Vol 7, No 2, March/April 1999 101 Nerve injury after total hip arthro- plasty (THA) can be a devastating complication. Upper-extremity nerve-traction palsies related to positioning have been reported with a higher incidence of ulnar nerve damage in arthroplasty pa- tients with inflammatory arthritis. 1 Obturator nerve injuries are ex- ceedingly uncommon but have been the subject of case reports. Superior gluteal nerve and femoral nerve injury are relatively rare and in general have a better prognosis than sciatic nerve injuries. Most of the literature focuses on the per- oneal division of the sciatic nerve. The physiologic demands as well as the structural anatomy of the nerves about the hip predispose these structures to injury during surgical procedures. In this article, we will review clinical studies on the diagnosis, treatment, and prog- nosis of this infrequent complica- tion and focus attention on the importance of prevention. Peripheral Nerve Anatomy and Physiology Nerves are uniquely designed to quickly transmit electrical impulses, or action potentials, over long dis- tances. Each nerve cell is composed of four regions (Fig. 1). Dendrites are the thin processes that collect signals from other nerve cells. The cell body contains the nucleus and organelles, which tend to the meta- bolic needs of the neuron. A single axon branches off each cell body and acts as a transport tube for pro- teins and a conduit for transmission of action potentials. Sensory nerves are afferent fibers that transmit action poten- tials from nociceptors or mechano- receptors toward the dorsal root ganglia of the central nervous sys- tem. Motor nerves are efferent fibers that carry action potentials to special motor end-plates on muscle spindles. Presynaptic terminals are the specialized nerve endings that transmit information to dendrites of other nerves or to a neuromus- cular junction. The model hypoth- esized for transmission of proteins and transport of neurotransmitter vesicles from cell body to axon involves carrier proteins, micro- tubules, and an adenosine triphos- phate (ATP)Ðdependent protein called kinesin. Retrograde axoplas- mic transport serves to recycle empty vesicles and has been linked to the transmission of herpes sim- plex, rabies, and polio viruses and tetanus toxin. A number of other cell types are associated with the primary nerve cells. Glial cells surround nerve cell bodies and their axons. Microglia are the phagocytes that mobilize after injury. Oligodendrocytes and Schwann cells are macroglia cells that produce the myelin sheath, which enhances the speed of signal transmission. Oligodendrocytes are found only in the central ner- vous system; each one myelinates many different axons. Schwann Dr. DeHart is Attending Orthopaedic Surgeon, Huebner Medical Center, San Antonio, Tex; and Clinical Assistant Professor, Department of Orthopaedic Surgery, University of Texas Health Science Center at San Antonio. Dr. Riley is Professor of Orthopaedic Surgery, Johns Hopkins University, Baltimore. Reprint requests: Dr. DeHart, Suite 390, 9150 Huebner Road, San Antonio, TX 78240. Copyright 1999 by the American Academy of Orthopaedic Surgeons. Abstract Nerve injury occurs in 1% to 2% of patients who undergo total hip arthroplas- ty and is more frequent in patients who need acetabular reconstruction for dys- plasia and those undergoing revision arthroplasty. Injury to the peroneal divi- sion of the sciatic nerve is most common, but the superior gluteal, obturator, and femoral nerves can also be injured. Nerve injury can be classified as neu- rapraxia, axonotmesis, or neurotmesis. The worst prognosis is seen in patients with complete motor and sensory deficits and in patients with causalgic pain. Prevention is of overriding importance, but use of ankle-foot orthoses and prompt management of pain syndromes can be useful in the treatment of patients with nerve injury. Electrodiagnostic studies hold promise in complex cases; however, their intraoperative role requires objective, prospective, con- trolled scientific study before routine use can be recommended. J Am Acad Orthop Surg 1999;7:101-111 Nerve Injuries in Total Hip Arthroplasty Marc M. DeHart, MD, and Lee H. Riley, Jr, MD Nerve Injuries in Total Hip Arthroplasty Journal of the American Academy of Orthopaedic Surgeons 102 Nerve Injury Seddon classified nerve injuries into three types (Fig. 5). Neura- praxia is a conduction block of anatomically intact nerves caused by minor injury. A period of loss of sensation may occur, but recov- ery is likely to be complete. Axon- otmesis is a more severe injury in which axons are disrupted but the investing connective tissue sur- vives. Wallerian degeneration is the process of disintegration of the axon and myelin sheath over the entire axon distal to the site of injury. Preservation of the endo- neurial supporting structures can cells are found in the peripheral nervous system; each provides myelination for only a 0.1- to 1.0- mm segment of the axon of one neuron (Fig. 2). The spaces between Schwann cells, or nodes of Ranvier, are the sites of action potential initiation. The internal milieu of a resting nerve cell is electronegative com- pared with the extracellular fluid. This electrical difference is main- tained by energy from ATP and a sodium-potassium