevent thrombosis and flows should not be decreased below such levels unless immediate decannulation is planned.
Although the bedside intensivist is unlikely to be involved in the decision to place long-term, durable support devices, a knowledge of their function, management, and common complications is necessary during their postimplantation ICU stay. LVADs draw blood from an inflow cannula at the LV apex and reinject into an aortic outflow cannula. They may be placed as a bridge to heart transplant (BTT), bridge to recovery, or destination therapy (DT). While first generation LVADs produced pulsatile flow, modern third generation devices produce continuous flow via a centrifugal pump utilizing a magnetically suspended impeller. The HeartMate 3 (Abbott, Abbott Park, IL) and Heartware HVAD (Medtronic, HeartWare, Miami Lakes, FL) are the two centrifugal devices currently on the market. The HeartMate II is a second generation, nonpulsatile, axial flow device, which until recently was the preferred device for DT. Currently, HeartMate 3, HVAD and HeartMate II are all devices approved by the Food and Drug Administration for both BTT and DT. One notable complication of LVAD placement is right heart failure, which may occur as a result of both altered morphology and contractility from intraventricular septal shifting, as well as because of increased right ventricular preload secondary to an increase in CO and thus increased venous return. Careful attention should be given to CVP, PA pressure, and signs of venous congestion in the immediate postimplantation period. Treatment is through inotropes, diuresis, pulmonary vasodilators, and reduction of pump speed to minimize septal shift. Other complications include device thrombosis, infection, and hemolysis.
Patients undergoing cardiac surgery with CPB frequently have a variety of derangements of hemostasis (with striking interpatient variability poorly explained by clinical, procedural, biologic, and genetic markers) that frequently result in transfusion of allogeneic blood products. More than 12 million units of allogeneic red blood cells (RBCs) are administered in the United States annually, with more than 2 million units administered to patients undergoing cardiovascular surgery. Institutions vary significantly in perioperative blood conservation and transfusion practices for cardiac surgery patients. Depending on the institution, between 27% and 92% of CABG patients are transfused, and the number of units of packed cells received ranges from 0 to 4. Patients undergoing repeat sternotomy for cardiac surgery are approximately three times more likely to receive a perioperative transfusion than those undergoing primary cardiac surgery. Transfusion-related complications have decreased significantly in the last 10 to 15 years but still remain a concern.
The transfusion of allogeneic RBCs has recently been described as a risk factor for decreased long-term survival after CABG surgery. Another recent study showed that the storage duration of perioperatively transfused RBCs is associated with an increased risk of both short-term in-hospital and long-term out-of-hospital mortality, independent of the number of transfusions administered and other confounding factors. Other risks include hepatitis C infection (approximately 1 in 50,000), transmission of the human immunodeficiency virus (less than 1 in 500,000), major ABO group incompatibility (less than 1 in 33,000), and minor transfusion reactions (approximately 1 in 5).
According to the American Society of Anesthesiologists (ASA), the cost of transfusion therapy approaches $5 to $7 billion per year in the United States, with up to 25% of RBC transfusions judged to be unnecessary. The ASA strongly advocates judicious use of blood products and has published a set of practice guidelines. Although practice guidelines are imperfect and must not replace clinical judgment, the ASA guidelines provide a scientifically based model to assist decision making. However, in the cardiac surgical patient population, the ASA guidelines fail to account for platelet dysfunction known to occur after CPB. When major bleeding occurs in this patient group, platelet transfusion is indicated even when platelet count is greater than 100,000.
Unfortunately, practice guidelines such as those from the ASA and the College of American Pathologists do not seem to change old transfusion habits and inappropriate use of blood products has not been curtailed, especially in the cardiac surgical population. One possible reason for lack of success is that the time delay until coagulation results return from the laboratory makes directed transfusion therapy difficult in a bleeding patient. In response to this problem, several rapid assay devices are being developed. When combined with goal-directed transfusion algorithms, such point-of-care tests may improve care and limit unnecessary transfusions. These algorithms usually incorporate some measure of platelet function, coagulation factor activity, and fibrinolysis as the stimulus for therapy. Although not perfect, thromboelastography appears to be a key component of such algorithms. Because its negative predictive value is greater than 90%, a normal TEG implies with greater than 90% certainty that a bleeding patient needs to return to the operating room and a surgical bleeder will be found. Incorporation of TEG produced a significantly reduced rate of reoperation for bleeding in a single-center study.
