Following a brief stay in the intensive care unit, most patients undergoing cardiac surgical procedures follow a routine pattern of recovery. The use of fast‐track protocols and critical care pathways ensures that the healthcare team and the patient have a clear understanding of what to expect at different junctures during recovery. These pathways are designed to standardize care and identify variances from the expected. However, they are not a substitute for careful patient evaluation, which may identify problems that might otherwise be ignored by rigid adherence to protocols.
Most patients are transferred to an intermediate care unit or the postoperative cardiac surgical floor on the first postoperative day. Invasive monitoring is no longer utilized, although bedside telemetry should be considered for several days to identify arrhythmias. It should be remembered that patients are still in an early phase of recovery from surgery with many physiologic derangements still present. Restoring the patient to a normal physiologic state requires careful attention to the prevention, identification, and treatment of complications that may develop at any time during the hospital stay. A detailed daily examination of the patient must be performed, with particular attention paid to each organ system. Although pre‐printed or computerized order sets are available upon transfer, orders must be thought out carefully and individualized to ensure the best possible postoperative care.
Although postoperative complications are more common in elderly patients and those with comorbidities, they may still develop unpredictably in low‐risk, healthy patients despite an uneventful surgical procedure and early postoperative course. Problems such as atrial arrhythmias are very common and quite benign, with little influence on the patient’s hospital course or long‐term prognosis. In contrast, less common complications, such as stroke, mediastinitis, tamponade, renal failure, or an acute abdomen, may be devastating, resulting in early death or prolonged hospitalization with multisystem organ failure.
II. Transfer from the ICU and Postoperative Routines
The patient recovering uneventfully from open‐heart surgery is usually extubated within 6–8 hours and off all inotropic support by the first postoperative morning. The following interventions represent standardized steps in a critical care pathway which are applicable to most patients (Table 13.1). In more critically ill patients who may require an additional period of ventilatory or pharmacologic support, adherence to these time‐related recommendations may need to be modified, and withdrawal of “intensive care” must be carefully evaluated and not rushed. Typical orders for transfer to the postoperative floor are noted in Table 13.2 and Appendix 5.
Table 13.1 Critical Pathway for Coronary Artery Bypass Grafting
____ units regular insulin (Novolin R or Humulin R) SC ____ qAM ____ qPM
____ units NPH insulin (Novolin N or Humulin N) SC ____ qAM ___ qPM
Sliding scale: treat fingerstick/glucometer glucose according to the following scale at 06:00 AM, 11:00 AM, 3:00 PM, and 8:00 PM 150–160, give 2 units regular insulin SC (Novolin R or Humulin R) 161–200, give 4 units regular insulin SC 201–250, give 6 units regular insulin SC 251–300, give 8 units regular insulin SC 301–350, give 10 units regular insulin SC >350, call house officer
Acetaminophen 650 mg PO q3h prn temp >38.5 °C
Ascorbic acid 1 g PO qd × 5 days
Zolpidem 2.5–5 mg PO qhs prn sleep
Melatonin ____mg PO qhs prn sleep (1.5–3 mg usual dose)
Furosemide ___ mg IV/PO q _ h
Potassium chloride ___ mEq PO bid (while on furosemide)
Albuterol 2.5 mg/5 mL NS via nebulizer q4h prn
Levalbuterol (Xopenex) 0.63 mg in 3 mL NS q8h via nebulizer or two inhalations q4–6h through a pressured MDI
Duoneb inhaler q6h
Postoperative day and night
Wean vasoactive medications
Wean from ventilator and extubate
Remove nasogastric tube
Remove Swan‐Ganz and arterial lines
Get patient out of bed (OOB) in a chair
Initiate β‐blocker therapy and aspirin
Start warfarin for valve patients if minimal chest tube drainage
Remove chest tubes if minimal drainage
Transfer to floor; place on telemetry and pulse oximetry × 72 hours
Get patient out of bed and ambulating
Remove Foley catheter
Start warfarin for valve patients if not started night before
Remove chest tubes if minimal drainage
Stop antibiotics (after 48 hours maximum)
Advance diet to achieve satisfactory nutrition
Increase activity level
Continue diuresis to preoperative weight
Consider heparin for patients receiving mechanical valves
Commence planning for home services or rehabilitation
Assess potential discharge location (home vs. rehab)
Initiate discharge teaching
Carefully review discharge medications and instructions with patient and family
Discharge home or to rehab facility
III. Differential Diagnosis of Common Postoperative Symptoms
The development of chest pain, shortness of breath, fever, or just feeling “plain lousy” with a poor appetite and fatigue during the early convalescent period is not unusual, especially in elderly patients. Although the cause of these signs and symptoms may be benign, they should not be taken lightly because they may indicate the presence of potentially serious problems that warrant investigation. Careful questioning and examination of the patient on a daily or more frequent basis can prioritize diagnoses, direct the evaluation, and lead to prompt and appropriate treatment.
Differential diagnosis. The development of chest pain following cardiac surgery often raises the suspicion of myocardial ischemia, but the differential diagnosis must include several other potential causes. The greatest fear to a patient is that the recurrence of chest pain indicates a failed operation; the surgeon meanwhile may purposely try to provide an alternative explanation. Although musculoskeletal pain is the most common cause of chest discomfort, significant problems that must be considered include:
Sternal wound infection
Evaluation. Careful physical examination (breath sounds, pericardial rub, sternal wound), a chest x‐ray, and 12‐lead ECG will usually provide the appropriate diagnosis and direct additional testing. Differentiation of ST‐segment elevation related to ischemia vs. pericarditis is important and can be difficult to make (Figures 8.2 and 8.3, pages 384 and 385). Consultation with the cardiology service is essential in managing patients with a suspected cardiac origin to their chest pain. Stress imaging or even coronary angiography may be warranted. Other diagnostic modalities include echocardiography, computed tomography (CT) pulmonary angiography to rule out pulmonary embolism, and sternal wound aspiration.
Shortness of breath
Differential diagnosis. Shortness of breath is usually caused by splinting from chest wall discomfort and is not uncommon in the anemic patient with underlying lung disease. However, significant shortness of breath, its acute onset, or deterioration in pulmonary status should raise awareness of a significant problem. The source may be of a primary pulmonary nature, but it may also be the consequence of cardiac dysfunction or oliguric acute kidney injury (AKI). Diagnoses to be considered include:
Atelectasis and hypoxia from mucus plugging or poor inspiratory effort
Pneumonia (possibly aspiration)
An enlarging pleural effusion
Cardiopulmonary problems – low cardiac output states or acute pulmonary edema caused by:
Acute myocardial ischemia or infarction
Residual or new‐onset mitral regurgitation (ischemic, associated with systemic hypertension) or a recurrent ventricular septal defect
Fluid overload from surgery or heart failure, often associated with oliguric AKI
Severe diastolic dysfunction
Atrial or ventricular tachyarrhythmias
Compensatory response to metabolic acidosis (low cardiac output state)
Evaluation. Careful lung examination may reveal absent breath sounds or diffuse rales/rhonchi, suggesting a parenchymal process, a pneumothorax, or pulmonary edema. Clinical evidence of cardiac tamponade (muffled heart sounds, orthostatic blood pressure changes, pulsus paradoxus) should be sought. An arterial blood gas (ABG), chest x‐ray, and ECG should be obtained. An echocardiogram gives an assessment of ventricular function, detects valve dysfunction or recurrent shunting, and may also identify a large left pleural or pericardial effusion or tamponade. A CT pulmonary angiogram should be performed if pulmonary embolism is suspected.
Differential diagnosis. Fever is very common during the first 48–72 hours after surgery. Initially, it may be related to the systemic inflammatory response or cytokine release from cardiopulmonary bypass (CPB), but subsequently is usually ascribed to atelectasis from poor inspiratory effort after extubation. Thorough evaluation of recurrent fevers is warranted after the first 72 hours.1,2 Potential causes of postoperative fever include:
Atelectasis or pneumonia
Urinary tract infection (UTI)
Wound infections: sternum or leg
Clostridium difficile colitis or other intra‐abdominal process
Sinusitis (usually in patients with indwelling endotracheal or nasogastric tubes)
Endocarditis (especially on a prosthetic valve)
Deep venous thrombosis (DVT) and pulmonary embolism
Postpericardiotomy syndrome (PPS)
Evaluation. The lungs, chest, and leg incisions should be examined carefully. A CBC with differential, chest x‐ray, urinalysis, and appropriate cultures should be performed. A stool sample for C. difficile should be obtained if the patient has abdominal pain or diarrhea. Indwelling central and arterial lines should be cultured and removed if in place for more than five days or if cultures return positive. If the WBC is normal or there is an eosinophilia, a drug fever may be present. Occult sternal infections may be investigated with a chest CT scan, but results are usually nonspecific; needle aspiration may be performed if suspicion is high. Head CT scans can identify sinusitis. A transesophageal echocardiogram can evaluate the heart valves for vegetations consistent with endocarditis.
Treatment. It is best to defer antibiotic therapy until an organism has been identified. However, a broad‐spectrum antibiotic may be initiated based on the presumed source and organisms involved as soon as cultures have been obtained. This is especially important in patients who have received prosthetic material (valves, grafts). A more narrow‐spectrum antibiotic may be substituted subsequently. Empiric oral vancomycin (125 mg four times a day) may be started for suspected C. difficile colitis. Occasionally a patient will have a fever and elevated WBC with no evident source, but will respond to a brief course of antibiotics. Further comments on nosocomial infections and sepsis can be found on pages 776–778.
Respiratory function is still impaired when the patient is transferred to the postoperative floor, with many patients exhibiting shortness of breath with some splinting from chest wall discomfort. Arterial desaturation is not uncommon, and all patients should have an arterial saturation measured several times daily by pulse oximetry until the SaO2 remains above 90%. It is not uncommon to see significant desaturation when the patient becomes more ambulatory. Most patients have some degree of fluid overload and require diuresis, and steps must be taken to overcome a poor inspiratory effort and atelectasis. Potential complications, such as pneumonia, bronchospasm, pleural effusions, or pneumothorax, can be identified by examination and a chest x‐ray (Table 13.3). Standard orders should include:
Supplemental oxygen via nasal cannula at 2–6 L/min on the postoperative floor. In the ICU, high flow oxygen systems or BPAP are useful for hypoxemia, but after transfer to the floor, if the patient requires more than supplemental oxygen by nasal cannula or facemask to achieve an acceptable oxygen saturation, they should probably be transferred back to an ICU setting for more intensive respiratory care.
Frequent use of incentive spirometry to encourage deep breathing
Provision of adequate, but not excessive, analgesia. Patient‐controlled analgesia (usually morphine) is particularly beneficial for one or two days following surgery, and may be supplemented with other pain medications, such as ketorolac (Toradol) 15–30 mg IV q6h for a few days. In patients with abnormal renal function, IV acetaminophen is beneficial in the ICU. Most patients obtain adequate analgesia with oral medications 2–3 days after surgery and seem to do better with regular, rather than prn, pain medications.
Bronchodilators administered via nebulizers should be used if copious secretions or bronchospasm are present (see pages 503–504). These commonly include albuterol, levalbuterol (Xopenex), or a combination of albuterol and ipratropium (Duoneb). Chest physical therapy may benefit patients having difficulty raising secretions.
