Incidence and Etiology of Rare Saphenous Nerve Injury After Knee Surgery with Adductor Canal Block: A Retrospective Observational Study

Abstract

Background:

Adductor canal block (ACB) is commonly included in multimodal analgesia regimens for knee surgery. Nonetheless, the incidence, etiology, and procedure-specific risk of saphenous nerve injury after knee surgery with ACB have not been established.

Purpose:

We sought to identify the risk of saphenous nerve injury during knee surgery with ACB.

Methods:

We conducted a retrospective cohort study of patients at a single institution who underwent elective knee surgery with ultrasound-guided ACB between January 1, 2014, and December 31, 2018, and had subsequent saphenous nerve injury. The primary outcome was the incidence of saphenous nerve injury within 3 months of surgery, by surgical type and approach. Secondary outcomes included attribution of the most likely etiology and clinical outcome of the injury.

Results:

In 28,196 cases of knee surgery with ACB, we identified 18 cases (0.06%) of saphenous nerve injury. The most common surgery associated with saphenous nerve injury was anterior cruciate ligament (ACL) reconstruction with autograft (8/18 cases); 3 cases of injury were seen after TKA, 2 after medial patellofemoral ligament reconstruction, 2 after arthroscopy/meniscal surgery, and 1 after patellar fixation. The most likely etiology of nerve injury was attributed to ACB in 5 of 18 cases (28%) and to non-ACB cause in 13 of 18 (72%). Prognosis was rated as unknown in 11 of 18, poor in 2 of 18, favorable in 3 of 18, and full recovery in 2 of 18.

Conclusions:

This 5-year retrospective, single-institution cohort study found a low overall incidence of saphenous nerve injury after knee surgery with ACB, but the injury likelihood varied based on surgery and approach. Although not statistically significant, ACL reconstruction with hamstring autograft and ACB performed for postoperative rescue analgesia were most frequently associated with nerve injury.

Introduction

Adductor canal block (ACB) is frequently included as a component of multimodal analgesia for surgery of the knee. The value of ACBs has been suggested in major orthopedic procedures, including total knee arthroplasty (TKA) and anterior cruciate ligament (ACL) reconstruction, where benefits include effective, opioid-sparing analgesia and lower risk of falls compared with femoral nerve blocks [7,13]. In contrast, a recent meta-analysis concluded that ACB may not offer analgesic benefit over placebo for ambulatory and arthroscopic knee procedures [21], although calls for more high quality studies have been made [15]. Given these data, an evaluation of procedure-specific risks of ACB for knee surgery should be weighed against potential benefits.

The incidence of saphenous nerve injury after knee surgery specifically with ACB has not yet been established, hampering the ability to inform patients of their risk, and to guide decision-making whether to perform the block on specific patients presenting for specific surgeries. The etiology of nerve injury after surgery with regional techniques is likely multifactorial, including anesthetic factors (mechanical or chemical trauma), surgical factors (incision or tissue retraction), and other contributors (thigh tourniquet use) [17]. Any or all of these risks can be further exacerbated by patient factors and comorbidities including diabetes, obesity, or peripheral vascular disease [6,11,12,17].

A recent retrospective analysis of the incidence of new postoperative neurologic symptoms (PONS) found 20 cases in 26,251 surgeries performed with peripheral nerve blocks (PNBs): 5 were associated with ACB for TKA, but none of these were attributed to the block [11]. These data are provocative and raise the question whether there are procedure-specific risks associated with performing ACB, and how the most likely etiology of nerve injury should factor into the decision to perform a regional technique. Furthermore, the incidence of PONS may not translate into permanent nerve injury, and the latter remains to be determined after knee surgery. These questions are particularly relevant considering that ACB is recommended in the ambulatory setting and as part of enhanced recovery protocols to facilitate rapid recovery and discharge [1,24].

This study’s objective was to establish the incidence, most likely etiology, and prognosis of clinically important saphenous nerve injury after knee surgery with ACB. In addition, we analyzed the incidence of injury by surgical type, and hypothesized that risk of saphenous nerve injury would differ based on surgical type and approach.

