Fluid Management, Renal, Metabolic, and Endocrine Problems
12 Fluid Management, Renal, Metabolic, and Endocrine Problems
Perioperative renal dysfunction is a major determinant of both operative and long‐term mortality following cardiac surgery.1–3 Even patients with mild renal dysfunction prior to surgery are more likely to experience acute kidney injury (AKI) afterwards, with compromised short‐ and long‐term outcomes.4,5 Therefore, it is essential to identify patients at high risk for developing postoperative AKI who may benefit from specific interventions aimed at optimizing renal function. Although many risk factors for AKI cannot be modified, measures can be taken preoperatively, during cardiac catheterization procedures, during surgery, and in the postoperative period to minimize the risk of developing AKI.6 If it does develop, careful medical management and, if necessary, early aggressive use of renal replacement therapy may reduce the high mortality associated with postoperative AKI.
I. Body Water Distribution
An understanding of body water distribution is important when administering fluids to patients after open‐heart surgery. Approximately 60% of the body’s weight (50% in women) is water, with two‐thirds of this residing in the intracellular space and one‐third in the extracellular space. In the latter, two‐thirds is in the interstitial space (the so‐called third space), and one‐third constitutes the intravascular volume.
Water moves freely among all three compartments and shifts so as to normalize serum osmolality (which generally reflects the serum sodium concentration).
Sodium moves freely between the intravascular and interstitial spaces but does not move passively into cells. Therefore, if a patient receives a hypotonic sodium load (e.g. 0.45% saline) which would lower the serum osmolality and sodium concentration, water will move from the extracellular space into the intracellular space to normalize these values. The presence of a low serum sodium concentration in the postoperative patient usually indicates total body water overload.
Starling’s law governs the influence of hydrostatic and oncotic pressures on fluid shifts. The primary determinant of oncotic pressure is serum protein, which remains within the intravascular space. Elevated hydrostatic pressure (e.g. increased pulmonary capillary wedge pressure, PCWP) or lower intravascular colloid oncotic pressure (e.g. very low serum albumin, usually <2 g/dL) will shift fluid from the intravascular space into the interstitial space, contributing to lung and tissue edema. Conversely, raising the intravascular oncotic pressure with colloid (e.g. 25% albumin) in patients with hypoalbuminemia will tend to draw fluid from the lung interstitium back into the intravascular space.
It should be kept in mind that Starling’s law describes fluid shifts in the absence of abnormalities in membrane integrity. However, extracorporeal circulation is associated with a systemic inflammatory response, characterized by increased membrane permeability and a transient capillary leak. When this leak is present, administered fluid will shift more readily into the interstitial space. Clinically, one may note impaired oxygenation and decreased pulmonary compliance (higher peak pressures on the ventilator) associated with increased extravascular lung water. This can produce the picture of noncardiogenic pulmonary edema. Expansion of the interstitial space may also contribute to cerebral edema (mental obtundation), hepatic congestion (jaundice), and splanchnic congestion (ileus).
Because of the capillary leak associated with cardiopulmonary bypass (CPB), patients will be total body fluid overloaded, but can have low filling pressures with compromised cardiac output. Thus, until the capillary leak ceases, which is generally within 12 hours of surgery, fluid administration, whether colloid or crystalloid, will be required to optimize preload to maintain the cardiac output.
II. Effects of CPB and Off‐Pump Surgery on Renal Function
The influence of CPB on renal function is multifactorial.7–11 It involves nonpulsatile perfusion with hemodilution and variable degrees of hypothermia. A number of factors can adversely affect renal vasomotor tone, triggering a reduction in renal blood flow (RBF). The longer the duration of CPB, the more protracted the exposure to these adverse elements, increasing the risk of AKI.
CPB increases levels of hormones (endogenous catecholamines, vasopressin) and induces the renin–angiotensin–aldosterone cascade, altering vascular tone, RBF, glomerular filtration rate (GFR), filtration fraction, and electrolyte balance.
Hemodilution reduces the hematocrit (HCT) and oxygen‐carrying capacity of blood.12–15
Vasodilation may also lower blood pressure during CPB, increasing fluid requirements and the need for vasopressor support to maintain an adequate mean arterial pressure. ACE inhibitors and ARBs may attenuate the effects of vasoconstrictors such as norepinephrine, so stopping them just prior to surgery may reduce the risk of hypotension during bypass and arguably will lower the risk of AKI.16–21 Judicious use of other vasodilators during surgery (propofol, narcotics, inhalational anesthetics, nitroglycerin) may minimize hypotension during surgery as well.
Extracorporeal circulation evokes an inflammatory response with activation of complement and neutrophils, release of cytokines, and production of oxygen free radicals (“oxidative stress”).
Aortic cannulation and clamping can generate atheroembolism.
Low‐grade hemolysis from CPB may cause release of iron leading to oxidation from reactive oxygen species.22
CPB is associated with an increase in virtually all kidney‐specific proteins that are markers for tubular damage.23 Some of these, such as neutrophil gelatinase‐associated lipocalin (NGAL), cystatin C, Kidney Injury Molecule‐1 (KIM‐1), and interleukin‐18 (IL‐18), have been shown to be early biomarkers of AKI that correlate with the severity and duration of AKI.24–26 However, very few centers routinely evaluate renal function other than by serum creatinine (SCr) levels, and in the vast majority of cases, there is little significance to subtle changes in tubular function as long as the kidneys produce a satisfactory urine output with or without diuretics with minimal change in the SCr.
The potential benefit of avoiding CPB by performing off‐pump coronary artery surgery (OPCAB) to reduce the risk of postoperative AKI is controversial.
Some studies have reported a reduction in AKI with OPCAB, but with no clear impact on the requirement for renal replacement therapy.27–31 Others suggest that OPCAB reduces the incidence of postoperative AKI only in patients with normal renal function, but not those with preexisting chronic kidney disease (CKD).32,33
In theory, avoidance of CPB might preserve RBF and glomerular function better by maintaining a higher systemic pressure. Tubular epithelial function may be better preserved because of decreased complement activation and a lessened inflammatory response.34 However, off‐pump surgery is associated with significant fluid administration, use of comparable anesthetic and vasoactive medications, cytokine release that can damage proximal tubules, and alterations in perfusion pressure (lower systemic pressures with elevated venous pressures during exposure of the posterior heart and lower systemic pressures during construction of proximal anastomoses), all of which can adversely affect renal function.
Because postoperative renal dysfunction is related more to preexisting renal disease or significant hemodynamic alterations than to the inflammatory response, particular attention to fluid and hemodynamic management remains paramount no matter whether CPB is used or not.
III. Routine Fluid Management in the Early Postoperative Period
Hemodilution on CPB produces a state of total body sodium and water overload, expanding the body weight by about 5% (estimated at 800 mL/m2/h, but quite variable in amount). Cardiac filling pressures usually do not reflect this state of fluid overload because of a capillary leak from the systemic inflammatory response, decreased plasma colloid osmotic pressure, impaired myocardial relaxation (diastolic dysfunction) from ischemia/reperfusion after cardioplegic arrest, and vasodilation.
Low filling pressures are consistent with hypovolemia despite the presence of body water overload, and additional fluid administration may be necessary to maintain satisfactory hemodynamics.
High filling pressures may suggest hypervolemia or ventricular dysfunction, but they may also be noted in the presence of hypovolemia, especially in patients with diastolic dysfunction or marked vasoconstriction, and additional fluid administration may be indicated in that situation. Use of Swan‐Ganz monitoring (and its correlation with echocardiographic findings) is helpful in providing a scientific basis for fluid management after surgery, especially in patients with significant right or left ventricular dysfunction or after very complex operations with long durations of CPB, although it may not be necessary in low‐risk patients.
Giving fluids to optimize preload and cardiac output in the early postoperative period may be required whether urine output is adequate or marginal (<1 mL/kg/h). During the first 4–6 hours after surgery, cardiac output is often depressed, and the achievement of satisfactory hemodynamics to optimize renal perfusion is dependent on both preload and inotropic support. Thus, fluid must invariably be administered to maintain intravascular volume and cardiac hemodynamics at the expense of expansion of the interstitial space. It should be noted that early extubation is helpful in reducing fluid requirements because it eliminates the adverse effects of positive‐pressure ventilation on venous return and ventricular function.
It can be difficult to decide which fluid to administer to maintain filling pressures. Clearly, any fluid infused during a period of altered capillary membrane integrity will expand the interstitial space, but those that can more effectively expand the intravascular space while minimizing expansion of the interstitial space are preferable. Nonetheless, clinical outcomes are fairly comparable with colloid or crystalloid administration in critically ill patients, and this most likely holds true in most cardiac surgery patients who exhibit a systemic inflammatory response.35,36 Most patients have enough pulmonary reserve to tolerate the volume overload until it can be diuresed, but those developing AKI with oliguria will have a more tenuous course.
Blood and colloids are superior to hypotonic or even isotonic crystalloid solutions in expanding the intravascular volume.37,38 Although a rapid infusion of crystalloid is effective in increasing intravascular volume acutely, this benefit is transient.39 For example, in the absence of a capillary leak, after a five‐minute infusion of one liter of lactated Ringer’s, the intravascular volume expands approximately 630 mL. Yet, due to rapid redistribution into the interstitial space, barely 20% of this volume is retained within the intravascular compartment after an hour. Similarly, only 25% (250 mL) of one liter of infused normal saline (NS) is retained in the intravascular compartment after one hour. In contrast, after a five‐minute infusion of one liter of 6% hetastarch, the intravascular volume expands by 1123 mL with more long‐lasting effects. Five percent albumin can expand the plasma volume five times more than a comparable volume of normal saline.40
In general, it is reasonable to initially administer a moderate amount of inexpensive crystalloid (up to a liter) if the patient is oxygenating well. Infusing greater amounts may contribute to tissue edema, commonly impairing oxygenation. Colloids should be selected if additional volume is required, although at some centers, they are given first. The selection of colloid should be based on the patient’s pulmonary and renal function and the extent of mediastinal bleeding.
Albumin (5%) provides excellent volume expansion (approximately 400 mL retained per 500 mL bottle administered), has a half‐life of 16 hours, and leaves the bloodstream at a rate of about 5–8 g/h. It has primarily dilutional effects on clotting parameters and preserves coagulation better than the hydroxyethyl starches.41 It has oxygen free‐radical scavenging and anti‐inflammatory properties, which may exert protective effects on the kidney. However, it will leak into the interstitial space due to the capillary leak and may cause movement of fluid out of the intracellular space. Furthermore, 5% albumin is a saline‐based colloid with a high chloride load, and studies suggest that use of albumin increases the risk of AKI in a dose‐dependent manner after cardiac surgery.42 Therefore, although some groups use 5% albumin as the preferential fluid after surgery, it can be inferred that large volumes of 5% albumin must be used with caution.
Hydroxyethyl starch (HES) preparations are nonprotein colloid volume expanders that provide excellent volume expansion in excess of the volume infused.
The high‐molecular‐weight solutions, Hespan (6% hetastarch in saline) and Hextend (6% hetastarch in balanced electrolyte solution), maintain volume expansion for about 24 hours, but may cause renal dysfunction and produce a coagulopathy by binding to the von Willebrand/factor VIII complex, causing platelet dysfunction, and also by causing fibrinolysis.43 This risk may be slightly less with Hextend.44–46 Thus, despite a recommended maximum infusion of 20 mL/kg, their use has been discouraged in the early postoperative period and should absolutely be avoided in the bleeding patient.37,38,43
The low‐molecular‐weight solutions include pentastarch (Pentaspan [DuPont Pharma]), tetrastarch in 0.9% saline (Volvuven), and tetrastarch in a balanced electrolyte solution (Volulyte [Fresenius Kabi, Canada]). These also produce excellent volume expansion, but for shorter periods of time (18–24 hours for pentastarch, six hours for tetrastarch). The risks of renal dysfunction and coagulopathy may be slightly less than with the high‐molecular‐weight solutions, but they are still present.47 Therefore, these products are also not recommended in the bleeding patient. Otherwise, infusion volumes should be limited to 28 mL/kg (2 L/day maximum) for pentastarch and 50 mL/kg for tetrastarch (3.5 L/day maximum).
