Chapter 54: Postcardiothoracic Surgery Care

The care of patients after cardiac surgery, particularly in the immediate postoperative period, frequently involves a period of physiologic volatility as the body adapts to the cardiac intervention and recovers from the effects of cardiopulmonary bypass (CPB) and anesthesia. The successful management of this patient population can be facilitated by addressing common issues such as fast-track eligibility and extubation, hypotension and low cardiac output (CO), postoperative bleeding, dysrhythmias, renal function, and glucose control. This chapter aims to provide an overview of these topics as well as a framework for approaching problems in these areas.

When the postcardiac surgery patient arrives in the intensive care unit (ICU), the optimal transition of care is achieved by direct communication between the ICU team and the surgeons and anesthesiologists involved in the case. If a coronary artery bypass graft (CABG) was performed, the report from the surgeons should include the number of bypasses, the locations of the grafts, and whether the quality of the targets were good, as this has implications for the degree of protection achieved against future myocardial ischemia. Patients very commonly will have had coronary stents placed prior to their CABG. If the grafts did not bypass these stents, then antiplatelet agents such as clopidogrel may need to reinstated as soon as hemostasis is assured. When radial artery or bilateral internal mammary grafts are utilized, some institutions utilize low-dose nitroglycerin or nicardipine to protect against vasospasm, and the need for these agents should be communicated to the ICU team. If a valvular procedure has been performed, the differentiation the surgeon's report should include the valve of interest, whether it was repaired or replaced, and it was replaced, the type of valve that was implanted, that is, bioprosthetic or mechanical. Mechanical valves necessitate anticoagulation once hemostasis is achieved, and a plan for implementation should be made clear. The surgeon should communicate the need for any particular blood pressure goals. A lower blood pressure may be requested in situations where the risk of bleeding is higher than usual, such as when a friable aorta was encountered during the procedure. A higher blood pressure goal may be needed in patients with chronic hypertension or in those with known vascular stenoses of the carotid or renal vasculature. Finally, a plan for devices such as intra-aortic balloon pumps should be communicated.

The transition of care should also include a thorough report from the anesthesiologist. The difficulty of mask ventilation and intubation should be communicated as this has implications for whether a patient might be safely extubated later on. If the intubation was difficult, the details on how intubation was achieved should be obtained. The report should also comment on the difficulty or ease of line placement. Information about CPB should be communicated as well, including prebypass stability, duration of bypass, duration of circulatory arrest, difficulty in arresting the heart and difficulty of weaning off bypass including the number of attempts and the vasoactive medications required for separation. Knowledge of the findings on transthoracic echocardiography are paramount for postoperative management in the ICU including left and right ventricular functions, thickness of the ventricle, and the presence of any valvular abnormalities. Other details such as rhythm issues and the requirement for pacing as well as baseline hemodynamic measurements such as pulmonary artery pressures and cardiac index (CI) should be obtained. Finally, the number and type of blood products should be communicated to the ICU team.

Once report has been obtained from the surgical and anesthetic teams, the decision of whether to fast track the patient should be made. Fast-track care or fast tracking involves a strategy of early extubation with the aim of promoting an expedient discharge for the purpose of reducing the length of stay in the ICU and hospital and to realize cost savings. It typically starts as an anesthetic strategy of decreased opioid dosage (eg, intraoperative fentanyl usage of less than 20 micrograms/kg vs greater than 20 micrograms/kg) to promote timely awakening and return of respiratory drive, and is executed as an ICU team strategy of achieving extubation within a predetermined timeframe, often with the help of a time-directed extubation protocol, typically within 8 hours of arrival to the unit. One meta-analysis found that fast tracking resulted in a significant mean reduction in time to tracheal extubation of 8.1 hours as well as achievement of a statistically significant decrease in overall ICU length of stay of approximately 5 hours, though not in overall hospital length of stay,¹ a finding echoed by a Cochrane Review on fast-track cardiac care (decreased time to extubation 3.0 to 10.5 hours, decreased length of stay in ICU 0.4-8.7 hours).² An earlier trial by Cheng et al found that fast tracking (intraoperative fentanyl 15 g/kg vs 50 g/kg) led to a decrease in time to extubation by approximately 15 hours and a decrease in ICU length of stay of approximately 5 days; in this context, use of fewer higher cost ICU days resulted in a cost savings of per CABG of 25% as well as a decrease in the rate of case cancellations (0.3% vs 2%) due to improved ICU throughput.³ Fast tracking appears to be safe without evidence of increased mortality or postoperative complications such as myocardial infarction, reintubation, acute renal failure, major bleeding, and stroke.^1,2

Patient selection for fast tracking must be considered. For example, Cheng et al excluded patients who were older than 75 years of age and who experienced a recent myocardial infarction (within 3 weeks), inotropic therapy within 24 hours before the study, or had an intra-aortic balloon pump as well as patients with severe hepatic disease, significant renal insufficiency, or severe chronic obstructive pulmonary disease (COPD), among other criteria. Every postcardiac patient should be evaluated for eligibility for fast tracking, but the decision to proceed with this strategy should be assessed on a case-by-case basis and may be modified by patient comorbidities or situational factors.

Delays in extubation may be warranted with certain comorbidities. For example, the patient with a low ejection fraction (EF) due to left ventricular dysfunction should be carefully assessed for signs of low CO prior to liberation from the ventilator. CPB induces a temporary deterioration of myocardial contractility^4,5,6 (see section “Hypotension and Low Cardiac Output”), which may significantly impact a patient with low myocardial reserve. Extubation during this time period, with loss of left ventricular afterload reduction due to withdrawal of positive pressure ventilation, may further compromise a tenuous state of contractility.^7,8,9 Patients who maintain an adequate CO despite a starting low EF and demonstrate signs of adequate perfusion should be extubated when ready. However, those with myocardial dysfunction resulting in hemodynamic instability with a rising vasopressor requirement, especially with concurrent acidosis, low mixed venous saturation (Svo₂) and oliguria should be kept intubated until stabilized with therapies such as initiation or escalation of inotropic therapy or placement of an intra-aortic balloon pump. Another example is the patient with end-stage renal disease who is dependent on dialysis who may require renal replacement therapy in the immediate postoperative period. In this case, extubation may be better delayed after completion of a therapy that will induce fluid shifts in the early postoperative period.