Measures to reduce the risk of venous thromboembolism (antiembolism stockings, sequential compression devices, subcutaneous heparin or low‐molecular‐weight heparin [LMWH]) should be considered depending on the patient’s mobility and risk (see section E on pages 750–752).
Patients with a history of obstructive sleep apnea should utilize CPAP machines at night or any other device that has been helpful to them prior to surgery.
Patients with preexisting lung disease and a history of heavy smoking often have a tenuous respiratory status postoperatively with borderline oxygenation, and acute decompensation can occur with little provocation, including ambulation. Mucus plugging, atelectasis from poor inspiratory effort, mobilization of “third space” fluid, or even a minor cardiac event can cause arterial desaturation and respiratory distress. In patients without significant underlying lung disease, acute decompensation usually indicates the presence of a significant process, such as a pleuropulmonary event (significant pneumothorax, pneumonia, pulmonary embolism), myocardial ischemia, worsening mitral regurgitation, cardiac tamponade, or acute fluid overload from AKI with oliguria.
The management of respiratory insufficiency, pneumothorax, pleural effusions, and bronchospasm is discussed in Chapter 10. Other complications, including diaphragmatic dysfunction from phrenic nerve paresis and pulmonary embolism, are discussed below.
Diaphragmatic dysfunction from phrenic nerve injury has been noted in 10–20% of patients following open‐heart surgery.5
Etiology and prevention
Cold injury to the phrenic nerve from use of iced saline slush in the pericardial well is the primary cause of this problem. Systemic hypothermia may also be contributory. Use of insulation cooling pads that protect the phrenic nerve from cold solutions, minimizing systemic hypothermia, intermittently pouring cold saline over the heart (the “shallow technique”), and avoiding iced slush reduce the incidence of phrenic nerve paresis.6–8
The phrenic nerve may be injured directly during dissection of the internal thoracic artery (ITA) in the upper mediastinum, especially on the right side. It may also be damaged when making a V‐incision in the pericardium to allow for better lie of the ITA pedicle. Phrenic nerve devascularization with compromise of the pericardiophrenic artery may also be contributory, especially in patients with diabetes.9
Most patients with unilateral phrenic nerve paresis have few respiratory symptoms and are extubated uneventfully. Difficulty weaning, shortness of breath, and the requirement for reintubation may be noted in patients with severe chronic obstructive pulmonary disease (COPD).
Bilateral phrenic nerve palsy usually produces tachypnea, paradoxical abdominal breathing, and CO2 retention during attempts to wean from mechanical ventilation.
A chest x‐ray will demonstrate an elevated hemidiaphragm at end‐expiration during spontaneous ventilation, most commonly on the left side. This will not be evident during mechanical ventilation. An elevated hemidiaphragm may be difficult to appreciate if basilar atelectasis or a pleural effusion is present. Therefore, when planning a thoracentesis or tube thoracostomy for a pleural effusion, one must always consider the possibility of an obscured, elevated hemidiaphragm. The position of the gastric bubble on chest x‐ray should identify the position of the diaphragm on the left. If the diaphragm is elevated, one might inadvertently insert a needle below the diaphragm, risking injury to intra‐abdominal structures.
Diaphragmatic fluoroscopy (“sniff test”) will demonstrate paradoxical upward motion of the diaphragm during spontaneous inspiration if unilateral paralysis is present.
Ultrasonography will show a hypokinetic, immobile, or paradoxically moving diaphragm during respiration.
Transcutaneous phrenic nerve stimulation in the neck with recording of diaphragmatic potentials over the seventh and eighth intercostal spaces can measure phrenic nerve conduction velocities and latency times.10 This is helpful in assessing whether phrenic nerve dysfunction may be a contributing factor to a patient’s respiratory problems.
Transdiaphragmatic pressure measurements can be used to make the diagnosis in patients with bilateral phrenic nerve palsies.11
Treatment is supportive until phrenic nerve function recovers, which may take up to two years. One study of patients with COPD found that nearly 25% of patients had persistent pulmonary problems with a decreased quality of life at midterm follow‐up.12 Diaphragmatic plication can provide significant symptomatic and objective improvement in patients with marked dyspnea. This can be performed robotically or via a thoracotomy (video‐assisted [VATS] or open) or laparoscopically.13 Ventilatory support is usually necessary for patients with bilateral involvement. Some patients can be managed at home with a cuirass respirator or a rocking bed.
Venous thromboembolism (VTE) is a term that describes both deep venous thrombosis (DVT) and pulmonary embolism (PE). Screening noninvasive studies have documented a 15–20% incidence of DVT and a 6–20% incidence of PE after cardiac surgical procedures, with both being more common after OPCAB.14,15 However, symptomatic VTE is noted in only about 1–2% of patients.16,17 One study of routine CT pulmonary angiography and lower‐extremity venous studies in patients undergoing elective CABGs and considered at low risk for VTE reported a 21% incidence of DVT, and in more than half of these patients, PE occurred in the absence of lower‐extremity DVT.18
It has been presumed, perhaps inappropriately so, that the risk of postoperative VTE is low because of heparinization and hemodilution during surgery and the presence of thrombocytopenia and platelet dysfunction in the early postoperative period. However, platelet activity is increased immediately after surgery, and elevated fibrinogen levels, thrombin generation, tissue factor activation, and reduced fibrinolysis are also noted. Increased platelet reactivity and aspirin resistance are particularly common after off‐pump surgery, although the risk of symptomatic VTE is still only 1%.19–22 A study of patients undergoing on‐pump CABGs found that the absorption of aspirin was reduced early after surgery, leading to reduced antiplatelet effect that could contribute to VTE.23 Thus low‐dose aspirin may not be sufficient to inhibit platelet aggregation early after surgery, although this may impact graft patency more than the risk of VTE.
Risk factors for VTE in the perioperative period include older age, obesity (BMI >30), right‐ or left‐heart failure, a history of VTE, prolonged bed rest and immobility, prolonged ventilation, multiple blood and blood product transfusions, and the occurrence of significant postoperative complications, including acute kidney injury, infection/sepsis, and neurological complications.17,24,25 VTE may result from heparin‐induced thrombocytopenia (HIT) and may occur several weeks later.
Prevention. Numerous studies have evaluated the use of mechanical or pharmacologic prophylaxis from which general recommendations can be made.25–27
Early mobilization is the most important factor in reducing the risk of VTE. Once the patient is stable, getting them out of bed and ambulating several times a day is very important. Having a patient sit in a chair is the least desirable position.
Elastic graduated compression stockings (GCS), such as T.E.D. stockings, should be placed after the initial leg dressing and ace wraps are removed, and should be placed on both legs. Use of sequential compression devices (SCDs) or intermittent pneumatic compression devices (IPCs) in well‐mobilized ambulatory patients provides little additional benefit to use of GCS alone.26,27
Although increased platelet reactivity and aggregation are present in postoperative patients, most patients are usually given aspirin 81 mg daily after surgery. This dose is insufficient to inhibit platelet aggregation and probably provides little benefit in reducing the risk of VTE.20 One study did demonstrate, however, that the addition of low‐dose aspirin to heparin 5000 units SC q8h reduced the risk of VTE nearly fivefold in patients undergoing OPCAB.28
ICU patients may be maintained at bedrest because of ongoing clinical issues and are usually poorly mobilized. For these patients as well as those with the risk factors listed above in section E.2, pharmacologic prophylaxis should be considered in addition to compression devices. One study found the combination of heparin plus one of these devices reduced the risk of PE by 60% compared with heparin alone,29 although another study of ICU patients found no added benefit of IPC when the patient was already receiving heparin.30 Nonetheless, the literature remains controversial on whether early initiation of pharmacologic prophylaxis with either SQ heparin (5000 units SC q12h) or LMWH (40 mg SC daily) should be considered in these patients and when it should be started. Some authors recommend initiating one of these medications on the first postoperative day once mediastinal bleeding has tapered off,24,25,31 but other guidelines have concluded that the bleeding risk outweighs the early benefits.32 The risk of developing a hemopericardium leading to cardiac tamponade must always be taken into consideration when deciding whether to initiate heparin early after surgery.
Manifestations. Pleuritic chest pain and shortness of breath with hypoxemia are usually present. The acute onset of these symptoms distinguishes them from typical postoperative respiratory symptoms. The new onset of atrial fibrillation (AF), sinus tachycardia, or fever of unknown origin may be clues to the diagnosis. Calf tenderness and edema are unreliable signs of DVT, especially in the leg from which the vein has been harvested. However, the new development of such findings several days to weeks after surgery should prompt further evaluation.
Assessment. ABGs, a chest x‐ray, ECG, and CT pulmonary angiography should be obtained. The presence of a low arterial oxygen saturation is nonspecific, but may be compared with values obtained earlier in the postoperative course. A positive venous noninvasive study of the lower extremities in association with respiratory symptoms and hypoxia is suggestive evidence of a pulmonary embolism and should prompt further evaluation. A falling platelet count with VTE mandates evaluation for HIT, for which alternative anticoagulation should be initiated.
Treatment. Traditionally, IV heparin has been recommended for 1 week (unless HIT is present), followed by warfarin for six months. However, equally effective therapy can be achieved with the use of a non‐vitamin K antagonist oral anticoagulant (NOAC), such as apixaban (10 mg bid × 1 week, then 5 mg bid) or rivaroxaban (15 mg bid × 3 weeks, then 20 mg daily) – doses conditional upon renal function – and these medications do not require blood monitoring. Early ambulation has not been shown to increase the risk of extending the DVT or developing pulmonary embolism, although bedrest was traditionally considered part of routine management.33,34 An inferior vena cava filter should be placed if anticoagulation is contraindicated. Systemic thrombolytic therapy should be avoided because of the recent sternotomy incision, although ultrasound‐assisted catheter‐directed thrombolysis using the EKOS system might be considered.35,36 Other interventional methods, including suction embolectomy and fragmentation therapy, may be beneficial in patients with massive PE,37 reserving surgery for salvage situations to avoid a redo sternotomy and pump run. However, surgical pulmonary embolectomy may be applicable to hemodynamically stable patients with massive PE and may be preferable in the rare patient with massive pulmonary embolism early after cardiac surgery, when thrombolytic therapy is contraindicated.38,39
Upon transfer to the postoperative floor, the patient should be attached to a telemetry system to continuously monitor the heart rate and rhythm for several days. Vital signs are obtained every shift if the patient is stable, but more frequently if the patient’s heart rate, rhythm, or blood pressure is abnormal or marginal.
The evaluation and management of complications noted most frequently in the intensive care unit are presented in Chapter 11. These include low cardiac output states, perioperative infarction, cardiac arrest, coronary spasm, hypertension, and arrhythmias. This section will discuss several cardiac problems commonly noted during subsequent convalescence (Table 13.4).
Arrhythmias and conduction problems
Atrial arrhythmias are the most common complication of open‐heart surgery and occur with a peak incidence on the second or third postoperative day. Although some patients become symptomatic with lightheadedness, fatigue, or palpitations, many have no symptoms and are noted to be in AF or flutter on telemetry monitoring. Treatment entails rate control, attempted conversion to sinus rhythm, and anticoagulation if AF persists or recurs. Management protocols are discussed in detail on pages 626–632 and in Table 11.15 (page 627).