Other strategies for limiting blood product requirements show some benefit in clinical trials. Antifibrinolytic therapy decreases the incidence of excessive postoperative bleeding caused by fibrinolysis. Effective antifibrinolytic agents include ɛ-aminocaproic acid (Amicar [numerous sources]), tranexamic acid, and aprotinin. Aprotinin is the only agent with class A level 1 evidence for reduction in rates of transfusion and return to operating room to control bleeding after heart surgery. However, because of its increased risk for thrombosis and renal dysfunction, aprotinin was withdrawn from the market in 2007. Amicar and tranexamic acid have been shown to reduce total blood loss and decrease the number of patients who require blood transfusion during cardiac surgery. In the 2011 update to The Society of Thoracic Surgeons and the Society of Cardiovascular Anesthesiologists Blood Conservation Clinical Practice Guidelines, antifibrinolytic agents were strongly recommended (level of evidence A) for blood conservation.
Intraoperative autologous hemodilution has also demonstrated a blood-sparing effect in some studies, but postoperative use of cell salvage has not proven to be effective in limiting the use of blood products in cardiac surgery. Reinfusion of shed mediastinal blood may be associated with a greater frequency of wound infection and is not recommended.
In current postoperative practice, hemostasis is assessed by measuring the amount of blood draining out of the chest tubes. Prothrombin time and activated partial thromboplastin time are normally slightly elevated after cardiac surgery. An activated clotting time may also be used as a point-of-care test for adequate reversal of heparin given during surgery. As outlined earlier, the incorporation of TEG as a point-of-care monitor into a specific transfusion algorithm allows a more specific diagnosis of bleeding problems and a reduction of indiscriminate transfusion practices. Fresh frozen plasma (FFP) is not required when there is no significant drainage from the chest tubes; if drainage from the chest tube exceeds 400 mL in the first 2 hours and 100 to 150 mL/h after the first 2 hours, FFP is probably required to stop the bleeding. Infusion of 10 mL/kg of FFP normally restores the coagulation factors to an adequate level. It is also important to consider platelet transfusion in a bleeding patient after CPB even if the platelet count is greater than 100,000/μL, because platelet function might be impaired as a result of exposure to CPB. Cryoprecipitate is added when the fibrinogen level is low.
In cases of severe ongoing bleeding despite transfusion of blood products, prothrombin complex concentrate (PCC) can be considered. The PCC contains relatively high levels of factors II, IX, and X, and in some preparations, inactive or activated form of factor VII. Compared with plasma, PCC is administered in much smaller volume, is free of transfusion reactions associated with plasma, and acts faster. PCC was originally approved for emergency reversal of warfarin or novel oral anticoagulants (NOAC) therapy when antidotes are unavailable. The off-label use of PCC for microvascular bleeding attributed to coagulation factor deficiencies has recently shifted from rescue therapy to earlier and even first-line therapy in some centers. Currently there is insufficient evidence to recommend a universal transition toward the use of PCCs in lieu of plasma therapy for microvascular bleeding associated with cardiac surgery. When used, the inactive PCC (Kcentra; CSL Behring, King of Prussia, PA) is recommended over activated factor concentrates such as recombinant activated factor VII (rFVIIa, NovoSeven) or anti-inhibitor coagulant complex (FEIBA, Shire, Dublin, Ireland) because of concern for thromboembolic complications associated with the latter.
The chest tubes need to be evaluated on a regular basis to avoid clogging by blood clots. When there is a sudden decrease in drainage from the chest tubes, the intensivist should exclude the possibility of concealed bleeding in the mediastinum and pericardial sac, which can lead to cardiac tamponade. If the response to blood transfusion is not satisfactory, exploratory sternotomy should be seriously considered.
Transient alterations in HR or conduction may contribute to postoperative hypotension or low CO. When BP or CO is marginal, augmenting the HR should be considered, even if it is already in the normal range. Abnormal patterns within the first 24 hours include transient bradyarrhythmia, sinus tachycardia (> 110 beats/min), and, in cases of valvular surgery, junctional tachycardia with atrioventricular interference or even heart block. Significant ventricular dysrhythmias are rare but if present may present a challenge to treat.
