Chapter 58. Regional Anesthesia & Systemic Disease

Patients with coexisting severe systemic disease may be at a higher risk for perioperative complications related to surgery and administration of anesthesia. Regional anesthesia is often touted as beneficial in many patients who have pulmonary, cardiac, renal, and other disease. However, the physiologic changes that occur with various regional anesthesia techniques must be understood and viewed within the context of an individual patient's pathophysiology so that the technique benefits the patient fully and reduces the risk of complications from the patient' disease. This chapter focuses on the pathophysiology of several common systemic diseases frequently encountered by the regional anesthesiologist and discusses the interplay between common regional anesthesia techniques and patient disease.

Surgical patients with coexisting pulmonary impairment are at risk for intraoperative or postoperative pulmonary complications, regardless of anesthetic technique.1 However, increasing evidence suggests that regional anesthesia may be associated with improved pulmonary outcomes compared with those associated with general anesthesia.2–4 A thorough understanding of respiratory physiology and the implications of regional anesthetic techniques is crucial to the safe and effective use of regional anesthesia in these patients.

Epidural & Spinal Anesthesia

Most of the pulmonary effects of neuraxial anesthesia are due to motor block of the intercostal and abdominal musculature. If a significant systemic uptake of local anesthetic occurs, some central and direct myoneural respiratory depression can also be seen, although this plays a minor role overall.5 Since neuraxial anesthesia produces a “differential” blockade of motor, sensory, and autonomic fibers, the degree to which respiratory function is impaired depends on the relative extent of segmental motor blockade. Using dilute concentrations of epidural local anesthetic may provide adequate sensory block as high as the cervical levels, while sparing the motor function of the respiratory muscles in the lower somatic segments.6 Achieving diaphragmatic paralysis via a phrenic nerve (C3 through C5) block in the absence of a total spinal anesthesia is difficult in practice, since even a sensory block as high as C3 will only produce a motor block at approximately T1 through T3.5 However, high neuraxial blocks may precipitate hypotension sufficient to decrease blood flow to the respiratory center in the medulla, leading to respiratory arrest.

With higher levels of epidural or spinal anesthesia, chest wall musculature, and in particular the intercostal muscles, become segmentally weakened and contribute less to the respiratory effort. This in turn, may eventually lead to altered chest wall motion during spontaneous respiration. For instance, some studies have suggested that during high neuraxial anesthesia, the more compliant chest wall is retracted during inspiration and may actually display paradoxical rib cage motion.7,8 Others, however, have found that epidural blockade to sensory levels of T6 or even T1 do not lead to rib cage constriction with inspiration and may in fact increase the contribution that chest wall expansion makes to tidal volume.9,10 This may be explained by an incomplete motor block of high, intercostal muscles or the compensatory role played by the “accessory” muscles of respiration such as the anterior and middle scalene muscles.11

Lumbar epidural anesthesia does not impair resting minute ventilation, tidal volume, or respiratory rate.12–14 Furthermore, functional residual capacity (FRC) and closing capacity appear to be relatively unchanged during lumbar epidural anesthesia.15–17 Effort-dependent tests of respiratory function, such as forced expiratory volume in one second (FEV1), forced vital capacity, and peak expiratory flow rate, do exhibit modest decreases in the setting of lumbar epidural blockade, reflecting the reliance of these indices on intercostal and abdominal musculature.18 In contrast, the effect of thoracic epidural anesthesia on pulmonary mechanics is less clear, with studies showing both a decrease14,19 and increase20 in minute ventilation and tidal volumes. One volunteer study found that high thoracic epidural anesthesia (T1 sensory level) led to an increase in FRC of approximately 15% with no change in tidal volume or respiratory rate.10 This somewhat surprising finding may be explained by two mechanisms offered by the investigators. First, most volunteers exhibited a decrease in their intrathoracic blood volume, a physiologic occurrence confirmed by Arndt and colleagues.21 Second, the study also found that the end-expiratory position of the diaphragm was shifted caudally, which is possibly related to a relative increase in diaphragmatic tonic activity or a reduction in intraabdominal pressure. Thoracic epidural anesthesia results in a modest decrease in vital capacity (VC), FEV1, total lung capacity, and maximal midexpiratory flow rate.15

The ventilatory response to hypercarbia and hypoxia is preserved with neuraxial anesthesia.9,14 Partial pressures of both oxygen (Po2) and carbon dioxide (Pco) are essentially unchanged during epidural or spinal anesthesia.9,10 In addition, bronchomotor tone is not altered to any significant degree, despite theoretical concerns of bronchoconstriction secondary to sympatholysis.21 Indeed, epidural anesthesia has been used successfully for high-risk patients with chronic obstructive pulmonary disease and asthma undergoing abdominal operations.22,23