Methods

This was a single-center, retrospective observational study. The study was conducted in accordance with the ethical principles of the Helsinki declaration and approved by the institutional review board of Hospital for Special Surgery. Written informed consent was waived.

Records of all patients who underwent any knee surgery with ultrasound-guided ACB between January 1, 2014, and December 31, 2018, were reviewed to identify cases of saphenous nerve injury. The total number and type of knee surgery performed with ACB were identified by an internal billing database (using CPT codes 01400, 01402, 27310, 27403, G0289, and 64447). Cases of saphenous nerve injury were identified by searching 2 institutional databases: a Quality Assurance (QA) registry, and a radiology database. For both databases, search terms included “knee,” “saphenous nerve,” “adductor canal,” “injury,” “neuropraxia,” and “neuritis.” The radiology database was queried for any magnetic resonance imaging (MRI) report that included any search term of interest. Where ambiguous, patient electronic medical records (EMRs) were additionally queried for any documented diagnosis or description of saphenous nerve injury. Cases of saphenous nerve injury were defined as meeting the following criteria: the saphenous nerve injury was reported within 3 months of surgery on the ipsilateral knee; an ACB was provided as part of the analgesic regimen, either preincision or in the postanesthesia care unit (PACU); and a postoperative MRI (report and images) was available. Cases with the following criteria were excluded from analysis: surgery outside of the date range, no knee surgery performed, no ACB performed, no saphenous nerve injury, and saphenous nerve injury documented more than 3 months after surgery.

Each case of saphenous nerve injury was then evaluated by 2 independent attending anesthesiologists to determine the most likely etiology of the injury. The major decision was whether the injury was more likely to be secondary to ACB or to another cause. An a priori set of criteria was established by the study team to guide case review and decision-making: nerve injuries proximal to the knee or within the adductor canal were deemed more likely to be secondary to ACB. Injuries localized to the infrapatellar branches or associated with fat pad scarring were more likely secondary to other factors. In cases of disagreement, a third reviewer (an attending neurologist) independently decided the most likely cause.

Data were retrospectively extracted from the EMR and included patient demographics (age, sex, body mass index [BMI]), American Society of Anesthesiologists [ASA] class, presence and degree of valgus deformity, spinal stenosis, diabetes mellitus, peripheral neuropathy, surgical type, anesthetic type, timing of ACB (preincision or in the postanesthesia care unit [PACU]), surgical duration, and use and duration of an intraoperative thigh tourniquet.

The primary outcome was the incidence of saphenous nerve injury after any knee surgery with ACB and the incidence of injury according to surgical type (total knee arthroplasty, ACL or medial patellofemoral ligament (MPFL) reconstruction, other surgeries performed via arthroscopy or arthrotomy and trauma/open reduction internal fixation). Secondary outcomes included the most likely etiology (secondary to ACB or non-ACB), the presence of patient and perioperative risk factors for postsurgical neuropraxia, and outcome of injury (unknown, poor, favorable, full recovery). Outcome of injury was decided after review of the EMR and rated “unknown” where notes were incomplete or otherwise lacked sufficient information to permit conclusion; “poor” prognosis was concluded by documentation of ongoing symptoms without improvement and/or progression to postsurgical chronic regional pain syndrome (CRPS), according to diagnostic criteria (pain disproportionate to the inciting event, the presence of symptoms and signs of sensory, vasomotor, sudomotor and motor/trophic changes) [23]. Favorable recovery was concluded by qualitative description of improvement in symptoms and “full recovery” where complete resolution of symptoms was documented. Follow-up of patient outcomes was until 2 years postoperatively unless the patient had complete recovery documented sooner in EMR.

Statistical Analysis

Results are presented descriptively, with categorical data represented as counts and percentages (%) and continuous data as mean (M) and standard deviation (SD) or median (MED) and interquartile range (IQR). Baseline characteristics and outcome variables among patients where nerve injury was attributed to ACB versus non-ACB etiology were summarized and compared using Mann-Whitney U test for continuous variables and Fisher’s exact tests for categorical variables, respectively. Significance was set at P < .05.