Note: there is concern that saline‐based solutions (0.9% saline, 5% albumin, Hespan, Pentaspan and tetrastarch in saline) provide a high chloride load that, given in high doses, can produce progressive renal vasoconstriction, a decrease in GFR, and a hyperchloremic metabolic acidosis. Studies show that use of chloride‐restricted solutions, such as lactated Ringer’s, tetrastarch in balanced electrolyte solution, and Plasma‐Lyte, is associated with a lower risk of AKI.48–51 The development of a metabolic acidosis related to the use of high chloride solutions might raise the specter of poor tissue perfusion, prompting unnecessary interventions.
Hypertonic solutions are effective in augmenting intravascular volume by extracting fluid from the interstitial and intracellular spaces. They may reduce the amount of fluid required to maintain intravascular volume when there is total body fluid overload. Twenty‐five percent albumin can increase the intravascular volume by 450 mL for every 100 mL administered. Other solutions are available that may increase intravascular volume without providing excessive free water, but they are usually used only in the setting of hyponatremia. Hypertonic saline (3%) can produce neurologic problems if it causes acute hypernatremia. Studies from Europe have shown that hypertonic saline (7.5%) can produce renal vasodilation, increase GFR, and produce a diuresis.52 It should be noted that use of these hypertonic colloids can produce hyperoncotic renal failure in dehydrated patients because the glomerular filtration of hyperoncotic colloid molecules may cause hyperviscosity and stasis of tubular flow, resulting in tubular obstruction. This effect may also be one of the mechanisms of high‐molecular‐weight HES‐induced renal dysfunction.
An ideal solution for volume expansion would be a commercially available hemoglobin (Hb) based oxygen carrier. Thromboelastographic studies of Oxyvita, a polymerized bovine‐Hb‐based oxygen carrier, have shown similar effects on the coagulation profile as Hespan at doses up to 23 mL/kg, but with minimal coagulopathic effects at the recommended dose of 2–3 mL/kg.53 These products enhance fibrinolysis and must be avoided in the bleeding patient.54
It cannot be overemphasized that the objective of postoperative fluid management is to maintain adequate intravascular volume to ensure satisfactory cardiac output and tissue perfusion. Administration of excessive volume to maintain high filling pressures and the highest possible cardiac output will increase extravascular water, which will be primarily manifested by pulmonary edema that will delay extubation. The amount of fluid to administer can be confusing in patients with diastolic dysfunction who already have high filling pressures, but often a marginal cardiac output. In addition, the hemodilution caused by intravascular volume expansion may decrease the hematocrit and also reduce the level of clotting factors, possibly precipitating bleeding and necessitating homologous blood or blood product transfusions.
When cardiac function is satisfactory, but there is an ongoing volume requirement to maintain filling pressures or blood pressure, often from a combination of the capillary leak, vasodilation, and an excellent urine output, “flooding” the patient with volume should be resisted. After 1.5–2 L of fluid is given, norepinephrine or vasopressin should be used to maintain filling pressures and improve the systemic blood pressure. Norepinephrine may provide some cardiac support, will improve RBF, and may lower renal vascular resistance by lowering renal sympathetic tone.55,56 Vasopressin (0.01–0.1 units/min) is very effective in restoring the blood pressure to within the renal autoregulatory range (generally a mean pressure >80 mmHg) in the vasodilated “vasoplegic” patient with a good cardiac output.57,58 In conditions of low cardiac output, however, it may cause splanchnic vasoconstriction, inducing bowel ischemia. Phenylephrine should be utilized only when the cardiac output is satisfactory, because it provides a pure α effect on systemic vascular tone, causing renal arteriolar vasoconstriction.
If both cardiac output and urine output remain marginal after adequate filling pressures have been achieved, inotropic support must be considered first, with use of vasoconstrictor drugs only if systemic resistance remains low. Use of α‐agents at substantial doses is always of concern with a marginal cardiac output because they may produce renal vasoconstriction and compromise renal function.
Generally, diuretics are best avoided in the first six hours after surgery unless pulmonary edema with borderline oxygenation is present. They may be beneficial if the pulmonary edema is cardiogenic in origin, but noncardiogenic pulmonary edema may be present even if the patient is hypovolemic. When the patient has achieved a stable core temperature and the capillary leak has ceased, usually after the first 6–12 hours, filling pressures will stabilize or rise with little fluid administration. By this time, myocardial function has usually recovered, inotropic support can be gradually withdrawn, and the patient can be extubated. Diuresis may then be initiated to excrete the excess salt and water administered during CPB and the early postoperative period. Patients who have undergone operations that require long periods of CPB (usually >3 hours) or who have persistent low output syndromes may experience a longer period of “capillary leak” that requires further fluid administration to maintain filling pressures. In either circumstance, if the patient has low filling pressures despite fluid overload, initiation of diuretics should probably be delayed.
Diuresis can be augmented most efficiently by the use of loop diuretics.59,60
Loop diuretics inhibit sodium reabsorption in the ascending limb of the loop of Henle and increase solute (sodium) presentation to the distal tubules. By inhibiting tubular sodium and chloride reabsorption, they increase natriuresis and diuresis. To a lesser extent, they may also act as renal vasodilators, increasing RBF and GFR, and they may improve medullary oxygenation.
Most patients with preserved renal function respond to furosemide (Lasix) 10–20 mg IV. In the absence of renal insufficiency, furosemide has a half‐life of 1.5–2 hours, and thus it can be repeated every four hours, if necessary. Not infrequently, the diuresis persists after one dose. Some patients with advanced CKD appear to respond better to bumetanide 1–2 mg IV and then may be given torsemide 10–20 mg orally (10–20 mg can be given IV, but this preparation is not available in the USA).
A gentle continuous diuresis may be obtained in patients with significant fluid overload and hemodynamic instability using a 40–60 mg IV bolus dose of furosemide followed by a continuous infusion of 0.1–0.5 mg/kg/h (usually 10–20 mg/h).61 This may decrease the total dosage requirements and usually improves the diuretic response, especially in patients who are diuretic “tolerant”. This benefit is also seen in patients with CKD. The addition of a thiazide (chlorothiazide 500 mg IV) is beneficial in overcoming this problem of tolerance, which may be caused by compensatory hypertrophy of the distal nephron segments in response to increased exposure to solute from chronic use of loop diuretics.
Diuretics are continued in IV or oral form until the patient has achieved their preoperative weight. This is a common practice, although most patients with normal renal function will auto‐diuresis several days after surgery in the absence of diuretics. One study did in fact show no clinical benefit to initiating early diuresis in low‐risk patients with normal renal function.62 Another study suggested that not only intraoperative furosemide but also use of any diuretic postoperatively increased the risk of AKI.63
“Renal‐dose” dopamine (2–3 μg/kg/min) increases RBF and GFR in patients with normal renal function, resulting in effective diuresis and natriuresis and possibly reducing the need for diuretics. However, dopamine initiated during surgery and continuing afterwards for 24 hours is not renoprotective and, in fact, may cause a deterioration in renal function. This has led to the recommendation that dopamine should not be used in postoperative cardiac surgical patients.64–67
Guidelines for the hemodynamic and fluid management of typical postoperative scenarios are presented in Chapter 8.
IV. Identifying Risk for Acute Kidney Injury
The risk of developing postoperative AKI is very low when a patient with normal renal function undergoes an uneventful operation and maintains satisfactory postoperative hemodynamics. In contrast, the presence of any degree of preoperative renal dysfunction increases the risk of postoperative AKI and mortality.1–5 Therefore, it is important to identify patients with preoperative renal dysfunction and those with other risk factors for developing postoperative AKI.
Definition of preoperative renal dysfunction. The staging of CKD should be based on GFR, rather than SCr, as GFR provides a better estimate of renal reserve and the ability of the kidneys to tolerate surgical stress. It also correlates with both in‐hospital mortality and the long‐term prognosis.68,69 SCr may be in the normal range even when there is a greater than 50% reduction in GFR, which reflects the number of functional neurons. The stages of CKD are defined as follows (in mL/min/1.73 m2)
Stage 5<15 or on dialysis
A GFR <60 mL/min/1.73 m2 (CKD stages 3–5) represents evidence of significant CKD and is the level below which there is an increased risk of postoperative AKI and increased mortality.1 However, a higher risk of adverse outcomes, including the need for renal replacement therapy and mortality has even been confirmed in patients with “occult” kidney disease (i.e. a low GFR but normal SCr), who account for about 13% of patients with a normal SCr.4,5 In fact, one study showed that patients with a normal GFR but impaired renal functional reserve demonstrated after a high oral protein load are also at risk for postoperative AKI.70
A 24‐hour urine collection to precisely measure the creatinine clearance is considered to be no more reliable than an estimate based on formulas using the SCr.
Using the Cockcroft‐Gault equation, the GFR can easily be calculated at the bedside and is indexed to the patient’s age and weight. This may be the best GFR equation to predict in‐hospital mortality.68
Other formulas include the Modification of Diet in Renal Disease (MDRD) formula and the 2009 Chronic Kidney Disease Epidemiology Collaboration (CKD‐EPI) equation which may be more accurate in calculating GFR, especially at higher levels of GFR.71,72 Hospital laboratories utilize one of these equations to provide GFR results routinely.
Risk factors for postoperative AKI and predictive models (Table12.1)3,8–11,73–77
The basic pathophysiology of postoperative AKI involves renal ischemia and other phenomena related to CPB, including inflammation, reperfusion injury, oxidative stress, and hemoglobinuria. Hemodynamic compromise in the pre‐, intra‐, or postoperative period will also heighten the risk of AKI and potentially affect its duration. Thus, more complex procedures requiring longer durations of CPB and the use of vasoactive drugs to support a low cardiac output are the major perioperative factors predisposing to AKI.
AKI is usually categorized as prerenal (reduced renal perfusion), renal (intrinsic renal insults), or postrenal (obstructive uropathy). Mechanisms contributing to the first two categories in the perioperative period are noted in Table 12.1. When the kidneys have sustained an acute preoperative insult, either from a cardiac catheterization or more ominously from decompensated heart failure (HF) or cardiogenic shock from an acute ischemic event, they seem to be particularly sensitive to the abnormal physiology of CPB and to tenuous postcardiotomy hemodynamics. This is especially true in patients with “acute on chronic” renal dysfunction. The BUN and SCr should therefore be allowed to return towards baseline, if possible, before proceeding with surgery.
Most predictive models for AKI include similar risk factors, although several additional risk factors, such as a long duration of CPB and recent use of contrast, have been identified in other studies but are not included in many of the risk models.3,75–78 Common risk factors include:
Older age (2.5‐fold increase in risk for each 10‐year increment in one study)79
Use of preoperative diuretics, nephrotoxic drugs, or medications that interfere with renal vasomotor tone
More tenuous hemodynamic status (shock, recent myocardial infarction, preoperative IABP, low EF), reoperations, and urgent or emergent surgery
One simple risk model found that four factors were able to predict AKI with great accuracy (age >65, GFR <80 mL/min/m2, cross‐clamp time >50 minutes, and more complex surgery).80
In virtually all of the risk models predicting AKI and the need for dialysis, the most significant risk factor is that of preexisting renal dysfunction, and most models use SCr rather than GFR as the marker of CKD.73 Risk models are useful to individualize the risks of AKI and RRT based upon an assessment of multiple contributing risk factors.