pump that con- stantly pumps sodium out of the cell. Chemical, mechanical, or volt- age changes can cause sodium gates to open, allowing an influx of positive sodium ions, a process known as depolarization (Fig. 3). The result is an action potential Ñ a brief explosive change in the nerve cell to a very positive value. Nodes of Ranvier have high con- centrations of voltage-gated sodi- um channels that ease initiation of action potentials. The insulating myelin allows action potential cur- rent to flow quickly with little attenuation to the next node of Ranvier. This jumping of the ac- tion potential down the neuron from node to node is very effective in increasing the conduction speed of the signal down the axon. De- myelinating diseases, such as mul- tiple sclerosis and Guillain-BarrŽ syndrome, interfere with the prop- agation of the signal, slowing or completely stopping conduction. As many as several thousand axons and their accompanying Schwann cells are bundled together by a loose endoneurial connective tissue into fascicles (Fig. 4). Edema of this endoneurium is the hallmark of irreversible nerve damage. Peri- neurium is the expansion of dense connective tissue that surrounds each fascicle. It has high tensile strength and serves as the major barrier for endoneurial edema. Fascicles are bundled together into nerves by epineurium, which is a loose meshwork composed of colla- gen and elastin that protects the nerve from compressive forces. Spinal nerves do not have peri- neurium or epineurium and are more vulnerable to tensile and com- pressive forces. 2 Because of the high metabolic demands of the nerve tissue, blood supply is crucial for effective signal transmission. The vascular supply to the nerves is derived from both a segmental extrinsic system of superficial vessels, which give off perforating branches, and an intrin- sic system of epineurial, perineu- rial, and endoneurial plexuses with their communicating branches. Fig. 1 A typical neuron. (Adapted with permission from Bodine SC, Lieber RL: Peripheral nerve physiology, anatomy, and pathology, in Simon SR (ed): Orthopaedic Basic Science. Rosemont, Ill: American Academy of Orthopaedic Surgeons, 1993, p 327.) Cell body Nucleus Myelin sheath Dendrites Axon hillock Axon Node of Ranvier Terminal branches Presynaptic terminal Postsynaptic neuron Nodes of Ranvier Nucleus Inner tongue Axon Layers of myelin Fig. 2 The Schwann cell wraps lipid-rich myelin around the axon to enhance con- duction of electrical impulses. (Adapted with permission from Bodine SC, Lieber RL: Peripheral nerve physiology, anatomy, and pathology, in Simon SR (ed): Ortho- paedic Basic Science. Rosemont, Ill: Amer- ican Academy of Orthopaedic Surgeons, 1993, p 328.) Marc M. DeHart, MD, and Lee H. Riley, Jr, MD Vol 7, No 2, March/April 1999 103 sis. The effect of ischemia with compression in the tissue under a tourniquet was compared with ischemia alone in more distal tissue. Evidence of endoneurial edema was found after 2 to 4 hours of com- pression. The amount of stretch a nerve will tolerate depends on whether the nerve is freely mobile in supple soft tissue or whether it is bound down by osseous prominences, fascia, or scar. Rabbit sciatic nerve showed conduc- tion failure with 25% lengthening. Histologic changes were found after lengthening of nerves by 4% to 11%. Nerve microcirculation was found to be impaired after 8% stretch and stopped after 15% stretch. 5 The rup- ture of axon fibers precedes the fail- ure of fascicles and may occur with stretch of as little as 4% to 6%. 2 Four major factors that increase the probability of mechanical dis- ruption include increased load due to compression or stretch, increased rate of loading, increased duration of loading, and uneven application of load to tissues. Clinical experi- ence and LaplaceÕs law (tension is proportional to the pressure and radius of a cylindrical structure) demonstrate that large-diameter axons are more susceptible to dam- guide the slow (1 mm per day) regeneration of sprouts to their original end-organ. Atrophy and scarring of muscle can prevent effective function if axons fail to reach motor end-plates within 2 years of the time of injury. 3 This accounts for the variability in de- gree of recovery. Complete disrup- tion of the nerve, or neurotmesis, carries the worst prognosis for recovery and may lead to abortive efforts at regeneration, which can result in painful neuromas. Damage to nerves and their blood supply can be caused by var- ious combinations of compression, stretch, ischemia, and transection. Compressive trauma affects both the structure of the nerve and the neural vascular supply. Large, closely packed fasciculi have less cushioning epineurium and are more vulnerable than nerves with smaller fasciculi and a greater amount of epineurial tissue. 