Situational factors such as bleeding, hemodynamic instability, delayed emergence, difficult intubation, hypothermia, or intraoperative complications may also prevent fast tracking. The bleeding patient with copious output from chest tubes may require a return to the operating room for achievement of surgical hemostasis. The hemodynamically unstable patient with a rising pressor requirement will require assessment for and treatment of hypovolemia or bleeding, tamponade, loss of myocardial contractility, or sepsis. Delayed emergence despite lower doses of narcotic utilization during the operation due to pharmacologic sensitivity may prevent the patient from achieving extubation criteria by 8 hours due to excessive sleepiness and apnea. Although a difficult intubation in and of itself should not be a deterrent to fast tracking as long as extubation criteria are achieved, a difficult airway resulting in trauma due to multiple intubation attempts may warrant a delay in removal of the endotracheal tube in order to institute measures to decrease airway edema (eg, elevation of the head of the bed and steroids), especially if the patient does not demonstrate an air leak with the endotracheal tube balloon deflated. In addition, the patient with the difficult airway with marginal achievement of extubation criteria may benefit from resources that may be more easily available during the daytime rather than the nighttime such as advanced airway equipment and the immediate availability of an anesthesiologist. Hypothermia (core temperature < 36°C) promotes arrhythmias, impairs coagulation, and may cause shivering, which increases oxygen consumption and myocardial oxygen demand and should be rectified prior to extubation. Finally, intraoperative events such as a prolonged (> 2.5 hours) bypass run, bleeding requiring massive transfusion, or difficulty in repairing the aortic cannulation site may merit exemption from fast tracking. In the latter case, avoidance of hypertension both pharmacologically and by preventing rapid awakening with immediate sedation on arrival to the ICU may be requested to decrease the probability of dehiscence of the repair and resultant surgical bleeding.

In general, regardless of whether the patient qualifies for the fast-track strategy, sedatives and analgesics should be held after arrival in the ICU until the patient awakens. Documentation of emergence from the anesthetic with movement of all extremities is an important milestone in the postoperative course, as postoperative stroke may manifest as delayed emergence or hemiplegia. If the patient is to be fast tracked, subsequent administration of sedatives and analgesics should be done judiciously to keep the patient comfortable while intubated but to avoid oversedation and respiratory depression that might delay extubation. In those patients who arrive to the ICU awake or awaken shortly thereafter, a number of agents are available for sedation and analgesia (Table 54–1).^10,11 Dexmedetomidine, an α₂ agonist that decreases sympathetic outflow, produces sedation without respiratory depression and is useful in the patient who expected to be fast tracked. Its cost as well as the potential for inducing bradycardia and resultant hypotension are limiting factors. When dexmedetomidine is not available, propofol provides rapid onset and offset of sedation at low doses; at higher doses propofol will induce respiratory depression. Benzodiazepines such as midazolam are slower in offset and are less favored in the patient expected to be extubated. However, they may provide greater stability in the hemodynamically tenuous patient. The patient in pain may benefit from an array of available analgesics. Intravenous acetaminophen may provide pain relief without concern for inducing apnea. Fentanyl, a μ opioid agonist, provides rapid pain relief with limited duration. Longer acting analgesics such as hydromorphone and morphine should be given judiciously to prevent respiratory depression that may delay extubation. Morphine should not be dosed in patients with significant kidney dysfunction, as the active metabolite morphine-6-glucuronide has additive sedative effects. Certain patients may be candidates for IV ketorolac, a potent nonsteroidal anti-inflammatory drug (NSAID). However, this analgesic should be avoided in patients with renal insufficiency, a history of gastric ulcers, severe or aspirin-sensitive asthma, and in those in whom bleeding is a concern. Lidocaine patches may give relief to the patient who is experiencing great discomfort from chest tubes. The patient who is shivering may benefit from meperidine, although this agent should not be given to the patient on monoamine oxidase inhibitors (MAOIs) or selective serotonin-reuptake inhibitors (SSRIs). Together, these agents should keep the patient comfortable while intubated; ongoing weaning should occur concurrent to completion of an appropriate postoperative evaluation (electrocardiogram, chest X-ray, laboratory values, and continual evaluation of hemodynamics as well as chest tube and urine outputs).

In general, extubation of the postcardiac surgery patient can be considered once the patient is warm (temperature > 36.0°C) and has demonstrated hemodynamic stability with regards to heart rate (HR), CI (ideally > 2.2 L/min/m²), and blood pressure. Inotropic and vasopressor requirements should be stable or decreasing and parameters associated with adequate perfusion such as the Svo₂, lactate, and urine output should be acceptable. In addition, chest tube outputs should not be excessive (consider a threshold of > 200 cc/h). As the postoperative evaluation is being completed, the patient should be actively weaned to a fraction of inspired oxygen (Fio₂) 40% and positive end-expiratory pressure (PEEP) 5 cm H₂O while maintaining a peripheral capillary oxygen saturation (Spo₂) greater than 92%. As the spontaneous respiratory drive returns, pressure support ventilation should be instituted and weaned to 5 cm H₂O. The final assessment of extubation should be performed on minimal ventilator settings with evaluation of the following parameters: oxygenation, ventilation and strength, and airway protection (Table 54–2).

Common issues such as atelectasis and pain control may provide barriers to timely extubation. Some patients become atelectatic during the bypass run and develop an alveolar to arterial (A-a) gradient that precludes full weaning of Fio₂. As hemodynamics allow, recruitment maneuvers (ie, PEEP 20 cm H₂O for 20 s) can open up collapsed alveolar units and improve oxygenation. Pain control can pose a challenge, as well. Some patients find inadequate pain relief with opioids and as a result, may splint and breathe shallowly on the ventilator. Other patients demonstrate dose-limiting sleepiness and apnea only to wake up in persistent pain. To those who qualify, adjuncts such as intravenous acetaminophen or ketorolac can be beneficial.