Figure 12.1 provides an estimate of the risk of developing severe AKI in patients with occult CKD.81
Figure 12.2 provides a model to predict AKI (defined as SCr >2.0 mg/dL and a 50% increase in SCr) based upon pre‐, intra‐, and postoperative parameters.82
Figures 12.3 and 12.4 provide the Cleveland Clinic and STS risk models for the prediction of RRT after cardiac surgery.83,84
These models do not predict operative mortality, which may be related to other contributing factors. For example, patients undergoing urgent or emergent operations of high complexity will have a higher predicted mortality at any level of SCr. It is best to use the STS short‐term risk calculator available online at www.sts.org to incorporate other parameters to determine the risk of AKI and mortality. Generally, mortality risk is greater with increasing SCr levels, correlates with the degree and duration of oliguria,85 and averages about 10% for patients with non‐dialysis‐dependent CKD.1
The development of postoperative AKI is associated with about a fourfold increase in operative mortality.73,86 Although patients already on dialysis have an operative mortality of about 10–15% in most series,87–89 but as high as 37% in others,90 the operative mortality for patients who develop AKI that requires de novo dialysis is quite high, ranging from 25–58% in a few series.91,92 One study even estimated that the risk of dialysis exceeded 30% if the preoperative SCr exceeded 2.5 mg/dL.89These alarming statistics emphasize the crucial importance of taking any steps possible to minimize renal insults and preserve renal function in the perioperative period, especially in patients at increased risk. The presence of any degree of preoperative renal insufficiency should therefore lead to a search for potentially treatable causes that might lower the risk of AKI postoperatively. Identifying and correcting these contributing factors before surgery and using measures during and after surgery to optimize renal perfusion and tubular function to try to prevent AKI may ameliorate the complications associated with the development of oliguric renal failure. These may include electrolyte abnormalities, pulmonary and cardiac dysfunction, bleeding, delayed return of gastrointestinal (GI) function affecting nutrition, and infection from immune dysfunction, not to mention the possibility of requiring dialysis and its attendant complications.
Table 12.1 Factors Contributing to Pre‐ and Postoperative Acute Kidney Injury
Demographics: advanced age, female gender Comorbidities: CKD stage 3+, diabetes, vascular disease, hypertension, hyperlipidemia Low cardiac output states/hypotension (cardiogenic shock from acute MI, mechanical complications of MI) Medications that interfere with renal autoregulation (ACE inhibitors, NSAIDs) Nephrotoxins (contrast‐induced nephropathy, especially in diabetics), medications (aminoglycosides) Diuretics Anemia Renal atheroembolism (catheterization, IABP) Interstitial nephritis (antibiotics, NSAIDs, furosemide) Glomerulonephritis (endocarditis)
Cardiopulmonary bypass (nonpulsatile, low flow, low pressure perfusion with reduced renal perfusion, systemic hypotension, impairment of autoregulation) Low cardiac output syndrome/hypotension after CPB Blood transfusions Profound anemia (hematocrit <21%) Hemolysis and hemoglobinuria from prolonged duration of CPB
Low cardiac output states (decreased contractility, hypovolemia, absent AV synchrony in hypertrophied hearts) Hypotension Blood transfusions Intense vasoconstriction (low flow states, α‐agents) Atheroembolism (IABP) Sepsis Medications (cephalosporins, aminoglycosides, ACE inhibitors)
V. Prevention of Acute Kidney Injury
Prior to and during cardiac catheterization, consider the following interventions (Table 12.2):
Avoid medications the day of the catheterization that may have adverse effects on renal function, including diuretics.
Adequately hydrate with normal saline before, during, and after catheterization. Leaving a patient NPO all day and performing a catheterization late in the day without hydration is a set‐up for contrast‐induced nephropathy. When this occurs, studies have shown a fourfold increase in operative mortality.93 Fluid administration may be based on the patient’s estimated left ventricular end‐diastolic pressure (LVEDP). A common protocol is to give:
3 mL/kg × 1 h
LVEDP <13 mm Hg
LVEDP 13–18 mm Hg
LVEDP >18 mm Hg
1.5 mL/kg/h × 4 h or 1 mg/kg × 6 h
Although some studies have shown benefits of sodium bicarbonate or n‐acetylcysteine alone or together in minimizing the risk of contrast‐induced nephropathy, most groups have found hydration protocols to be the most beneficial.94–96
Use low volumes of iso‐ or low‐osmolar nonionic contrast
Repeat the SCr after contrast studies and defer surgery, if possible, until it has returned to baseline. Unfortunately, it can take up to 24–36 hours for the SCr to rise, so even delaying on‐pump (but probably not off‐pump) surgery for 24 hours after catheterization may not be sufficient to reduce the risk of postoperative AKI in patients with preexisting CKD.78,97–99 Some studies suggest that surgery should be delayed for five days after coronary angiography, especially when high‐contrast doses (>1.4 mL/kg) are used.100–102 However, surgery should not be delayed in critically ill patients with hemodynamic compromise and worsening renal function, since delay will often lead to less‐reversible renal failure and multisystem organ failure.
Stop any medication with potential nephrotoxic effects, such as the NSAIDs, which impair autoregulation of RBF. Whether ACE inhibitors and ARBs should be withheld the morning of surgery remains controversial. They may cause refractory hypotension on CPB, which could contribute to kidney injury. However, studies have shown differing effects on the incidence of AKI, with evidence of both higher and lower risks of AKI in patients taking ACE inhibitors prior to surgery.16–21
Fluid overload should be avoided as it is associated with a higher risk of AKI and worse surgical outcomes, most likely because of its association with decompensated HF and hemodynamic instability.103 In these patients, diuretics may need to be given to improve the clinical picture. Generally, in stable patients, diuretics should be withheld and consideration given to some preoperative hydration in patients with CKD.104
Statins are routinely given to patients undergoing coronary bypass surgery, and often to those undergoing valvular surgery as well. There is some evidence that they lower levels of kidney biomarkers after surgery, suggesting that they might provide some degree of renoprotection.105 Although some individual studies show this benefit,106 a meta‐analysis and subsequent studies of high‐dose atorvastatin failed to do so.107,108 A study of rosuvastatin actually showed it increased the risk of AKI.109 Another meta‐analysis suggested that statin use did not influence the risk of AKI but did reduce the need for RRT.110 One review suggested continuation of statins might be beneficial for renoprotection, but starting them in statin‐naive patients may not.111
Optimize hemodynamic status. Patients with poor cardiac function, fluid overload, and decompensated HF will often achieve substantial improvement in renal function with appropriate medical care which should lower the risk of surgery. Patients in cardiogenic shock have a high mortality rate, which might be lessened by emergency surgical intervention. Consideration may be given to mechanical circulatory support if multisystem organ failure has developed, and this may allow for improvement in renal function. If surgery is performed in these high‐risk cases, postoperative AKI is inevitable but hopefully transient and reversible.
Preoperative anemia is associated with an increased risk of AKI as is the use of multiple transfusions.6,112–114 Erythropoietin plus iron can be recommended to improve the hematocrit in patients undergoing elective surgery.115 In other patients in whom the calculated on‐pump hematocrit will be less than 21% based on the patient’s size and blood volume, it is not unreasonable to transfuse patients preoperatively. It has been suggested that this strategy may be associated with a lower risk of AKI.116
Correct acid–base and metabolic abnormalities that are often seen in patients with CKD. These patients are more susceptible to fluid overload, metabolic abnormalities (hyponatremia, hyperkalemia, hypomagnesemia, and hyperphosphatemia), and metabolic acidosis or alkalosis (from diuretics) in the perioperative period.
Patients on chronic dialysis should be dialyzed within the 24 hours before and after surgery. The overall mortality rate for patients on chronic dialysis undergoing open‐heart surgery is approximately 10–15%, but even higher in those with advanced NYHA class and those undergoing urgent or emergent surgery and requiring complex surgery.
Preoperative dialysis should also be considered in non‐dialysis‐dependent renal failure (stage 4–5). This approach to patients with a preoperative SCr ≥2.5 mg/dL has been shown to reduce the need for postoperative dialysis, with less morbidity and significantly less mortality.117–119
Intraoperative measures should be taken to try to augment renal reserve by improving RBF, enhancing the GFR, and preventing tubular damage in patients with known renal dysfunction or risk factors for its development (Table 12.2).
Consider performing off‐pump coronary surgery, especially in diabetic patients with preoperative renal dysfunction. Whether this reduces the risk of AKI is controversial.27–33
Maintain optimal hemodynamic performance before CPB. This may require fluid administration, treatment of ischemia, or use of vasoactive drugs to support myocardial function or systemic resistance.
Use dexmedetomidine for sedation during surgery and afterwards. This may produce more hemodynamic stability and has been shown to reduce the incidence of AKI.67,120,121
Use antifibrinolytic drugs to minimize the bleeding diathesis that commonly accompanies renal dysfunction (uremic platelet dysfunction). ε‐aminocaproic acid (Amicar) is commonly used and is generally safe, although it is associated with some degree of renal tubular dysfunction without a significant change in creatinine clearance.122 Tranexamic acid is a good alternative.
Pharmacologic means to optimize renal perfusion have been studied with variable results.121,123
Fenoldopam (0.03–0.1 μg/kg/min) may reduce the risk of AKI without any reduction in the need for renal replacement therapy or mortality, but this has not been uniformly demonstrated.123,124
Diltiazem (0.1 mg/kg bolus followed by an infusion of 2 μg/kg/min) reduces renal vascular resistance by dilating afferent arterioles, resulting in an increase in RBF and GFR. It may limit calcium influx into renal tubular cells, preserving their integrity. It has been shown to increase sodium excretion and improve SCr and free water clearance by a direct effect on tubular reabsorption. However, the vasodilatory effect of diltiazem may adversely affect renal function during CPB, and in some studies has increased the risk of AKI.79 Studies do suggest that diltiazem is more likely to improve rather than reduce glomerular function in patients with mild–moderate renal dysfunction.125 One study showed that not using a calcium‐channel blocker perioperatively increased the risk of AKI.126 Another showed that the combination of dopamine and diltiazem started 24 hours before surgery and continued for 72 hours afterwards improved SCr and free water clearance compared with use of either drug alone.127
Regimens using sodium bicarbonate, N‐acetylcysteine, and statins have shown mixed results and none can be recommended.10,121,123,128,129 Specifically, renal‐dose dopamine (3 μg/kg/min) may increase urine output during CPB, but is not renoprotective.65,66 Spironolactone, a mineralocorticoid receptor blocker, may reduce ischemia‐reperfusion injury, but has also not been shown to reduce the risk of AKI.130
Mannitol is commonly added to the pump to increase tubular flow and produce a diuresis. It increases oncotic pressure, reduces tissue edema, and may reduce cell swelling after cardioplegic arrest. Usually 25–50 g is added to the pump prime. However, mannitol has not been demonstrated to provide any benefit in preserving renal function in patients with both normal and abnormal preoperative renal function.131,132
Furosemide is often given during surgery to augment urine output and is beneficial in treating patients with significant volume overload, severe oliguria, or hyperkalemia. However, although it may increase urine output, most studies have not demonstrated a renoprotective effect.66,133–135 There is one study that showed some benefit in patients with more than mild CKD.136 Although a lower urine output during CPB may be associated with an increased risk of AKI, this does not necessarily imply that pharmacologically increasing the urine output will reduce that risk.137
Considerations during CPB
Use heparin‐coated and/or miniaturized circuits, and leukodepletion during the pump run, which may lower the incidence of postoperative AKI by reducing the systemic inflammatory response.138–140
Use a non‐potassium‐containing crystalloid prime to reduce the risk of hyperkalemia induced by use of cardioplegia.
Maintain a higher mean perfusion pressure on bypass (around 80 mm Hg) by increasing the systemic flow rate. If this does not raise the blood pressure to adequate levels, a vasopressor (phenylephrine, norepinephrine, or vasopressin) can be added. Autoregulation of RBF occurs down to a pressure of about 80 mm Hg, but below that, flow is pressure‐dependent.141 Nonetheless, one study did indicate that the incidence of postoperative AKI was no different whether mean pressures <60, 60–69, or >70 mm Hg were used during CPB, although urine output was less at lower pressures.77
Avoid more than mild hypothermia for routine cases and do not overwarm the patient prior to terminating CPB. Mild–moderate hypothermia probably has little impact on renal function, but rewarming to 37 °C may be deleterious.142,143 Therefore, it is generally recommended that patients only be warmed to 36.5 °C since the duration of a rewarming temperature above 37 °C contributes to AKI.144 This may occur because the kidneys may rewarm more rapidly than other organs, including the brain, resulting in hyperthermia‐related exacerbation of renal injury. For procedures involving deep hypothermic circulatory arrest, the incidence of AKI correlates with the duration of DHCA.145
Keep the “pump run” as short as possible. Do what needs to be done and do it expeditiously. Generally, the longer the duration of CPB, the greater the incidence of AKI.75–77
Avoid extreme hemodilution on CPB. Studies have suggested that there is a correlation between the lowest hematocrit on pump (usually <21%, but <24% in one study)14 and the incidence of AKI, especially in obese patients.12,13 However, the hematocrit is only one factor in oxygen delivery. Since AKI is more common below a critical level for oxygen delivery (272 mL/min/m2), lower hematocrits may be acceptable as long as oxygen delivery is maintained above the critical level with increased pump flow rates.146 There is a delicate balance between tolerating a lower hematocrit and administering blood transfusions, as administration of >2 units of blood has been associated with an increased risk of AKI.114 It has been suggested that patients who are chronically anemic tend to tolerate lower hematocrits on pump, and this might reduce the need to transfuse.147
Avoid excessive use of cardioplegia to minimize the potassium load. Use low‐K+ reinfusions of standard blood cardioplegia, consider del Nido cardioplegia, which requires less frequent administration, and substitute intermittent cold blood (without cardioplegia) to minimize the risk of hyperkalemia.
