Clinical spectrum of neuropathy after primary total knee arthroplasty: a series of 54 cases

Abstract

Introduction:

Neuropathy after total knee arthroplasty (TKA) can cause significant morbidity but is inconsistently reported.

Methods:

We reviewed the clinical, electrodiagnostic and perioperative features of all patients who underwent primary TKA at our institution and developed a new neuropathy within 8 weeks postoperatively.

Results:

Fifty-four cases were identified [incidence 0.37% (95% CI 0.28-0.49)] affecting the following nerve(s): peroneal (37), sciatic (11), ulnar (2), tibial(2), sural (1), and lumbosacral plexus (1). In all cases with follow-up data, motor recovery typically occurred within 1 year and was complete or near-complete.

Discussion:

Post-TKA neuropathy is uncommon, typically does not require intervention and usually resolves within 1 year. Post-TKA neuropathy most often affects the nerves surgically at risk. Anesthesia type does not correlate with post-TKA neuropathy. An inflammatory etiology for post-TKA neuropathy is rare but should be considered in specific cases.

Introduction

Total knee arthroplasty (TKA) is among the most common orthopedic surgeries performed in the United States with 700,000 procedures performed annually, projected to reach 3.48 million by 2030¹. The neurologic impact of this volume is important as surgery has long been identified as a potential precipitant of new neuropathies. Most often, new neuropathy after surgery is attributed to mechanical forces in the intraoperative setting (traction, compression or transection) as well as positioning and anesthesia² with recent reports suggesting female gender and prior lumbar disease as potential risk factors for post-TKA neuropathy³. Still other studies^{4, 5} have shown that in rare cases, pathologically proven immune-mediated inflammation is the cause for the development of new postoperative neurologic deficit, and can occur in the same limb as operative site of orthopedic procedures. Given these potential explanations for new neuropathy after surgery and the high frequency of TKA surgeries, we sought to 1) identify the frequency with which post-TKA neuropathies occur, 2) clarify the clinical and electrophysiological features of post-TKA neuropathy, and 3) answer whether both mechanical and inflammatory etiologies are of concern.

Methods

This retrospective study was approved by the Mayo Clinic Institutional Review Board. The Advanced Cohort Explorer (ACE) program developed at Mayo Clinic was used to perform a retrospective review of the electronic medical record. All patients 18 years and older who underwent TKA at Mayo Clinic Rochester between January 1^st, 1996 and September 30^th, 2016 who also had a possible neuropathy within 8 weeks of surgery were identified. Neuropathy was ascertained via ICD9 or ICD10 diagnosis codes chosen to capture the spectrum of lower extremity neuropathies (peroneal, tibial, sural, sciatic, femoral, obturator, lumbosacral neuropathy). Patients were included regardless of anesthesia type. Patients with definitive pre-existing neuropathy, active radiculopathy, or central nervous system processes which limited reliability of the physical examination were excluded. We performed further comprehensive chart review of 471 cases meeting these criteria including review of anesthesia, clinical, electrodiagnostic, and radiologic findings. We also extensively reviewed tourniquet time and the time to motor recovery as a marker for neuropathy severity. Based on prior studies which demonstrated that a tourniquet time of greater than 100 minutes was associated with an increased risk for complications⁶; we used this value as our cutoff in our study.

Results

During the interval reviewed, there were 14,450 TKAs. Fifty-four cases of new neuropathy (53 patients) following TKA were identified resulting in a neuropathy incidence of 0.37% (95% CI 0.28-0.49). The mean age of post-TKA neuropathy patients was 65.2 years (range 32-86 years); 41 patients were female (Table 1). Nearly all of the post-TKA neuropathies were ipsilateral lower limb mononeuropathies. One patient was diagnosed with both peroneal and tibial mononeuropathies. Mean time to neuropathic symptom onset was postoperative day (POD) 2 (range 0-28 days). Patients were clinically followed for a mean of 41.9 months (range 2-136 months). Mean time to complete or near complete motor recovery was 10.1 months (range 1-30 months) with 4 patients lost to follow-up and only 1 patient without any recovery of motor function (Table 2, Supplementary Table 1).

Table 1.

Demographic variables and medical history

	Total
Number of Patients	53
Mean Age (years)	65.2
Female (%)	41 (77.4%)
Male (%)	12 (22.6%)
Family History of Neuropathy	4 (7.5%)
History of Diabetes	7 (13.2%)
History of Autoimmunity*	6 (11.3%)
Prior Surgery without Neuropathic Complication	37 (69.8%)
Pre-surgical Valgus Deformity	24 (45.3%)

^*Diagnoses of rheumatoid arthritis, systemic lupus erythematosus, and Grave’s disease

Table 2.