2 LundborgÕs classic experiments involving tourniquet application 4 demonstrated that normal circula- tion returned within seconds when inflation time was 2 hours or less. For tourniquet times between 4 and 6 hours, it took 2 to 3 minutes for the circulation to return, and a peri- od of hyperemia was observed. After 8 to 10 hours, blood flow took 5 to 20 minutes to return, and there was evidence of microvascular sta- Fig. 3 A, When the cell is at rest, the passive fluxes of Na + and K + into and out of the cell are balanced by the energy-dependent sodium-potassium pump. (Adapted with permis- sion from Bodine SC, Lieber RL: Peripheral nerve physiology, anatomy, and pathology, in Simon SR [ed]: Orthopaedic Basic Science. Rosemont, Ill: American Academy of Orthopaedic Surgeons, 1993, p 332.) B, An action potential can be the result of voltage (top), chemical (center), or mechanical (bottom) changes that open axon sodium channels. The resulting sodium influx causes an explosive positive change in the nerve cell. (Adapted with permission from Kandel ER, Schwartz JH, Jessell TM: Principles of Neural Science, 3rd ed. Norwalk, Conn: Appleton & Lange, 1991, pp 75-76.) Vessel Epineurium Perineurium Endoneurium Fascicle Nerve fiber Epineurium Perineurium Endoneurium Fig. 4 Cross-sectional anatomy of a nerve. (Adapted with permission from Wilgis EFS, Brushart TM: Nerve repair and grafting, in Green DP (ed): Operative Hand Surgery. New York: Churchill Livingstone, 1993.) A B Closed Ligand binding Stretch Open P i ∆Vm Nerve Injuries in Total Hip Arthroplasty Journal of the American Academy of Orthopaedic Surgeons 104 age than smaller fibers. Clinical factors, which will be discussed later, also play a role in nerve injury and repair. Anatomy of the Peripheral Nerves About the Hip The superior gluteal nerve arises from the L4, L5, and S1 nerve roots and exits at the sciatic notch to sup- ply the gluteus medius, gluteus minimus, and tensor fascia lata. It travels with the superior gluteal artery deep to the gluteus maximus and medius but superficial to the gluteus minimus. In one study, 6 subclinical superior gluteal nerve injury was found in 77% of cases in which a transtrochanteric lateral or posterior hip approach was used. Anterior branches supplying the distal tensor fascia lata may be sac- rificed during the anterolateral approach when dissecting the inter- val between the gluteus medius and the tensor. It is also at risk if the 3- to 5-cm Òsafe areaÓ proximal to the greater trochanter is violated during direct lateral approaches to the hip. 7 When this safe area is respected, clinical deficits are rare. 8 A positive Trendelenburg sign or Trendelenburg gait and weak ab- duction can indicate damage to this nerve. The obturator nerve arises from the L2, L3, and L4 nerve roots within the posterior psoas and then emerges medially at the sacral ala to travel along the ilio- pectineal line (Fig. 6). It is rarely injured, 9 but case reports docu- ment a risk of injury when cement, screws, or reamers penetrate the anterior quadrants of the acetabu- lum (Fig. 7). The obturator nerve exits the pelvis at the superior aspect of the obturator foramen, where it supplies the adductor muscles and a medial patch of thigh skin. Persistent pain in the groin or thigh, adductor weakness after placement of intrapelvic screws, or allograft or cement visi- ble on radiographs after hip arthro- plasty suggests obturator nerve in- jury. 10 Hip disease can cause re- ferred pain to the knee through the obturator distribution. The femoral nerve arises from the L2, L3, and L4 nerve roots; passes through the psoas major muscle; and then travels between the psoas and the iliacus to enter the thigh as the lateralmost structure of the femoral triangle. The femoral nerve supplies motor impulses to the mus- cles of knee extension (the quadri- ceps) and sensation to most of the medial thigh and calf. Prolonged hyperextension of the thigh can cause femoral-nerve traction in- juries. Iliacus hematomas in patients who have bleeding disorders or are taking anticoagulants are well- known causes of femoral nerve palsy. The femoral nerve is rarely injured after hip arthroplasty (0.04% to 0.4% of cases) 1,9,11 ; it is most at risk during placement of anterior acetabular retractors when anterior or anterolateral approaches are used (Fig. 8). Simmons et al 12 reported the highest rate of femoral nerve palsy after hip arthroplasty (2%); however, all patients had full func- tional recovery by 12 months. The presence of thigh pain, anteromedial Fig. 5 Spectrum of injury to a nerve as it is stretched. (Only one axon in its connec- tive tissue layers is shown.) A = Neura- praxia, with all anatomic structures intact. B = Axonotmesis, with disruption of the axon but intact connective tissues. C = Neurotmesis, with complete disruption of all layers. (Adapted with permission from Sunderland S: Nerve Injuries and Their Repair: A Critical Appraisal. New York: Churchill Livingstone, 1991, p 148.) A B C Fig. 6 Cross section of the pelvis at the level of the hip joint shows location of the obtura- tor, femoral, and sciatic nerves. (Adapted with permission from Weber ER: Peripheral neuropathies associated with total hip arthroplasty. J Bone Joint Surg Am 1976;58:66-69.) Obturator nerve Femoral nerve Sciatic nerve Femoral artery Femoral vein Marc M. DeHart, MD, and Lee H. Riley, Jr, MD Vol 7, No 2, March/April 1999 105 thigh and medial leg numbness, quadriceps weakness, and difficulty climbing stairs suggests femoral nerve injury. The sciatic nerve arises from the L4, L5, S1, S2, and S3 nerve roots and is composed of the preaxial anterior tibial and postaxial poste- rior peroneal divisions. These divi- sions usually travel together in a single sheath, but in 10% to 30% of cases they are separate as