Ventricular arrhythmias are always of concern because they may be attributable to myocardial ischemia or infarction and may herald cardiac arrest. Low‐grade ectopy or nonsustained ventricular tachycardia (VT) with normal ventricular function does not require aggressive therapy and may be managed with a ß‐blocker. In contrast, VT with impaired LV function requires further evaluation and may benefit from placement of an implantable cardioverter‐defibrillator (ICD). An echocardiogram should be considered and may identify new regional wall motion abnormalities attributable to a perioperative infarction, which may account for the arrhythmia. On occasion, when epicardial wires are still connected to a temporary pacemaker, VT may develop due to improper sensing, with inadvertent firing on the T wave triggering the malignant arrhythmias (“R on T” phenomenon).
Conduction abnormalities and heart block (see pages 614–618). Temporary pacemaker wires are routinely removed by the third postoperative day unless there is evidence of symptomatic sinus bradycardia, lengthy sinus pauses, advanced degrees of heart block, or a slow ventricular response to AF. If these issues are present, medications that reduce atrial automaticity or reduce AV conduction (β‐blockers, amiodarone, calcium‐channel blockers [CCBs], and digoxin) should be stopped. If they persist, permanent pacemaker implantation should be considered.
New bundle branch or fascicular blocks have been noted in up to 45% of patients after CABG and appear to correlate with longer aortic cross‐clamp times. However, they resolve in more than half of the patients before discharge.40,41
Patients with sick sinus or tachycardia/bradycardia syndrome may have a rapid ventricular response to AF intermixed with a slow sinus mechanism that limits use of ß‐blockers. A permanent pacemaker (PPM) system should be considered if these problems persist beyond three days. Studies have shown, however, that pacemaker dependence usually resolves within a few months, unless the indication was complete heart block.42
Patients undergoing aortic valve surgery are more prone to conduction disturbances and heart block because of debridement, edema, hemorrhage, or suturing near the conduction system. The average incidence of PPM is 5–9% after AVR,43,44 but it is greater in patients receiving rapid deployment valves.45 Preoperative conduction system disease (first‐degree block, left anterior hemiblock, right bundle branch block [RBBB], or left bundle branch block [LBBB]) is generally the major risk factor for the development of heart block using both traditional and rapid deployment valves. Both the development of a new postoperative LBBB and the necessity for a permanent pacemaker early after AVR have adverse prognostic significance with a compromise in long‐term survival.46,47
When pacemaker wires are removed, there is always the potential for bleeding and the development of tamponade. It is recommended that they be removed when the INR is less than 2, but this does not eliminate the possibility of bleeding. If the INR remains persistently elevated, the wires may be cut and left behind. After removal, the patient should remain at bedrest for one hour and vital signs should be taken every 15 minutes for the first hour and then hourly for a few hours to monitor for orthostatic changes. Tamponade can occur within minutes or hours and can prove fatal unless the possibility is entertained and addressed on an urgent, if not emergent, basis. If concern is raised because of hypotension or a complaint of chest pain, a STAT echocardiogram may be helpful. Emergency thoracotomy at the bedside may be life‐saving.
A transfemoral temporary pacing wire may be retained at the conclusion of a TAVR procedure if the heart rate is very slow or there is evidence of advanced heart block. The latter is more likely to develop in patients with a preexisting RBBB and first‐degree AV block.
If the heart rate stabilizes above 50–60/min and there is no evidence of complete heart block, the wire may be removed several hours later or the following morning. Despite this common practice, some patients with these preexisting conduction abnormalities may still develop delayed heart block (>48 hours later). If heart block persists, permanent pacemaker implantation may be necessary. Guidelines for management of conduction disturbances after TAVR are noted on page 615.48 One study found that only 21% of patients receiving a PPM after TAVR were pacer‐dependent at one year; however, pacer dependence was more common with use of self‐expanding valves and post‐balloon dilatation, or when complete heart block was the indication for the PPM.49
When the transvenous pacing wire and the 6 Fr transfemoral sheath are removed from the femoral vein after a TAVR, persistent bleeding may occur despite a short period of manual pressure, because antiplatelet therapy is routinely administered within a few hours of the procedure. If this persists, simply suturing the skin entrance site and applying a sandbag are usually effective in controlling the bleeding.
Hypertension. When the patient is transferred to the postoperative floor, oral antihypertensive medications must be substituted for the potent intravenous drugs used in the ICU. Blood pressure tends to return to its preoperative level several days after surgery once myocardial function has returned to baseline, the patient has been mobilized, and chest wall pain improves with moderate analgesia. Aggressive patient‐specific management is important to prevent issues related to blood pressure variability. For example, a patient with renal dysfunction may need a slightly higher blood pressure to ensure renal perfusion. In contrast, more strict control of blood pressure may be essential in an elderly patient with fragile tissues or in patients with perioperative bleeding. Not only can hypertension increase cardiac wall stress and cause myocardial ischemia, but it may increase any residual mitral regurgitation and can precipitate an aortic dissection from graft or cannulation sites.
A decrease in systolic blood pressure from preoperative levels may be noted in patients who are hypovolemic or anemic, or have experienced a perioperative infarction. In these patients, preoperative antihypertensive medications can be withheld and then restarted at lower doses when the blood pressure increases. In contrast, patients who have ongoing pain issues or have undergone an AVR for aortic stenosis may develop significant systolic hypertension.
ß‐blockers are recommended for virtually all patients after surgery as prophylaxis against AF and can be titrated up to control the heart rate and blood pressure as well. Resumption of the patient’s preoperative medications should then be instituted to optimize blood pressure control. Other considerations when selecting an antihypertensive medication include the following:
Poor ventricular function (EF <40%): use one of the ACE inhibitors or ARBs. A ß‐blocker, preferably carvedilol, should also be given in this situation, but must be used cautiously if the patient has decompensated heart failure or a low output syndrome.
Sinus tachycardia with good LV function, or with evidence of residual myocardial ischemia: use higher doses of a ß‐blocker (metoprolol) or labetalol.
Coronary spasm or use of a radial artery graft: use a nitrate or CCB (amlodipine, diltiazem, or nicardipine).
Sinus bradycardia with good LV function: if initial use of an ACE inhibitor or ARB is insufficient, amlodipine should be considered.
Hypotension may develop after transfer to the floor and should be evaluated using the differential diagnosis of shock or a low cardiac output state (see Chapter 11).
Etiology. The possibility of a significant clinical condition should always be considered in a patient with hypotension, although transient hypotension is usually of a benign etiology. It is best to correlate the patient’s cuff pressure with an arterial line pressure in the ICU before the catheter is removed, because a significant discrepancy can confound interpretation of cuff pressures on the postoperative floor. Concerns must always include the possibility of hypoxemia, myocardial ischemia/infarction with cardiogenic shock, an aortic dissection (if discrepant upper extremity pulses), sepsis, and especially delayed tamponade (see section G, pages 757–761). However, the more common causes of hypotension several days after surgery include:
Hypovolemia, usually from aggressive diuresis
β‐blockers or amiodarone used prophylactically to prevent AF
Arrhythmias, especially a rapid ventricular response to AF/flutter and then the medications used to treat them, which usually lower the blood pressure (β‐blockers, CCBs, amiodarone)
Vasoplegia from the residual effects of the systemic inflammatory response from CPB or from a diabetic autonomic neuropathy
Initiating too high a dose of the patient’s preoperative medications. Often the patient’s initial postoperative hypertension is related to pain and sympathetic overactivity, and once these resolve, hypotension may result, necessitating fluid resuscitation and unnecessary transfusions.
Assessment and management
Review of the patient’s medications, fluid status, orthostatic blood pressure measurements, heart rate and rhythm, pulse oximetry reading, 12‐lead ECG, and hematocrit should be sufficient to delineate the mechanism for hypotension. If the patient appears warm and well perfused, administration of a moderate amount of volume and modification of the medical regimen should suffice.
Management of hypotension associated with AF can be problematic, in that medications that slow the rate tend to cause vasodilation and lower the blood pressure further (especially metoprolol and diltiazem). Amiodarone is an excellent alternative, and usually causes hypotension only with rapid IV infusion. However, in most patients, rate control will improve left ventricular filling and the blood pressure. If the patient has refractory hypotension with a fast rate that is difficult to control pharmacologically, cardioversion may be indicated.
If the patient has refractory hypotension and does not appear well perfused, the likelihood of tamponade is increased and a STAT echocardiogram should be performed.
Myocardial ischemia. The development of recurrent angina or new ECG changes (usually ST segment elevation) postoperatively always requires careful evaluation for evidence of ischemia or myocardial infarction. Manifestations may include a low output state, hypotension, heart failure and pulmonary edema, ventricular arrhythmias, or cardiac arrest.
Coronary hypoperfusion from:
Acute thrombosis of a graft after CABG, either from an anastomotic problem, a kinked graft, or occasionally grafting to an incompletely endarterectomized artery
Reduced flow from anastomotic narrowing (technical issue) or hypoperfusion through a small conduit (for example, using a very small ITA graft or replacing a moderately diseased vein graft with a small ITA at reoperation)
Coronary spasm (graft or native vessel)
Nonbypassed, diseased coronary arteries either due to failure to locate the vessel, small vessel size, or severe calcification (incomplete revascularization)
Coronary ostial narrowing, occlusion or kinking after AVR or aortic root procedures with coronary button reimplantation
Circumflex artery compromise during mitral valve surgery
Poor myocardial protection during surgery
Careful review of the ECG may indicate whether the ECG changes are consistent with ischemia or pericarditis (see Figures 8.2 and 8.3 on pages 384 and 385).
Empiric use of a nitrate and/or CCB may be helpful for ischemia or spasm and can be diagnostic.
Urgent coronary arteriography should be considered when there are significant ECG changes.50 It may identify a technical problem with a graft or confirm the diagnosis of spasm.
In less urgent situations, a nuclear stress imaging study can be performed to identify the presence of myocardial ischemia and differentiate between ischemic and nonischemic causes of chest pain.
Intensification of a medical regimen with nitrates and ß‐blockers is indicated.
Placement of an intra‐aortic balloon pump (IABP) is beneficial for ongoing ischemia or evidence of hemodynamic compromise, especially in the immediate postoperative period. However, it should only be considered a supportive measure until the etiology of the problem is identified.
If a technical problem with a graft is identified by coronary angiography, or on occasion, there is failure to graft the correct artery or the graft is placed proximal to the most significant stenosis, percutaneous coronary intervention (PCI) is often the best treatment because it can be performed most expeditiously.50,51 If this is not feasible, but a major area of myocardium is in jeopardy and the patient has not suffered a significant perioperative myocardial infarction (PMI), reoperation should be considered. In contrast, medical management may be indicated if the coronary vessels supplying an ischemic zone are small and diffusely diseased, if they were bypassed but graft flow was limited by vessel runoff, or if they were not bypassable. Inability to address small vessels should only leave a minor area of the heart potentially ischemic. PCI of larger, more proximal stenotic segments may provide some benefit by improving inflow. If the patient has sustained an extensive infarction and has a delayed assessment, surgical intervention may not prove beneficial and may be high risk.