The goal mean arterial pressure (MAP) is typically set as greater than 65 mm Hg. A higher goal may be needed with chronic hypertensives in whom the autoregulatory curve of the brain and kidney have shifted to accommodate long-standing elevated systemic pressures. In addition, patients with known vascular stenoses of the carotid or renal arteries, as well as those with unrepaired coronary artery disease may require a higher perfusion pressure to overcome these potentially flow-limiting lesions. A lower MAP may be required for a time in patients with a friable aorta to decrease the risk of bleeding.

MAP is defined as MAP = CO × SVR, where SVR denotes systemic vascular resistance. As CO itself is defined as HR times stroke volume (SV), the equation may be further defined as: MAP = HR × SV × SVR. Thus, hypotension may be conceptualized as being affected by 1 of these 3 major variables.

Heart Rate

HR may be a factor in hypotension. The process of cooling and arresting of the myocardium during cardiac surgery renders the muscle less compliant and induces diastolic dysfunction.¹² This resultant increase in the “stiffness” of the ventricles can limit the stroke volume and render the CO more dependent on HR. Thus, a HR that is too slow, even if not strictly bradycardic, may limit CO and decrease MAP. The responsivity of the blood pressure to pacing with epicardial wires to a HR of 90 to 100 beats/minute may be tested. If both atrial and ventricular wires are present, pacing can be accomplished with maintenance of atrial kick. However, if only ventricular wires are available, the hemodynamic benefit of increasing HR with ventricular pacing may be diminished by loss of contribution to stroke volume from loss of atrial kick. In fact, an increased pressor requirement may be observed in patients who lose atrioventricular (AV) synchrony and the atrial contribution to SV due to the development of an arrhythmia such as atrial fibrillation.

Stroke Volume

SV is affected by preload, contractility, and afterload. Hypotension, especially in the patient with a labile blood pressure, may be due to decreased preload and low SV from hypovolemia. Low preload may be due to an obvious process such as chest tube drainage or to a more occult occurrence such as retroperitoneal bleeding in the patient who has undergone femoral arterial cannulation. The volume responsiveness of the patient can be tested empirically by 250 to 500 cc fluid boluses with subsequent assessment of the effect upon blood pressure. Both normal saline and lactated ringers may be used for resuscitation although 5% albumin is frequently utilized with the intention of administering an agent with a greater intravascular half-life. Transfusion with packed red blood cells may be indicated in the setting of significant anemia, especially with a concurrent vasopressor requirement. The data on hydroxyethyl starch (HES) agents discourage its use in the cardiac surgical ICU. One study randomized 45 patients to receive either 15 mL/kg HES solution or 4% albumin after admission to the cardiac ICU and found that the HES group demonstrated thromboelastometry tracings indicative of impaired fibrin formation and clot strength.¹³ Although the 2 groups demonstrated similar chest tube outputs, a subsequent meta-analysis of 18 trials comparing HES solutions with albumin for fluid management in CPB surgery found that HES solutions increased blood loss, reoperation for bleeding, and the rate of blood transfusion.¹⁴ Given these results, it seems prudent to avoid the use of HES solutions for resuscitation in the postcardiac surgery patient. Hypotension from decreased preload and low SV may be secondary to mechanical obstructive process rather than hypovolemia. Conditions such as tamponade and tension pneumothorax prevent adequate venous return, depress SV and CO, and can result in hypotension and shock with rising pressor requirement and lactate and oliguria. Tamponade is further discussed under section “Bleeding.”

Hypotension may also be secondary to low SV from decreased contractility. A temporary decrease in myocardial function is to be expected in the postcardiac surgery patient.^4,5,6 One study examining 24 patients undergoing elective coronary artery bypass surgery with serial hemodynamic measurements and radionuclide evaluation of ventricular function found that 96% experienced left- and right-sided myocardial depression postprocedure. The left-sided EF dropped from 58% preoperative to 37% with the nadir occurring 4.4 hours after coronary bypass. Recovery back to 55% occurred approximately 7.1 hours after coronary bypass.⁶ An earlier study of 22 patients undergoing coronary artery bypass surgery found that those with an EF of more than 55% recovered ventricular function within 4 hours of bypass but that those with an EF more than 45% demonstrated a more profound decrease in function and a delayed recovery to baseline even after 24 hours.⁵ Thus, the immediate postoperative cardiac patient may demonstrate a temporary inotropic or vasopressor requirement until myocardial contractility has fully recovered. However, a high index of suspicion for differentiating between this known and expected decrease in function and concurrent detrimental processes should be maintained. Other independent causes of decreased myocardial contractility include new onset myocardial infarction from unrepaired disease or from CABG thrombosis or vasospasm. Diagnosis of a new ischemic insult may be accomplished with an electrocardiogram or with echocardiography demonstrating a new wall motion abnormality. In addition to hemodynamic instability, the onset of new arrhythmias such as ventricular tachycardia (VT) may be a manifestation of myocardial supply-demand mismatch. Troponins are of limited utility in detecting ischemia as they are typically elevated postcardiac surgery; however, some advocate trending the values. Management may include the administration of nitroglycerin to vasodilate the coronary arteries and to increase coronary perfusion pressure by decreasing preload and hence left ventricular end-diastolic pressure, cardiac catheterization to remove thrombus, or reexploration in the operating room to visually inspect the grafts.

Systemic Vascular Resistance

Hypotension may also be caused by insufficient systemic vascular resistance. Some patients arrive to the ICU hypothermic despite adequate rewarming on CPB due to a phenomenon called afterdrop—a redistribution hypothermia secondary to a cool periphery. These patients require active rewarming, which results in vasodilation of the peripheral vasculature and can uncover or worsen a vasopressor requirement. CPB causes a systemic inflammatory response that can lead to decreased SVR and a form of postbypass hypotension known as vasodilatory shock,^15,16 generally described as a MAP less than 70 mm Hg in the setting of an adequate CI of more than 2.5 L/min/m². Increased levels of vasodilators such as nitric oxide and bradykinin have been implicated in the pathogenesis of this phenomenon. The increase in bradykinin may be related to lack of deactivation by angiotensin converting enzyme (ACE) due to decreased blood flow to the lungs during CPB.^17,18 Vasodilatory shock may also be caused by a relative vasopressin deficiency.^19,20 One prospective study of 145 patients measured vasopressin levels 5 minutes after weaning from CPB. Those with vasodilatory shock demonstrated lower levels of vasopressin than those with postbypass hypotension due to decreased CI, and infusion with exogenous vasopressin improved MAP.¹⁹ This study found the incidence of vasodilatory shock to be 8%, although the incidence was higher (27%) in those with an EF of less than 35%; in addition, preoperative ACE-inhibitors, commonly part of a heart failure regimen, were found to be a risk factor for the development of vasodilatory shock. A retrospective study of 2823 patients found a 20% rate of vasoplegia; clinical risk factors included observation of a clinically significant decline in MAP at the onset of CPB, length of bypass, and preoperative use of β blockers and ACE-inhibitors among others.²¹