Prevent hyperglycemia or large variations in blood glucose levels during the pump run with intravenous (IV) insulin.148 Despite some concerns that a very strict hyperglycemia protocol may be associated with episodes of hypoglycemia, it has been shown that maintaining a blood sugar (BS) between 80 and 110 mg/dL is associated with a significant reduction in AKI and the need for dialysis in nondiabetic patients.149 There may be additional potential benefits of insulin administration related to its anti‐inflammatory and antioxidant properties.
Initiate hemofiltration towards the end of the pump run to reduce the positive fluid balance and increase the hematocrit. This is especially helpful in patients with preoperative HF and hypoxemia who require urgent surgery.150
One study suggested that the use of sodium nitroprusside during rewarming on pump improved renal function in patients undergoing elective CABG, although this must not be allowed to occur at the expense of unacceptable hypotension.151 In addition, any hemolysis and free hemoglobin release during CPB may accelerate the release of free cyanide from SNP.152
Considerations upon termination of CPB
Use Swan‐Ganz monitoring and frequent evaluations of transesophageal echocardiography to determine optimal filling pressures and the use of inotropes, vasopressors, or an IABP to support hemodynamics.
Carefully monitor serum potassium in patients with preexisting CKD. Levels greater than 6 mEq/L may need to be treated with dextrose/insulin before coming off pump.
Use dexmedetomidine for sedation post‐pump and during the early ICU stay.
Desmopressin may be considered in uremic patients who manifest platelet dysfunction.153 In patients with coagulopathic bleeding after CPB, consider using factor concentrates (prothrombin complex concentrate or fibrinogen concentrate), rather than plasma, to minimize volume infusions.
Consider placement of a central double‐lumen dialysis catheter if concerned about the possibility of needing dialysis in the early postoperative period.
Table 12.2 Preoperative and Intraoperative Measures to Reduce the Risk of Acute Kidney Injury
Hydrate before, during, and after cardiac catheterization
Use low‐volume, low‐osmolar contrast
Avoid preoperative use of diuretics unless clinically indicated
Repeat SCr if preoperative CKD, especially in diabetics, and defer surgery, if possible, until it has returned to baseline
Delay surgery if feasible at least 24 hours after catheterization
Withhold use of ACE inhibitors and ARBs the day of surgery; stop NSAIDs several days in advance
Optimize hemodynamic status
Treat profound anemia: erythropoietin plus iron for elective cases and preoperative transfusions for more urgent cases
Perform emergency surgery, if feasible, for cardiogenic shock to reverse organ system dysfunction or consider mechanical circulatory support in very high‐risk patients
Correct all acid–base and metabolic problems
Perform hemodialysis the day prior to surgery in dialysis‐dependent patients and consider preoperative dialysis in patients with stage 4–5 CKD
Perform off‐pump surgery if possible
Optimize hemodynamics prebypass
Use antifibrinolytics (ε‐aminocaproic [Amicar] or tranexamic acid) to minimize bleeding
Pharmacologic renoprotection: possible use of fenoldopam or diltiazem
Considerations during CPB
Use heparin‐coated circuits, miniaturized if possible
Prime pump with non‐potassium‐containing crystalloid (NS rather than lactated Ringer’s, Normosol or Plasma‐Lyte)
Consider use of a leukocyte‐reducing filter
Maintain a high perfusion pressure (75–80 mm Hg) on bypass
Minimize extent of hypothermia and do not overwarm
Keep the pump run as short as possible
Maintain a hematocrit at least >20%
Be conservative with high‐potassium cardioplegia; consider using del Nido cardioplegia
Use hemofiltration to remove excess fluid
Optimize postbypass hemodynamics (drugs, IABP)
Use dexmedetomidine for sedation
Treat bleeding with factor concentrates, desmopressin (if suspect uremic platelet dysfunction) to minimize volume
Consider placement of a dialysis catheter for patients at high risk for requiring dialysis
Table 12.3 The RIFLE Criteria for Classification of Renal Failure
Increase in SCr × 3 or Decrease in GFR >75% or if baseline SCr >4 mg/dL, an acute rise in SCr of >0.5 mg/dL)
<0.3 mL/kg/h × 24h or anuria for 12 h
Persistent acute renal failure with complete loss of kidney function >4 weeks
End‐stage kidney disease >3 months
VI. Postoperative Oliguria and Acute Kidney Injury
The use of hemodilution during CPB expands the extracellular volume and usually produces an excellent urine output in the immediate postoperative period. Oliguria is considered present in the postoperative cardiac surgical patient when the urine output is less than 0.5 mL/kg/h. Transient oliguria is commonly noted in the first 12 hours after surgery and usually responds to a volume infusion or low‐dose inotropic support. However, the persistence of oliguria is usually a manifestation of an acute renal insult caused by a prolonged pump run, prolonged hypotension, or a low cardiac output state. The SCr will frequently be lower immediately after CPB and the following morning due to hemodilution, and may have a delayed rise despite a marked reduction in GFR, as it takes time for SCr to accumulate in the bloodstream. Thus, it is important to recognize that AKI may be associated with an abrupt and sustained decrease in urine output and/or a decline in GFR, prior to noting an increase in SCr. Measurement of kidney biomarkers may be the most sensitive means of early detection of AKI.23–26
Definition of acute kidney injury. The incidence of postoperative AKI depends on its definition, but it is estimated to occur in 30–50% of patients and is associated with a fourfold increase in mortality.73,86,154 Even very slight increases in postoperative SCr contribute to increased mortality.155
The RIFLE system was devised in 2004 to classify progressively worsening degrees of renal dysfunction (Table 12.3).156 Since the function of the kidneys is both elimination of nitrogenous waste products and the production of urine, either the SCr level/GFR criteria or the urine output criteria were used for classification in this model as well as in two additional models noted below. It should be noted that urine output is determined by the difference between the GFR and the rate of tubular reabsorption. Therefore, if the GFR is low from CKD in association with poor tubular absorption, the patient can have good urine output initially. With AKI, tubular absorption is initially normal with a low GFR and then falls.
In the RIFLE system (Table 12.3), there is a seven‐day window to assess changes in SCr. As the degree of AKI worsens, the risk of dialysis increases and short‐ and intermediate‐term mortality rates increase.157 In fact, once RRT is necessary, the mortality rate is usually about 50%.
In 2007, a slight modification of the RIFLE system was devised by the Acute Kidney Injury Network (AKIN) (Table 12.4).158 This did not utilize the GFR, did not require a baseline SCr, but did require two SCr determinations within 48 hours. It provided a fairly comparable assessment of AKI, but would not recognize AKI developing after 48 hours.159,160
In 2012, another classification system was proposed, termed the Kidney Disease Improving Global Outcomes (KDIGO) criteria.161 Acute kidney injury was defined as a 0.3 mg/dL increase in SCr from baseline within 48 hours of surgery, a 50% increase in SCr within seven days of surgery, or a decrease in urine output below 0.5 mg/kg/h for six hours.
Using each of these systems, the prognosis was worse if both the elevated SCr and oliguria criteria were met.8,162
The STS definition of AKI (version 2.42, available in 2020) was an increase in SCr to three times greater than baseline, or to >4 mg/dL with an acute rise >0.5 mg/dL, or the requirement for dialysis. This essentially corresponds to the “failure” category of RIFLE and stage 3 of the AKIN. However, oliguria was not taken into consideration.
Diagnosis of AKI. The early diagnosis of AKI can be difficult to make because of the delay in elevation of SCr and the maintenance of urine output despite nephron damage.
The SCr is influenced not just by glomerular function but also by tubular function and the generation of SCr. It is also influenced by patient gender, age, and muscle mass. Thus, it tends to underestimate the degree of renal dysfunction because, as the GFR falls, SCr secretion increases, minimizing the rise in SCr, upon which the GFR calculation is based. Although the SCr may not rise for several days after tubular injury has occurred, the eventual elevation in SCr does reflect changes in GFR, and is therefore valuable in confirming the diagnosis of AKI. A fall in SCr is an indicator of renal recovery from AKI, with the percentage decrease in SCr within 24 hours having the strongest correlation with long‐term outcomes.163
Elevation in plasma and urinary levels of kidney‐specific biomarkers, such as NGAL, cystatin C, KIM‐1, and IL‐18, may be noted within 2–6 hours of surgery and correlates with the extent and duration of AKI.23–26 These are valuable early indicators of AKI that precede elevation in SCr levels. Although cystatin concentration reflects baseline GFR more accurately than SCr and is independent of muscle mass, NGAL is rapidly induced in renal tubular cells in response to ischemic injury, and although its early appearance is independent of GFR, it is generally predictive of a subsequent decline in GFR. Because elevated NGAL levels may provide for earlier diagnosis and intervention of incipient AKI than elevations in SCr, some authors have recommended that all patients have baseline NGAL levels obtained for comparison with serial postoperative values.164
Nonoliguric renal failure, defined as a rise in SCr with a urine output >400 mL/day, is the most common form of AKI and may occur after an uneventful operation in a patient with preexisting renal dysfunction or risk factors for its development, and occasionally without any precipitating factors. This condition usually reflects less renal damage and is associated with a mortality rate of about 5–10%. Most patients can be managed by judicious fluid administration, hemodynamic support as indicated, and high‐dose diuretics to optimize urine output while awaiting recovery of renal function. Patients experiencing AKI have a higher rate of hospital readmission, which is more frequent with more advanced levels of AKI.165
Oliguric renal failure may occur in patients with varying degrees of reduction in GFR, but, when the urine output is <0.3–0.5 mL/kg/h for 12–24 hours and there is a twofold increase in SCr, significant AKI is present with a higher likelihood of requiring RRT and a high mortality rate. It is estimated that the mortality of patients requiring dialysis is three times higher than those with nonoliguric AKI.166 This high mortality rate has not changed much over the past 10–15 years despite the early institution of various forms of RRT and general improvements in postoperative care. This reflects the higher‐risk population undergoing surgery and the morbidity of conditions frequently associated with renal failure, such as low cardiac output states, respiratory failure, infection, and stroke.
Etiology and pathophysiology of postoperative AKI8–11,167
In patients with preexisting renal dysfunction, the complex effects of extracorporeal circulation will often induce some degree of AKI. Mechanisms include renal hypoperfusion from low‐flow, low‐pressure nonpulsatile perfusion with hemodilution and hypothermia, as well as an inflammatory response that may maintain afferent arteriolar constriction. The duration of CPB is therefore a major risk factor for the development of AKI.3,75–77 Using most of the recommendations delineated in Table 12.2, it may be possible to minimize the intraoperative insult and allow renal function to return to baseline within a few days if no additional insult occurs. However, the most common cause of a prolonged renal insult is a low cardiac output syndrome which may be present at the termination of CPB and may extend well into the early postoperative period in the ICU. An additional contributing factor is intense peripheral vasoconstriction, often related to use of α‐agents. Oliguria occurring as a consequence of reduced GFR is most clinically significant early after surgery when fluid overload and hyperkalemia can lead to pulmonary and myocardial complications and impair recovery from surgery.