Neuraxial anesthesia has been shown in a number of settings to lead to reduced postoperative pulmonary complications compared with general anesthesia. The reasons behind this are probably multifactorial, owing in part to superior analgesia, reduced diaphragmatic impairment, altered stress response, and a decreased incidence of postoperative hypoxemia.24,25 Epidural anesthesia provides better pain control than general anesthesia for abdominal and thoracic surgery, which leads to reduced splinting, a more effective cough mechanism, and preserved postoperative lung volumes, including FRC and VC.26 One study directly comparing epidural and general anesthesia in high-risk patients concluded that overall outcomes, including the need for prolonged postoperative ventilation, were improved with the regional technique.27 Another trial in patients undergoing lower limb vascular surgery reported a greater than 50% reduction in the incidence of respiratory failure in the group randomized to epidural anesthesia.28 A more recent meta-analysis of 141 randomized trials (including over 9000 patients) comparing regional and general anesthesia for hip surgery showed a risk reduction for pulmonary embolism, pneumonia, and respiratory depression of 55%, 39%, and 59%, respectively, with the regional anesthesia.25 Interestingly, these outcomes were unchanged regardless of whether neuraxial anesthesia was continued into the postoperative period, illustrating that the beneficial effect of epidural and spinal anesthesia on pulmonary physiology occurs, at least in part, at the time of surgical insult.

Brachial Plexus Block

In the absence of inadvertent complications such as pneumothorax, alterations in respiratory mechanics seen with brachial plexus block are due primarily to phrenic nerve blockade and hemidiaphragmatic paralysis. This has been shown to occur in 100% of patients receiving interscalene blockade,29 and leads to a reduction by 27% in both FVC and FEV1.30 While the clinical significance of this reduction in healthy patients is not entirely clear, it may be useful to risk-stratify patients about to undergo interscalene blocks as one would a patient undergoing lung resection. In other words, ask the question: “Will this patient tolerate a perioperative FEV1 reduction of 27%?” Some investigators have attempted to reduce the incidence of phrenic nerve palsy by decreasing the volume of local anesthetic; however, volumes as little as 10–20 mL still result in diaphragmatic paralysis.31,32 In fact, one case report illustrated clinically significant respiratory compromise requiring tracheal intubation following an interscalene block using a volume of 3 mL of 2% mepivacaine.33

The risk of phrenic nerve blockade decreases as one moves more distally along the plexus. Axillary nerve block has no effect on diaphragm function and presents a good choice for those patients with marginal pulmonary reserve (ie, cannot tolerate a 27% reduction in lung function). On the other hand, the supraclavicular block is associated with a 50–67% incidence of hemidiaphragm paralysis.34–36 The infraclavicular approach is probably sufficiently distant from the course of the phrenic nerve so as to spare the diaphragm,37,38 although there are case reports of phrenic nerve involvement.39 These discrepancies probably relate to the different approaches to the infraclavicular block—for instance the “coracoid block” is performed with a relatively lateral or distal puncture site, whereas the vertical infraclavicular block begins at a more medial location. Although the infraclavicular or axillary blocks may be desirable for their relative pulmonary-sparing profiles, they carry the disadvantage of providing incomplete anesthesia for the upper arm and shoulder. However, creative solutions have been employed to get around this issue. Martinez and coworkers combined an infraclavicular block with a suprascapular nerve block for emergent humeral head surgery in a patient who was acutely asthmatic and had a baseline FEV1 of 1.13 L (32% predicted). Therefore, a knowledgeable combination of peripheral nerve blocks can provide complete anesthesia of the upper limb while avoiding respiratory complications in patient with a pulmonary disease.40

Continuous brachial plexus blocks with perineural catheters are an attractive method of maintaining the advantages of plexus blockade into the postoperative period and have been shown to reduce postoperative pain, oral opioid requirements and their side effects, and sleep disturbances after shoulder surgery.41 However, there have been reports of complications attributed to the prolonged phrenic nerve paresis that invariably occurs with this technique. These have included chest pain, atelectasis, pleural effusion, and dyspnea.42,43 This is of particular concern because many patients are being discharged home with catheters and may not have access to timely intervention should these complications arise. On the other hand, the degree of clinically significant respiratory impairment with continuous interscalene blockade varies among patients and, in fact, may be well tolerated, especially if using relatively dilute concentrations of local anesthetic that only provide a partial phrenic paresis.44

Maurer and associates reported a case of a patient with no preexisting pulmonary disease who underwent bilateral shoulder arthroplasty under combined bilateral continuous interscalene blockade and general anesthesia.45 Postoperative analgesia was maintained in the hospital for 72 h via the catheters using infusions of 7 mL/h of 0.2% ropivacaine for each side (total 14 mL/h). Despite a marked postoperative reduction in FVC (60%) from baseline as well as sonographic evidence of diaphragmatic impairment, the patient had an uneventful postoperative course (with excellent analgesia) and good recovery. This anecdotal example illustrates that the clinical significance of phrenic paresis in patients with good respiratory function is questionable. Regardless the use of continuous brachial plexus techniques should be carefully considered in patients with preexisting pulmonary disease, especially if they are to be discharged home with the catheters in situ.