Summary of Post-operative Neuropathies and Clinical Course

Neuropathy	# of cases (%)	EMG Referral	Neurology Consult	Pre-operative Valgus Deformity	Motor Weakness	Complete/ Near Complete Motor Recovery
Peroneal	37 (68.5%)	10	8	21	34	23/7*
Sciatic	11 (20.4%)	10	9	1	9	2/6
Tibial	2 (3.7%)	1	1	0	0	0
Ulnar	2 (3.7%)	1	0	1	0	0
Sural	1 (1.9%)	0	0	0	0	0
Lumbosacral Plexopathy	1 (1.9%)	1	1	1	1	near complete
Totals	54	23 (42.6%)	19 (35.2%)	24 (44.4%)	44 (81.5%)	25/14*

^*4 patients lost to follow-up

Peroneal neuropathy presenting as a postoperative foot drop, often while still hospitalized, was the most common presentation of post-TKA neuropathy (Table 2). Given the relatively recognizable clinical picture of peroneal neuropathy after knee arthroplasty, only a minority of patients were referred for further evaluation with electromyography (EMG) and the rest observed clinically. Of those that underwent electrodiagnostic testing, non-focal peroneal neuropathies were identified in 8 of 10 patients; 1 study demonstrated both peroneal and tibial mononeuropathies. Needle examination was deferred by the remaining patient, which limited localization. In 2 of 37 cases of peroneal neuropathy, ultrasonography identified hematomas causing compression of the nerve. One patient had rapid improvement following hematoma evacuation with motor recovery at 1 month follow-up while the second case went on to recover over 10 months. In 21 of 37 patients with post-operative peroneal neuropathy, a preoperative valgus deformity was present. Three patients with peroneal neuropathy underwent peroneal nerve release postoperatively. One patient was lost to follow-up and the remaining 2 patients recovered motor function in 3 and 8 months respectively.

Sciatic neuropathy was the second most common neuropathy clinically identified following TKA. Additional evaluation was more readily pursued in these sciatic neuropathy cases with nearly all patients being referred for EMG. Four sciatic neuropathies localized proximal to the short head of the biceps femoris, including 1 neuropathy proximal to the long head of the biceps femoris. Three additional sciatic neuropathies were localized distally. Two studies were limited to nerve conduction studies only and 1 study was normal. One patient with proximal localization underwent MRI of the lumbosacral plexus which showed asymmetry of the sciatic nerves at the level of the ischial tuberosity with noted enlargement and T2 hyperintensity of the affected sciatic nerve.

Tibial neuropathies were also identified, 1 of which also had a peroneal neuropathy by EMG previously noted above. Both patients presented with paresthesias and/or hyperesthesia in the sole of the foot or toes, along with reduced ankle reflexes and mild inversion weakness. One case of sural neuropathy was also clinically identified.

We also discovered new cases of ulnar neuropathy in the post-operative period. One of the patients was referred for EMG which demonstrated focal ulnar neuropathy at the elbow and subsequently underwent transposition. The other patient had resolution of symptoms at 8 weeks postoperatively without intervention.

A single case of lumbosacral plexopathy was noted in a 67 year-old female patient. She developed progressive neuropathic pain and diffuse weakness involving the left iliopsoas, quadriceps, peronei and posterior tibialis. Due to continued severe pain refractory to opiates, intravenous methylprednisolone at was commenced at 11 weeks postoperatively. The patient noted significant improvement in pain with this regimen over the treatment period and was able to taper off opiates completely.

Our review of tourniquet time revealed the mean time to motor recovery was 11.8 months ±7.7 (95% CI 7.9-15.7) for patients with tourniquet time greater than 100 minutes and 8.1 months ±7.3 (95% CI 5.1-11.) for patients with tourniquet time less than 100 minutes. Due to a large standard deviation in each cohort, the difference in recovery time was not significantly different [t(36)=1.5, p=0.14].

Discussion

Our incidence of 0.37% of patients undergoing TKA (95% CI 0.28-0.49) is in line with prior studies looking specifically at peroneal neuropathy^{7, 8}. Others have also examined the neuropathic sequelae of TKA specifically looking at mechanical and electrocautery nerve damage⁹, risk factors related to female gender and prior lumbar pathology¹⁰, malpractice claims resulting from such injury¹¹ and reporting sacral plexopathy and sciatic neuropathy following total knee arthroplasty¹², but there has not been a study aimed at analyzing the full spectrum of post-TKA neuropathy. We confirmed that the peripheral nerves at risk in the periarticular region (peroneal more than tibial and distal sciatic nerves), were the most commonly affected. As a result, we suspect that intraoperative mechanical forces such as traction and compression are the most likely etiology of these neuropathies. The large number of valgus knees in the study, confirm that traction at the time of correction of a fixed deformity is a relevant risk factor. Importantly, our data suggests that patients with new, postoperative ipsilateral neuropathy after TKA typically have a monophasic course with complete (n=25; 64.1%) or near complete (n=14; 35.9%) motor recovery usually seen within 1 year from surgery and do not usually require surgical intervention.