Results

After review of the internal billing database, we identified 28,196 knee surgeries with ACB over the 5-year interval of interest. Of these, 20,351 were TKA, 4322 were ACL, Posterior cruciate ligament, or MPFL reconstruction or open reduction/internal fixation, and 3523 were knee arthroscopy (meniscectomy; debridement; removal of loose body).

We identified 1426 cases of potential saphenous nerve injury through search of the radiology database, and 69 cases of saphenous nerve injury through search of the hospital’s QA database (Supplemental Figure 1). After removing duplicate cases found in both databases, a total of 1426 cases of potential saphenous nerve injury were screened for inclusion criteria (ie, all 69 cases found in the QA database were also found in the radiology database). Of these, 1348 patients were excluded due to patients not having surgery (n = 1206) or surgery occurring prior to 2014 (n = 142). Each of the remaining 78 patient records were independently reviewed by 2 attending anesthesiologists to determine the presence of a saphenous nerve injury. Sixty records were excluded, due to no documented saphenous nerve injury (n = 45), other nerve injury (n = 6), no ACB performed (n = 5), or documentation of injury onset >3 months after surgery (n = 4), leaving 18 cases of clinically important saphenous nerve injury after knee surgery with ACB.

The overall incidence of clinically relevant saphenous nerve injury after any knee surgery with ACB was 0.06% (18:28,196 or 6:10,000; Table 1). The highest incidence of saphenous nerve injury was found following surgery performed for ligament reconstruction, revision, or open reduction/internal fixation (0.3%; 13:4322; 30:10,000). The most common surgery associated with ACB and subsequent saphenous nerve injury was ACL reconstruction (8/18 cases). Of these cases, 5 were performed with hamstring autograft and 3 with bone-patellar tendon-bone (BTB) graft. The incidence of injury after arthroscopic surgery was 0.06% (2:3523; 6:1000). The lowest incidence of injury was found among patients undergoing TKA (0.02%; 3:20,351; 2:10,000).

Table 1.

Patient, anesthesia, and surgical details for saphenous nerve injury after knee surgery.

	All	Etiology: nonblock	Etiology: block	P a, b
Sample size (n)	18	13	5
Median age (IQR), years	37 (27, 50)	33 (24, 53)	45 (30, 55)	.492
Sex, n (%)
Male	3 (17)	3 (23)	0 (0)	.522
Female	15 (83)	10 (77)	5 (100)
ASA classification, n (%)
I	14 (78)	10 (56)	4 (22)	.632
II	4 (22)	3 (17)	1 (6)
Median BMI (IQR), kg/m2	24 (23, 26)	25 (23, 27)	24 (23, 28)	.842
Valgus knee, n (%)
Yes	1 (6)	0 (0)	1 (20)	—
No	8 (44)	5 (38)	3 (60)
Unable to determine	9 (50)	8 (62)	1 (20)
Diabetes mellitus, n (%)
Yes	6 (33)	6 (46)	0 (0)	.114
No	12 (67)	7 (54)	5 (100)
Lumbar stenosis, n (%)
Yes	1 (6)	1 (8)	0 (0)	—
No	10 (56)	6 (46)	4 (80)
Unable to determine	7 (39)	6 (46)	1 (20)
Peripheral neuropathy, n (%)
Yes	0 (0)	0 (0)	0 (0)
No	15 (83)	10 (77)	5 (100)	—
Unable to determine	3 (17)	3 (23)	0 (0)
Anesthetic type, n (%)
Neuraxial	18 (100)	13 (100)	5 (100)	—
Timing of ACB, n (%)
Preincision	16 (89)	12 (92)	4 (80)	.490
PACU	2 (11)	1 (8)	1 (20)
Median surgical duration (IQR), minutes	61 (35, 98)	93 (46, 115)	44 (26, 60)	.073
Median tourniquet time (IQR), minutes	38 (30, 69)	58 (30, 87)	32 (19, 38)	.084
Incidence of nerve injury by surgical type (%)	18:28,196 (0.06)	13:28,196 (0.05)	5:28,196 (0.02)
Total knee arthroplasty	3:20,351 (0.02)	2:20,351(0.01)	1:20,351 (.001)	.895
Reconstruction (ACL, MPFL, and ORIF)	13:4322 (0.3)	10:4322 (0.23)	3:4322 (0.07)
Arthroscopy	2:3523 (0.06)	2:3523 (0.06)	0:3523 (0.0)
Incidence of nerve injury by surgical type, n (%)	18 (100)	13 (72)	5 (28)
Total knee arthroplasty	3 (17)	2 (15)	1 (20)
ACL reconstruction	8 (44)	6 (46)	2 (40)
Hamstring autograft	5 (63)	5 (100)	0 (0)
BTB autograft	3 (38)	1 (33)	2 (66)
ACL revision	2 (11)	2 (15)	0 (0)	.723
MPFL reconstruction	2 (11)	1 (8)	1 (20)
Arthroscopy/meniscectomy	2 (11)	2 (15)	0 (0)
ORIF, patella	1 (5)	0 (0)	1 (20)