Another way to conceptualize the hemodynamic changes of the postoperative period, especially in patients with valvular lesions, is to consider whether the myocardium is pressure-overloaded or volume-overloaded. Lesions such as aortic stenosis give rise to pressure-overloaded ventricles. In this situation, chronic obstruction to left ventricular ejection results in concentric hypertrophy as a compensatory response to increased wall stress. The increase in ventricular wall thickness results in a ventricle with decreased diastolic compliance or greater “stiffness.” As a result, the resultant limitation in SV can result in a decrease in CO with lower HRs. Conversely, excessively high HRs may limit diastolic filling time, SV, and may also compromise CO. The stiffness of the ventricle impedes filling and makes the myocardium more dependent on atrial kick for stroke volume. Hence, the development of a nonsinus rhythm such as atrial fibrillation may result in a decrease in SV that decreases CO and MAP. The pressure-overloaded ventricle is more prone to ischemia than the normal ventricle. The increase in muscle mass increases the basal metabolic rate of the entire ventricle. In addition, the increase in stiffness of the pressure-overloaded myocardium results in an increase in the left ventricular end-diastolic pressure (LVEDP). This results in a decrease in the gradient for coronary perfusion pressure (CPP = aortic diastolic pressure [AoDP] – LVEDP/coronary vascular resistance) rendering the myocardium more vulnerable ischemia with decreases in blood pressure. Tachycardia is particularly detrimental not only because it limits SV by decreasing diastolic filling time but also because it reduces the duration of diastolic coronary perfusion and increases myocardial oxygen demand. The hemodynamic considerations in the pressure-overloaded ventricle are summarized in Table 54–3.

Although hypotension is to be carefully avoided in the pressure-overloaded myocardium, patients with preserved LV function often present to the ICU hypertensive, as the left ventricle has not yet had time to adapt to the removal of the stenotic aortic lesion. In this case, institution of a vasodilator such as nicardipine can control blood pressure in the acute setting. The patient with decompensated aortic stenosis may have decreased contractility and may have presented with decompensated heart failure.

Lesions such as aortic and mitral regurgitation result in volume-overloaded ventricles. Chronic volume overload results in eccentric hypertrophy where overall muscle mass is increased but chamber enlargement is greater than the increase in ventricular wall thickness. This gradual increase in LV size and compliance usually causes little change in the LVEDP. Chronic volume overload may result in a dilated cardiomyopathy with decreased LV systolic function and may require the use of inotropes to separate from bypass and their continued use in the ICU until myocardial stunning resolves. Decreased CO after correction of mitral regurgitation may also be secondary to decreased volume loading of the left ventricle from the resolution of regurgitation. Reports of the ejection prior to and after resolution of mitral regurgitation may be somewhat misleading. For example, a preoperative EF of 55% with a postoperative EF of 30% may not reflect any actual change in the contractility. The estimated EF prior to surgery reflects a SV that includes both forward flow and retrograde flow into the left atrium, and the effective EF taking only forward flow into account may have been 30% from the start. Because left ventricular afterload increases acutely after correction of mitral regurgitation, the left ventricle may benefit from pharmacologic afterload reduction. Chronic volume overload differs greatly from the volume overload resulting from acute aortic or mitral insufficiency where LVEDP increases dramatically and can result in sudden pulmonary edema and decreased CO prior to surgical correction.

Mitral regurgitation and stenosis may result in left atrial hypertension and dilation that can predispose to atrial fibrillation. In addition, pulmonary hypertension secondary to chronic venous congestion may be present with concomitant right ventricular dysfunction due to long-standing increases in RV afterload.

The hemodynamic problem of low CO (CO = HR × SV) may be broken down by examining the HR as well as the factors affecting SV (preload, afterload, and contractility). In the situation where low CO is accompanied by bradycardia, the CO may be augmented by increasing the HR. This can be accomplished either by electrical pacing with an external epicardial wire, reprogramming the rate of an internal pacemaker, or increasing the HR chemically with an agent with chronotropic properties. If a low CO is accompanied by a low afterload state and hypotension, the blood pressure should first be supported with vasopressors until more definitive therapies can be instituted. If the patient is volume-responsive, the SV and CO may be augmented with additional preload. If the patient does not demonstrate improvement with volume, then the depressed CO may be secondary to decreased contractility. Epinephrine, an inoconstrictor, provides inotropic support along with vasoconstriction that should increase MAP. The inodilators dobutamine and milrinone may also be used but because they vasodilate the peripheral vasculature, vasopressor initiation or increase may be necessary to counteract a further drop in blood pressure. SV and CO may also be limited due to excessive afterload. The patient with high MAP and depressed CO may benefit from afterload reduction with a systemic vasodilator.