high as the greater sciatic notch. The nerve is located deep to the piriformis inside the pelvis and then travels distally deep to the gluteus muscles and superficial to the external rota- tors at the level of the hip joint. It is the nerve most frequently in- jured during hip arthroplasty and is at risk for injury either from placement of posterior acetabular retractors or from anterior or lateral traction on the femur. Distal to the level of the lesser trochanter and the ischial tuberosi- ty, the sciatic nerve passes between the adductor magnus and the long head of the biceps femoris, just medial to the insertion of the gluteal sling. All medial branches of the sciatic nerve arise from the tibial division and supply the ham- strings. The short head of the biceps is the only thigh muscle supplied by the peroneal division of the sci- atic nerve. At the superior aspect of the popliteal fossa, the two divi- sions split into the tibial nerve and the common peroneal nerve. The peroneal nerve innervates the dor- siflexors and evertors of the foot and ankle. The tibial nerve inner- vates the plantar flexors and inver- tors. The sural nerve provides sen- sation to most of the lateral aspect of the calf and arises from both the medial sural cutaneous branch of the tibial nerve and the lateral sural cutaneous branch of the common peroneal nerve. Sciatic Nerve Injury in THA As mentioned previously, the sciat- ic nerve is the nerve most common- ly injured during THA. It was in- volved in over 90% of the 53 nerve injuries reported by Schmalzried et al 11 in their series of more than 3,000 cases. The incidence of sciatic nerve injury in primary THA has been reported to be between 0.6% and 3.7%, with most large series citing a rate of about 1.5%. 9,11 Overall rates are elevated by the relatively higher incidence in revi- sions (3% to 8%) and in patients with developmental dysplasia of the hip (DDH) (5.8%). 11 In uncom- plicated primary THA, sciatic nerve injury occurs in fewer than 1% of cases. Weber et al 9 noted that since only severe injury pre- sents as a clinical problem, the con- dition may be more common than is generally appreciated. They reported that in a study utilizing preoperative and postoperative electromyography, 70% of THA patients had subclinical sciatic nerve injury. The etiology of nerve injury is protean. Direct trauma from scalpel, electrocautery, retractors, wires, reamers, Gigli saw, bone fragments, or cement protrusion; constriction by suture, wire, or cable; heat from the polymeriza- tion of cement; compression from dislocation; excessive lengthening; and subfascial hematoma have all been reported. However, the cause of 50% of all sciatic nerve palsies is unknown. 11,13 Fig. 7 The acetabular-quadrant system (ASIS = anterosuperior iliac spine). Screws originating in the anterosuperior quadrant threaten the external iliac artery and vein. Screws from the anteroinferior quadrant threaten the iliac vessels and the obturator nerve. There is less risk of neurovascular injury with placement of short (<25 mm) screws in the posterosuperior quadrant. (Adapted with permission from Wasielew- ski RC, Cooperstein LA, Kruger MP, Rubash HE: Acetabular anatomy and the transacetabular fixation of screws in total hip arthroplasty. J Bone Joint Surg Am 1990; 72:501-508.) ASIS Postero- superior Postero- inferior Antero- superior Antero- inferior Fig. 8 The close proximity of the femoral nerve and vessels to an anteriorly placed acetabular retractor. (Adapted with permis- sion from Shaw JA, Greer RB: Compli- cations of total hip replacement, in Epps CH [ed]: Complications in Orthopaedic Surgery. Philadelphia: JB Lippincott, 1994, p 1035.) Nerve Injuries in Total Hip Arthroplasty Journal of the American Academy of Orthopaedic Surgeons 106 The peroneal division of the sci- atic nerve is more susceptible to injury than the tibial division. Schmalzried et al 11 found that 94% of the sciatic nerve injuries in their study involved the peroneal divi- sion. The tibial division was only rarely involved by itself (2% of cases). The superficial position of the common peroneal nerve as it wraps around the neck of the fibula makes it vulnerable to compres- sion. The peroneal division may be more susceptible to stretch injuries because it is relatively more fixed between the sciatic notch and the fibular head. Another explanation is based on morphologic differences between the tightly packed fascicles of the peroneal division and those of the tibial division, which has relatively more connective tissue (Fig. 9). The fact that the peroneal division is more lateral may also increase its vulnerability to trauma. Generally accepted risk factors for sciatic nerve injuries include THA in patients with DDH and revision THA. Johanson et al 13 noted increased blood loss and time of surgery in their patients with nerve injury and considered all fac- tors to be related to the level of diffi- culty of the case. Whether the risk is higher in women and to what degree leg lengthening increases risk remain controversial. Some hypothesize that the increased risk in women is due to less soft-tissue mass, 9 while others believe that the increased risk is related to the prevalence of DDH. 3 Pritchett 14 suggests that a Òdouble crushÓ phe- nomenon, as described in carpal tunnel syndrome, may