Although we do not have pathologic confirmation of inflammation in our 1 case of suspected post-surgical inflammatory neuropathy following TKA, this case met previously proposed criteria for this entity: the affected nerves, predominantly the lumbar plexus, were distant from the operative site, the patient experienced significant neuropathic pain, symptoms progressed for an extended period following surgery and there was marked response to intravenous methylprednisolone after failing to improve with conservative measures⁵. This case highlights that in patients with this clinical phenotype, an inflammatory etiology for new neuropathy following TKA should be considered as these patients may benefit from immunotherapy.

An important clinical feature is the valgus deformity associated with TKA. The greater the angle of deformity (i.e. greater than 20 degrees) and presence of concomitant flexion contracture, the higher the risk for a post-TKA peroneal neuropathy^{13, 14}. During surgery the knee is brought from a fixed flexion and valgus deformity to an anatomic position, achieving full extension and correcting the valgus deformity. This puts the lateral aspect of the knee under stretch and if the peroneal nerve has been compressed due to scar tissue because of the progressive long-standing deformity this could lead to peroneal palsy after surgery. Recent literature has suggested decompression of the peroneal nerve in the immediate post-operative period may be of benefit^15-17. Although 3 of our patients underwent postoperative peroneal nerve decompression, none had pre-operative valgus deformities. A focused future study of this group of patients is warranted.

As would be expected for the standard of practice during the study period, this cohort received peripheral nerve blockade, unless contraindicated by pre-existing neuropathy or anticoagulation therapy, for postoperative analgesia following TKA. The peripheral nerve blockade consisted of both a lumbar plexus catheter (e.g., iliopsoas or femoral nerves) with, or if the surgeon preferred, without a single-injection proximal sciatic nerve blockade for postoperative analgesia. Four patients had injuries in a similar anatomic distribution to the peripheral nerve block (e.g., proximal sciatic nerve injury after proximal sciatic nerve blockade). Given the proximal nature of the nerve block (greater sciatic notch), EMG is limited in its ability to localize to this location. We had 1 case of sciatic neuropathy that convincingly localized proximal to the long head of the biceps by EMG, with the remainder of cases being more distal and unlikely block-related. Eighteen of our cases had neurologic injury in a distribution that might be related to the peripheral nerve block (e.g., isolated peroneal nerve injury after proximal sciatic nerve blockade), but these cases were not convincing for a causal relationship. Thirty-one patients had neurologic deficit in a distribution unrelated to the peripheral nerve block (e.g., proximal or distal sciatic nerve injury after femoral nerve blockade) including 1 patient not receiving a peripheral nerve block at all. Thus, our findings are in line with previous work suggesting that peripheral nerve blockade is not correlated with increased risk for peripheral nerve injury^18-20. We found no significant difference in time to motor recovery/neuropathy severity based on tourniquet time.

Limitations of our review include its retrospective design limiting the full scope of neuropathies identified and ability to identify all mild neuropathies following surgery that may have not come to clinical attention or that resolved quickly after surgery. Some limitations of the case ascertainment software are seen by the identification of 2 ulnar neuropathies despite ICD codes limited to lower extremity neuropathies. Given the inclusion of these cases as a result of software limitations, our study may underestimate the number of upper extremity neuropathies seen in a post-TKA population. Secondly, our 1 patient with a suspected post-surgical inflammatory neuropathy did not have pathologic confirmation. Finally, while all cases demonstrated complete or near complete motor recovery, persistent sensory symptoms may have been present but not clearly documented given the more robust focus on motor deficits by the examining physicians. A case control study evaluating the medical comorbidities, peri-operative factors and post-operative course of patients who develop neuropathy following TKA would be beneficial in further understanding this disease entity.

Conclusion

This systematic study demonstrates that new neuropathy following total knee arthroplasty is rare, with an incidence of 0.37% in patients undergoing TKA. When it occurs, it most commonly presents as a mononeuropathy affecting the nerves most likely to be at risk in the periarticular region, but not clearly related to regional anesthesia or tourniquet time. Although these patients may have marked weakness initially, they typically have a monophasic course with complete or near complete recovery within 1 year from surgery without intervention. Additionally, although very uncommon, a post-surgical inflammatory neuropathy is possible and these patients may benefit from further testing and treatment.

Presented as “Clinical spectrum of neuropathy after total knee arthroplasty: a series of 54 cases” at American Association of Neuromuscular and Electrodiagnostic Medicine Annual Meeting. Scottsdale, AZ, September 12–16, 2017

Abbreviations

TKA

Total knee arthroplasty

EMG

Electromyography

ACE

Advanced Cohort Explorer