Paravertebral & Intercostal Nerve Blocks

Several studies investigated the effects of paravertebral and intercostal blocks on pulmonary function in patients with rib fractures or those undergoing thoracotomy. Intercostal blockade has been shown to improve arterial oxygen saturation (Sao2) and peak expiratory flow rate (PEFR) in patients with traumatic rib fractures associated with severe pain.46 Likewise, Karmakar and investigators found that continuous paravertebral blockade over a period of 4 days in patients with multiple fractured ribs led to significant improvement in respiratory rate, FVC, PEFR, Sao2, and the Pao2 fraction of inspired oxygen ratio.47 These findings are probably related to the favorable effect of analgesia on respiratory efforts by the patient and improved respiratory mechanics.

Paravertebral blocks are very effective for management of pain following thoracotomy and can significantly improve postoperative spirometry. One review of 55 randomized, controlled trials of analgesic techniques following posterolateral thoracotomy revealed that paravertebral blockade was the method that best preserved pulmonary function compared with either intercostal or epidural analgesia.48 The combined results showed an average preservation of approximately 75% of preoperative pulmonary function when paravertebral analgesia was used versus 55% for both intercostal and epidural analgesia. It is unclear why paravertebral blockade might result in improved PEFR and Sao2 compared with epidural analgesia in this and other studies, but it may be related to increased utilization of opioids, higher incidence of nausea and vomiting, and the presence of bilateral intercostal muscle blockade (and therefore diminished chest wall mobility) in the epidural cohorts.49

Pulmonary Complications Not Related to Conduction Blockade

Pulmonary complications related to the use of regional anesthetic techniques fall into two categories. The first is those related directly or indirectly to the physiologic changes that occur with the blockade itself. Examples include atelectasis and pneumonia resulting from an inability to mobilize secretions. The second category comprises those that are independent of the effect of blockade, and although there are sporadic reports of rare complications such as pulmonary hemorrhage50 and chylothorax,51 the most common of these is pneumothorax. Not surprisingly, pneumothoraces occur most frequently when the puncture site overlies the pleura, and especially when performing supraclavicular and intercostal blocks. The overall incidence is low, however,52–55 and refinements of previously published techniques based on MRI studies and ultrasound guidance can decrease the incidence further.56,57 Nevertheless, techniques with a risk of pneumothorax should be carefully considered or avoided in patients with borderline pulmonary function.

Renal dysfunction is commonly present in the surgical population. Perioperative renal failure accounts for approximately 50% of all patients requiring acute hemodialysis in the United States. Patients with chronic renal insufficiency frequently present for procedures, such as creation of vascular shunts and revascularization of the lower limbs. Regional anesthetic techniques are frequently ideal options to provide anesthesia for these patients and procedures.

Effect of Regional Anesthesia on Renal Function

The treatment of patients at risk for perioperative renal dysfunction should focus on two principles: avoiding nephrotoxic agents and maintaining kidney perfusion. Local anesthetics do not possess any nephrotoxic properties per se, and in fact the coadministration of procaine has been shown to mitigate some of the nephrotoxic effects of cisplatin in rats.58 Of greater relevance is the effect of anesthetic-induced hypotension on renal blood flow. The kidneys are capable of autoregulation over a wide variety of mean arterial pressures (approximately 80–180 mm Hg) and maintain glomerular filtration rate (GFR) by autonomous changes in renal vascular resistance.59 Below the so-called lower limit of autoregulation, the kidney begins to shut down its energy-dependent physiologic processes, and the GFR and urinary output fall as a result. Ultimately, if left unchecked, renal ischemia develops, especially in the sensitive renal medulla. Although neuraxial anesthesia and the concurrent sympathectomy can reduce mean arterial pressure (MAP), renal blood flow is often preserved.60 This is believed to reflect an increase in left ventricular stroke volume in response to the drop in systemic vascular resistance (SVR). Rooke and investigators studied hemodynamic responses and abdominal organ perfusion (as measured by scintigraphy) in 15 patients undergoing lidocaine spinal anesthesia with a sensory block ranging from T1 to T10.61 Whereas the MAP and SVR fell on average by 33% and 26%, respectively, blood volume in the kidneys increased by approximately 10%. There may be limits to the degree of compensation afforded by cardiac output, however. One study using a primate model showed that although renal blood flow was minimally affected by T10 spinal anesthesia, it was significantly reduced by a T1 sensory block.62 This illustrates again that lumbar and low thoracic levels of neuraxial anesthesia in patients with renal disease are well tolerated physiologically and significant changes do not begin to manifest until higher levels are achieved.