play a role in hip arthroplasty patients who also have spinal stenosis. Of 16 spinal stenosis patients with footdrop of unknown cause diagnosed after hip arthroplasty, 12 had improvement of the nerve palsy after spinal decompression with laminectomy. The role of leg lengthening in cases of nerve injury is unclear. Nerves will tolerate only a finite amount of acute stretch. A soft- tissue bed that is scarred and has compromised vascularity should in- crease the potential for injury with lengthening. Both the presence of scar from previous operations and the desire to increase leg length must be considered in revisions and DDH cases. Edwards et al 15 noted that 6 of 10 patients with nerve palsy had leg lengthening of more than 3 cm. Lengthening of less than 3.8 cm was associated with only peroneal- division palsies; lengthening by more than 3.8 cm was associated with both peroneal- and tibial-division palsies. A Mayo Clinic review of DDH pa- tients who underwent THA demon- strated a 13% incidence of sciatic nerve palsy. 16 Of those patients with lengthening of 4 cm or more, 28% had palsy. No nerve injuries oc- curred in those with lengthening of less than 4 cm. Kennedy et al 17 de- monstrated acute progressive wors- ening of somatosensory evoked potential (SSEP) changes when reduction of the femoral head was achieved after increasing the neck length. All four patients in whom causalgia developed postoperatively had SSEP reductions in amplitude greater than 75% after component reduction. Nercessian et al 18 report- ed on 66 patients with lengthening between 2.0 and 5.8 cm without neu- rologic deficit. They calculated the amount of lengthening as a percent- age of the length of the femur and concluded that lengthening of up to 10% was safe. Diagnosis Clinical assessment alone underesti- mates the true incidence of nerve injury after hip arthroplasty. 6,9,19 The diagnosis of significant nerve Transection Fig. 9 Multifascicular nerves with abundant connective tissue (as seen in the tibial divi- sion) are less vulnerable to transection or compression than nerves with tightly packed fas- cicles (as seen in the peroneal division). (Adapted with permission from Bodine SC, Lieber RL: Peripheral nerve physiology, anatomy, and pathology, in Simon SR (ed): Orthopaedic Basic Science. Rosemont, Ill: American Academy of Orthopaedic Surgeons, 1993, p 361.) Tibial Division Peroneal Division Marc M. DeHart, MD, and Lee H. Riley, Jr, MD Vol 7, No 2, March/April 1999 107 injuries is usually not challenging and should begin with adequate pre- operative documentation of motor strength and sensation. Thorough evaluation of patients before revision procedures may help demonstrate minor nerve damage caused by prior procedures. Complaints of weak- ness, numbness, or paresthesias may indicate a compromised nerve that is at increased risk during the next operation. Electrodiagnostic tests can indicate nerve impairment and the presence of preexisting neu- ropathies of diabetes, chronic alcohol abuse, or hypothyroidism. Weakness of the muscles of ankle dorsiflexion can indicate damage to the peroneal division of the sciatic nerve. Its sensory distribution in- cludes the dorsum of the foot, with the deep peroneal nerve supplying the web space between the first and second toes. The short head of the biceps is the only muscle above the knee that is supplied by the peroneal division. Electromyographic (EMG) recordings of this muscle can show whether peroneal division injury is at the level of the hip or at another site of vulnerability, the fibular head. Damage to only the tibial division is rare and results in weakness of all knee flexors except the short head of the biceps. Weakness of the posteri- or muscles of ankle and toe plantar flexion should also be seen. Prognosis The prognosis of a nerve injury is related to factors specific to the injury and clinical factors related to the patient (Table 1). In the case of nerve injuries related to THA, the patient is often elderly, with multi- ple concurrent medical problems. The damaged nerve is large, and the distance from the end-organ is great. In revision surgery, the previous operation may leave binding scar tissue and an altered vascular sup- ply to the nerve. All these factors decrease the likelihood of successful nerve regeneration. In a large series of primary THA procedures, as many as 2% of patients had tran- sient neurologic problems, and 0.5% had permanent nerve damage. 20 Although 80% of patients with nerve injuries have some persistent neurologic dysfunction, 11,13 causal- gic pain most highly predicts major disability. 