A high CO in the presence of a low HR does not need any intervention. If accompanied by low MAP, the low SVR state may be counteracted with vasoconstrictors such as norepinephrine and vasopressin. A high CO state with high MAP should be treated, as a high MAP may increase the risk of bleeding and will increase the oxygen demand of the myocardium; a decrease in blood pressure should, however, always be made in the context of baseline blood pressure and with an evaluation of the risk of hypoperfusion to pressure-dependent organs. The various possible therapies for disordered blood pressure and CO are detailed in Table 54–4.²²

The details of various vasopressors and inotropes are included in Tables 54–5 and 54–6.²³ The vasopressors are utilized for their ability to increase MAP by increasing SVR (MAP = CO × SVR). Norepinephrine, an α₁ agonist that provides β₁ inotropy is typically chosen in favor of phenylephrine, a pure α₁ agonist that only increases afterload and does not provide inotropic assistance. Vasopressin may be useful in the vasoplegic patient with low SVR postbypass as a supplement to norepinephrine. In addition, as it has minimal pulmonary vasoconstrictive effects at low doses (< 4 U/h), vasopressin may be useful in patients with significant pulmonary hypertension or right ventricular dysfunction. Dopamine at low doses produces renal and splanchnic vasodilation, an effect that has not been shown to impact outcomes. Dopamine is the precursor to norepinephrine; at higher doses, the effect of dopamine becomes more similar to that of norepinephrine with the addition of inotropic and then vasoconstrictive properties. Epinephrine, dobutamine, and milrinone provide inotropy with varying effects upon the blood pressure. Epinephrine, due to greater α₁ constriction than β₂ vasodilation, has the overall effect of increasing blood pressure. Dobutamine and milrinone, both inodilators, increase inotropy but at the cost of increased vasodilation and the potential need for initiation or escalation of vasopressors. Compared with dobutamine, milrinone produces a greater decrease in SVR. Dobutamine and milrinone both increase levels of cyclic adenosine monophosphate (cAMP) (and subsequently calcium and contractility) in the cardiac myocyte but by different mechanisms: through the β₁ receptor, dobutamine activates the G_s/adenylate cyclase pathway while milrinone inhibits phosphodiesterase III, an enzyme that degrades cAMP. These 2 agents, hence, are synergistic and lower levels of each medication may be coadministered with fewer side effects. Although all β₁ agonists demonstrate some degree of chronotropy, isoproterenol demonstrates a particularly strong ability to increase HR and is colloquially known as the chemical pacemaker.

In addition to chronotropy and inotropy, the different β₁ agonists also demonstrate varying degrees of bathmotropy (arrhythmogenicity) and dromotropy (increased conduction through the AV pathway). High doses of norepinephrine, dopamine, epinephrine, and dobutamine may induce tachycardia or arrhythmias such as atrial fibrillation with rapid ventricular response. Milrinone also has a proarrhythmogenic tendency but is less than that of dobutamine.

The half-lives of these inotropes and vasopressors, with the exception of vasopressin and milrinone, are on the order of minutes, thus enhancing their titratability. Vasopressin has a half-life of approximately 10 to 20 minutes. Milrinone's terminal elimination half-life is approximately 2 hours, which renders it less titratable than other more rapid-offset inotropes such as dobutamine. It accumulates with prolonged infusions as well as in patients with renal insufficiency.

Epinephrine has been associated with the development of hyperglycemia and a type B lactic acidosis, that is, lactate generation in the absence of clinical evidence of tissue hypoperfusion (in contrast to type A which is associated with poor oxygen delivery). One study randomized 36 patients to receive norepinephrine or epinephrine for hypotension after CPB. They found that 6 of 19 patients who received epinephrine developed increased lactate levels peaking between 6 and 10 hours after CPB (lactate to ~3.8 mmol/L, pH 7.34) while 0 of 17 patients who received norepinephrine demonstrated the same phenomenon. The lactate normalized by 22 to 30 hours after bypass. There was no difference in the oxygen-delivery index calculated between the 2 groups. The authors postulated that because epinephrine inhibits the pyruvate dehydrogenase complex, the increased levels of pyruvate get shunted to generate more lactate in order to preserve the normal balance of lactate:pyruvate needed to maintain the oxidized form of nicotinamide adenine dinucleotide (NAD).²⁴ Another study conducted in a medical ICU studying patients with dopamine-resistant cardiogenic shock randomized patients to receive norepinephrine-dobutamine or epinephrine. This study found that patients in the epinephrine group demonstrated an increase in lactate that peaked at 4.9 mmol/L at 6 hours with a pH of 7.26; this finding was notably absent in the dobutamine-norepinephrine group.²⁵ Thus, lactic acidosis in the context of epinephrine therapy may not necessarily indicate hypoperfusion as it may reflect generation of endogenous lactate by another unrelated mechanism, especially if oxygen delivery is calculated to be adequate; however, use of epinephrine may obscure the utility of lactate as a marker of hypoperfusion.

The use of a vasodilator may be necessary with excessive hypertension that cannot be resolved with the usual doses of sedative and analgesics (Table 54–7). Nicardipine is a dihydropyridine calcium channel blocker whose main effect is vasodilation of the systemic, coronary, and cerebral circulations with essentially no AV nodal blocking activity or negative inotropy. Both nitroglycerin and nitroprusside release nitric oxide, a potent vasodilator that activates guanylyl cyclase, increase cyclic guanosine monophosphate (cGMP), and results in decreased intracellular calcium and vascular smooth muscle relaxation. Nitroglycerin predominantly causes venodilation, which decreases preload and the volume work of the heart. It is particularly useful when myocardial ischemia is suspected as CPP (AoDP – LVEDP/coronary vascular resistance) is improved by 2 mechanisms: by decreasing preload which decreases LVEDP and by decreasing coronary vascular resistance by promoting coronary vasodilation. Nitroprusside causes both venodilation and arterial dilation. High doses may result in cyanide toxicity.

These 3 vasodilators provide rapid onset of action within minutes. Nicardipine has an elimination half-life of approximately 45 minutes while the elimination half life of both nitroglycerin and nitroprusside is approximately 5 minutes. Vasodilators may result in a reflex tachycardia. Tachyphylaxis can be seen, as well as rebound hypertension on sudden discontinuation of the drip. All vasodilators can reverse hypoxic vasoconstriction and worsen hypoxia by increasing shunt in atelectatic areas of the lung. In addition, those with intracranial hypertension may experience an increase in intracranial pressure due to an increase in cerebral blood flow secondary to cerebral vasodilation.