The kidneys have a tremendous capacity to autoregulate and maintain RBF, GFR, filtration fraction, and tubular reabsorption in the face of reduced renal perfusion pressure. Intrinsic renal mechanisms that maintain autoregulation include a reduction in afferent arteriolar resistance and an increase in efferent arteriolar resistance. However, when a low cardiac output state or hypotension persists or potent vasopressor medications are used, these compensatory reserves gradually become exhausted, filtration reserve is exceeded, and endogenous and/or exogenous vasoconstrictors increase afferent arteriolar resistance, resulting in a fall in GFR. At this point of prerenal azotemia, oliguria may occur, but tubular function may still be intact. Aggressive management to optimize renal perfusion at this time is essential to try to avoid tubular damage.
However, a more protracted period of ischemia will eventually cause structural tubular injury with sloughing of cells that may obstruct the tubules with back leakage of fluid into the circulation. Impaired sodium absorption and increased sodium concentration in the distal tubules polymerizes proteins, contributing to cast formation. Oxidant injury and inflammatory phenomena result in further hypoperfusion and damage to tubular cells. Some of this damage is reversible and some results in apoptotic cell death. The term “acute tubular necrosis” (ATN) has commonly been applied to this condition, although it is a somewhat misleading term; therefore it is more commonly referred to as “acute kidney injury” (AKI). Thus, what usually originates as a prerenal “hypoperfusion” picture soon causes intrinsic renal damage.
It should be noted that an acute ischemic renal insult is a hypoperfusion injury that may be undetected in a normotensive patient.168 If autoregulation is impaired, the kidney may be more susceptible to lesser degrees of hypoperfusion. Factors to consider are:
Renal arteriolar disease, notably in elderly patients and those with hypertension, CKD, or renal artery stenosis.
Failure of afferent arterioles to dilate appropriately. NSAIDs and Cox‐2 inhibitors decrease prostaglandin synthesis and allow endogenous vasoconstrictors to act unopposed; sepsis and liver failure increase afferent arteriolar vasoconstriction.
Use of vasoconstrictors during a low output state, which tend to reduce RBF despite achieving systemic normotension.
Failure of efferent arterioles to constrict, noted with ACE inhibitors, ARBs, or direct renin inhibitors, such as aliskiren (Tekturna).
Systemic venous hypertension, often as a result of right ventricular failure, tamponade, or an abdominal compartment syndrome, which may reduce renal perfusion.169
The acute development of oliguria and a rising SCr several days after surgery should always raise the specter of cardiac tamponade. The combination of systemic venous hypertension and a low output state can compromise renal perfusion even if hypotension is not evident. The patient may have nonspecific systemic symptoms, and a compensatory tachycardia may be absent with the use of β‐blockers.
Conditions of impaired oxygen delivery (profound anemia from bleeding, hypoxemia from respiratory failure) may contribute to renal ischemia if there is borderline hypoperfusion.
Three patterns of acute renal failure were described following open‐heart surgery over 30 years ago and in principle still hold true today (Figure 12.5).170 In the first, termed “abbreviated ARF”, a transient intraoperative insult occurs that causes renal ischemia without tubular damage. The SCr peaks on the fourth postoperative day and then returns to normal. In the second pattern, termed “overt ARF”, the acute insult is followed by a more prolonged period of cardiac dysfunction and is associated with mild tubular damage. The SCr usually rises to a higher level and gradually returns towards baseline over the course of 1–2 weeks once hemodynamics improve and tubular cell regenerate. The third pattern (“protracted ARF”) is characterized by an initial insult followed by a period of cardiac dysfunction that resolves. Just as the SCr begins to fall, another insult, often from sepsis or a period of hypoperfusion or hypotension, occurs that triggers a progressive, often irreversible, rise in SCr. A fourth pattern of “protracted ARF” that may be added to this description is that of acute AKI that results from a very severe initial insult that may occur intraoperatively and/or during the early postoperative period that causes extensive tubular damage from the outset and does not improve for quite some time, if at all.
Assess cardiac hemodynamics (filling pressures, cardiac output). If the patient is no longer being intensively monitored, insertion of a Foley catheter is helpful in assessing urine output. Evidence of jugular venous distention or orthostatic vital signs raise the specter of tamponade. An echocardiogram may be considered to assess ventricular function and the presence of a significant hemopericardium.
Identify any drugs being prescribed with potential adverse effects on renal function (ACE inhibitors, ARBs, NSAIDs, nephrotoxic antibiotics).
Obtain a serum BUN, SCr, and electrolytes. Note: an elevation in SCr with minimal or parallel rise in BUN is frequently noted with AKI. In contrast, a disproportionate rise in BUN with little rise in SCr may reflect a prerenal process or increased protein intake, total parenteral nutrition (TPN), GI bleeding (often associated with prerenal azotemia), hypercatabolism, or steroid administration, which increase urea production.
Examine the urinary sediment. Tubular epithelial or granular (“muddy brown”) casts are indicative of tubular injury, whereas hyaline casts are seen in low perfusion states. The sediment is important to examine because tests of tubular function, such as the urine sodium and osmolality, may be inaccurate with use of diuretics.
Measure the urine sodium (UNa) and creatinine (UCr) concentrations. These tests can differentiate prerenal from renal causes, but their interpretation will be influenced by the use of diuretics. A UNa <20 mEq/L is strongly suggestive of prerenal disease, but it could be elevated if the urine output is low. The fractional excretion of sodium (FENa) is a better marker of renal sodium handling because it normalizes sodium handling against the secretion of creatinine, thus being independent of urine concentration. This is calculated as:
where U and P refer to the urinary and plasma concentrations, respectively, of sodium and creatinine.
In the oliguric patient, an FENa <1% reflects retained tubular function with absorption of sodium and water, consistent with a prerenal problem, except in some cases of contrast nephrotoxicity, HF, and hepatorenal syndrome. In contrast, an FENa >2% is usually caused by ATN with tubular damage. However, this may also be noted when a prerenal process is superimposed on CKD, with which the kidneys at baseline cannot conserve water and sodium appropriately. A rise in FENa may be noted during recovery of renal function due to sodium mobilization.
Monitor other electrolytes (especially potassium), blood glucose, and acid–base balance frequently.
Obtain a renal ultrasound to assess kidney size and rule out obstruction. A renal scan may be performed if a renal embolus is suspected.
Management of oliguria and AKI (Table 12.6).171–173 Early aggressive intervention in patients with oliguria and early evidence of AKI may prevent progressive tubular injury and worsening of renal function. However, once AKI is established, very little can be done to promote recovery of renal function except to prevent additional insults. There is little evidence that strategies that increase RBF or increase urine flow to reduce tubular obstruction have any impact on enhancing tubular epithelial cell proliferation and recovery of function. Generally, attention should be directed towards maintaining urine output to reduce tissue edema and treating electrolyte or metabolic problems as they arise.
Ensure that the Foley catheter is within the bladder and is patent (this may rule out an obstructive uropathy). Irrigate with saline if necessary or consider changing the catheter empirically. If the Foley catheter has been removed, a bladder scan may indicate whether oliguria is real or spurious. Significant urinary retention may provide evidence of a post‐obstructive uropathy as the cause of an elevated SCr. Either way, replacement of the catheter may be helpful in further assessing the urine output.
Discontinue all potentially nephrotoxic drugs (ACE inhibitors, ARBs, NSAIDs, nephrotoxic antibiotics) and avoid any diagnostic studies requiring IV contrast.
Optimize hemodynamics. Although augmenting the cardiac output may not be able to expedite recovery of renal function, it is clear that any additional insult that causes hypotension or hypoperfusion may contribute to a state of “protracted ARF”. These insults include hypovolemia (often GI bleeding), low cardiac output states (tamponade), arrhythmias (rapid atrial fibrillation, ventricular tachycardia), antihypertensive medications, or sepsis. Thus, there is little downside to optimizing hemodynamics to increase urine output even if the rate of renal recovery is not hastened.
Hemodynamic monitoring with a Swan‐Ganz catheter may be indicated if a low cardiac output state is suspected. If the diagnosis is not clear, echocardiography can differentiate ventricular failure from tamponade or significant hypervolemia. Otherwise, assessment of fluid balance, strict I & O’s, and/or a careful physical examination may give an overall assessment of the patient’s fluid balance and intravascular volume.
Optimize preload without being overzealous with fluid administration. Remember that, in a state of capillary leak (often seen following surgery with a long duration of CPB, with a persistent low output state, or with sepsis) or with reduced oncotic pressure (as noted from hemodilution or poor nutritional condition), excessive fluid administration may produce noncardiogenic pulmonary edema. Fluid administration with solutions with high chloride content (5% albumin, normal saline, Hespan, Volvuven) should be minimized as they may increase the risk of AKI.42,49–51
Optimize heart rate and treat arrhythmias. Increasing the heart rate with atrial or AV pacing (V pacing only if there is no atrial capture) above 80/min to augment the cardiac output might prove beneficial in improving renal perfusion and GFR. Successful electrical or pharmacologic conversion of atrial fibrillation will improve cardiac output.
Improve contractility with inotropes if a low cardiac output state is present.
Reduce afterload with vasodilators to improve cardiac function, but do so carefully; eliminate drugs that can cause renal vasoconstriction; avoid ACE inhibitors and ARBs.
Do not be overly aggressive in the reduction of systemic blood pressure in patients with preexisting hypertension and CKD. They usually require a higher blood pressure (130–140 mm Hg systolic) to maintain renal perfusion. In fact, while the patient is in the ICU, adding an α‐agent to increase the blood pressure to that range often results in a significant improvement in urine output.
If inotropic drugs with vasodilator properties are used, such as milrinone or dobutamine, an α‐agent may be necessary to maintain systemic blood pressure. Vasopressin can be used in vasodilated states with a good cardiac output (“vasoplegia”). Norepinephrine is preferable if the cardiac output is borderline because it will also provide some inotropic support. Use of a pure α‐agent, such as phenylephrine, is more likely to cause renal vasoconstriction unless the cardiac output is excellent.
If the cardiac output remains marginal despite the use of multiple inotropes, consider the placement of an IABP. This may result in an abrupt and dramatic increase in urine output.
If oliguria persists despite optimization of hemodynamics, the next step is selection of a diuretic, conditional upon the patient’s volume status. The majority of studies have shown that loop diuretics do not prevent AKI, do not improve renal functional recovery or alter the natural history of AKI, do not decrease the need for renal replacement therapy, and in fact may increase operative mortality and delay recovery of renal function.174–177 If the patient is euvolemic or mildly fluid overloaded, use of diuretics to simply improve urine output is not indicated. However, loop diuretics may improve urine output and can often convert oliguric to nonoliguric renal failure if administered early after the onset of renal failure. An improvement in urine output (diuretic‐responsive AKI) suggests that the extent of renal injury is less severe, coincidentally leading to an earlier decrease in SCr. Thus, it is potentially beneficially to administer diuretics to reduce fluid overload, primarily to optimize pulmonary function, although this will not hasten recovery of renal function.
Furosemide is given in incremental doses starting at 10 mg IV. However, once acute renal failure is established, a dose of 100 mg IV is commonly required and should be given over 20–30 minutes to minimize ototoxicity. If urine output fails to increase within a few hours, the following steps can be taken:
Increase the dose of furosemide up to 200 mg IV (limiting the cumulative daily dose to 1 g).
Use a continuous infusion of IV furosemide. Give a loading dose of 40–100 mg, and then initiate an infusion of 10–20 mg/h. Rebolus before an increase in the infusion rate. This may be the best means of maintaining an adequate urine output.
Alternatively, bumetanide can be given either as a bolus dose of 4–10 mg IV or as a 1 mg load followed by a continuous infusion of 0.5–2 mg/h depending on the estimated creatinine clearance. There is little evidence that one loop diuretic is better than any other, but some patients respond better to one than the other.
Various combinations of medications may be effective in improving diuresis.
Add a thiazide diuretic to the loop diuretic. These include chlorothiazide 500 mg IV, metolazone (Zaroxolyn) 5–10 mg PO or via a nasogastric tube, or hydrochlorothiazide 50–200 mg PO qd. Thiazides block distal nephron sites and act synergistically with the loop diuretics to increase exposure of the distal tubules to solute. This combination is particularly effective in patients who tend to be diuretic‐resistant.178 The thiazide should be given 20 minutes before the loop diuretic to prime the distal tubules; otherwise, there may be compensatory sodium reabsorption in the distal tubules.