The renin–angiotensin system, which is initiated in the kidney in response to a reduction in renal perfusion, plays an important role in blood pressure homeostasis. It serves as a complementary humoral mechanism to the sympathetic nervous systems. Hopf and colleagues conducted a study to determine if thoracic epidural anesthesia suppressed the renin response to induced hypotension.63 Plasma renin and vasopressin concentrations were measured before, during, and after a hypotensive challenge with nitroprusside in patients with and without thoracic epidural anesthesia (sensory level T1 through T11). With an intact sympathetic nervous system (ie, no epidural), plasma renin levels doubled in response to the hypotensive challenge lasting 15 min. In contrast, there was no change in the renin concentration when hypotension was induced to the same MAP in the epidural cohort. This suggests that sympathetic fibers play a key role in the renin–angiotensin system and that thoracic epidural anesthesia interferes with the functional integrity of that system.

For obvious reasons, postoperative renal function is of foremost concern when administering anesthesia for recipients of renal transplantation. Several studies have looked at the effect of general versus regional (or combined epidural/general) anesthesia on postoperative renal function in this setting. The choice of anesthetic technique was not shown to have an effect on graft outcome in either adult or pediatric populations.64,65 Also, the choice of anesthetic technique for living donors was shown to be independent of recipient graft outcome.66 Other nontransplant outcomes data, including those from the large meta-analysis by Rodgers and coworkers, indicate that regional anesthesia is associated with a lower risk for postoperative renal failure than general anesthesia. However, the authors cautioned that the confidence intervals were wide and were compatible with both no effect and a two-thirds risk reduction.25 Overall, it appears that a well-conducted regional anesthetic does not negatively affect perioperative kidney function and renal outcome compared with general anesthesia.

Considerations for Regional Anesthesia in Chronic Renal Failure

Patients with chronic renal failure often manifest a large number of pathophysiologic changes that may influence regional anesthetic care. These may include the presence of an anion-gap metabolic acidosis, electrolyte disturbances such as hyperkalemia, and coagulopathies due to uremia-induced platelet dysfunction. Plasma concentrations of local anesthetic following peripheral nerve blocks are often high enough to cause central nervous system or cardiac toxicity in any patient, even when no obvious intravascular injection has occurred. This is probably related to the large dose used when performing “high-volume blocks” such as brachial plexus blocks. Some authors have recommended that dosages be adjusted in patients with chronic renal insufficiency based on observations of toxicity presumed to be related to concurrent acidosis or hyperkalemia.67,68 Indeed, experimental evidence suggests that acidemia decreases the protein binding of bupivacaine, thereby increasing the free fraction and risk of toxicity.69 In addition, hyperkalemia (5.4 vs 2.7 mEq/L) has been shown in dogs to halve the dose of bupivacaine required to induce cardiotoxicity.70 Interestingly, the potassium level had no effect on seizure dose in the same animals. This is an ominous finding, as it suggests that the so-called safety margin of plasma levels between CNS and cardiac toxicity, which is already relatively narrow with bupivacaine, is even less reliable in the presence of hyperkalemia. Even in the absence of acid–base or electrolyte disturbances, plasma levels of local anesthetics following peripheral nerve block are often higher in patients with chronic renal failure.71,72 The reason for this is not entirely clear, but may relate to increased blood flow (and hence uptake at the injection site) due to the hyperdynamic circulation often seen in uremic patients.73 On the other hand, α1-acid glycoprotein (AAG) levels are increased in uremia74 and may lend a protective effect by binding more local anesthetic in the bloodstream.75 The increased levels of AAG also result in both a reduced free fraction available for hepatic metabolism and in a reduced volume of distribution. These two pharmacokinetic consequences appear to balance each other so that the serum half-life is not significantly changed.71 Hemodialysis is ineffective in removing lidocaine from plasma, and therefore cannot be relied on to treat toxicity.76

No significant difference exists between patients with chronic renal failure and healthy patients with respect to peripheral nerve block latency, duration, or quality.71,77,78 In one study of spinal anesthesia properties in patients with chronic renal failure versus healthy patients, Orko and associates found that block quality was similar, but that both onset time and duration of the block were reduced in the patients with uremia.79 The authors postulated a volume-contracted intrathecal space in uremic patients as a mechanism for the quicker onset, but the actual cause remains speculative. The shorter duration of sensory block may again be related to enhanced uptake in the setting of a hyperdynamic circulation.

Uremic coagulopathy is characterized by a defect of platelet aggregation that is probably due to a toxic effect by uremic substances on the binding of fibrinogen to the platelet glycoprotein IIb/IIIa receptor.80 This often manifests in clinically appreciable bleeding, and at least one case of subarachnoid hematoma leading to paraplegia after a spinal anesthetic in a chronic renal failure patient has been published.81 Patients undergoing hemodialysis require intermittent anticoagulation and may present to the operating room with an unclear coagulation status. Care must be taken to delineate heparin or other anticoagulant regimens. Despite this platelet dysfunction, uremic patients are at higher risk for thrombotic events.82 One case of hypoxia following a brachial plexus block in a uremic patient was later found to be secondary to pulmonary embolism.83 The authors of the report suggested that a likely mechanism was dislodgement of a preexisting thrombus from the proximal arm, facilitated by block-related manipulation and vasodilation of the upper extremity.