Postoperative bleeding is tracked by output from pleural and mediastinal chest tubes. A bleeding rate of 100-400 mL/h may be considered “medical” bleeding that may be successfully treated by correcting postoperative coagulopathy. Cumulative bleeding of more than 200 mL for 2 hours usually mandates intervention for resolution of “medical bleeding.” One study of 31 patients found that exposure of blood to the CPB circuit as well as hypothermia-induced activation of platelets resulting in a transient impairment of function and an increase in bleeding time that increased with the duration of bypass. This phenomenon resolved generally within 2 to 4 hours after CPB. However, 10 patients with platelet counts greater than 100,000/μL demonstrated substantial post-CPB bleeding and were found to have persistent platelet dysfunction even to 8 hours postbypass. Platelet transfusion resolved the bleeding in 6 of 10 patients.²⁶ The patient who has been on aspirin or loaded with clopidogrel due to a recent coronary catheterization (within 5 days) would be expected to have an even more profound platelet plug defect post-CPB due to irreversible inactivation of platelet activation. Significant bleeding in these patients should improve with platelet transfusion. Packed red blood cells should be administered as needed for significant anemia. Fresh frozen plasma may be administered empirically if bleeding is unremitting. Cryoprecipitate for hypofibrinogenemia (< 100 mg/dL) may be warranted. Desmopressin (DDAVP) may aid with bleeding in the context of uremia or in the patient suspected of having a von Willebrand factor deficiency (0.3 μg/kg IV). Recombinant factor VIIa (rFVIIa) may be used as a last resort for life-threatening and intractable bleeding. It activates factor X directly and promotes a thrombin burst that aids in the formation of clot. A phase II dose escalation study was perform in 2009 as an international randomized, double-blind, placebo-controlled study of 172 patients testing placebo against 40 and 80 μg/kg of rFVIIa. The rFVIIa groups experienced a lower rate of reoperation for bleeding (40 μg/kg: 14%, 80 μg/kg: 12% vs placebo: 25%) and a decreased rate of transfusion (40 μg/kg: 640 mL, 80 μg/kg: 500 mL vs placebo: 825 mL). The rate of critical serious adverse events was not statistically significant between the groups, but study was underpowered to assess this outcome; the authors emphasized the trend toward increased serious adverse events in the rFVIIa group and the need for further investigation.²⁷ Viscoelastic whole blood tests such as thromboelastography (TEG) may pinpoint the hemostatic defect (ie, coagulation factor deficiency, quantitative or qualitative platelet disorders, excessive fibrinolysis) and provide the ability to provide more targeted transfusion therapy.

Bleeding of more than 200 mL for 4 hours or more than 4 to 500 mL in 1 hour that does not respond to blood-product administration is typically designated “surgical bleeding” and may warrant a return to the operating room. Sources of surgical bleeding may include cannulation sites, graft suture lines, vein graft side branches, and small chest wall arterial bleeders.

Copious bleeding increases the risk of tamponade, especially if sudden cessation of chest tube output occurs due to a clot. Tamponade can also occur with removal of pacing wires several days after the procedure or weeks later with the start of anticoagulation for mechanical valves or atrial fibrillation. Cardiac tamponade reduces stroke volume due to restricted diastolic filling and decreased preload ultimately resulting in equalization of filling pressures between the right heart and left heart (including the central venous pressure [CVP], pulmonary artery diastolic pressure, and pulmonary artery wedge pressure). Tamponade should be suspected in the setting of a rising CVP, worsening pressor requirement, oliguria, and acidosis. The likelihood of other similarly presenting clinical entities such as right heart failure, pulmonary embolism, tension pneumothorax, and left heart failure should be assessed. Transthoracic or transesophageal echo may aid greatly in establishing the diagnosis. However, difficulties may be encountered with clotted blood, which is echo-dense and appears similar to myocardium or with effusions localized to a particular area not easily accessible by surface ultrasonography.²⁸ Ultimately, the patient with suspected tamponade will require a return to the operating room for exploration. Until that occurs, the compensatory tachycardia (increasing the CO due to limited SV) should not be treated and a high afterload should be maintained. To that end, fluid and inoconstrictors such as epinephrine and norepinephrine may be able to temporize the patient's hemodynamics until the mechanical obstruction can be relieved.

Atrial fibrillation after cardiac surgery is a common phenomenon, occurring in 10% to 65% of patients.²⁹ It occurs more frequently in older patients and those undergoing valvular surgery. Those with mitral pathology tend to be at risk for atrial fibrillation due to atrial distension. Agents such as β₁ agonists (dopamine, norepinephrine, epinephrine, dobutamine) as well as milrinone increase the arrhythmogenic potential of the myocardium and may precipitate the onset of atrial fibrillation. The peak incidence of atrial fibrillation is in the second and third postoperative days. Its occurrence increases morbidity, as the rate of stroke is higher, as well as 30-day mortality. In addition, the occurrence of atrial fibrillation increases the length of stay both in the ICU and the hospital.²⁹ In hemodynamically stable patient who has been weaned-off of inotropes and vasopressors, β-blockade with intravenous or oral (PO) metoprolol is frequently initiated to decrease adrenergic tone that may precipitate atrial fibrillation.

When atrial fibrillation occurs, the stability of the patient should immediately be assessed. If the patient is unstable and is hypotensive or symptomatic (mental status changes, chest pain, or shortness of breath), the patient should be emergently cardioverted. If the patient is stable, the decision with how to proceed often differs based on institution and surgeon. The Atrial Fibrillation Follow-up Investigation of Rhythm Management (AFFIRM) trial³⁰ showed that there is no mortality benefit to pursuing rhythm control (ie, attempting to convert the patient back to sinus) over HR control. In addition, the natural history of postoperative atrial fibrillation appears to be that of spontaneously converting back to sinus rhythm within 6 weeks. Many take the approach of pursuing rate control and attempting only pharmacologic cardioversion. However, some choose to treat atrial fibrillation aggressively with electrical cardioversion to avoid the need for anticoagulation and to potentially to decrease the risk of stroke, although the risk of stroke does persist for a time after sinus rhythm is achieved (whether via electrical, pharmacologic, or spontaneous cardioversion) due to atrial stunning; more than 80% of thromboembolic events occur during the first 3 days after cardioversion.³¹