Although dopamine has not been shown to be effective in preventing AKI, reducing its severity and duration, or lowering the risk of dialysis, one study found that the combination of mannitol (500 mL of 20% Osmitrol) + furosemide (1 g) + dopamine (2–3 μg/kg/min) started within the first six hours of oliguria produced a significant diuresis with early restoration of renal function.179
Fenoldopam has shown equivocal results in the treatment of established AKI after cardiac surgery. One study of patients receiving a 72‐hour infusion for early acute AKI (a 50% rise in SCr) showed a trend towards reduction in mortality and the need for dialysis in nondiabetic patients, but other studies have not shown much benefit.180,181
Diltiazem given intra‐ and postoperatively (primarily to prevent radial artery graft spasm) has shown an insignificant improvement in creatinine clearance, so whether it has any role in the management of AKI is not clear.182
Note: mannitol is an osmotic diuretic that is frequently used during surgery to increase serum osmolality during hemodilution to minimize tissue edema. It improves renal tubular flow, reduces tubular cell swelling, and also improves urine output. In patients with early postoperative AKI, it increases RBF by decreasing renal vascular resistance, but does not affect filtration fraction or renal oxygenation.183 Nonetheless, it is best avoided in the postoperative period because its oncotic effect mobilizes fluid into the intravascular space. This could theoretically lead to pulmonary edema if fluid overload is present and urine output does not improve. In fact, a significant increase in serum osmolality can cause renal vasoconstriction and induce renal failure.
Management of established renal failure
Once oliguric renal failure is established, treatment should be directed towards optimizing hemodynamics while minimizing excessive fluid administration, providing appropriate nutrition, and initiating early renal replacement therapy to hopefully reduce morbidity and improve survival. The blood pressure should be maintained at a higher level than usual in hypertensive patients whose kidneys may require higher perfusion pressures.
Restrict fluids with mL/mL of fluid replacement (i.e. input = output) plus 500 mL D5W/0.2% normal saline/day (about 200 mL/m2/day). Daily weights are helpful in assessing changes in day‐to‐day fluid status, but must also take into consideration the influence of nutritional status on body mass.
Monitor electrolytes and blood glucose
Avoid potassium supplements and medications that increase potassium levels (β‐blockers, ACE inhibitors). Correct hyperkalemia as described on pages 709–711.
Hyponatremia, if associated with inappropriate ADH (SIADH), should be treated with fluid restriction. The hypovolemic patient with hyponatremia may need isotonic saline.
Metabolic acidosis is common with acute/chronic kidney injury and does not need to be corrected if the serum bicarbonate is >15 mEq/L. If it is lower, a potential contributing cause to a hypoperfusion issue should be sought and corrected.
Correct hyperglycemia and abnormalities of calcium, phosphate, or magnesium metabolism.
Eliminate drugs that impair renal perfusion or are nephrotoxic (ACE inhibitors, ARBs, aminoglycosides, NSAIDs).
Avoid or adjust doses of medications that are excreted or metabolized by the kidneys (particularly low‐molecular‐weight heparin, and renally excreted antibiotics) (see Appendices 12 and 13). Unfractionated heparin may be used for patients with mechanical heart valves until the INR becomes therapeutic on warfarin. NOACs (especially apixaban) can be used in reduced dosage for other patients with indications for anticoagulation other than mechanical heart valves.
Give antacid medications (proton pump inhibitors [PPIs]) to minimize the risk of GI bleeding, but avoid magnesium‐containing antacids and laxatives. However, PPIs have been linked to hypomagnesemia, interstitial nephritis, and other renal issues when used for more than one week.184
Remove the Foley catheter and catheterize daily or prn depending on the urine output. Culture the urine if clinically indicated.
Improve the patient’s nutritional state with enteral nutrition if possible.185,186 If the patient is able to eat, an essential amino acid diet should be used. Protein should not be restricted if the patient is on hemodialysis, which can result in the loss of 3–5 g/h of protein. Patients on dialysis should receive approximately 25–30 Kcal/kg of nutrition with a minimum of 1.5 g/kg/day of protein.
If a patient on dialysis is unable to eat but has a functional GI tract, a high nitrogen tube feeding can be used. For most patients with acute renal failure, there is no need to alter the amount of protein, and standard tube feedings can be used unless hyperkalemia is present. In patients with CKD who do not require dialysis, a low‐protein supplement can be used to provide 0.5–0.8 g/kg/day of protein.
If the patient is unable to tolerate enteral feedings, total parenteral nutrition using a 4.25% amino acid/35% dextrose solution that contains no potassium, magnesium, or phosphate is recommended.
Consider the prompt initiation of renal replacement therapy.
Table 12.4 AKIN Classification System of Acute Kidney Injury
Increased Scr × 1.5 or an increase ≥0.3 mg/dL
<0.5 mL/kg/h × >6 h
Increased Scr × 2
<0.5 mL/kg/h × >12 h
Increased Scr × 3 or Scr ≥4 mg/dL with an acute rise >0.5 mg/dL
Ensure that Foley catheter is in the bladder and is patent
Optimize cardiac function
Treat hypovolemia: minimize use of chloride‐liberal solutions
Control arrhythmias and pace for slow rhythms
Reduce elevated afterload, but allow BP to drift up to 130–140 mm Hg
Diuretics or other medications
Give increasing doses of furosemide (up to 500 mg IV) or a continuous infusion of 10–20 mg/h
Add chlorothiazide 500 mg IV to the loop diuretic
Consider bumetanide 4–10 mg or 1 mg bolus, then a 0.5–2 mg/h infusion
Consider use of fenoldopam 0.1 μg/kg/min, especially if hypertensive
If above fail
Limit fluid to insensible losses
Readjust drug doses
Avoid potassium supplements
Nutrition: essential amino acid diet
High nitrogen tube feeds if on dialysis
Total parenteral nutrition with 4.25% amino acid/35% dextrose
Consider early renal replacement therapy
VII. Renal Replacement Therapy
Various forms of renal replacement therapy (RRT) can be used to remove excessive fluid and solute to improve electrolyte balance and remove other nitrogenous waste products (Table 12.7).187
Indications. The most important indications for initiation of RRT are fluid overload, hyperkalemia, and metabolic acidosis. Other signs of uremia, such as a change in mental status, pericarditis, or GI bleeding, should also prompt initiation of RRT, although these are uncommon in the acute setting. However, a very important and sometimes difficult decision to make is whether RRT should be initiated at the first sign of persistent oliguria or a rising SCr, especially since the latter tends to lag behind the extent of renal dysfunction. Some studies of cardiac surgical patients have shown that early and aggressive dialysis, before the patient develops signs and symptoms of renal failure and before a marked elevation in SCr occurs, improves outcomes.188,189 Other studies in ICU patients have shown conflicting results.190,191 Certainly, when marked oliguria is present early after surgery in a patient with significant fluid overload, and when there is a poor response to diuretics, a delay in initiating RRT may lead to respiratory compromise and prolonged ventilation with its attendant risks. On the other hand, some patients may recover renal function without the need for dialysis, which has numerous risks, including hypotension, which can prolong renal failure.
An additional consideration in patients with moderate renal dysfunction (SCr >2–2.5 mg/dL) is use of “prophylactic” preoperative dialysis. Several studies have demonstrated that this reduces the need for postoperative RRT, with a reduction in overall morbidity and mortality.117–119
RRT involves several processes to remove volume, solute, or waste products. A transmembrane gradient (basically hydrostatic pressure) drives plasma water (ultrafiltration) and solute (convection) across a semipermeable membrane. Solute removal by diffusion is driven by the difference in the solute concentration between plasma and an electrolyte solution on the other side of the membrane.
The most common forms of RRT used in cardiac surgery patients are those of intermittent hemodialysis and continuous veno‐venous hemofiltration. Selection of the appropriate modality depends on the indications for its use (whether primarily for volume or solute removal), and the hemodynamic stability of the patient. Both of these approaches are associated with comparable outcomes, such as recovery of renal function and survival.189–191
Intermittent hemodialysis (HD)
Principle. Solute passes by diffusion down a concentration gradient from the blood, across a hollow‐fiber semipermeable membrane, and into a dialysate bath. Fluid removal by ultrafiltration and solute transport by convection also occur due to the differences in hydrostatic pressure on either side of the membrane.
Indications. HD is indicated for the management of hyperkalemia, acid–base imbalances, fluid overload, or a hypercatabolic state in the hemodynamically stable patient. It is the most efficient means of removing solute (urea, creatinine) and correcting severe acid–base abnormalities. It can be combined with ultrafiltration to remove excess volume.
Access. Standard intermittent HD is performed using a single 12 Fr double‐lumen catheter (such as the Mahurkar [Medtronic] or Niagara, DuoGlide and Power‐Trialysis catheters [Bard]) preferentially placed in the internal jugular vein for short‐term dialysis. Placement in the subclavian vein can cause venous thrombosis, which could impair ipsilateral fistula maturation if that is required in the future and should be avoided. Femoral venous access is feasible but impairs patient mobility. To reduce the risk of infection in patients requiring more extended periods of dialysis, a double‐lumen Permcath (Medtronic) or HemoGlide (Bard) long‐term hemodialysis catheter can be placed into the internal jugular vein and brought through a subcutaneous tunnel as a “midline” catheter. Subsequently, a fistula can be created for permanent dialysis. When recovery appears unlikely and a fistula is being considered, the arm vessels on one side should be protected from use as much as possible.
Technique. Intermittent HD is performed over a three‐ to four‐hour period and is usually performed at least three times per week until renal function recovers. The blood is pumped into the dialysis cartridge at a rate of 300–400 mL/min, while the dialysate solution is infused at a rate of 500–700 mL/min in a direction countercurrent to blood flow. Although heparin is commonly used, heparin‐free HD is possible in patients with bleeding problems or heparin‐induced thrombocytopenia.
Circulatory instability with hypotension from a blunted sympathetic reflex response to hypovolemia is the most common complication of HD, especially if large volumes are being removed in a short period of time. Rapid removal of volume can cause myocardial stunning, and ultrafiltration exceeding 13 mL/kg/h is associated with increased cardiovascular mortality.192 Too high a blood flow during initiation of dialysis will reduce plasma osmolality, prompting water movement into cells, exacerbating the depletion of extracellular volume. This problem has been mitigated to some degree by use of biocompatible membranes, bicarbonate baths, initial high dialysate sodium, cool temperatures, and volumetric control during dialysis. Colloid or blood transfusions and hemodynamic support (usually with α‐agents or PO midodrine) are frequently necessary. About 20–30% of patients develop hypotension during HD, and about 10% of patients cannot tolerate it because of hemodynamic instability. This is probably more common in patients early in the postoperative period, especially if hemodynamic compromise was a major contributing factor to their developing AKI. Therefore, HD is best avoided in the hemodynamically unstable patient.
Dialysis machines are complex and costly, and require special expertise.
Continuous veno‐venous systems can be used in a variety of ways, including slow continuous ultrafiltration (SCUF), continuous veno‐venous hemofiltration (CVVH), and continuous veno‐venous hemodialysis or hemodiafiltration (CVVHD).
Principle. An occlusive pump is included in a circuit that actively withdraws blood at a designated rate from the venous system, pumps it with hydrostatic pressure through the membrane of a hemofilter, and then returns the blood to the venous system. This circuit achieves filtration or convection of plasma water. For CVVHD, dialysis fluid runs countercurrent to the direction of blood flow within the ultrafiltration membrane. Solute then passes by diffusion down a concentration gradient across a hemofilter into the dialysate solution.
Indications. These systems are indicated for the management of fluid overload, especially in the hemodynamically unstable or hypotensive patient. Slow correction of electrolyte imbalance can be achieved with CVVH using a crystalloid solution of different composition for replacement fluid. Severe electrolyte imbalance and hypercatabolic state are better managed with CVVHD.
Access is obtained using a 12 Fr double‐lumen catheter (12 gauge for each lumen) placed in the internal jugular or femoral vein.