IQR, interquartile range, ASA American Society of Anesthesiologists, BMI body mass index, ACB adductor canal block, PACU post anesthesia care unit, ACL anterior cruciate ligament, MPFL medial patellofemoral ligament, ORIF open reduction/internal fixation, BTB bone-patellar tendon-bone.

^aP compares variables between cases attributed to ACB versus non-ACB etiology.

^bMann-Whitney U test was used to compare continuous variables and Fisher’s exact tests for categorical variables.

Thirteen of 18 cases (72%) were determined as most likely to be unrelated to ACB and 5 (28%) were determined to be most likely related to ACB (Table 2). There was agreement between 2 independent reviewers in 15 of 18 cases (83%) with a third reviewer required for 3 of 18 cases (17%). Of these, the third reviewer considered the most likely etiology of nerve injury to be non-ACB related in 2 of 3 cases (67%, Table 2).

Table 2.

Most Likely Etiology and Prognosis of Saphenous Nerve Injury After Surgery on the Knee With Adductor Canal Block.

Case number	Surgery	Reviewer 1	Reviewer 2	Reviewer 3	Prognosis
1	ACL/hamstring	NB	NB		Unknown
17	TKA	U	NB	NB	Unknown
19	Arthroscopy/meniscectomy	U	NB	NB	Unknown
23	ACL/hamstring	NB	NB		Unknown
29	ACL/BTB	U	B	B	Unknown
34	ACL/hamstring	NB	NB		Unknown
35	ACL/BTB	NB	NB		Poor: progression to CRPS
38	Arthroscopy/meniscectomy	NB	NB		Favorable: improvement after surgical exploration and nerve decompression
42	ACL/hamstring	NB	NB		Favorable: improvement in pain and numbness in saphenous distribution
48	ACL/hamstring	NB	NB		Favorable: improvement in pain and numbness in saphenous distribution
49	TKA	NB	NB		Full recovery
54	Open reduction/internal fixation, patella	B	B		Full recovery
60	ACL revision	NB	NB		Unknown
62	TKA	NB	NB		Unknown
69	ACL/BTB	B	B		Unknown
70	MPFL	B	B		Poor: progression to CRPS
76	ACL revision	NB	NB		Unknown
77	MPFL	B	B		Unknown

B, Nerve injury most likely due to adductor canal block, NB, Nerve injury most likely unrelated to adductor canal block, U, Uncertain most likely etiology.

ACL/hamstring anterior cruciate ligament reconstruction with hamstring autograft, TKA total knee arthroplasty, ACL/BTB anterior cruciate ligament reconstruction with bone-patellar tendon-bone autograft, CRPS chronic regional pain syndrome, MPFL medial patellofemoral ligament reconstruction.

There was incomplete or unclear documentation in the EMR to facilitate a decision regarding outcome in 11 of 18 cases (61%, Table 2). Of the remaining 7 cases, 3 (17%) had documented evidence of improvement/ongoing recovery and 2 (11%) were documented as full recovery. Documented evidence of poor recovery was found in 2 cases (11%), with progression to chronic regional pain syndrome. Of these, 1 case was deemed most likely related to ACB and 1 was deemed unrelated to ACB. In both cases, the ACB was performed in the PACU for rescue analgesia, after ACL reconstruction via arthrotomy. We did not find any significant differences in the incidence of nerve injury among cases judged secondary to ACB versus non-ACB etiology, when analyzed by demographic variables, surgical duration, or surgical subtype (Table 1).