Several studies have compared anesthetic techniques for the creation of arteriovenous fistulae, a procedure that is common in patients with end-stage renal disease and is well suited to brachial plexus block.84 Although some investigators have concluded that little difference exists in outcome between general, local, and brachial plexus anesthesia for this operation,85 Mouquet and colleagues specifically studied the effects of these three techniques on brachial artery blood flow and concluded that both general anesthesia and brachial plexus block improved blood flow through the fistula during surgery, whereas local infiltration did not.86

Liver injury or dysfunction can range from a mild, asymptomatic “transaminitis” to frank hepatic failure. There are many causes of liver disease, both acquired and congenital, but all manifest as either failure of parenchymal cell function (ie, acute and chronic hepatitis, cirrhosis) or cholestasis.87 Considerations for regional anesthesia in patients with liver disease include the potential for altered disposition and metabolism of local anesthetics, the effect of regional anesthesia on hepatic perfusion, and concerns regarding coagulopathy related to liver dysfunction.

Pharmacokinetics of Local Anesthetics in Liver Disease

Amide local anesthetics are metabolized in liver microsomes by the cytochrome P-450 system.88 A decrease in microsomal function, as may be seen in acute or chronic liver disease, can lead to a reduction in biotransformation and clearance of these drugs, putting the patient at risk for local anesthetic toxicity. As with other drugs that are metabolized in the liver, the hepatic extraction ratio determines the relative importance of hepatic perfusion versus intrinsic enzyme activity in the overall clearance of the drug. For example, bupivacaine has a low extraction ratio (ie, its clearance is more sensitive to alterations in hepatic enzyme activity), whereas etidocaine exhibits a relatively high extraction ratio and depends on adequate liver perfusion for clearance.89 Lidocaine has an intermediate hepatic extraction ratio and therefore relies on both perfusion and enzymatic activity. Severe hepatic disease such as cirrhosis can affect both liver perfusion and intrinsic enzyme function. In this scenario, the clearance of all amide local anesthetics, regardless of their extraction ratio, is likely to be reduced. Because the volume of distribution of local anesthetics (and many other drugs) is increased in hepatic disease, the actual plasma levels may not differ significantly from healthy patients with a single dose, despite the diminished clearance.90–92 The altered distribution may be related to decreased levels of plasma AAG, which are reduced in proportion to the severity of liver disease.93 Clinically, it appears that single-dose peripheral nerve blocks with amide local anesthetics probably do not require a dosage adjustment in this population, whereas continuous infusions have the potential to accumulate to toxic levels.94

The cytochrome P-450 enzyme system is subject to induction or inhibition by a variety of drugs and dietary nutrients. This may play a role in the subsequent metabolism of amide local anesthetics. For example, substances that inhibit microsomal enzymes, such as cimetidine or grapefruit juice, may lead to an accumulation of local anesthetic, augmenting the risk of toxicity, especially in the setting of preexisting hepatic dysfunction.95

Ester local anesthetics are cleared by plasma cholinesterases in the blood and liver. Severe hepatic disease may result in decreased levels of cholinesterase and result in prolonged plasma half-lives of esters such as procaine.96 On the other hand, red cell esterases remain intact during liver disease and are able to provide some hydrolytic function.97 Because plasma cholinesterase is extremely efficient, it is unlikely that an enzyme deficiency secondary to hepatic disease could impair the hydrolysis of ester-type local anesthetics to a degree sufficient to cause toxicity.89

Effect of Regional Anesthesia on Hepatic Blood Flow

The hepatic blood supply is unique in that it relies on both portal venous return and hepatic artery blood flow, which make up approximately 75% and 25% of the total flow, respectively. The regulation of hepatic blood flow is complex. The portal system is passive and not subject to autoregulation, whereas the hepatic artery can increase or decrease its contribution to flow in response to alterations in portal venous flow.98 The hepatic artery also autoregulates in response to MAP, in much the same way as cerebral or renal vessels do, but may rely on an intact sympathetic response for effect.99

General anesthesia has been shown to cause a decrease in hepatic blood flow, which may lead to ischemia and postoperative liver dysfunction.100 Less is known about the effects of regional anesthesia on hepatic perfusion. Grietz and coworkers performed a high epidural (block level T1 through T4) in 16 dogs and examined the effect on systemic and hepatic hemodynamics.101 MAP and portal venous flow were both reduced compared with control values, by 52% and 26%, respectively. In contrast, hepatic artery flow was unchanged, probably relating to a reduction in hepatic artery resistance of 51%. In addition, hepatic oxygen uptake was preserved through an increased oxygen extraction. Another study by Vagts and colleagues found that thoracic epidural anesthesia in anesthetized pigs was associated with decreased mean arterial blood pressure and hepatic artery blood flow, but no change in hepatic oxygen delivery or uptake, or tissue oxygen partial pressure compared with pigs receiving general anesthesia alone.102 Taken together, these findings should reassure the clinician that high neuraxial anesthesia may be well tolerated with respect to hepatic oxygenation, despite a modest reduction in MAP. Care should be taken to maintain cardiac output and perfusion pressure during anesthesia to ensure the adequate perfusion of all the vital organs.