13 Edwards et al 15 found that patients who had palsy of only the peroneal division did well, but that patients with injury to both the tibial and the peroneal divisions had less optimal recovery of func- tion. In a review of over 3,100 hip arthroplasty operations, Schmalz- ried et al 11 found that patients who recovered neurologic function usu- ally did so by 7 months. All pa- tients who had evidence of some motor function immediately after the operation or who recovered some motor function during their hospital stay had a good recovery. No patient with dysesthesias had satisfactory recovery of function. Treatment If no specific cause is identified, often no immediate treatment to decrease compression or stretch of the nerve is indicated. Serial exam- inations may demonstrate nerve recovery. An advancing Tinel sign distal to the site of injury signifies regeneration of axons and at least partial nerve continuity. Electro- myograms and nerve conduction velocity measurements may pro- vide a more objective measure of the level of injury, the degree of injury, and evidence of recovery of motor function. When transection of the nerve is discovered intraoperatively, an attempt at nerve repair seems war- ranted. Sunderland et al 21 reported a single case of a sharply transected sciatic nerve in a young patient that was repaired early with a good result. The results of nerve repair in elderly patients are not promising. In the rare case in which a mechani- cal source of compression can be identified, it should be removed. Removal of long-skirted femoral heads or more superior placement of the acetabular component can relieve tension on the nerve; how- ever, this may require trochanteric advancement to restore soft-tissue tension. Delayed onset of progres- sive neurologic symptoms after a normal postoperative check should alert the physician to consider evac- uation of a subfascial hematoma. In patients who have received anti- coagulation therapy, motor and sen- sory deterioration with increasing thigh circumference demands cor- rection of coagulation status and surgical decompression. Motor deficits can often be man- aged with physical therapy to strengthen ankle dorsiflexors and stretch antagonist muscles so as to prevent joint contractures. Ankle- foot orthoses can be used to treat footdrop, allowing clearance dur- ing the swing phase and prevent- Table 1 Clinical Factors in Nerve Injury and Repair Injury factors Nerve injured Degree of injury Size of zone of injury Distance of zone of injury from an end-organ Local tissue condition at injury site (tension, vascularity, presence of scar and infection) Patient factors Age Preexisting neuropathies (e.g., those due to diabetes, alco- holism, hypothyroidism, spinal stenosis) General medical condition (e.g., history of smoking or cortico- steroid use) Nerve Injuries in Total Hip Arthroplasty Journal of the American Academy of Orthopaedic Surgeons 108 ing the steppage gait indicative of weak dorsiflexors. Orthoses may also help prevent equinus defor- mity. The presence of sensory deficits requires diligence from the patient in preventing inadvertent trauma to the extremity, as may occur in patients with diabetic neuropathy. Dysesthesia and causalgic pain postoperatively are best treated with antidepressants as well as with early and repeated sympa- thetic nerve blocks as needed. 13 Surgical correction of late equino- varus deformities may be necessary in rare cases to provide stability of the ankle joint, to release contrac- tures, or to provide active dorsi- flexion. Prevention The best treatment of any complica- tion is prevention. The first step in prevention is to identify the pa- tients who are most at risk. Patients with hip dysplasia and those who will undergo revisions are clearly at increased risk. Minimizing the amount of leg lengthening during preoperative planning and using leg-length measurement techniques may decrease the risk to the nerve. There is no strong evidence favor- ing any one approach for prevent- ing nerve injury. Whether it is prudent to expose the nerve in high-risk cases is still debated. Stillwell 22 performed neurolysis to free the sciatic nerve from binding scar and to allow mobilization during revision cases. Simon et al 23 recommend exposure, macroneurolysis, and protection routinely in every posterior ap- proach and noted only one instance of sensory loss in 400 cases. In- creased intraoperative attention with palpation of the nerve before and after arthroplasty and limited sciatic neurolysis of nerves that are tethered or difficult to locate may help to decrease the prevalence of nerve injury. 24 However, direct exposure of the nerve may damage the anastomotic blood supply and can lead to increased scarring. Technical factors that may help decrease the incidence of nerve damage include wide exposure and meticulous hemostasis to ensure visualization of the anatomy, con- stant attention to nerve position, and careful placement and replace- ment of retractors. Deliberate con- trol of scalpels and reamers is essential to avoid unwanted injury to soft tissue. Careful placement of fixation screws and attention to drill-bit depth are essential. Use of anterior-quadrant screws predis- poses to nerve injury (Fig. 7). Internal rotation of the femur dur- ing placement of cerclage wires and cables is recommended to help visualize the soft tissue of the pos- terior femur. 