The long‐term results of coronary bypass surgery are influenced by the development of atherosclerotic disease in bypass conduits, nonbypassed native arteries, or native arteries beyond the bypass sites. Factors that can improve these results include use of arterial grafting (one or both ITAs and a radial artery), aggressive control of risk factors, including abstinence from smoking, statins for dyslipidemia, optimal control of hypertension and diabetes, and use of aspirin for at least one year.52 Supplemental use of a P2Y12 inhibitor may be beneficial in patients undergoing surgery for acute coronary syndromes or after OPCABs. On rare occasions, the late development of ischemia has been attributed to a coronary steal syndrome, either from a coronary–subclavian steal or an ITA–pulmonary artery fistula.53
Pericardial effusions and delayed tamponade. Pericardial effusions have been noted in about 60–75% of patients undergoing routine echocardiography in the first 1–2 weeks following surgery but usually resolve completely.54,55 Several large studies have shown that about 1–2% of patients may develop symptomatic effusions that gradually increase in size, leading to a low cardiac output state and tamponade that requires drainage.56,57 However, smaller series have suggested that invasive treatment for pretamponade or tamponade is indicated in about 5% of patients.58,59 This problem may be noted within the first week of surgery or weeks later. Suspicion must remain high because symptoms may develop insidiously and can be difficult to differentiate from those noted in patients recovering slowly from surgery. This is one of the most serious yet most potentially correctable of all postoperative problems.
Risk factors for development of pericardial effusions and delayed tamponade include:57–60
Use of perioperative antiplatelet drugs (aspirin, P2Y12 inhibitors) or other anticoagulants (heparin, LMWH, NOACs)
Early postoperative bleeding that requires blood product administration or re‐exploration
Early initiation of anticoagulation for VTE prophylaxis, mechanical valves, or the development of AF: a decision that must be made with caution as slow intrapericardial bleeding may occur despite minimal early postoperative bleeding.
Comorbidities including larger body surface area, hypertension, immunosuppression, chronic kidney disease (CKD), hepatic dysfunction
More advanced heart failure
Urgent surgery, more complex operations requiring longer durations of CPB (more coagulopathy), and surgery for endocarditis
Acute hemorrhage may occur during removal of ventricular pacemaker wires from laceration of a superficial artery or vein overlying the right ventricle or from the right ventricle itself. Some surgeons request that pacing wires be removed prior to chest tube removal, but most commonly they are removed a few days after the chest tubes have been removed. Because bleeding might be exacerbated by anticoagulation, it is advisable to remove pacing wires before the INR becomes therapeutic, by withholding 1–2 doses of a NOAC, or by stopping heparin for a few hours to minimize that risk. If the patient is therapeutically anticoagulated, it may be advisable to cut the wires rather than remove them (see also page 609). Bleeding during withdrawal of atrial pacing wires may occur if the wires are directly attached to the atrial wall rather than placed into a plastic sleeve that is sewn to the heart (the Medtronic model 6500 wires). Rarely, a patient may develop a delayed rupture of an infarct zone or LV rupture from a mitral valve prosthesis.
In patients with the above risk factors, especially those taking pre‐ or postoperative anticoagulants or antiplatelet drugs, a progressively worsening hemopericardium may develop and the presentation may be insidious. In a large series from the Mayo Clinic, nonspecific symptoms often led to performance of an echocardiogram, but 42% of patients had hemodynamic compromise consistent with tamponade. This report also noted that half of the patients who had insignificant effusions on echo after valve surgery were readmitted with tamponade within two weeks of discharge.56
Acute pericarditis is occasionally noted on a postoperative ECG and can be difficult to distinguish from an evolving MI (see Figure 8.3). This may be related to epicardial hemorrhage, undrained blood, or an early inflammatory response, but it is often of unclear etiology. It may be asymptomatic and noted only on an ECG and may contribute to the formation of serous or serosanguineous effusions.
Late serous or serosanguineous effusions may develop from PPS, which is considered one type of “postcardiac injury” syndrome.61
Acute hemorrhage will present with refractory hypotension and the clinical picture of acute cardiac tamponade.
Acute pericarditis may cause chest discomfort, but in the early postoperative period may be indistinguishable from the pain of a sternotomy incision.
The classic picture of delayed tamponade is a low output state manifest by malaise, shortness of breath, chest discomfort, anorexia, nausea, or a low‐grade fever. These symptoms are frequently ascribed to medications or simply a slow recovery from surgery. Jugular venous distention, a pericardial rub, progressive orthostatic hypotension, tachycardia (often masked by use of β‐blockers), and a pulsus paradoxus are often noted. Occasionally, the first sign is a decrease in urine output with a rise in the BUN and creatinine caused by progressive renal dysfunction from the low output state, arterial hypotension, and systemic venous hypertension.
A chest x‐ray may reveal enlargement of the cardiac silhouette, but this may be attributed to obtaining an AP portable film during a poor inspiratory effort. However, the chest x‐ray is often normal, depending on the site and rapidity of blood accumulation.
Two‐dimensional echocardiography can identify the pericardial effusion, confirm tamponade physiology (>40% inspiratory increase in tricuspid valve flow and >25% inspiratory decrease in mitral valve flow), and also assess the status of ventricular function. It is important to recognize that tamponade may be caused by selective compression of individual cardiac chambers, often by small effusions, and not necessarily by large circumferential effusions.57,62A transthoracic echocardiogram often has limitations in obtaining certain acoustic windows, which may be related to the patient’s body habitus. Thus, it will occasionally not identify an effusion.
If the clinical suspicion remains high and transthoracic imaging is suboptimal, CT scanning can be used to identify and localize a significant effusion. This is useful in patients who are several days out from surgery, hemodynamically stable, and on the postoperative floor.63 However, in unstable patients and those in the ICU, a transesophageal echocardiogram (TEE) is preferable and is more sensitive than a TTE in detecting posterior fluid collections (Figure 13.1).64
Stopping antiplatelet and anticoagulant medications at the appropriate time prior to surgery should not be overlooked.
Meticulous attention to obtaining hemostasis at the conclusion of surgery is essential to minimize the risk of bleeding, the use of blood products, and the development of a coagulopathy that may lead to delayed tamponade.
Performance of a small posterior pericardiotomy incision at the conclusion of surgery reduces the incidence of posterior pericardial effusions.65 It is also likely that placing one of the mediastinal tubes below the heart, rather than two tubes anteriorly, might improve drainage and reduce the incidence of residual effusions.
Chest tubes can be removed when drainage for the preceding eight hours is <100 mL. One study found that leaving chest tubes in at least until the second postoperative day and removing them only after the drainage was <50 mL/4h more than halved the incidence of delayed tamponade.66
Early initiation of anticoagulation following valvular heart surgery, for VTE prophylaxis, or for the management of AF always requires careful judgment as to whether the patient might be at higher risk for mediastinal bleeding and the development of delayed tamponade.
Prophylactic use of NSAIDs may reduce the incidence of postoperative pericardial effusions.67 A number of medications have been successful in reducing the risk of postoperative PPS (see next section), but none can be recommended for routine use for that purpose. However, colchicine, anti‐inflammatory medications and/or steroids might be utilized in treating effusions due to pericarditis or PPS which have not produced hemodynamic compromise.
Remove pacing wires only in patients who are not therapeutically anticoagulated.
Emergency mediastinal exploration is indicated for active bleeding. If the patient is bleeding massively and/or is very unstable, this should be performed in the ICU or even at the bedside on the postoperative floor with equipment available in an “open chest kit”. If the patient can be stabilized, but there is suspicion of an active bleeding source, transferring the patient to the operating room for re‐exploration is preferable.
Pericardiocentesis is the least invasive means of draining a progressively enlarging effusion and is effective in about 50% of cases, usually when the effusion is anterior or circumferential.56,57 This is usually performed in the cardiac catheterization laboratory under ECG or two‐dimensional echocardiographic guidance. This will not be completely effective if the blood has clotted with loculated strands.
Subxiphoid exploration should be considered when the echocardiogram suggests that the fluid collection cannot be successfully approached percutaneously (usually a posterior collection) or when it is loculated. If this approach is ineffective in draining the effusion, the entire sternal incision may need to be opened.
A pericardial “window” or limited pericardiectomy through a left thoracotomy approach can be considered for loculated posterior effusions or recurrent effusions several weeks after surgery.
Postpericardiotomy syndrome (PPS) has been reported in 10–20% of patients following open‐heart surgery and is considered to represent an autoimmune inflammatory response to injury, often associated with the occurrence of perioperative bleeding and the requirement for blood transfusions.68,69 It is associated with an elevation in inflammatory markers, including cytokines, markers of neutrophil activation, and oxidative stress mediators.70 It is one of several forms of “postcardiac injury” syndrome which may also occur after pacemaker implantation, PCI, transmural myocardial infarctions, or radiofrequency arrhythmia ablation, presumably as a response to cardiac injury.61 It may occur within the first week of surgery or several weeks to months later and appears to have a different etiology than the pericarditis and pericardial effusions noted early after surgery. The development of PPS may contribute to cardiac tamponade, early vein graft closure, or constrictive pericarditis.71
Risk factors for PPS include younger age, nondiabetics, CKD, lower hematocrits or platelet counts preoperatively (which most likely accounts for more perioperative bleeding and the need for transfusions), and valve or aortic surgery.68,69,72 A low preoperative level of interleukin‐8 is a high‐risk marker for its development.73
Presentation. Fever, pleuritic chest pain, a pericardial friction rub, or a new or worsening pleural or pericardial effusion may be present, with the diagnosis of PPS being made if two or more of these are present. Malaise and arthralgias may also be present.
Prevention. Colchicine (1 g bid × 1 day, then 0.5 mg bid × 1 month) started after surgery has seen the most promise in preventing PPS.74,75 The efficacy of prophylactic intraoperative steroids is uncertain76 – one study showing a benefit of methyprednisolone77 and another showing no benefit of dexamethasone.78 Diclofenac (representative of the NSAIDs) has also been effective in reducing the occurrence of postoperative PPS.79
Evaluation. Lymphocytosis, eosinophilia, and an elevated ESR are noted, but a fever work‐up is negative. Effusions are usually demonstrable by chest x‐ray and echocardiography.
The best initial treatment for PPS is a combination of aspirin (750 mg tid × 2 weeks) and colchicine (0.5 mg bid × 6 months), given with a PPI.61 If there is minimal symptomatic relief or as an alternative to aspirin, a one‐ to two‐week course of an NSAID, such as ibuprofen 600 mg tid, can be given with cessation of aspirin. It should be noted that the regular use of NSAIDs (especially ibuprofen) may inhibit the cardioprotective antiplatelet effect of aspirin.80
For patients with recurrent PPS, a one‐month tapering course of prednisolone starting with 0.5 mg/kg daily can be added to the regimen of colchicine and aspirin.61
Colchicine (0.5 mg bid × 1 month) is recommended for the treatment of acute and recurrent pericarditis.75,81
Pericardiocentesis may be necessary to drain a large symptomatic pericardial effusion.
Pericardiectomy is recommended for recurrent large effusions or constriction, which is more common in younger patients with the early onset of PPS.