Bradyarrhythmia is common after cardiac surgery, and atrial and ventricular epicardial pacing wires are typically placed to facilitate temporary pacing in the immediate postoperative period. Optimal pacing modalities have been investigated. Dual-chamber pacing maximizes CO over a wide range of AV delay (100 to 225 milliseconds). Conditions impairing diastolic filling (ventricular hypertrophy, cardiomyopathy, fibrosis) may benefit by extending the AV delay.
The rate of pacemaker dependency after cardiac surgery varies between studies and is approximately 1% after CABG, 2% after primary valve replacement, 7% after repeat valve replacement, and 10% after orthotopic heart transplant (OHT). Identified risk factors include annular calcification, older age, preoperative left bundle branch block, and increased CPB time. The etiology of bradyarrhythmia includes ischemia, edema, and irreversible surgical destruction of the conducting system. Recovery from the reversible causes can be considerably delayed. Permanent pacemakers are typically implanted for symptomatic sinus node dysfunction or AV block lasting beyond the fifth postoperative day. In the long term, up to 40% of patients with sinus node dysfunction and up to 100% of patients with complete AV block past day 5 remain pacemaker dependent.
For heart transplant patients, chronotropic medication may avoid the need for a permanent pacemaker. Sinus node dysfunction is five times more likely than AV block to result in permanent pacemaker implantation in this group. Both types of dysfunction can be significantly reduced with bicaval rather than biatrial anastomoses.
Atrial dysrhythmias (primarily atrial fibrillation [AF]) are by far the most common complication after cardiac surgery, with an incidence consistently reported to range between 27% and 40% and with little change over the past 2 decades. AF most commonly occurs 2 to 3 days after surgery and can increase hospital stay, risk of stroke, risk of neurocognitive deficits, and mortality. Those who have undergone valvular surgery are at greatest risk. Patients with preexisting AF may return from the operating room in sinus rhythm and soon revert back to AF. If an atrial epicardial lead is in place, an atrial electrocardiogram can be conducted when the rhythm or conduction cannot otherwise be diagnosed. Use of adenosine is another option for diagnosis and possible therapy when the atrial arrhythmia is not AF. The exact pathogenesis of postoperative AF is unclear. Structural changes in the atria, reduced threshold for dysrhythmia generation, and a hyperadrenergic state after CPB and surgery are all suggested as mechanisms. Associated risk factors for the development of AF include advanced age, history of atrial fibrillation or chronic obstructive pulmonary disease, prolonged CPB and aortic cross-clamp times, left atrial enlargement, cardiomegaly, valve surgery, and postoperative withdrawal of a beta blocker or an angiotensin-converting enzyme (ACE) inhibitor.
Pharmacologic prophylaxis against AF is controversial. Prophylactic therapy includes supplemental magnesium, potassium, and use of beta blockade in the perioperative period to reduce the incidence of AF. The American College of Cardiology and American Heart Association task force consensus recommendations for prevention and management of AF are summarized in Box 31.2 . In the largest trial of amiodarone in patients undergoing CABG surgery or valve replacement or repair surgery reported to date, prophylactic amiodarone was shown to reduce effectively the incidence of postoperative AF. Management of postoperative supraventricular tachycardia (SVT) depends on the patient’s clinical condition. If appropriate, a 12-lead ECG and an atrial rhythm strip via the atrial pacing electrodes aid diagnosis of the exact rhythm. Postoperative therapy of AF includes the correction of electrolyte abnormalities, empirical administration of magnesium and potassium, and amiodarone or beta blockade. If pharmacologic restoration of sinus rhythm fails or if hemodynamic instability merits immediate synchronized direct current (DC) cardioversion, electrical cardioversion is attempted with an initial energy of 200 J for AF and 50 to 100 J for atrial flutter. Restoration of sinus rhythm rather than simple rate control is the ultimate goal. After AF lasts more than 48 hours (sustained or paroxysmal AF), or AF of an unknown duration, it is reasonable to initiate anticoagulation despite the recent cardiac surgery.