25 Proper placement of compo- nents helps minimize dislocations and the need for revisions that put the nerve at increased risk. Black et al 26 recommend extreme care when revising only the acetabular com- ponent in patients with monolithic stems, which can rest directly on the sciatic nerve. When cementing the cup in acetabular revision cases, the use of bone graft may help prevent intrapelvic extravasa- tion of cement. Electrodiagnostic Studies Several studies have reported the use of electrodiagnostic tools, such as evoked potentials (Fig. 10) and electromyography (Fig. 11), to warn surgeons of impending damage to peripheral nerves during surgery. Evoked potentials, first described in 1875 by Caton, are voltage changes in sensory fibers after stimulation of peripheral nerves. Damage or irritation of nerves can alter electri- cal signals by decreasing the size Fig. 10 Examples of SSEP recordings. A, Baseline. B, Increased latency and decreased amplitude associated with retractor compression of the sciatic nerve. C, Recovery of trac- ing when the retractor is removed. (Reproduced with permission from Stone RG, Weeks LE, et al: Evaluation of sciatic nerve compromise during total hip arthroplasty. Clin Orthop 1985;201:26-31.) Postincision P1 = 39.2 msec N1 = 46.4 msec P1−N1 = 1.5 µV Retractor on sciatic nerve P1 = 46.4 msec N1 = 54.4 msec P1−N1 = 0.7 µV Retractor removed P1 = 40.8 msec N1 = 52.0 msec P1−N1 = 1.2 µV A B C 1.5 µV 10 msec + − Marc M. DeHart, MD, and Lee H. Riley, Jr, MD Vol 7, No 2, March/April 1999 109 (amplitude) or increasing the trans- mission time (latency) of the evoked potential. Somatosensory evoked potentials (SSEPs) are recorded over the somatosensory cortex and are monitored by an electroen- cephalographlike device to give feedback to the operating surgeon. Somatosensory potentials are used commonly in spine surgery. The American Electroencephalographic Society guidelines recommend using a decrease of 50% or more in amplitude or an increase of 10% or more in latency to identify neuro- logic compromise. In an animal study, 27 statistically significant SSEP changes were seen prior to damage that caused postoperative motor changes. However, com- plete motor palsy in one division can be caused while normal SSEP tracings are seen in the other divi- sion. Amplitude changes can be influ- enced by changes in patient tem- perature, blood pressure, P CO 2 , level of anesthesia, and the electri- cal noise in the operating room. Cortical SSEPs cannot be recorded during spinal anesthesia. For these reasons, Kennedy et al 17 recom- mend placing the monitoring elec- trode directly on the most proximal extent of the sciatic nerve. Stone et al 28 first used SSEP monitoring of the peroneal nerve during total hip arthroplasty and found a 20% incidence of intraoper- ative signal changes. Changes in SSEPs have been noted with retrac- tor placement, leg positioning for femoral reaming and cement re- moval, anterior or lateral retraction of the femur, and hip reduction. 29,30 In a nonrandomized, unmonitored control group, 31 2 of 35 patients (6%) had postoperative incomplete sciatic nerve palsy, while none of the 25 patients in the monitored group had neurologic compromise. Intraoperative monitoring was con- sidered to be a valuable method for use in revisions and reoperations. Black et al 26 found no reduction in sciatic nerve palsy in monitored patients compared with an unmon- itored historical control group from the same institution. They felt that monitoring may be more appropri- ate in selected high-risk groups. In a follow-up study, Rasmussen et al 29 found no difference in the inci- dence of sciatic nerve injury be- tween 290 monitored patients and a historical control group of 450 unmonitored patients (2.8% vs 2.7%). When they compared only revision cases, no statistically sig- nificant difference between groups was found (6.7% vs 5.3%). They concluded that SSEP monitoring was not effective in predicting or preventing nerve injuries. They also reported that 2 patients who were found to have postoperative palsies had no SSEP changes dur- ing the procedure. Another method for monitoring nerve function is to record EMG responses. Intraoperative electro- myography has been used to moni- tor nerve function during opera- tions that place the recurrent laryn- geal nerve, facial nerve, spinal nerves, and sciatic nerves at risk. Electromyographic responses can be recorded either as an averaged motor evoked potential or as the individual contraction of a muscle. If an averaged motor evoked poten- tial is recorded, either the spinal cord or the nerve must be stimulat- ed proximal to the site of surgery. After changes through the site of surgery, the elicited neurologic activity is recorded as an EMG trac- ing from a peripheral muscle. Multiple stimulations are made, and the muscle responses are aver- aged by a computer. Yet another technique is the mechanically elicited, or sponta- neous, electromyogram. Mechan- ical irritation of the nerve results in an action potential that produces a muscle contraction measured by an EMG response. The EMG response elicited by mechanical irritation provides the surgeon immediate real-time feedback of exploration. Muscle relaxation must be kept to a minimum of two twitches of a train-of-four stimulation for EMG recording to be possible. Because the site of hip surgery and the site of calf muscle recording Fig. 11 Example of EMG tracings during hip arthroplasty. Baseline is on the left. Repetitive firing of anterior tibialis muscle is represent- ed during compression of the peroneal division of the sciatic nerve. LT ANT TIB LT GASTROC Nerve Injuries in Total Hip Arthroplasty Journal of the American Academy of Orthopaedic Surgeons 110 are both peripheral, anesthetic effects on the brain and spinal cord do not interfere with the perfor- mance of mechanically elicited elec- tromyography. This allows use of spinal or epidural anesthesia with- out signal interference. Sutherland et al 32 used spontaneous electro- myography in 44 consecutive revi- sion and complex hip arthroplasty procedures. In 5, intraoperative EMG activity resolved with retrac- tor or limb adjustments. None of the patients in that study had post- operative nerve deficits. Intraoperative electromyography may be a helpful adjunct in the pre- vention of nerve palsy in high-risk patients. However, larger prospec- tive trials are necessary to demon- strate a reduction in the overall rate of sciatic nerve palsy between mon- itored and unmonitored patients. Summary Preservation of neurologic struc- tures is important to maintain high levels of limb function and patient satisfaction after hip arthroplasty. Fortunately, nerve injury in THA is an uncommon complication, occur- ring in only 1% to 2% of patients who undergo primary hip arthro- plasty. The peroneal division of the sci- atic nerve is the nerve most fre- quently injured during revision cases and in the treatment of demanding cases of hip dysplasia. There is a trend in the literature that supports an increase in sciatic nerve injuries when the leg is lengthened by more than 4 cm or by more than 10% of the length of the femur. Postoperative footdrop or weakness of ankle dorsiflexors results from peroneal division palsy and causes a steppage gait. Often, an ankle- foot orthosis is all that a patient requires to manage the deficit. Complete loss of neurologic function or severe causalgic pain carries the worst prognosis. The role of electrodiagnostic studies intraoperatively requires further study before recommendations for routine use can be made. The importance of prevention is best summarized by Schmalzried et al, 11 who stated, ÒNo amount of preop- erative discussion or postoperative consultation decreased the high degree of dissatisfaction that was expressed by these patients.Ó Acknowledgments:The authors would like to thank Taryn Tuinstra, Kym Palatto, PAC, and Jeffery H. Owen, PhD, for their valuable assistance in preparing this manu- script. References 1.Nercessian OA, Macaulay W, Stinch- field FE: Peripheral neuropathies fol- lowing total hip arthroplasty. J Arthroplasty1994;9:645-651. 2.Bodine SC, Lieber RL: Peripheral nerve physiology, anatomy, and path- ology, in Simon SR (ed): Orthopaedic Basic Science. Rosemont, Ill: American Academy of Orthopaedic Surgeons, 1994, pp 325-396. 3.Nuwer MR, Schmalzried TP: Nerve palsy: Etiology, prognosis, and pre- vention, in Amstutz HC (ed): Hip Arthroplasty. New York: Churchill Livingstone, 1991, pp 415-427. 4.Lundborg G: Structure and function of the intraneural microvessels as related to trauma, edema formation, and nerve function. J Bone Joint Surg Am1975;57:938-948. 5.Lundborg G, Rydevik B: Effects of stretching the tibial nerve of the rabbit: A preliminary study of the intraneural circulation and the barrier function of the perineurium. J Bone Joint Surg Br 1973;55:390-401. 6.Jacobs LG, Buxton RA: The course of the superior gluteal nerve in the lateral approach to the hip. J Bone Joint Surg Am1989;71:1239-1243. 7.Abitbol JJ, Gendron D, Laurin CA, Beaulieu MA: Gluteal nerve damage following total hip arthroplasty: A prospective analysis. J Arthroplasty 1990;5:319-322. 8.Comstock C, Imrie S, Goodman SB: A clinical and radiographic study of the Òsafe areaÓ using the direct lateral approach for total hip arthroplasty. J Arthroplasty1994;9:527-531. 9.Weber ER, Daube JR, Coventry MB: Peripheral neuropathies associated with total hip arthroplasty. J Bone Joint Surg Am1976;58:66-69. 10.Siliski JM, Scott RD: Obturator-nerve palsy resulting from intrapelvic extru- sion of cement during total hip re- placement: Report of four cases. J Bone Joint Surg Am1985;67:1225-1228. 11.Schmalzried TP, Amstutz HC, Dorey FJ: Nerve palsy associated with total hip replacement: Risk factors and prognosis. J Bone Joint Surg Am1991; 73:1074-1080. 12.Simmons C Jr, Izant TH, Rothman RH, Booth RE Jr, Balderston RA: Femoral neuropathy following total hip arthro- plasty: Anatomic study, case reports, and literature review. J Arthroplasty 1991;6(suppl):S57-S66. 13.Johanson NA, Pellicci PM, Tsairis P, Salvati EA: Nerve injury in total hip arthroplasty. Clin Orthop1983;179: 214-222. 14.Pritchett JW: Lumbar decompression to treat foot drop after hip arthroplas- ty. Clin Orthop1994;303:173-177. 15.Edwards BN, Tullos HS, Noble PC: Contributory factors and etiology of sciatic nerve palsy in total hip arthro- plasty. Clin Orthop1987;218:136-141. 16.Hanssen AD, Mariani EM, Kavanagh BF, Coventry MB: Resection arthroplas- ty (Girdlestone procedure), nerve palsies, limb length inequality, and osteolysis following total hip arthro- plasty, in Morrey BF (ed): Joint Re- placement Arthroplasty. New York: Churchill Livingstone, 1991, pp 897-905. 17.Kennedy WF, Byrne TF, Majid HA, Pavlak LL: Sciatic nerve monitoring during revision total hip arthroplasty. Clin Orthop1991;264:223-227. 18.Nercessian OA, Piccoluga F, Eftekhar NS: Postoperative sciatic and femoral nerve palsy with reference to leg lengthening and medialization/later- alization of the hip joint following total hip arthroplasty. Clin Orthop 1994;304:165-171. 19.Weale AE, Newman P, Ferguson IT, Bannister GC: Nerve injury after pos- terior and direct lateral approaches for