Hepatic Coagulopathy

Severe hepatic disease is associated with abnormalities of the coagulation system. The cause is multifactorial and may include the decreased synthesis of procoagulant proteins, impaired clearance of activated coagulation factors, nutritional deficiency (eg, vitamin K, folate), the synthesis of functionally abnormal fibrinogen, splenomegaly secondary to portal hypertension (sequestrational thrombocytopenia), qualitative platelet defects, and bone marrow suppression of thrombopoiesis (eg, by alcohol, hepatitis virus infection) (Figure 58–1).103 Because of the potential complexity of the coagulopathy, it may be necessary to perform additional laboratory tests such as clotting factor and fibrinogen assays to completely delineate the nature of the problem. Clotting factor deficiencies can be treated with vitamin K supplementation or fresh frozen plasma transfusion or both. Platelet transfusion may be necessary in the case of thrombocytopenia. Newer therapies such as recombinant factor VIIa have also been used to correct bleeding associated with liver failure.104 Because the vitamin K–dependent clotting factors are more susceptible to hepatocellular disease, the prothrombin time (PT) and international normalized ratio (INR) are often used as markers of coagulation system integrity. However, the predictive value of PT/INR on hemorrhage during bedside procedures such as lumbar puncture or central line placement has been shown to be poor.105 As such, it is important to carefully weigh the risks and benefits of a neuraxial anesthetic technique in a patient with suspected hepatic-induced coagulopathy. Although an INR of less than 1.5 “should be associated with normal hemostasis” according to the ASRA consensus guidelines on anticoagulation,106 this statement applies primarily to warfarin-induced anticoagulation and may not be a reliable indicator of the likelihood of problematic bleeding in liver failure. The risks associated with performing peripheral nerve blocks in patients with abnormal coagulation parameters is less clear. Obviously the risk of bleeding is increased with techniques in which the needle is laced in the vicinity to a major blood vessel. Deep blocks, such as the sciatic nerve block, should probably be avoided in a coagulopathic patient due to the risk of deep muscular hemorrhage. Likewise, care should be taken when performing blocks in the vicinity of noncompressible blood vessels (ie, the subclavian artery in the case of supraclavicular block) in patients who have coagulation abnormality. The risks of regional anesthesia in the setting of a coagulable disorder are elaborated in Chapter 7.

Diabetes is a multisystem disease characterized by carbohydrate intolerance and insulin dysregulation that has many implications for the regional anesthesiologist.107 Besides the usual anesthetic concerns such as the presence of coronary artery, cerebrovascular, and renal disease, diabetics have a high incidence of preexisting peripheral neuropathy, which has implications for block performance, success, and risk for neurologic complications. Other considerations are the effect of regional anesthesia on glucose homeostasis and the increased risk of infection in diabetic patients.

Peripheral Neuropathy in Diabetics

Diabetic neuropathy is one of the most common neurologic diseases, affecting up to 100% of patients with long-standing disease.108 Patients can be asymptomatic, but in affected patients symptoms are typically described as paresthesias, sensory loss, or neuropathic pain. The mechanism of diabetic neuropathy is thought to be related to either a direct metabolic and osmotic effect of chronic hyperglycemia on neurons or a microvascular insult leading to nerve ischemia.109

Performing nerve blocks on diseased nerves has been a source of concern. Kalichman and Calcutt studied sciatic nerve histology in rats following blockade with lidocaine and found significantly more nerve edema in diabetic versus control cohorts.110 The reason for the edema is probably multifactorial and may include the presence of an altered blood–nerve barrier or decreased uptake of local anesthetic, leading to a longer duration of nerve bathing. An increase in endoneural fluid pressure due to edema may constrict small transperineural vessels, precipitating ischemia in an already compromised nerve. This may translate to an increased incidence of postoperative paresthesias following nerve blocks, including neuraxial blocks, in diabetics. Al-Nasser reported a case of a prolonged (>8 weeks) bilateral lower limb paresthesias and pain following lumbar epidural analgesia with 0.2% ropivacaine in a diabetic patient undergoing radical prostatectomy.111 Postoperative electromyographic studies showed widespread sensory neuropathy of both upper and lower limbs, indicating that the patient, although asymptomatic, had preexisting neuropathy that may have predisposed to this rare complication. The actual prevalence of neurologic complications in diabetics receiving nerve blocks is unknown, but is probably quite low. Diabetes is a common disease, and reports of this complication in the literature are sparse, suggesting that in the vast majority of cases, recovery from peripheral nerve blocks is uneventful.