Most continuous veno‐venous circuits are “integrated” systems that include a blood pump, pressure monitors, air detector with shut‐off controls, and fluid balancing systems for ultrafiltrate control. A high‐efficiency biocompatible hemodialysis cartridge is attached downstream to an occlusive pump and heparin is infused into the inflow portion of the circuit to maintain a dialyzer output (venous) PTT of 45–60 seconds. Alternatively, regional citrate anticoagulation can be infused into the inflow limb instead of heparin, especially in patients with bleeding or heparin‐induced thrombocytopenia.193,194 When citrate is used, steps must be taken to avoid hypocalcemia and metabolic alkalosis. Calcium should be infused in the venous return line postfilter and alkaline buffers must be reduced in replacement fluids for CVVH or in the dialysate solution for CVVHD.
SCUF removes fluid by ultrafiltration, but does not remove much solute. The blood flow rate is set at 50–80 mL/min and the ultrafiltrate rate is set at the desired amount (about 5 mL/min), which can potentially achieve a net negative fluid balance of up to 7 L/day. The filter is more prone to clotting due to the slow flow and because the postfilter hematocrit is high. Since no replacement fluid is given and minimal solute is removed, SCUF is used primarily to treat volume overload since it is ineffective for uremia or hyperkalemia.
With CVVH, fluid and solute are transported by ultrafiltration and convection, respectively, and no dialysate solution is used. The pump is usually set to deliver blood at a rate of 50–300 mL/min and the ultrafiltrate rate is usually set at a preselected rate around 16.7 mL/min or 1 L/h (range of 0.5–4 L/h). The blood then passes through a bubble trap air detector and is returned to the patient. Replacement fluid (alternating 1 L of 0.9% NS plus 1 ampoule [10 mL of a 10% solution] of calcium gluconate with 1 L of 0.45% NS plus 1 ampoule [50 mEq/50 mL] of 8.4% NaHCO3) is infused into the outflow (return) circuit or into the venous chamber to correct electrolyte and acid–base imbalances. The amount administered is dictated by the desired negative fluid balance per hour. This technique can achieve a moderate amount of solute and fluid removal due to the high ultrafiltration rate. Clotting of the system is less likely because the dialyzer can be prediluted with a large volume of fluid.
CVV hemodialysis (CVVHD) transports solute by diffusion with variable degrees of ultrafiltration to remove fluid, whereas CVV hemodiafiltration combines the convective solute removal of CVVH with the diffusive solute removal of hemodialysis, along with ultrafiltration. The blood flow is set at 50–300 mL/min and the dialysate (Dianeal 1.5% with 4 mL of 23% NaCl per 2 L bag) is infused into the dialysis cartridge at a rate of 1–2 L/h or even higher for hemodiafiltration. These techniques are most useful in the highly catabolic patient to remove a large solute load. The effluent flow rate (hemofiltration rate + the dialysate flow rate) should be about 20–25 mL/kg to achieve adequate solute clearance.
Advantages and limitations
Citrate anticoagulation eliminates concerns about use of heparin.
High flow rates of CVVH/CVVHD reduce the potential for clotting of the filter noted with SCUF.
Use of the blood pump enables CVVH to be performed when the patient is hypotensive or hemodynamically unstable.
Because CVVH removes so much fluid, it is essential to carefully monitor electrolytes and modify the replacement solutions as necessary to maintain electrolyte balance.
The pump adds some complexity and cost to the system compared with arteriovenous hemofiltration.
Continuous arteriovenous hemofiltration (CAVH) is used during surgery to remove excessive fluid before terminating CPB. It is beneficial in improving hemodynamics, hemostasis, and pulmonary function in higher‐risk patients. Postoperatively, its use is limited by the need for arterial access, heparinization to minimize clotting of the hemofilter, and the requirement for satisfactory arterial pressure to provide the hydrostatic pressure to achieve hemofiltration. Because of these drawbacks, CAVH has been replaced by CVVH in most units.
Peritoneal dialysis (PD) is rarely used in cardiac surgical patients, because it produces abdominal distention and glucose absorption that can compromise respiratory status, and it carries the risk of peritonitis. Its use is usually limited to patients on chronic peritoneal dialysis.
Table 12.7 Techniques of Renal Replacement Therapy (Hemofiltration and Hemodialysis)
High‐volume, high‐potassium cardioplegia solutions used in the operating room. The potassium load is usually eliminated promptly by normally functioning kidneys, but hyperkalemia can be problematic in patients with acute or chronic renal dysfunction or oliguria from other causes.
Low cardiac output states associated with oliguria. Potassium levels may rise with alarming and life‐threatening rapidity.
Severe tissue ischemia, whether peripheral (from severe peripheral vascular disease or complication of an IABP) or intra‐abdominal (mesenteric ischemia). Hyperkalemia is often the first clue to the existence of these problems.
Acute and chronic renal insufficiency
Medications that impair potassium excretion or increase potassium levels (ACE inhibitors, ARBs, potassium‐sparing diuretics [triamterene, spironolactone, amiloride, eplerenone], NSAIDs, β‐blockers)
Severe hyperglycemia, which increases potassium release from cells into the extracellular fluid
Note: remember that hyperkalemia is exacerbated by acidosis, which often accompanies low‐output or ischemic syndromes. A 0.2 unit change in pH produces about a 1 mEq/L change in serum potassium concentration. However, in conditions of organic acidosis, the potassium is more likely to rise from tissue breakdown and release of potassium from cells (lactic acidosis) or from insulin deficiency and hyperglycemia (ketoacidosis), than from a change in pH.
Manifestations are predominantly electrocardiographic due to depolarization of cardiac cell resting membrane potentials, which decreases membrane excitability. An asystolic arrest may occur when the potassium rises rapidly to a level exceeding 6.5 mEq/L. The ECG changes of hyperkalemia do not always develop in classic progressive fashion and are more related to the rate of rise of serum potassium than to the absolute level. These changes include:
Peaked T waves
Smaller R waves
Prolonged PR interval
Loss of P waves
QRS widening, bradycardia, ventricular fibrillation, and asystole. When the heart is being paced, hyperkalemia may result in failure to respond to the pacemaker stimulus.
Treatment entails stabilizing the cell membrane, shifting potassium into cells, and increasing its excretion from the body (Table 12.8).195,196 A hyperkalemic emergency is present when there are ECG changes, a serum potassium >6.5 mEq/L, or a level >5.5 mEq/L with acute renal impairment and ongoing tissue breakdown or potassium absorption (GI bleeding). These patients need an emergency reduction in their potassium levels. In other patients, a slower reduction is feasible.
It is essential to identify and remove any potential source of potassium intake or medications that may increase the potassium level (as above); use a low‐potassium diet in patients with renal dysfunction and persistent hyperkalemia.
Address cardiac toxicity:
If there is evidence of advanced cardiac toxicity or ECG changes, usually when the potassium is >6.5 mEq/L, administer calcium gluconate 10 mL of a 10% solution (1 g) IV over 2–3 minutes to stabilize the cell membranes. The beneficial effect should be noted within minutes, but lasts only about 30 minutes. Note that calcium potentiates the cardiotoxic effects of digoxin, so hyperkalemia in patients on digoxin should preferentially be treated with digoxin‐specific antibody fragments (Digibind), although a slow infusion of calcium gluconate can safely be given over 30 minutes.
Alternatively, calcium chloride 5–10 mL of a 10% solution (0.5–1 g) can be infused through a central line over several minutes. This may produce hemodynamic changes, but provides four times more elemental calcium than calcium gluconate.
Shift potassium into cells:
Regular insulin 10 units in 50 mL of a 50% dextrose solution IV. This should lower the potassium about 0.5–1.5 mEq/L within 15 minutes and last for several hours. Another regimen is 10–20 units in 500 mL of 10% dextrose given over 60 minutes. This can be considered first if there is marked hyperkalemia but no ECG changes.
Sodium bicarbonate (NaHCO3) administration in patients with acidosis results in hydrogen release from cells in exchange for potassium movement into cells. Thus, correction of acidosis should lower potassium levels. Bicarbonate also has a direct effect on hyperkalemia independent of a change in pH.197 However, boluses of bicarbonate are not that effective in reducing serum potassium, so in hyperkalemic emergencies, an infusion (150 mEq in one liter of D5W given over 2–4 hours) is recommended. In hyponatremic patients, the sodium load may reverse some of the ECG changes of hyperkalemia.
β2‐agonists activate the Na+–K+–ATPase system to drive potassium into cells and can lower the potassium level by 0.5–1.5 mEq/L. The only recommended drug is albuterol 10–20 mg in 4 mL NS by nebulizer over 10 minutes, which has a peak effect in about 90 minutes. Epinephrine (0.05 μg/kg/min) may also be effective, but it is best avoided in the cardiac patient unless there is an inotropic requirement.
Enhance potassium excretion because the effects of calcium and glucose‐insulin infusions are short‐lived.
Furosemide 20–40 mg IV is very effective in reducing serum potassium in patients without significant renal impairment. Diuretics may be given alone in patients with hypervolemia, but should be combined with a saline infusion in patients who are euvolemic or hypovolemic. Continuous infusions may also be used. In patients with impaired renal function, IV saline or isotonic bicarbonate should be given along with the diuretic, usually in higher doses, since sodium delivery to the distal tubules is essential to promote potassium excretion.
Hemodialysis is indicated in patients with hyperkalemia and severe renal dysfunction. It can remove up to 50 mEq of potassium per hour.
Sodium polystyrene sulfonate (Kayexalate) is a cation‐exchange resin that may be given orally (15–30 g in 60–120 mL of 20% sorbitol) or as an enema (50 g in 150 mL of tap water, but NOT with sorbitol). Each gram may bind up to 1 mEq of potassium. It may be considered if the other measures have failed to reduce the potassium or are not immediately available. However, it should not be given to patients with an ileus or receiving narcotics, which includes most postoperative patients. Kayexalate has been associated with the occurrence of colonic necrosis which may be related to decreased colonic motility, and may occur with or without use of sorbitol.198
Sodium zirconium cyclosilicate (ZC‐9) is a selective cation‐exchanger that entraps potassium in the intestinal tract in exchange for sodium and hydrogen. It is given in doses starting at 2.5 g three times a day and produces a rapid and dose‐dependent reduction in potassium levels within 48 hours.199,200
Patiromer (8.4 g orally) is a GI cation‐exchanger that binds potassium ions in the GI tract in exchange for calcium ions, so less potassium is absorbed and more is excreted in the stool. Potassium starts falling in seven hours and continues to decrease for 48 hours. It should be considered if hemodialysis is not immediately available in a patient with severe renal impairment, but is usually used for the chronic, rather than acute, treatment of hyperkalemia.200
Profound diuresis without adequate potassium replacement. Potassium excretion parallels the urine output after CPB, which tends to be copious because of hemodilution. The use of potent diuretics may produce a significant diuresis and kaliuresis in the early postoperative period.
Insulin to treat hyperglycemia, which redistributes potassium from the extracellular fluid into cells (“redistributive hypokalemia”)
Hypomagnesemia (usually from diuresis) which can cause refractory hypokalemia
Alkalosis (metabolic or respiratory) (the “L” in alkalosis tells you the potassium is “lowered”)
Significant nasogastric tube drainage
Manifestations. The primary concern with hypokalemia in the cardiac patient is the induction of cardiac reentrant arrhythmias due to enhanced cardiac automaticity and delayed ventricular repolarization. Hypokalemia may produce atrial, junctional, or ventricular ectopy (PACs, PVCs), paroxysmal atrial and junctional tachycardias, AV block, and VT/VF. The ECG may demonstrate flattened ST segments, decreased T‐wave amplitude, and the presence of “u” waves. Factors that may promote the development of hypokalemic arrhythmias include myocardial ischemia, enhanced sympathetic tone (often from epinephrine or β2‐agonists), digoxin, and low magnesium levels (commonly seen after CPB). Hypokalemia can also result in weakness involving the respiratory muscles, the GI tract (producing an ileus), or the skeletal muscles.
Treatment is indicated for any potassium level below the normal range, although ECG changes do not become evident until the level is <3 mEq/L.201,202 Most patients with diuretic‐induced hypokalemia have a metabolic alkalosis associated with chloride depletion. Low chloride levels enhance bicarbonate reabsorption in the kidneys, and sodium is exchanged for potassium rather than excreted with chloride. Thus, failure to replenish chloride leads to more potassium wasting.