Once electrolyte repletion of inadequate potassium and magnesium levels are addressed, first-line medical therapy with hemodynamically stable atrial fibrillation should be β-blockade with an agent such as metoprolol in an attempt to achieve a HR of less than 120. As a class II antiarrhythmic, β-blockade may achieve conversion back to sinus rhythm. Diltiazem, a benzothiazine calcium channel blocker with AV nodal blocking properties, is frequently used in the general hospital setting to control atrial fibrillation with rapid ventricular response; however, its negative inotropic property may induce unwanted hypotension, especially in the immediate postoperative period where postcardiac patients are still recovering from postbypass myocardial stunning, and is not utilized with great frequency in the cardiac ICU. Amiodarone may achieve rate control in the patient with a tenuous blood pressure or who is on vasopressors or inotropes or has a low EF and may achieve chemical cardioversion, as well. It has a long terminal elimination half-life of about 58 days. The patient with a difficult to control HR or with recurrent bouts of atrial fibrillation may benefit from multiple doses of β-blocker or an additional load of amiodarone. Digoxin, an inhibitor of the Na-K ATPase pump, decreases conduction velocity through the AV node and may aid in attaining rate control. In addition, digoxin increases inotropy and may be useful in the patient with a low EF or still on vasopressors. However, its onset of action is slow and requires loading over 1 day. In addition, lower doses must be used in those with renal insufficiency.

The side effects of these aforementioned therapies for atrial fibrillation with rapid ventricular response are hypotension and bradycardia. Thus, those being treated for atrial fibrillation should not have their epicardial pacing wires removed until the therapeutic response is established. Should bradycardia occur, pacing may be required and the therapeutic approach should be reevaluated. Postoperative new onset atrial fibrillation with a slow ventricular response often precludes use of rate-lowering agents. Amiodarone, while ubiquitously used, may with long-term use induce hepatotoxicity, pulmonary disease, and thyroid dysfunction (hyperthyroidism and hypothyroidism). Thus avoidance in patients with liver dysfunction or pulmonary insufficiency may be merited.

Anticoagulation should be considered in those with more than 48 hours of atrial fibrillation to decrease the risk of the development of cardiac thrombus that may embolize to cause stroke. Characteristics that increase the risk of stroke include hypertension, an EF less than 35%, diabetes mellitus (DM), and age 75 years and older. The benefit of anticoagulation must be weighed against the risk of inducing bleeding in the postoperative patient. If the risk is deemed too high, anticoagulation should be held until the risk of bleeding diminishes.

VT is also seen commonly in the postcardiac surgical population. VT may be an expression of myocardial ischemia and its onset should spur an investigation into the presence of a potential myocardial supply-demand imbalance. As with atrial fibrillation, cardiotonic medications such as the β₁ agonists and milrinone may precipitate VT. Prior myocardial infarction and the presence of scar frequently act as anatomic foci of arrhythmogenicity. Finally, VT may manifest in the setting of an unfavorable milieu such as hypoxia, electrolyte imbalances, and acidosis. As with atrial fibrillation, an unstable VT should be electrically cardioverted immediately.

The actual diagnosis of VT may be difficult in the patient with a preexisting bundle branch block. In this situation, a supra-VT (SVT) with aberrancy may be indistinguishable from VT. Provided that the rhythm is regular, adenosine may be utilized to assess whether the tachycardia breaks, in which case the arrhythmia may be diagnosed as SVT.³² The diagnostic distinction is important as it may have consequences for whether the patient may need invasive therapies such as ablation and an implantable cardioverter defibrillator (ICD). Adenosine should not be given to patients with an irregular wide-complex tachycardia because such arrhythmias may become unstable after administration.

The patient with nonsustained VT (NSVT) or stable VT should be evaluated for the presence of correctable conditions such as hypokalemia and hypomagnesemia, hypoxia, and acidosis. Amiodarone may be administered in an attempt to suppress the arrhythmia. Lidocaine may be used as an antiarrhythmic adjunct. If ischemia is a component of the onset of VT, a trial of a short acting β₁ blocker such as esmolol may be worthwhile to decrease myocardial oxygen demand. Esmolol is degraded by red cell esterases and has a short elimination half-life of 6 minutes. Overdrive pacing may be used in an attempt to suppress reentry circuits at work in the pathogenesis of VT. The patient in VT may ultimately need electrical cardioversion to get back into their native rhythm. Finally, consultation from electrophysiological experts should be obtained as the patient may be a candidate for ablation or may need placement of an ICD. The medications used in the management of dysrhythmias are detailed in Table 54–8.³³

Renal dysfunction is not uncommon after cardiac surgery. Mangano et al conducted a prospective observational study of 2222 patients undergoing CABG and found that the rate of renal dysfunction, defined as a postoperative creatinine of 2.0 mg/dL or more or an increase in creatinine of 0.7 mg/dL or more was about 7.7%. The incidence of oliguric renal failure requiring dialysis was 1.4%. Preoperative risk factor for renal dysfunction included age 70 years or more, New York Heart Association (NYHA) class 3 or 4 heart failure, previous CABG, chronic kidney disease with creatinine 1.4 to 2.0, type I diabetes, and glucose more than 300. Intraoperative and postoperative factors include CPB lasting 3 hours or more, and a low output state with a CI of less than 1.5 or requiring insertion of an intra-aortic balloon pump.34 Other additional risk factors include preoperative cardiac angiography or angiotensin-converting enzyme inhibitors, angiotensin receptor blockade or diuretic therapy, and perioperative red blood cell transfusion. Additional intraoperative factors include the number of red blood cell units administered, low MAP during CPB, perioperative hemodilution, and postoperative use of norepinephrine.³⁵

With regards to outcomes, Mangano et al found that compared with those without renal dysfunction, those with renal dysfunction and failure experienced increased length of stay in the critical care unit (3.1 vs 6.5 and 15.4 days) and the hospital (7.4 vs 11.4 and 14.9 days). In addition, those with renal dysfunction experienced greater mortality (0.9% vs 19% and 63%).

Preventative strategies to decrease the risk of renal dysfunction after cardiac surgery are limited and include minimizing exposure to radiocontrast dyes, delaying the surgery to allow for renal recovery if preoperative renal injury was sustained, and avoidance of intraoperative anemia. Other protective strategies such as urinary alkalinization, glucose control, early continuous renal replacement therapy, statins, and fenoldopam still as of yet need more evidence to elucidate their role in renal protection.³⁵

Oliguria, defined as urine output < 0.5 mL/kg/h, as well as increases in creatinine reflective of a decrease glomerular filtration rate occur frequently in the cardiac ICU. Table 54–9 outlines the risk, injury, failure, loss, and end-stage renal disease (RIFLE) criteria and Table 54–10 presents the Acute Kidney Injury Network (AKIN) modifications. The AKIN classification, in particular, highlights how even small increases in creatinine of 0.3 mg/dL indicate some degree of renal injury.

The approach to the patient with oliguria and/or rising creatinine may be categorized as postrenal, prerenal, and renal processes.

Postrenal

In the immediate postoperative period, postrenal obstruction may occur if a clogged Foley is preventing the egress of urine. Ultrasonic visualization of a full bladder in the setting of lack of urine output confirms postrenal obstruction. Flushing or changing of the Foley should rectify the problem. Less likely forms of postrenal obstruction such as nephrolithiasis may be formally evaluated by looking for hydronephrosis on a renal ultrasound.

Prerenal

Prerenal causes of oliguria and rising creatinine are secondary to insufficient perfusion pressure or hypovolemia or both. An insufficient MAP may cause oliguria, especially in the patient with chronic uncontrolled hypertension or a renal vascular stenosis and may be caused purely by a vasodilated state with decreased ability of the smooth muscle to constrict or be due to decreased CO due to hypovolemia and insufficient SV. MAP may be supported pharmacologically with vasopressors with concurrent volume loading to address any contribution of hypovolemia. The patient with an adequate MAP may still be hypovolemia. In general, aside from the patient with tenuous pulmonary or right heart function, small boluses of 250 to 500 mL generally incur little harm and can interrogate the volume responsiveness of the patient. Aside from history and calculation of fluid balance, the prerenal state of the patient may be interrogated with the fractional excretion of sodium (FENa) or fractional excretion of urea (FEUrea) if the patient has recently received furosemide.

Renal

Prerenal acute kidney injury, when prolonged, can result in acute tubular necrosis (ATN). Some patients, such as those who underwent cardiac arrest or precipitous operating room courses may have sustained prolonged periods of hypotension and hypovolemia, and will have progressed into ATN. At this point, the patient should be assessed as to whether they are nonoliguric or oliguric. If the patient demonstrates a rising creatinine indicative of ongoing loss of kidney function but are able to generate urine, then not much intervention may be needed except monitoring of volume and electrolyte status and for plateauing of the creatinine level. If the patient has oliguric kidney injury with a poor urine output, they should be assessed for diuretic responsivity. A trial of a loop diuretic such as furosemide (40-100 mg IV) or bumetanide (1-2 mg IV) with or without augmentation with a thiazide diuretic such as metolazone (5-10 mg PO) may be made. If the patient is diuretic responsive, then medical management may be sufficient to manage the volume and electrolyte status of the patient. If the loss of kidney function can be managed pharmacologically, the patient might be able to be maintained only diuretics for management of volume and electrolytes. If the patient is not diuretic-responsive, then renal replacement might be indicated, especially if the patient demonstrates hyperkalemia with peaked T waves on the electrocardiogram, acidosis with pH less than 7.2, blood urea nitrogen (BUN) more than 80 with clinical signs of uremia (bleeding, encephalopathy, pericardial rub or effusion), or volume overload resulting in pulmonary edema and dyspnea/hypoxia.

As mentioned earlier, length of CPB is an intraoperative risk for acute kidney injury. Most patients sustain a small amount of injury from what might be delineated postbypass ATN. The injury incurred is typically small although not infrequently achieves at least stage I by AKIN criteria. However, especially with prolonged bypass runs, the injury to the kidney can be significant. After ruling out prerenal causes of kidney dysfunction, evaluation and treatment of the patient should include an assessment of concurrent oliguria or nonoliguria, diuretic responsiveness, and whether renal replacement therapy is indicated. ATN may be more formally assessed by FENa/FEUrea as well as urine microscopy showing muddy brown casts.

If the patient has not sustained a prerenal injury by history and if MAP and volume status have been deemed to be adequately addressed, oliguria and a rising creatinine be secondary to diuretic dependence. Some patients with preoperative chronic kidney disease with diminished renal function at baseline may be dependent on diuretics for generation of adequate urine. As long as the patient is not suffering from low MAP (absolute or relative) or hypovolemia, reinstating these medications may restore urine production and by providing excretion of creatinine, restore plasma creatinine levels back to baseline.

Other causes of renal injury include toxins (exogenous and endogenous) and sepsis. Exogenous toxins include contrast dye, antibiotics such as the aminoglycosides, and NSAIDs such as ketorolac and ibuprofen. Endogenous toxins include myoglobin from rhabdomyolysis. The septic patient may demonstrate a septic nephropathy that resolves only with source control and antibiotics.

One landmark single center study of 1548 surgical ICU patients (63% of whom were postcardiac surgical patients) suggested that intense glucose control between 80 and 110 mg/dL versus 180 and 215 mg/dL resulted in a relative reduction in mortality (8% to 4.6%).³⁶ However, the results of this study could not be replicated in multiple subsequent investigations. In particular, the Normoglycemia in Intensive Care Evaluation and Surviving Using Glucose Algorithm Regulation (NICE-SUGAR) study, a multicenter trial of 6104 medical and surgical patients, randomized the patients to 81 to 108 mg/dL versus 144 to 180 mg/dL and not only found increased mortality with intensive glucose therapy with a 90-day mortality of 27.5% in the treatment group and 24.9% in the control group but also an unacceptably high incidence of severe hypoglycemia (blood glucose ≤ 40 mg/dL) with 6.8% in the intensive group and 0.5% in the conventional group.³⁷ The optimum level of glucose control, thus, is as yet to be defined, and may be better elucidated with the development of technology that may be able to allow for the safer administration of continuous insulin regimens.

Multiple issues arise in the care of the immediate postoperative cardiac patient. A decision must be made as to the appropriate timing of extubation. In addition, knowledge of normal physiology as well as the pathophysiology resulting from CPB informs the understanding of hypotension and low CO, bleeding, and renal dysfunction in the postcardiac patient. Management of these common postsurgical problems also requires a thorough understanding of the characteristics of the pharmacological tools available to us. This knowledge, as well as good communication with the surgical team, allows for successful management of the postcardiac surgical patient.