A perhaps more practical issue for regional anesthesiologists is the effect of diabetic neuropathy on electrolocation of nerves while using a nerve stimulator. Diabetic patients may have altered response to nerve localization. In one case report, authors were unable to elicit motor response in a patient with diabetes even with current intensity as high as 2.5–5 mA.112 Indeed, nerve conduction studies in diabetics with neuropathy consistently show a reduction in conduction velocity and amplitude, for both motor and sensory nerves, which may help explain this phenomenon.113 However, in our tertiary center where we perform nearly all foot and amputation surgery exclusively under lower extremity nerve blocks, the overwhelming majority of patients still have a normal response to nerve stimulation, and low current stimulation (0.5 mA) is almost always obtained even in patients with overt diabetic neuropathy. Nevertheless, with the advent of ultrasound-guided nerve blocks, it has been shown that in certain circumstances if the needle is directly adjacent to a nerve, it may be possible to obtain little or no motor response to stimulation, even when injection of local anesthetic through that needle results in a clinically sound block.114 For these reasons, in this population it is probably acceptable to inject local anesthetic even when stimulation cannot be obtained with currents <0.5–0.7 mA as long as the obtained motor response is well defined and specific for the nerve being blocked. Where applicable, ultrasound-guided nerve blocks should also be beneficial in these patients.

Effect of Regional Anesthesia on Glucose Homeostasis

It is well known that surgery performed in combination with general anesthesia provokes a counterregulatory response that significantly increases plasma levels of glucose, as well as levels of cortisol and catecholamines. This so-called stress response has long been considered a homeostatic defense mechanism that is important in an organism's adaptation to harmful stimuli. However, recent evidence has suggested that the avoidance or even treatment of hyperglycemic states in the perioperative and critical care settings can lead to improved morbidity and mortality rates. For example, Van Den Berghe and coworkers randomized over 1500 critically ill surgical patients to receive either aggressive (goal blood glucose 80–110 mg/dL) or standard (goal blood glucose 180–200 mg/dL) insulin therapy while in the intensive care unit.115 At 12 months, the mortality rate with aggressive therapy was markedly reduced from 8.0% to 4.6%. These outcomes data were supported in a similar study by Krinsley involving a heterogeneous group of critically ill patients, both surgical and medical.116 Cardiac surgical patients have been shown to have a 30% higher risk of an adverse outcome (including cardiac, neurologic, pulmonary, infectious, or death) when intraoperative mean glucose levels are 20 mg/dL higher than those in controls.117 The exact mechanism by which tight glycemic control improves outcomes in these populations is unclear, but may involve decreased hepatic glucose production, positive influences on immune function with alteration in inflammatory cytokines, or enhanced glucose transport to intracellular sites.118

Regional anesthesia has been shown to ameliorate the hyperglycemic response to surgery and therefore may play a role in this protective phenomenon. An intraoperative glucose tolerance test resulted in markedly elevated plasma glucose levels in patients receiving general versus epidural anesthesia for such procedures as inguinal herniorrhaphy and hysterectomy.119–120 Likewise, abdominal hysterectomy performed under spinal anesthesia is associated with lower intra- and postoperative glucose levels compared with those for neurolept anesthesia.121 Retrobulbar block reduces the hyperglycemic stress response to both cataract122 and scleral buckle123 surgery.

Glucose homeostasis is rather complex and several factors likely contribute to the salutary action of regional anesthesia on glycemic control. These may include the inhibition of hepatic gluconeogenesis, as well as the inhibition of catecholamine and cortisol responses to surgery.124 In addition, the “absence of general anesthesia” may be a causative factor in glycemic control, as volatile agents such as halothane and enflurane have been shown to impair glucose tolerance in dogs.125 It seems clear from the available data that to improve outcomes from major surgery anesthesiologists must prevent as much nociceptive input from reaching the central nervous and neuroendocrine system as possible.126 The use of regional anesthesia can easily facilitate this and may be especially pertinent for “brittle” diabetics in whom tight glycemic control is difficult at the best of times.

Diabetes and uremia are the most common metabolic neuropathies; however, several other less common causes have implications for the regional anesthesiologist. These include those neuropathies that result from the use of certain medications; exposure to toxins; and those related to connective tissue, autoimmune, and vascular diseases.127 One of the most prevalent causes of metabolic neuropathy is that associated with overt hypothyroidism. Thyroid neuropathy is a largely sensory phenomenon that is poorly understood, but is present in approximately 40% of patients diagnosed with hypothyroidism.128 It is most obvious in frank myxedema, but nerve conduction studies have shown evidence of velocity impairment in subclinical hypothyroidism.129 Thyroid neuropathy is most likely to present as peripheral nerve entrapment, particularly the median nerve, and these patients are frequently referred for carpal tunnel decompression.130 Entrapment of the eighth cranial nerve leading to deafness is also not uncommon. Patients may complain of dysesthesias in a glove-and-stocking pattern, as well as lancinating pains suggestive of nerve root compression. “Hung” deep tendon reflexes (brisk reflex response with a delayed return to normal tone) are a hallmark of hypothyroidism and probably relate to both neuropathy and myopathy. Pathologically, affected nerves exhibit mucinous deposition and, in advanced cases, segmental demyelination with loss of large myelinated nerve fibers.131

Little data exist on the effect preexisting thyroid neuropathy may have on the management of regional anesthesia in this population. It may be reasonable to predict that electrolocation responses may be diminished, as is seen in diabetes, resulting in an inability to elicit twitches at desired levels (eg, 0.3–0.4 mA) of current, although no direct data support this. Another potential consequence of performing regional anesthesia in patients with nerve entrapment is what has been termed the “double-crush syndrome.”132 This refers to the enhanced susceptibility of nerves to injury or impairment at one anatomic location, when already compressed or otherwise injured at another, separate location. A classic example is the patient with symptoms of carpal tunnel syndrome after seemingly minor trauma or injury to the median nerve, who is later found to have compression of the C6 nerve root. Although originally described in terms of mechanical injury, it has been recognized that metabolic and pharmacologic factors can contribute to the double-crush syndrome,133 including hypothyroidism. Thus, it may be that patients with thyroid neuropathy are at increased risk for neurologic injury when receiving regional anesthesia blocks, as minor needle trauma to a susceptible nerve may produce functional neurologic deficits. Although this remains speculative at present, it reinforces the need for a detailed history and documentation of preexisting neurologic deficit in patients with hypothyroidism and the careful consideration of techniques in these patients. Finally, if suspected, thyroid neuropathy has been shown to be correctable in many cases by prompt treatment with thyroid replacement therapy, which may allow for some degree of risk reduction from this complication.134,135

Obesity is an increasingly prevalent problem worldwide. Between 1980 and 2000 in the United States, obesity rates doubled for adults and children and tripled for adolescents.136 In addition to the usual anesthetic considerations for morbidly obese patients, such as the presence of various cardiopulmonary, gastrointestinal, and endocrine comorbidities, the abundance of extra tissue can present a challenge to regional anesthesiologists. Obesity has been shown to impair the ability of anesthesiologists to correctly identify lumbar spinal interspaces.137 Outcomes are similarly affected in overweight patients. In a study of over 9000 blocks at a single institution, patients with a body mass index (BMI) >30 kg/m2 were 1.62 times more likely to experience a failed regional block than were those with a BMI >25 kg/m2.138 Not surprisingly, the investigators cited difficulty in landmark identification, patient positioning, and insufficient length of needle used as the main impediments to successful block placement. Despite this, regional anesthesia often remains an attractive option for obese patients, because it may reduce the incidence of cardiopulmonary and airway complications compared with those experienced during general anesthesia.

Several groups have sought to overcome the challenge of increased tissue mass with the use of novel radiologic techniques while performing regional blocks. Although fluoroscopy has been used extensively by interventional pain management physicians to increase the reliability and precision of various nerve blocks, its application in facilitating blocks for surgical anesthesia has been less practical. Nevertheless, fluoroscopy has been used in the placement of axillary brachial plexus catheters,139 the performance of sciatic nerve blocks,140 and as an aid in facilitating spinal anesthesia in morbidly obese patients.141 However, its use is limited by the need to relate neural anatomy to structures that appear radiodense, such as bones, needles, or contrast-injected vessels.

Ultrasound is another promising technology that is being utilized with increasing frequency in regional anesthesia and that may aid the placement of peripheral and neuraxial blocks in the morbidly obese. Proponents of ultrasound-guided regional anesthesia cite the potential advantage of direct visualization of both the needle and the nerve structures,142 which may be especially useful in obese patients with obscured surface landmarks. Studies conducted in obese parturients have verified the utility of ultrasound in identifying the epidural space and other spinal structures prior to introduction of an epidural needle.143,144 One potential limitation of ultrasound-guided blocks in the obese is the penetration depth of the transducer, which becomes smaller as the frequency (and thus the resolution) increases. To date, little evidence has been published that directly addresses the use of ultrasound for regional anesthesia in the obese population.

Obesity may have an effect on the dosing of spinal medications, although the issue is somewhat controversial. A common notion is that increased abdominal mass leads to compression of intrathecal volume by engorgement of epidural plexuses, leading to an increased, and potentially dangerous, block height during spinal anesthesia. This is supported by data correlating block height with patient weight during standardized spinal anesthesia for cystoscopy.145 Indeed, some authors have advocated the consideration of “low-dose” spinal anesthesia in the morbidly obese, due to wide variations in their dosing requirements. In one extremely overweight parturient (BMI = 66 kg/m2), cesarean section was completed successfully with a 5-mg spinal dose of bupivacaine as the sole anesthetic agent.146 However, patient weight does not necessarily correlate with degree of compression of the intrathecal sac, and many investigators have argued that weight alone is not a reliable predictor of block height during spinal anesthesia.147–149 It is probably reasonable to approach the spinal dosing of obese patients with a degree of caution, and when practical, incrementally adjust the anesthetic dose.