It is essential that renal function and urine output be evaluated before a potassium chloride (KCl) drip is started, because acute hyperkalemia can develop very rapidly when oliguria or renal dysfunction is present. A slower infusion rate is advisable in this situation, with frequent rechecking of the serum potassium level. Particular attention should be paid to sources of urinary or GI loss of potassium that may require more aggressive replacement.
Serum magnesium levels should be checked because hypomagnesemia can make the hypokalemia refractory to treatment; if low, magnesium should be replaced.
In the ICU setting, KCl is administered through a central line at a rate of 10–20 mEq/h (mix of 20–40 mEq/100 mL in 0.45% NS). An infusion pump should always be considered when the IV bag contains more than 40 mEq or the infusion rate exceeds 10 mEq/h, to avoid a catastrophic hyperkalemic event. A dextrose carrier should be avoided because it may lower the potassium level by stimulating secretion of insulin. The serum potassium rises approximately 0.1 mEq/L for each 2 mEq of KCl administered. Repeat potassium levels should guide therapy.
When a central line is not present, a concentrated potassium drip cannot be administered peripherally because it scleroses veins. The maximum concentration of KCl that can be administered peripherally is 60 mEq/L. IV bags containing 60 mEq/L or 10 mEq in a 100–200 mL bag are commonly used.
Once the patient is extubated, oral potassium (10–20 mEq tablets up to three to four times a day) will usually suffice to treat potassium levels of 3–4 mEq/L. However, doses of 40–60 mEq three to four times a day may be necessary to maintain normal potassium levels when the potassium is <3 mEq/L.
Potassium‐sparing diuretics should be considered in the chronic management of HF, but not acutely to raise serum potassium. The preferred drug is amiloride, which is an epithelial sodium channel blocker in cortical collecting ducts. The mineralocorticoid receptor antagonists (spironolactone and eplerenone) can be considered in HF patients with normal or mildly abnormal renal function (SCr <2.5 mg/dL).
Note that patients with uncontrolled diabetes may have movement of potassium out of cells due to insulin deficiency and hyperosmolality. Thus, the serum potassium may be elevated despite a marked potassium deficit. Aggressive treatment with insulin will lower the potassium, so potassium levels must be carefully evaluated to determine when to start replenishing potassium. If the patient has severe hyperglycemia and hypokalemia, potassium should be given prior to giving insulin, which will lower potassium levels.
Calcium plays a complex role in myocardial energetics and reperfusion damage. Ionized calcium (normal = 1.1–1.3 mmol/L) should be measured because total calcium levels, which are affected by protein binding, usually decrease during surgery because of hemodilution, hypothermia, shifts in pH, hypomagnesemia, and the use of citrated blood. Hypocalcemia is usually associated with prolongation of the QT interval on the ECG tracing. It reduces cardiac sensitivity to digoxin.
It is common practice to empirically administer a 500 mg bolus of calcium chloride at the termination of CPB to support systemic vascular resistance and possibly increase myocardial contractility.203 It may be given with protamine to offset its vasodilatory effects.204
It is questionable whether treatment of hypocalcemia identified in the ICU is of any value in improving cardiovascular function. In fact, calcium salts may attenuate the cardiotonic effects of catecholamines, such as dobutamine or epinephrine, although they have little effect on the efficacy of milrinone.205 Nonetheless, if the ionized calcium level is measured and found to be <1 mmol/L, calcium gluconate (10 mL of 10% solution in 50 mL D5W) may be given over 10–20 minutes, although there is no clear benefit to doing so. It may also be given as an infusion of 50 mL/h of a 10% solution (placing 100 mL in 1 L of D5W). Calcium chloride is best avoided for “asymptomatic” hypocalcemia to minimize any acute hemodynamic effects, but similar dosing is used. Because hypocalcemia can be difficult to correct if the serum magnesium level is low, an infusion of magnesium sulfate 2 g (16 mEq) of a 10% solution over 10–20 minutes may be given, with a subsequent infusion of 1 g (8 mEq) in 100 mL of fluid per hour.206
Calcium chloride (0.5–1 g IV) may be given in emergency situations to provide temporary circulatory support when a low cardiac output syndrome or profound hypotension develops suddenly. The transient improvement in hemodynamics allows time for analysis of causative factors and the institution of other pharmacologic support. It should not be given routinely during a cardiac arrest.
Magnesium plays a role in energy metabolism and cardiac impulse generation. Low levels have been associated with coronary spasm, low cardiac output syndromes, prolonged ventilatory support, a higher incidence of postoperative atrial and ventricular arrhythmias, perioperative infarction, and a higher mortality rate.207–210
Magnesium levels (normal = 1.5–2 mEq/L) are usually not measured during surgery, but are reduced in the majority of patients. This is usually the result of hemodilution during CPB as well as urinary excretion, but it is also very common in patients undergoing off‐pump surgery.211 Medications contributing to low magnesium levels include diuretics and PPIs.
Administration of magnesium sulfate (MgSO4) 2 g in 100 mL solution given over an hour to raise the serum level to 2 mEq/L is effective in reducing the incidence of postoperative atrial fibrillation and ventricular arrhythmias after both on‐pump and off‐pump surgery.211 Notably, magnesium has been found to inhibit the vasoconstrictive response to epinephrine, but not its cardiotonic effects.212 Administering 1–2 g of MgSO4 at the conclusion of bypass and on the first postoperative morning can be recommended.
Magnesium may be helpful in the treatment of torsades de pointes.
XII. Metabolic Acidosis
A low cardiac output state causing tissue hypoperfusion is usually the primary cause of metabolic (lactic) acidosis in the cardiac surgery patient. Poor peripheral and visceral perfusion will be exacerbated by use of vasopressors.
Intra‐abdominal catastrophes, such as mesenteric ischemia from a low‐flow state, should always be considered when progressive metabolic acidosis occurs.
Low‐dose epinephrine occasionally causes a metabolic acidosis out of proportion to its α effects when the cardiac output is satisfactory. This may reflect a metabolic type B lactic acidosis (not associated with tissue hypoxia) caused by metabolic factors that increase lactic acid production, such as hyperglycemia and lipolysis.213,214
High doses of sodium nitroprusside
Renal failure (which reduces acid excretion)
Acute hepatic dysfunction
Aggressive crystalloid infusions with high chloride content (normal saline)
Adverse effects of metabolic acidosis usually do not occur until the pH is less than 7.20.215,216 A primary metabolic acidosis (low serum bicarbonate with acidemic pH) may be noted in the heavily sedated patient in whom there is no respiratory compensation. However, compensatory hyperventilation to neutralize the acidosis will occur when the patient can breathe spontaneously, and for every 1 mEq/L fall in bicarbonate, the PCO2 is generally reduced about 1.2 torr. However, it is not uncommon to see incomplete compensation with a mixed respiratory/metabolic acidosis. Notably, some of the deleterious effects of metabolic acidosis may be related to the metabolic products associated with the acidosis, rather than the absolute level of pH, although they may be reversed by administration of sodium bicarbonate.
The presence of a progressive or significant metabolic acidosis (as assessed by the serum bicarbonate level) is often an indication of a serious ongoing problem that must be corrected before adverse consequences occur. Occasionally, this may be identified surreptitiously. For example, after the patient is extubated and the arterial line has been removed, the patient may develop tachypnea of unknown etiology. An ABG or basic metabolic panel may reveal that the patient is compensating for a profound metabolic acidosis. Some adverse consequences of a metabolic acidosis include:
Decreased contractility and cardiac output; reduction in hepatic and renal blood flow
Attenuation of the positive inotropic effects of catecholamines
Venoconstriction and arteriolar dilatation which increase filling pressures and decrease systemic pressures
Increased pulmonary vascular resistance
Sensitization to reentrant arrhythmias and reduction in the threshold for ventricular fibrillation
Dyspnea and tachypnea
Decreased respiratory muscle strength
Increased metabolic demands
Hyperglycemia caused by tissue insulin resistance and inhibition of anaerobic glycolysis
Decreased hepatic update and increased hepatic production of lactate
Increased protein catabolism
Inhibition of brain metabolism and cell volume regulation
Obtundation and coma
Type A lactic acidosis reflects impaired tissue oxygenation and anaerobic metabolism resulting from circulatory failure. The acidosis is self‐perpetuating in that excess lactate is being produced at a time when there is suppression of hepatic lactate utilization. The lactate ion, probably more than the acidosis, contributes to potential cardiovascular dysfunction. An elevated lactate level (>3 mmol/L) upon arrival in the ICU is associated with a worse outcome.217 This is more commonly noted in patients with preexisting renal dysfunction, after long pump runs, and with use of intraoperative vasopressors. It is likely that this reflects inadequate oxygen delivery during bypass that has contributed to splanchnic and renal ischemia, with the acidosis perpetuated by a low cardiac output syndrome. Needless to say, the presence of elevated lactate levels during bypass or upon arrival in the ICU requires prompt attention. The development of a metabolic acidosis several days after surgery raises the specter of mesenteric ischemia, especially in patients requiring additional days of ICU care.
Type B lactic acidosis occurs in the absence of tissue hypoxia.213,214 It may be a catecholamine‐induced metabolic effect (especially with epinephrine) caused by hyperglycemia and alterations in fatty acid metabolism that cause pyruvate accumulation and elevated levels of lactic acid. Acute hepatic failure may also be present with severe lactic acidosis due to failure to clear lactic acid. Metformin is associated with lactic acidosis in patients with renal insufficiency, low cardiac output states, and liver disease, and with use of contrast agents.
Measurement of the anion gap (AG) is important is sorting out the etiology of acidosis (AG = Na+ − [Cl− + HCO3−], with normal range being 10–12 mEq/L).
Although there are a number of factors that can influence the AG, an increase in AG generally reflects additional acid production, and high AG metabolic acidosis from lactic acid accumulation is most commonly noted after cardiac surgery. It may also be elevated in diabetic ketoacidosis due to production of hydroxybutyrate, in renal failure from decreased acid excretion and increased bicarbonate excretion, and from excessive ingestion of aspirin/methanol/ethylene glycol.
A normal or low AG represents loss of bicarbonate (diarrhea) or inability to excrete an acid load (renal tubular acidosis).
Treatment should be directed primarily towards reversal of the underlying cause. This will allow for oxidation of lactate and regeneration of bicarbonate to correct the acidosis. Whether correction of a primary metabolic acidosis (not one that compensates for a primary respiratory alkalosis) should be considered when the serum bicarbonate is less than 15 mEq/L (base deficit greater than 8–10 mmol/L) is controversial.218–222
Proponents of bicarbonate administration suggest that severe metabolic acidosis does have significant deleterious effects on cardiovascular function that can be corrected with a more normal pH. Furthermore, more responsiveness to catecholamines does seem to occur with a more normal pH. Thus, correction of the acidosis may be important when the etiology of the acidosis is unclear or not imminently remediable.
Others argue that the use of bicarbonate can cause metabolic derangements with little evidence of hemodynamic improvement.219,220 Sodium bicarbonate can cause fluid overload, hypernatremia, and hyperosmolarity, increased affinity of hemoglobin for oxygen (and thus less tissue release), and reduced ionized calcium, which may reduce cardiac contractility. It is proposed that bicarbonate may correct only the blood pH, not the intracellular pH, and that the increased production of CO2 that may not be eliminated in low output states may impair lactate utilization, perpetuating the elevated lactate levels.
A randomized study of ICU patients with severe metabolic acidosis found that use of sodium bicarbonate reduced one‐month mortality only in patients with acute kidney injury, and did cause more hypernatremia and hypocalcemia.221 A clinical literature review suggested that there was little documented clinical benefit of using bicarbonate in critically ill patients with metabolic acidosis.222
Nonetheless, if one elects to correct the pH, sodium bicarbonate is most commonly used. Tromethamine (THAM) is a good alternative in patients with hypernatremia with mixed metabolic/respiratory acidosis, but is no longer available in the USA.
Sodium bicarbonate is administered in a dose calculated from the following equation: