NERVE injury is a well-recognized complication of peripheral nerve blocks. Capdevila et al. 1reported an incidence of 0.21% in a large prospective multicenter study including more than 1,400 patients with perineural catheters. Beside factors related to the regional anesthesia technique,1other causes such as concomitant patient disease,2inadequate positioning or manipulation during surgery,3and the neurotoxicity of local anesthetics4,5have been implicated.
We report on a patient who developed a femoral neuropathy after continuous femoral block. The only risk factor found in this patient was a postoperatively discovered, preexisting subclinical polyneuropathy.
A 72-yr-old man (American Society of Anesthesiologists physical status II) was scheduled to undergo left total knee arthroplasty. His medical history was unremarkable. All laboratory results were in normal range. The anesthetic procedure included continuous femoral nerve block, single-shot sciatic nerve block, and spinal anesthesia. The patient was orally premedicated with 7.5 mg midazolam. In the preanesthetic room, he was monitored and venous access was established.
For femoral catheter placement, the patient was placed supine. The puncture point was located 2 cm distally of the inguinal ligament and 1.5 cm lateral of the femoral artery. After disinfection and local anesthesia of the skin, a 30° short beveled 21-gauge needle (Stimuplex A; Braun, Melsungen, Germany) connected to a nerve stimulator (Stimuplex HNS 11; Braun Melsungen AG) with initial settings of 1.4-mA current intensity, 0.1-ms impulse duration, and 2-Hz frequency was introduced. The needle was cranially directed with 45° to the skin. After a “dancing patella” response was elicited at a depth of 5 cm with 0.35 mA, a 23-gauge catheter (Polymedic; Te Me Na, Bondi, France) was placed with the cannula-over-needle technique. The catheter was secured by subcutaneous tunneling and fixed to the skin with adhesive tape. After negative aspiration, 30 ml ropivacaine, 0.5%, was administered slowly through the catheter. Nerve localization, catheter placement, and local anesthetic injection were performed without occurrence of pain or paresthesias/dysesthesias at any time.
The sciatic nerve block was performed using a proximal, lateral approach at the thigh. After localization of the puncture point 2 cm distal of the greater trochanter, a 22-gauge short beveled needle (Polymedic) was connected to the nerve stimulator and introduced parallel to the table. Needle placement was considered successful when plantar flexion could be elicited with current intensity of 0.4 mA and 0.1-ms impulse duration. After careful negative aspiration, 20 ml ropivacaine, 0.5%, was slowly injected.
The blocks were tested and considered successful after loss of cold perception in the distribution areas of the femoral, tibial, and peroneal nerves within 15 min after local anesthetic application.
Spinal anesthesia was performed in lateral decubitus position. After skin infiltration with 1% lidocaine, a 27-gauge spinal needle (Whitacre; Becton Dickinson AG, Basel, Switzerland) was introduced at the L4–L5 interspace by the paramedian approach. Correct needle placement was identified by free flow of cerebrospinal fluid, and 2.5 ml hyperbaric bupivacaine, 0.5%, was slowly injected. The spinal needle was removed, and the patient was placed supine. The upper level of the sensory block was determined bilaterally using loss of sensation to cold at the T4 level. The intraoperative course was uneventful. During cementation of the knee prosthesis, a thigh tourniquet was inflated for 23 min with a pressure of 280 mmHg. The surgical procedure was uneventful and lasted 150 min. Postoperative analgesia was provided with 0.2% ropivacaine given continuously through the femoral nerve catheter at a rate of 10 ml/h. One gram acetaminophen was administered every 6 h. Two hundred twenty minutes after intrathecal application of bupivacaine, complete recovery of sensibility in the right leg was documented. Postoperative analgesia in the first 48 h was efficient, and no additional narcotics were necessary. The degree of motor block was not specifically documented, but no adverse or unusual events were noted.
The continuous ropivacaine infusion was stopped 48 h after the start, and the femoral catheter was removed 6 h later. At that time, a persistent weakness of the left quadriceps was noted. Neurologic examination disclosed hyposensitivity in the medial aspect of the left thigh restricted to the L3 dermatome, and absence of patellar tendon reflex was observed. Ultrasonography in the area of the puncture point of the femoral catheter excluded the presence of a hematoma. Magnetic resonance imaging of the lumbar spine revealed neither a lesion of the nerve roots nor hematoma at the spinal anesthesia puncture site.
On the fourth postoperative day, an electrophysiologic investigation was performed (fig. 1). The needle electromyography showed low-voltage denervation activities, presenting as sharp waves and fibrillation potentials in the medial vastus muscles bilaterally. In the left and right anterior tibial muscles, signs of neurologic degeneration were also found. The neurography revealed normal conduction velocity of the right tibial nerve (43 m/s) and marginal decreased conduction velocity of the left tibial nerve (39 m/s). Latencies of the tibial somatosensory evoked potentials were in normal range on both sides.
On the ninth postoperative day, the patient was discharged with persistent weakness of the left quadriceps muscle, hyposensitivity of the medial aspect of the thigh, and abolished patellar tendon reflex.
After 3 months, moderate weakness of the left quadriceps muscle was still present, but the patellar tendon reflex could be triggered on both sides. A new electromyography was performed and revealed high-voltage dense-frequent denervation (fibrillations and sharp waves) within the left quadriceps muscle (fig. 2). Both tibial anterior muscles showed the same low-voltage denervation as before. After 6 months, complete sensorimotor recovery of the left quadriceps muscle was noted.
This case showed the occurrence of prolonged sensorimotor deficit of the femoral nerve after continuous femoral nerve anesthesia/analgesia in a patient with a preoperatively undetected polyneuropathy. None of the usual predisposing factors for the development of a neuropathy was obvious in this patient.
Postoperative neurophysiological studies were performed to investigate the left-sided weakness of the quadriceps muscle, the hyposensitivity in the medial aspect of the thigh, and the abolished patellar tendon reflex. The initial investigation showed signs of denervation in the medial vastus muscle, including low-voltage fibrillations potentials and positive sharp waves. These findings are consistent with a preexisting neuropathy and could not be explained by an acute lesion, because it usually takes 3–4 weeks after nerve damage for these signs to be detected.6Moreover, the presence of a preexisting neuropathy is supported by similar pathologic needle electromyographic findings on the nonoperated contralateral limb. The second examination, performed 6 weeks after surgery, revealed a partial lesion of the femoral nerve with high-voltage fibrillations and sharp waves, representing signs of acute denervation. No needle electromyographic change was observed on the other side. The second needle electromyography is consistent with worsening of the preexisting neuropathy after continuous application of local anesthetics, which is supported by the absence of new needle electromyographic modifications on the other side.
Little is known about the impact of continuously applied local anesthetics on predamaged nerves. This could possibly include risks for nerve lesions such as mechanical trauma, neuronal ischemia, or local anesthetic neurotoxicity.4Neurologic damage after local anesthetic block is explained in most cases by two or more of these factors.7A definite mechanism can rarely be stated. In our patient, a direct nerve trauma by the needle or intraneural injection seems unlikely because no pain or paresthesias/dysesthesias occurred during the block or the catheter placement, and no unusual resistance during local anesthetic injection was encountered.7Magnetic resonance imaging of the lumbar spine excluded the presence of a hematoma compressing the femoral nerve root. Nerve trauma due to other external factors such as surgical retractors was unlikely because the surgical site was not in the proximity of the femoral nerve. The thigh tourniquet was used for only 23 min with 280 mmHg and therefore could hardly be responsible for the damage.
It is conceivable that the application of local anesthetics on preinjured nerves may predispose these nerves to further damage. Preexisting pathology has been reported to play a role in the development of postoperative nerve injury.6,8,9Alvine and Schurrer8performed bilateral nerve conduction studies in patients in whom perioperative ulnar neuropathy occurred. They found abnormal slowing of the nerve conduction on the contralateral side despite the absence of clinical signs or symptoms of neuropathy. They suggested that subclinical neuropathy may become symptomatic after perioperative manipulations, such as traction and stretching.
Local anesthetics are potentially neurotoxic,4,5and the neuronal blockade can be seen as a reversible expression of this. The neurotoxicity parallels the anesthetic potency and has been identified to be dose and concentration dependent.5
Radwan et al. 10showed that ropivacaine leads to growth cone collapse in chick embryo dorsal root ganglion cells in a concentration below the common clinical concentrations used for regional anesthesia. The author concluded that local anesthetic application may interfere with the growth and regeneration of neuronal tissue and that the detrimental effects of local anesthetics on growing or regenerating neurons should be considered.10According to this statement, although 0.5% ropivacaine is usually not neurotoxic, the underlying neuropathy of our patient may have contributed to make the femoral nerve more susceptible to the potential neurotoxic effect of continuous local anesthetic infusion.
As a clinical implication, the occurrence of unexplained neuropathy after a continuous perineural catheter should warrant for the search of subclinical undetected preexisting neuropathy. The recognition of preexisting subclinical neuropathy is challenging. Few data are available on this subject, and future prospective studies looking at this issue will be needed to evaluate and anticipate the possible implications of its occurrence on the practice of regional anesthesia. Its presence will be helpful for the understanding of the complication and may also have medicolegal implications.