Constrictive pericarditis is a late complication of cardiac surgery that is extremely uncommon despite the development of adhesions that form within the mediastinum following surgery. It has been noted in patients with undrained early postoperative hemopericardium, use of warfarin, early PPS, and previous mediastinal radiation, yet has also developed in the absence of any of these factors. It has been theorized that residual blood and proinflammatory conditions in the pericardium may contribute to the formation of both PPS and later the formation of dense adhesions that more readily cause thickening of the epicardium. This may produce ventricular constriction (“constrictive epicarditis”) and also result in graft failure.70,82
Presentation. The patient will note the insidious onset of dyspnea on exertion, chest pain, and fatigue. Signs of right‐heart failure (peripheral edema, ascites, and jugular venous distention) are common, but pulsus paradoxus is infrequent.
The chest x‐ray is frequently normal in the absence of a pericardial effusion.
Two‐dimensional echocardiography will demonstrate signs of constriction, such as septal bounce and diminished respiratory variation in the inferior vena cava.
A CT or preferably an MRI scan usually documents a thickened pericardium, with delayed enhancement on MRI and occasionally a small pericardial effusion. However, in cases of constrictive epicarditis, pericardial thickening may not be seen.
Right‐heart catheterization provides the most definitive information. It will document the equilibration of diastolic pressures and demonstrate a diastolic dip‐and‐plateau pattern (“square‐root” sign) in the right ventricular pressure tracing (see Figure 1.32). On occasion, significant fluid overload will produce hemodynamics consistent with constriction, when in fact there is no anatomic evidence of a thickened pericardium or epicarditis other than standard postoperative scarring.
Treatment. If there is no clinical response to diuretics and steroids, a pericardiectomy is indicated to decorticate the heart. This is best performed through a sternotomy incision, which allows for adequate decortication of the right atrium and ventricle and much of the left ventricle. It also allows for the institution of CPB in the event of a difficult or bloody operation. Relief of epicardial constriction is difficult and may result in surgical damage to bypass grafts or significant bleeding. A “waffle” or “turtle shell” procedure is performed with crisscrossing incisions made in the epicardial scar to relieve the constriction.82 Results are suboptimal when there is poor LV systolic function, higher RV and LV diastolic pressures, persistent impaired diastolic filling, which correlates with the duration of symptoms prior to surgery, and when significant tricuspid regurgitation is present, which usually will not improve with pericardiectomy alone.
Table 13.4 Cardiac Complications of Cardiac Surgery
VI. Renal, Metabolic, and Fluid Management and Complications
Most patients are still substantially above their preoperative weight when transferred to the postoperative floor. Comparison of the patient’s preoperative weight with daily weights obtained postoperatively is a guide to the use of diuretics to eliminate excess fluid. Achievement of dry body weight may require more aggressive diuresis if heart failure (HF) was present before surgery. In the chronically ill patient, preoperative weight may be achieved after several days despite fluid overload, due to poor nutrition.
Dietary restriction (sodium and water) need not be overly strict in most cases. With the availability of potent diuretics to achieve negative fluid balance and the common problem of a poor appetite after surgery, it is more important to provide palatable food without restriction to improve the patient’s caloric intake.
If a patient required diuretics before surgery (especially valve patients and those with poor myocardial function), it is advisable to continue them upon discharge from the hospital even if preoperative weight has been attained.
Transient renal failure (see also Chapter 12). Patients with preoperative renal dysfunction, hypertension and diabetes, prolonged pump runs, postoperative low cardiac output syndromes, or those requiring substantial doses of vasopressors may develop postoperative AKI. Although diuretics are routinely used to reduce the immediate postoperative fluid overload, they have to be used with caution in patients developing AKI. They do not influence the course of AKI and can in fact exacerbate renal dysfunction by causing prerenal azotemia from intravascular volume depletion. Management can be very difficult on the postoperative floor when methods of monitoring intravascular volume are limited.
A common scenario with mild AKI is gradual elevation in serum creatinine (SCr) with or without elevation in the BUN, along with a low serum sodium, reflective of persistent total body water overload. Oliguria may improve and AKI may be transient with adjustment of medications. Antihypertensive medications should be reduced to allow the normally hypertensive patient’s blood pressure to rise to higher levels than normal. ACE inhibitors should be withheld, NSAIDs avoided, and diuretics used gently, if at all, to maintain adequate intravascular volume. If the patient was aggressively diuresed and has a poor appetite, additional hydration may be necessary to address prerenal azotemia which will elevate the BUN. In most patients, renal dysfunction is transient as long as the cardiac output remains satisfactory.
If a rising SCr is associated with significant fluid retention, compromise of pulmonary function with oxygen desaturation often results. A high dose of diuretics may improve urine output to address this problem even if it does not directly promote renal recovery. If not successful, the patient may need to return to the ICU for more invasive monitoring and possibly use of intravenous inotropic support, more aggressive noninvasive or even mechanical ventilation, and ultrafiltration or dialysis if renal dysfunction is significant. A rising BUN and creatinine of unclear etiology, especially when associated with new‐onset oliguria, should always raise the suspicion of delayed tamponade. An echocardiogram should be performed to assess myocardial function and look for possible cardiac tamponade.
Hyperkalemia usually occurs in association with renal dysfunction. Its manifestations and treatment are discussed on pages 708–711. Particular attention should be directed to stopping any exogenous potassium intake, ACE inhibitors ARBs, and NSAIDs, and reevaluating renal function.
Hyperglycemia in diabetics, and occasionally in nondiabetics, is a common postoperative problem. The blood glucose level may be elevated due to insulin resistance and residual elevation of the counterregulatory hormones (glucagon, cortisol) after surgery.83 Adequate, but not overly stringent, control of blood glucose during the early postoperative period with an IV insulin protocol has been shown to reduce not only the incidence of wound infection but also other morbidities and operative mortality (see Appendix 6).84–87 Once the patient is transferred to the floor, frequent fingersticks should be obtained (usually before meals and at bedtime) to assess the adequacy of blood glucose control.
Insulin resistance is commonly noted during the early postoperative period. Insulin‐dependent diabetics should have their insulin doses gradually increased back to preoperative levels depending on oral intake and blood glucose levels. It is preferable to use a lower dose of intermediate‐acting insulin initially and supplement it with regular insulin as necessary (see Table 12.9, page 721, for commonly used insulin preparations).
Oral hypoglycemics can be restarted once the patient has an adequate oral intake, usually starting at half the preoperative dose, and increasing the dose depending on oral intake and blood glucose.
Other electrolyte and endocrine complications are fairly unusual once the patient has been transferred to the postoperative floor. Chapter 12 discusses the evaluation and management of some of these problems.
VII. Hematologic Complications and Anticoagulation Regimens
Despite the obligatory hemodilution associated with use of CPB, effective blood conservation strategies and performance of off‐pump surgery have reduced the requirements for perioperative blood transfusions. Although the STS guidelines recommend transfusion for a hemoglobin (Hb) <6 g/dL on CPB in low‐risk patients,88 such a restrictive strategy may increase the risk of renal failure, stroke, and neurocognitive dysfunction, and should not be applied to patients who are elderly, diabetic, have cerebrovascular disease, or are at risk for end‐organ ischemia.
Most studies recommend maintaining a HCT above 20% on CPB and then at least 22% postoperatively. Multiple studies comparing a liberal (transfuse for a Hb <7 g/dL) and restrictive (transfuse for a Hb <9 g/dL) strategy after cardiac surgery have shown comparable outcomes.89,90 However, transfusion to a higher hematocrit postoperatively may be considered for elderly patients, those who feel significantly weak and fatigued, and those with ECG changes, hypotension, or significant tachycardia.
Although the HCT may rise gradually with postoperative diuresis, it frequently will not as fluid is mobilized into the bloodstream from extracellular tissues. Furthermore, the HCT may be influenced by the shortened red cell lifespan caused by extracorporeal circulation and the loss of 30% of transfused red cells within 24 hours of transfusion. In one study, “hemoglobin drift” was noted in virtually all patients, averaging about 1.1 mg/dL, was greater with longer durations of CPB, but was unrelated to whether the patient received a transfusion.91 However, with diuresis, the Hb level improved in nearly 80% of patients prior to discharge, and this concept should be taken into consideration when contemplating a transfusion.
Any patient with a HCT <30% should be placed on iron therapy (ferrous sulfate or gluconate 300 mg tid for one month) at the time of discharge. Exogenous iron may not be necessary if the patient has received multiple transfusions, because of the storage of iron from hemolyzed cells.
Consideration may also be given to use of recombinant erythropoietin (Epogen or Procrit) to stimulate red cell production (50–100 units/kg SC three times a week), especially in patients with CKD with adequate iron stores (transferrin saturation >20% and ferritin >100 ng/mL).
Thrombocytopenia is caused by platelet destruction and hemodilution during extracorporeal circulation, but platelet counts gradually return to normal within several days. Impaired hemostasis noted in the early postoperative period is caused more commonly by platelet dysfunction induced by CPB or use of antiplatelet medications, although it is attenuated somewhat by the use of the antifibrinolytic drugs.
Platelet activation or dilution during CPB
Excessive bleeding and multiple blood transfusions without platelet administration
Use of an intra‐aortic balloon pump
Heparin‐induced thrombocytopenia (HIT). Note: platelet counts must be monitored on a daily basis in any patient receiving heparin. A falling platelet count after the initial recovery of the platelet count should always raise the specter of HIT and is an indication for in vitro aggregation testing to identify HIT.
Other medications that may reduce the platelet count, such as furosemide, NSAIDs, and ranitidine
Thrombotic thrombocytopenic purpura (TTP), which is usually characterized by AKI, thrombocytopenia, and a microangiopathic hemolytic anemia with schistocytes on blood smear. Occasionally, fever and altered mental status may be present.92
Treatment. Platelet transfusions are indicated:
When the platelet count is <20,000–30,000/μL (<20−30 × 109/L).
For ongoing bleeding when the platelet count is <100,000/μL and sometimes higher if platelet dysfunction is suspected.
For a planned surgical procedure (such as percutaneous IABP removal) when the platelet count is <60,000/μL (<60 × 109/L).
Heparin‐induced thrombocytopenia (HIT) is a very serious problem that may result in widespread arterial and venous thrombosis, and carries a mortality rate of about 20%. Because of its high risk and the necessity for treatment with direct thrombin inhibitors, early suspicion, identification, and management are essential.93–96
HIT is an immune‐mediated phenomenon caused by the formation of IgG antibodies that bind to the heparin‐platelet factor 4 (PF4) complex, producing platelet activation. This results in release of procoagulant microparticles that lead to thrombin generation. This binding causes release of more PF4, promoting more platelet activation. Antibody binding to glycosaminoglycans on the surface of endothelial cells leads to endothelial cell damage and tissue factor expression. This procoagulant milieu promotes arterial and venous thrombosis in 30–50% of patients and may cause a stroke, myocardial infarction, mesenteric thrombosis, or deep venous thrombosis.
Suspicion of the diagnosis
The diagnosis of HIT requires the presence of heparin antibodies and thrombocytopenia. HIT is more common with use of bovine heparin and is 8–10 times more likely to occur with UFH than LMWH.
The suspicion of HIT is based on the “four Ts”, which include the degree of Thrombocytopenia, the Timing of its occurrence, the occurrence of Thrombotic events, and the likelihood of oTher potential causes. One scoring system found that independent risk factors for HIT after CPB were a biphasic response to the platelet count, an interval of >5 days since CPB, and a bypass run exceeding about two hours.97
Thrombocytopenia is extremely common postoperatively and may be related to hemodilution, platelet damage on pump, clearance of transfused platelets, and sepsis. The platelet count is usually reduced about 40% after surgery on CPB, but begins to rebound by the third or fourth day after surgery. If the platelet count does not improve after four days or there is a subsequent fall in platelet count >50%, HIT should be suspected.
Although immediate‐onset HIT (occurring within hours of giving heparin) may develop in patients who received heparin within 100 days due to residual circulating HIT antibodies, it is rare for a patient receiving short‐term preoperative heparin to develop HIT within the first four days after surgery. The general pattern is for HIT to occur 5–14 days after surgery.
Delayed‐onset HIT (occurring after heparin has been discontinued due to the presence of residual heparin‐PF4 platelet‐activating antibodies) is often manifested by venous thromboembolism, and is commonly not diagnosed, because platelet counts are rarely checked after heparin is stopped. If there is evidence of a thrombotic event (such as lower extremity VTE), a low platelet count should raise suspicion of this entity.
It is estimated that 30–50% of patients who develop HIT will develop evidence of thrombosis, mandating additional treatment beyond stopping heparin. Disturbingly, it is estimated that 25% of patients may develop thrombosis before developing thrombocytopenia.98 Thus, HIT can be a very difficult management problem because of the temptation to administer heparin for a thrombotic event when it might be contraindicated. More commonly, however, patients will develop thrombosis as the platelet count is falling and reaching its nadir, but, as noted, it may occur even after the heparin has been stopped (delayed‐onset HIT). Thus, there are clinical scenarios when HIT may be present but the suspicion is low, leading to complications from delayed recognition and management.
Other potential causes, especially medications, use of an IABP, and sepsis should be considered as possible causes of thrombocytopenia.
Diagnostic testing. ELISA serologic testing for IgG specific heparin‐dependent platelet antibodies has replaced prior testing that also detected non‐HIT‐causing IgM antibodies. Such testing was positive in about 20% of preoperative and in up to 50% of postoperative cardiac surgical patients, but was uncommonly associated with thromboembolic events.94,99–102 It is estimated that only 1–2% of patients undergoing cardiac surgery will develop HIT. Use of optical densities (ODs) with ELISA testing and antibody titers may be helpful in improving specificity as well, since most cases of HIT are associated with an OD >1.4 units.95 More specific functional assays of washed platelet activation (serotonin release assay and heparin‐induced platelet aggregation studies) are able to more accurately identify antibodies that trigger platelet activation. Preoperative testing for heparin antibodies is not recommended in the absence of thrombocytopenia or thrombosis.
If HIT is identified preoperatively with positive testing and thrombocytopenia, an alternative method of anticoagulation (usually bivalirudin) must be used during surgery (see Chapter 4, pages 251–253). Antibodies generally clear within three months of the last heparin administration, and heparinization can be performed safely if HIT testing is negative.
With any suspicion of postoperative HIT, even before testing results become available, all heparin administration must be stopped. This includes cessation of heparin flushes and removal of heparin‐coated pulmonary artery catheters. However, if the likelihood of HIT is low based upon clinical grounds, yet an anticoagulant is indicated for prophylaxis (such as for VTE), fondaparinux 5–10 mg qd SC (weight‐based) can be given safely.
Platelets should not be routinely administered, because they may promote thrombosis, but may be considered if the patient is bleeding or to prevent bleeding if the platelet count is very low (<30,000). The platelet count generally begins to increase within about a week after cessation of heparin.
Warfarin should not be started immediately, because tissue necrosis from microvascular thrombosis may occur due to depletion of the vitamin‐K‐dependent natural anticoagulant protein C. This has been noted in patients who rapidly develop a supratherapeutic INR. Warfarin may be started safely after the platelet count reaches 150 × 109 /L and should overlap the nonheparin anticoagulant for five days, even when the INR is in therapeutic range.
Alternative anticoagulation is indicated to minimize the risk of thrombotic events and also to provide protection for the process for which the heparin was originally indicated. This is essential because the risk of symptomatic thrombosis remains greater than 25% if only the heparin is stopped.
Argatroban is a synthetic direct thrombin inhibitor that is preferred in patients with renal dysfunction because it undergoes hepatic metabolism. It has a plasma half‐life of about 40 minutes. It is given starting at a dose of 2 μg/kg/min once heparin effect has been eliminated (usually four hours for UFH and 12 hours after the last dose of LMWH), and maintained at a rate of 0.5–1.2 μg/kg/min and adjusted to maintain a PTT of 1.5–3 times baseline. Lower doses are recommended in patients with liver disease or heart failure. Conversion from argatroban to warfarin can be somewhat problematic because both affect the INR. Warfarin should be started once the platelet count exceeds 150 × 109 and is given in doses of 2.5–5 mg for five days of overlapping treatment. At that point, the usual protocol is to stop the argatroban if the INR is >4 and recheck an INR in four hours. If the new INR is >2, it does not need to be restarted; if it is <2, the argatroban should be restarted. However, it has been noted that even when the INR is >4, the risk of thrombosis may still exceed the risk of bleeding.104
Bivalirudin is a direct thrombin inhibitor that produces reversible binding to thrombin and has a short half‐life of only 25 minutes. It can be used for the management of HIT with an initial bolus of 0.75 mg/kg followed by an infusion of 0.15 mg/kg/h to achieve a target PTT of 1.5–2.5 times baseline. Advantages include 80% enzymatic metabolism (although some modification is indicated in patients with a GFR <30 mL/min), non‐immunogenicity, and minimal effect on the INR.
Danaparoid is a heparinoid that has a low degree of reactivity with heparin antibodies. It is given in a bolus dose of 2250 anti‐Xa units, followed by 400 units/h × 4 hours, then 300 units/h × 4 hours, then a maintenance infusion of 200 units/h. It is monitored by anti‐Xa levels, trying to achieve a level of 0.5–0.8 units/mL. It is available from Aspen Pharmacare, but is not available in the USA.
Non‐vitamin K antagonist oral anticoagulants (NOACs) have had limited evaluation in the management of HIT, but may be considered off‐label as alternative medications using the doses recommended for acute VTE (see section IV.E, pages 750–752).
Duration of antithrombotic therapy is generally one month without and three months with the occurrence of a thrombotic event.
Coronary bypass surgery. Postoperative aspirin should be utilized to increase saphenous vein graft patency, although there is no documented benefit in improving the patency of arterial conduits.52
Class I recommendations for antiplatelet therapy after CABG in most guidelines (American College of Chest Physicians [ACCP], American College of Cardiology/American Heart Association [ACC/AHA], the European Society of Cardiology [ESC], and the Society of Thoracic Surgeons [STS]) are to start aspirin in doses ranging from 75 mg to 325 mg starting 6–24 hours after surgery and to also prescribe clopidogrel 75 mg after OPCAB.105–108 The use of higher doses of aspirin may offset the increase in platelet reactivity and the decreased absorption of aspirin noted postoperatively.23 Non‐enteric‐coated aspirin has better absorption and is recommended. The timing of initiation of therapy may be influenced by the degree of postoperative mediastinal bleeding. It is recommended to continue aspirin indefinitely due to its benefits in the secondary prevention of coronary disease, although it may not influence graft patency when used beyond one year.
Class IIa recommendations are to prescribe dual antiplatelet therapy (DAPT) with aspirin and a P2Y12 inhibitor (preferably ticagrelor) in patients who have undergone CABG for an acute coronary syndrome (ACS). Several meta‐analyses have suggested that DAPT after CABG for stable ischemic heart disease (SIHD) may reduce vein graft occlusion, with the greatest benefit noted following OPCABs (hence the level I recommendation), although it may be associated with more bleeding.109,110 However, DAPT did not improve cardiovascular outcomes in diabetic patients with multivessel disease undergoing CABG for SIHD or an ACS in the FREEDOM trial.111 DAPT is only a class IIb indication following on‐pump CABG per the ACC/AHA guidelines. If the patient is aspirin‐intolerant, clopidogrel should be given with a 300 mg loading dose, followed by 75 mg daily.
DAPT should be resumed after surgery if the patient has received a drug‐eluting stent within the past year and it remains patent.
There is no documented superiority in graft patency comparing aspirin to warfarin or to a NOAC (rivaroxaban) whether used alone or in addition to aspirin.112 Whether one of these might be considered after coronary endarterectomy is not known.
Prosthetic heart valves. All patients receiving mechanical heart valves must take a vitamin K antagonist (VKA) indefinitely following surgery, but the guidelines regarding the short‐term use of a VKA after tissue valve replacement provide differing recommendations.113 The following recommendations (summarized in Table 13.5) are a general consensus from the 2012 ACCP,114 2014/2017 ACC/AHA,115 and the 2017 ESC guidelines,116 all of which are updated every few years and are available online. The evolution and comparison of these guidelines are presented in a paper published in 2019.117
Aortic valves (surgical). Some reviews have concluded that use of either aspirin or a VKA provides similar outcomes after bioprosthetic AVR, and the slight decrease in thromboembolism and death from the combination of both is offset by an increased risk of bleeding.118 Although surgeons and patients often opt for a tissue valve to avoid a VKA, even for a short period of time, there is some evidence that patients not receiving VKAs are more susceptible to thromboembolic complications during the first three months after implantation, with a slightly increased risk of stroke.119 Furthermore, “subclinical leaflet thrombosis” has been identified on aortic tissue valves implanted surgically or with TAVR and has been associated with an increased risk of stroke after surgical AVR.120
The 2012 ACCP guidelines recommend aspirin 75–100 mg (usually 81 mg) indefinitely if the patient is in sinus rhythm, with the addition of warfarin to achieve a target INR of 2.5 if the patient is at higher risk for thromboembolism (AF, hypercoagulable disorder, EF <35%, previous thromboembolism, or a left atrial dimension >50 mm).114
The 2017 AHA/ACC guidelines give level IIa indications to the indefinite use of aspirin and to the use of a VKA for 3–6 months with a target INR level of 2.5 if the patient is at low risk for bleeding. However, the guidelines do not specify whether the recommendation is to use either alone or both together. After 3–6 months, only aspirin is recommended unless there is another indication for warfarin.115 New guidelines should be available online towards the end of 2020.
The 2017 ESC guidelines provide a class IIa indication for use of aspirin and a IIb indication for use of an oral anticoagulant (without concomitant use of aspirin), because the addition of a VKA will increase the bleeding risk despite a lower risk of thromboembolism.116 In contradistinction to the AHA/ACC guidelines, the ESC guidelines do not recommend the indefinite use of aspirin.
Limited data have shown that a NOAC is noninferior to the use of warfarin in patients with AF and prior bioprosthetic valve replacement or repair. Therefore, it may be feasible to consider use of a NOAC after three months if there is an indication for continuing anticoagulation, such as AF.121
Aortic valves (TAVR). Standard protocols following TAVR recommend clopidogrel 75 mg daily for six months and aspirin 81 mg daily indefinitely. For patients who are at high risk for bleeding on DAPT, there are data showing that outcomes are comparable with mono and dual antiplatelet therapy.118,122,123 For a patient in AF, either a VKA or a NOAC is used and may be combined with one antiplatelet medication, but not both.
Concerns about subclinical leaflet thrombosis causing an increased gradient across the valve and an increased risk of stroke have raised the question as to whether a VKA should be used in these patients. This has received an ACC/AHA level IIb indication in patients at low risk for bleeding.115 This is probably less of an issue in elderly patients in whom anticoagulation is more dangerous, but may be applicable to younger, lower‐risk patients undergoing TAVR.
A NOAC with aspirin cannot be recommended as an alternative approach based on the GALILEO study, which found a higher risk of bleeding, thromboembolism, and death with rivaroxaban + aspirin compared with clopidogrel + aspirin in patients with no other indication for anticoagulation.124 However, this study did not assess whether use of a NOAC alone might reduce subclinical leaflet thrombosis without increasing risks compared with warfarin or DAPT in patients without AF.
Mitral valves. Warfarin should be started by the first postoperative day and given along with aspirin (ACC/AHA guidelines level IIa) for 3–6 months to achieve a target INR of 2.5 (range 2.0–3.0). ESC guidelines provide a IIa recommendation for three months of a VKA without aspirin. The initiation of heparin can be considered in the hospital before the INR reaches the target level. The patient can then be discharged on LMWH until the INR reaches the therapeutic range. After three months, warfarin is stopped and aspirin 75–100 mg (usually 81 mg) is given if the patient is in sinus rhythm. However, warfarin should be continued indefinitely in patients at high thromboembolic risk, and aspirin should be added to the regimen. NOACs may be feasible after three months in patients with AF.113,121
Mitral rings. The benefit of using a VKA rather than aspirin after mitral valve repair is not clear. ACCP guidelines recommend aspirin alone if the patient is in sinus rhythm, whereas the ACC/AHA and ESC guidelines recommend use of a VKA for three months. This conclusion is based primarily on evidence that about 30% of patients discharged in sinus rhythm will experience AF shortly thereafter, but is not based upon thromboembolic risk from the prosthetic ring. However, some studies suggest that aspirin alone may be sufficient.125
Mechanical valves. Based on the RE‐ALIGN study published in 2013, dabigatran, and inferentially the factor Xa inhibitor NOACs, are contraindicated as anticoagulants in patients receiving mechanical valves.126
Aortic valves. Current‐generation bileaflet valves (Abbott St. Jude, Medtronic ATS) should receive warfarin indefinitely starting by the day after surgery to achieve a target INR of 2.5 (range 2.0–3.0). In patients with older valves, which carry higher thromboembolic risk (Starr‐Edwards or Bjork‐Shiley valves), and in patients with double mechanical valves, the target INR should be 3.0 (range 2.5–3.5). The ESC guidelines recommend increasing the target INR by 0.5 if the risk factors noted in section 1.a.i on page 771 are present. The target INR for On‐X aortic valves is also 2.5, but may be lowered to 1.5–2 after three months.
Mitral valves. Patients should receive warfarin indefinitely starting by the day after surgery to achieve a target INR of 3.0 (range 2.5–3.5).
The ACC/AHA and ACCP guidelines recommend the addition of aspirin 75–100 mg daily to warfarin for all patients receiving mechanical valves. Because the evidence supporting this recommendation is weak and the risk of bleeding is unequivocally increased, the ESC recommends the addition of aspirin only if vascular disease, coronary stenting, or recurrent embolism is present.116,117
In patients receiving mechanical valves or those in AF receiving tissue mitral valves, there is a potentially increased risk of early postoperative thromboembolism when the patient is not therapeutically anticoagulated. Therefore, use of heparin is recommended until the INR becomes therapeutic. However, the timing of initiation of heparin as a bridge is not well defined and must be individualized. Although some groups advocate initiating heparinization as early as the first postoperative day, this may increase the risk of bleeding. A safe approach is to initiate UFH or LMWH on the fourth or fifth postoperative day if the INR is less than 1.8. It is recommended that either of these be continued until the INR has been therapeutic for two days. The ESC recommends anti-factor Xa monitoring if LMWH is used to ensure adequate anticoagulation.
Dosing and overanticoagulation. Warfarin is a dangerous drug that requires thoughtful administration and careful monitoring to avoid overanticoagulation.127
Initiation of warfarin results in the more rapid depletion of factors VII, IX, and X than factor II (prothrombin), which has a longer half‐life. Thus, it exhibits an antihemostatic effect before it achieves an antithrombotic effect, the latter being attributable primarily to a reduction in factor II. Following cardiac surgery, it is essential that warfarin not be loaded and that doses be carefully individualized to avoid rapid overanticoagulation. An initial dose of 5 mg is given to most patients. However, 2.5 mg should be given to small elderly women, patients with hepatic dysfunction, chronic illness, and those receiving antibiotics or amiodarone (Table 13.6). The INR generally begins to rise in 2–3 days, but often takes 5–7 days to achieve a stable dosing level.
Potential dangers of overanticoagulation include cardiac tamponade from intrapericardial bleeding, and gastrointestinal (GI), intracranial, or retroperitoneal hemorrhage. Although there are antithrombotic benefits to the combined use of warfarin and aspirin, which is generally recommended for valve patients, this combination does increase the long‐term risk of bleeding for patients with a variety of indications for anticoagulation.128 An ACC expert consensus pathway was written in 2017 for the management of bleeding related to oral anticoagulation.127 A brief protocol for the management of overanticoagulation with warfarin is noted in Figure 13.2 and in Appendix 9.
If the patient has significant bleeding with a markedly elevated INR, warfarin should be held, and fresh frozen plasma (up to 15 mL/kg) should be given. Vitamin K 5–10 mg IV in 50 mL NS over 30 minutes should be given and may be repeated every 12 hours for persistent INR elevation. Use of four‐factor prothrombin complex concentrate (4F‐PCC) 25–50 units/kg or recombinant factor VIIa 40 μg/kg may also be considered.
If the patient has no evidence of bleeding, general recommendations for management of elevated INRs are as follows, although rapid elevation of the INR in the early postoperative period may warrant more aggressive therapy.
INR ≥9: hold warfarin, give vitamin K 2.5–5 mg orally (INR should fall within 24–48 hours)
INR ≥5 and <9: hold warfarin for 1–2 days and restart when INR is <4; alternatively, omit one dose of warfarin and give 1–2.5 mg of vitamin K orally
INR above therapeutic range but ≤5: lower or omit one dose until at therapeutic range
Vitamin K given in small oral doses can reduce the INR to a therapeutic level within a few days and is usually the best approach when withholding warfarin does not reduce the INR. Use of large doses of IV vitamin K will produce more rapid reversal of the INR, but it is associated with a small risk of anaphylaxis. It may also lead to warfarin resistance and generally should be avoided. If the INR becomes subtherapeutic (which is usually safer than a markedly elevated INR), heparin can always be given until the INR rises back to the therapeutic range. In patients on long‐term warfarin with variable INR responses, concomitant administration of vitamin K 100–200 μg/day orally helps stabilize the INR.129,130
For patients taking a NOAC, bleeding can be addressed using reversal agents:
For dabigatran, give idarucizumab 5 g IV (Praxbind); if not available, give 4F‐PCC 25–50 units/kg.
For factor Xa inhibitors (apixaban, rivaroxaban), give recombinant factor Xa (andexanet alpha [Andexxa]) 400–800 mg at 30 mg/min. If not available, give 4F‐PCC 25–50 units/kg.
Prothrombin complex concentrates contains the vitamin‐K‐dependent coagulation factors II, IX, and X (so‐called three‐factor PCC [Profilnine]) and may contain variable amounts of factor VII (four‐factor PCC [Kcentra].Another product called FEIBA (anti‐inhibitor coagulant complex or activated PCC) may be the best product to control bleeding. This contains nonactivated factors II, IX, and X, activated factor VII, factor VIII inhibitor bypassing activity, and some factor VIII coagulant antigen. It is usually given in a dose of 500–1000 units.
Table 13.5 Recommended Anticoagulation Regimens for Prosthetic Heart Valves
Adapted from Whitlock et al. Chest 2012;141(2 Suppl):e576S–600S114, Nishimura et al., Circulation 2017;135:e1159–95 and J Am Coll Cardiol 2017;70:252–89115, Baumgartner et al., Eur Heart J 2017;378:2739–91116.
INR target 2.5 for 3 months if:
Risk factors (ACCP)
All patients (AHA/ACC IIa) × 3–6 months
All patients (ESC IIb) × 3 months)
Possible use of a NOAC after 3 months
Aspirin 75–100 mg alone if no risk factors (ACCP) Aspirin 75–100 indefinitely (AHA/ACC IIa) Aspirin 75–100 mg alone x 3 months (ESC IIa)(ESC IIa)
INR target 2.5 indefinitely
Aspirin 75–100 mg (ACC/ACCP); only if atherosclerotic disease or history of thromboembolism (ESC)
Mitral valve repair
INR target 2.5 for 3 months (ACC/ESC)
Aspirin 75–100 mg alone (ACCP)
INR target 2.5 for 3–6 months (AHA/ACC IIa) or 3 months (ESC IIa) Continue indefinitely if risk factors Possible use of a NOAC after 3 months
Aspirin 75–100 mg with warfarin × 3 months (ACC IIa) Aspirin 75–100 mg after warfarin is stopped
INR target 3.0 indefinitely
Aspirin 75–100 mg (ACC/ACCP); only if atherosclerotic disease or history of thromboembolism (ESC)
INR target 3.0 for 3 months Possible use of a NOAC after 3 months
Aspirin 325 mg after 3 months
INR 3.0–4.5 indefinitely
Aspirin 75–100 mg
AF with any of above
Continue warfarin indefinitely Possible use of a NOAC if tissue valve after 3 months
Risk factors: hypercoagulable state, history of systemic thromboembolism, ejection fraction <35%, history of anteroapical infarction, atrial fibrillation
ACCP, American College of Chest Physicians recommendations 2012; ACC/AHA, American College of Cardiology/American Heart Association recommendations 2014; ESC, European Society of Cardiology recommendations 2017
Table 13.6 Protocol for Initiation of Warfarin Doses
Assess whether patient is at greater risk for sensitivity to warfarin – if so, use low‐dose protocol
Small, elderly females
Over age 75
Renal (creatinine >1.5 mg/dL) or hepatic dysfunction
Prophylactic antibiotics are indicated prior to cardiac surgical procedures to minimize the risk of postoperative mediastinitis. A first‐generation cephalosporin is usually chosen (cefazolin), although there is some evidence that a second‐generation drug (cefuroxime) may be more effective.131 Vancomycin is substituted if there is a penicillin allergy and is commonly selected for patients undergoing valve or graft placement because of its effectiveness against Gram‐positive organisms. The STS guidelines give a level IIb recommendation to give an antibiotic providing Gram‐negative coverage (gentamicin) when vancomycin is used, although this is not common practice, and many groups just use cefazolin along with vancomycin to accomplish that. Antibiotics are started within one hour (cephalosporins) or two hours (vancomycin) of surgery and continued for 24–48 hours. They should not be continued any longer even if invasive lines and catheters remain in place.132,133
Wounds closed with subcuticular sutures and covered with 2‐octyl cyanoacrylate adhesive (Dermabond) at the conclusion of surgery do not require dressing coverage. Closed incision negative pressure wound management systems (PREVENA, KCI) applied to a sternotomy wound should be left in place for 5–7 days. If neither of the above is used, the wound should be cleansed and covered with a dressing every day for the first three postoperative days. Subsequent coverage is not necessary unless drainage is noted. All drainage should be cultured and sterile occlusive dressings applied.
Nosocomial infections develop in 5–10% of patients undergoing cardiac surgery using CPB. Most infections involve the urinary or respiratory tracts, central IV catheters, or the surgical site. Infections are less common in patients undergoing OPCAB, presumably because of the reduced necessity for blood transfusions and the absence of the immunomodulatory effects of extracorporeal circulation.134,135 A nosocomial infection not only increases the length of stay but also significantly increases operative mortality (to about 20%) because of the frequent development of multisystem organ failure.135–137Staphylococcus is the most common organism noted in bacteremia and wound infections, whereas Gram‐negative infections are more common in the respiratory tract. The high mortality rates of mediastinitis (about 20–25%) and septicemia (30–50%) do not differ much between patients with methicillin‐resistant S. aureus (MRSA) and other organisms.138
Risk factors have been identified in multiple studies, most of which are included in the STS risk model for major infection shown in Figure 13.3.139,140 To a large degree, these overlap the risk factors for major mediastinal infections noted on pages 778–779.
Nasal carriage of S. aureus, which is associated with a significant incidence of postoperative MRSA infections (about 20%) with a mortality risk of 15–30%.141,142
Operative factors: long complex operations, urgent/emergent surgery, reoperations, use of an IABP
Postoperative factors (with specific relationships in parentheses)
Hyperglycemia (wound infections)
Prolonged duration of intubation or need for reintubation (pneumonia). The overall risk of ventilator‐associated pneumonia is 5–8%, but it has been estimated to be >50% in patients intubated over 48 hours.143
Prolonged duration of indwelling Foley catheter (UTI), with the risk of bacteriuria increasing 3–8%/day of catheter usage.144
Prolonged duration of central venous catheter placement (bacteremia). Central venous catheters with multiple lumens and those changed over guidewires increase the risk of bacteremia.145,146
Blood transfusions (pneumonia)
Reoperation for bleeding
Empiric use of broad‐spectrum antibiotics
Low cardiac output syndromes
Early development of postoperative renal failure
Preventative measures that may reduce the incidence of nosocomial infections include:
Perioperative use of intranasal mupirocin to reduce staphylococcal nasal carriage. Although selective use in carriers is most appropriate and can be accomplished using polymerase chain reaction (PCR) assessment, this is not always logistically possible and is not cost‐effective, so it is easier to treat all patients as soon as possible before surgery and for 3–5 days postoperatively.147–149
Chlorhexidine gluconate 0.12% (Peridex) oral rinse has been shown to reduce the rate of nosocomial respiratory infections, wound infections, and mortality.150
Use of a hyperglycemia protocol to maintain early postoperative blood glucose <180 mg/dL.
Early removal of invasive catheters, especially central lines.
Strict adherence to guidelines to avoid prolonged usage of prophylactic antibiotics.133 One study showed that early postoperative pneumonia was usually caused by organisms that colonized the respiratory tract prior to surgery. However, prolonged use of antibiotics was ineffective in reducing the incidence of pneumonia.151
Aggressive ventilatory weaning protocols to reduce the duration of mechanical ventilation and other steps to avoid ventilator‐associated pneumonia (see pages 492–493).
Raising the threshold for blood transfusions (transfuse only for a HCT <22% unless clinically indicated).89,90
The treatment of a nosocomial infection requires appropriate antibiotic selection for the organism involved and recognition of the appropriate time course of treatment. Prolonged treatment is often unnecessary and may lead to the development of resistant strains or fungal infections, and not infrequently to hepatic or renal dysfunction. When a Gram‐positive bacteremia occurs in a patient with a prosthetic heart valve, a six‐week course of treatment for presumed endocarditis may be indicated. In complex situations, infectious disease consultation is essential.
Clinical features. Sepsis resulting in hemodynamic compromise and multisystem organ failure is an uncommon, yet highly lethal, complication of cardiac surgery, with an estimated mortality rate of greater than 30%.152,153 It is usually noted in critically ill patients who remain in the ICU with multiple invasive monitoring lines, develop respiratory complications, and often have some element of renal dysfunction. It may be the first manifestation of an occult sternal wound infection.
Management. Basic principles of hemodynamic and ICU management should be initiated early to try to reduce the high mortality rate associated with sepsis. These should include:
Optimization of hemodynamics with fluid resuscitation, inotropic support, and selective use of vasoconstrictors (initially α‐agents, and then vasopressin, if necessary). This should be assessed by adequate hemodynamic monitoring (PA catheter and central or mixed venous oxygen saturations), aiming for an oxygen saturation >70%.
Initiation of broad‐spectrum antibiotic coverage after panculturing with prompt modification to cover the specific organism isolated
Low tidal volume ventilation if ARDS develops; minimizing sedation to promote early extubation
Early aggressive use of renal replacement therapies (CVVH)
Maintaining the blood glucose <180 mg/dL
Adequate nutrition, preferably by the enteral route
Low‐dose stress steroids to be considered in patients with documented inadequate response to an ACTH challenge
Sternal wound infections (SWI) complicate about 1% of cardiac surgical procedures performed via a median sternotomy and have been associated with significant hospital mortality (around 20%). Coagulase‐negative staphylococcus and S. aureus are the most common pathogens encountered despite the use of prophylactic antibiotics specifically directed at these organisms. An incidence of only 1% is amazingly low when one considers that open‐heart surgery tends to be performed in the sickest patients with multiple comorbidities, involves the use of CPB, is associated with a high prevalence of transfusions, and entails prolonged wound exposure. Nonetheless, sternal infections remain a major source of physical, emotional, and economic stress when they occur.154
Risk factors. Numerous models have been devised to predict the risk of developing mediastinitis, including that from the STS (available as part of the risk calculator at sts.org). Among the risk factors identified in numerous studies are the following:135,155,156
Comorbidities: obesity (BMI >40),157,158 diabetes (elevated HbA1c),159–161 smoking and COPD,162 HF, renal dysfunction, peripheral vascular disease (PVD), older age, impaired nutritional status (low serum albumin),158 use of steroids, and preoperative MRSA colonization142
Surgical considerations that increase risk include:
Performing hair removal with razors rather than clippers
Bilateral ITA usage in diabetics (controversial)163
Contaminated heater/cooler units with mycobacterium chimera, which give rise to latent sternal wound infections165
Excessive mediastinal bleeding, re‐exploration for bleeding, multiple transfusions
Prolonged ventilatory support (usually in patients with COPD who are actively colonized)
Low cardiac output states (cardiogenic shock) with use of an IABP
Refractory hyperglycemia in the ICU, independent of whether the patient has a history of diabetes
Acute kidney injury
Central venous line‐related bacteremia, which increases the risk fivefold145
Prevention.154,166 Because of the fairly universal adoption of the basic perioperative measures listed below, the risk of deep SWI in the STS database is about 1% – and this includes a high percentage of patients with multiple risk factors, especially diabetes, obesity, and smoking. The adoption of newer technologies, such as negative pressure management, has generally not been found to be cost‐effective, except perhaps in patients at highest risk for infection.167
Identify and treat preexisting infections.
Advise patients to stop smoking as soon as possible prior to surgery. Although it is suggested that smoking increases the risk of SWI,162 prominent coughing after surgery in such patients also increases the risk of sternal dehiscence.168
Optimize diabetic control after initial evaluation of the patient if HbA1C is >7.5.
Optimize nutritional status, if possible, in patients with a low serum albumin (<2.5 mg/dL).
Chlorhexidine chest and leg wash several times the night before and the morning of surgery to reduce the bacterial skin count.
Intranasal mupirocin (Bactroban) given at a minimum the morning of surgery and continuing for five days to reduce nasal carriage of staphylococcal organisms. This is not beneficial in patients who are MRSA negative, but, unless nasal swabs and PCR testing are performed in advance, it is recommended that all patients be treated with intranasal mupirocin.149 Although very effective against MSSA, mupirocin is only 50% effective in decolonizing patients with MRSA. One concern of universal decolonization is not just cost but the potential risk of developing antibiotic resistance.169
Hair clipping just prior to surgery
Appropriate timing and dosage of weight‐based prophylactic antibiotics:
Cephalosporins should be given within one hour of surgery. Cefazolin is recommended and should be given in a dose of 2 g to patients <120 kg and 3 g if >120 kg. IV bolus injection achieves a peak plasma concentration within 20 minutes and peak interstitial levels within 60 minutes, so ideally it should be given 20–30 minutes before surgery starts.135 An additional dose should be given on pump or may be repeated in four hours in off‐pump cases.
Vancomycin is usually selected for patients undergoing valve surgery or placement of prosthetic grafts and may be given as an alternative to cephalosporins in patients who are allergic to penicillin or cephalosporins. It should be started within two hours of surgery and given in a dose of 20 mg/kg. Although not common practice, the 2007 STS guidelines gave a level IIb indication for also giving a single preoperative dose of gentamicin, no more than 4 mg/kg, when vancomycin in used for prophylaxis.132
If there are contraindications to the above antibiotics, daptomycin (6 mg/kg IV) is usually selected.
Careful skin prep with chlorhexidine alcohol may be superior to povidone‐iodine (without alcohol).170–172 Any skin preparation can only reduce the bacterial count but cannot sterilize the skin. One study showed that 89% of subcutaneous wound cultures and 98% of adjacent skin cultures were positive just prior to skin closure after cardiac procedures.173
Consideration may be given to use of a microbial sealant that immobilizes bacteria (InteguSeal, Halyard Health). In some studies, this reduced the risk of SWI, but in others, it did not.174–176
Ensure a midline sternotomy and provide a secure sternal closure. Some studies suggest that figure‐of‐eight wires or rigid fixation systems are superior to simple cerclage wires in producing sternal stability and reducing the incidence of wound infection, although a meta‐analysis of closure methods found no difference.178–182 Use of Robicsek basket‐weaving sutures may be considered if the sternum is narrow, osteoporotic, or divided off midline.
Be selective in the use of bilateral ITAs in diabetic patients. Skeletonizing the ITA may be helpful, but avoidance of bilateral usage in patients with other risk factors, such as severe obesity and COPD, is prudent.163
Use meticulous surgical technique with respect for tissues and obtain adequate hemostasis to minimize mediastinal bleeding.
Avoid bone wax.
Use intravenous insulin to maintain intraoperative blood sugar <180 mg/dL.
Redose cephalosporins after four hours. One study reported that the risk of SWI was greater when the surgical duration approached six hours and there was a low cefazolin plasma concentration (<104 mg/L) during wound closure.183
Only gold members can continue reading. Log In or Register to continue