Discussion

In this 5-year retrospective, single-institution cohort study, we found the overall incidence of clinically relevant saphenous nerve injury after knee surgery performed with ACB was 0.06%. The incidence was lower among patients undergoing TKA (0.03%) compared with other procedures performed via arthrotomy (0.3%) or arthroscopy (0.06%). Furthermore, although not statistically significant, we found that the most common surgery preceding saphenous nerve injury was ACL reconstruction with hamstring autograft, and in 2 of 2 patients with poor recovery prognosis the ACB was performed in the PACU as rescue analgesia.

There are several important limitations of this study. First, this was a single-center study conducted at a specialty orthopedic surgery hospital with advanced regional anesthesia services, and the results may not be generalizable to other practice settings. Second, the retrospective design relied on several aspects to identify clinically important saphenous nerve injury that may have underestimated the true population incidence. We did not include direct patient follow-up to verify the trajectory of the recovery but rather relied on the MRI report and EMR documentation to identify cases of nerve injury, assuming that our quality and quantity of search terms would be sufficient. Also, we were restricted to analyzing patients who underwent follow-up at our institution, raising the risk of reporting bias. Thus, unreported, poorly documented, or even mis-spelled instances of saphenous nerve injury may not have been captured in screening. Finally, our definitions of the most likely source of injury (block or nonblock) fails to capture the multifactorial nature of saphenous injury and the possibility that in some cases both may have contributed to the outcome.

Strengths of our study include our use of 2 databases to identify cases of saphenous nerve injury and our use of standardized a priori definitions of recovery and etiology. These data add to the emerging body of work on PONS versus persistent nerve injury and help define intermediate and long-term risks of ACB and saphenous nerve injury after knee surgery.

This study advances our understanding of the procedure-specific risks associated with ACB. The benefits of ACB on pain and opioid-related outcomes, with motor sparing benefits (when compared with femoral nerve block) first emerged in the literature around 2013 [9,10]. However, since that time, we have identified just 1 study which attempted to estimate the incidence of saphenous nerve injury after ACB for knee surgery, and the surgical cohort was restricted to TKA [5]. The authors included 97 patients and found no instances of saphenous nerve injury caused by ACB performed at the mid-thigh. However, 84% of patients had signs of injury to the infrapatellar branch of the saphenous nerve in the operated limb, all of which were attributed to a surgical complication [5]. While our study did estimate the incidence of saphenous nerve injury attributable to ACB as higher than 0.0%, the number of patient cases examined was likewise significantly higher.

Looking more broadly at the incidence of peripheral nerve injury (PNI) after TKA, a 2011 retrospective study from Jacob et al [8] examined 12,329 TKAs over a 20-year period from 1998 to 2007 and found an overall incidence of PNI of 0.79%, although this data was derived prior to the description of ACB. Interestingly, the authors did not associate PNI with PNB, although patients who received PNB were less likely to experience full recovery after injury.

Numerous surgical and patient-related risk factors have been implicated as potential contributors to peripheral nerve injury after knee surgery, including preoperative valgus deformity or flexion contracture, preexisting neurologic deficits, lumbar spinal stenosis, tourniquet inflation times, female sex and patient age <50 years [2,8,22]. While the small number of cases identified in our study precluded analyses adjusting for these (known) risk factors, it is notable that 15 of the 18 patients (83%) with postoperative saphenous nerve damage in this study were female and the median age of the cohort was 37.

It is important to note that our inclusion criteria consisted of clinically significant instances of saphenous neuropraxia, which prompted further diagnostic workup. Accordingly, our estimate of nerve injury is likely to be lower than the incidence of postoperative neurological symptoms (PONS), which is a broader term encompassing a lesser degree of symptoms that are frequently transient and less noticeable or concerning to patients. A recent retrospective cohort study by Lam et al investigated the incidence of short term and prolonged neurological symptoms after PNB. The authors found that while the short-term incidence (defined as less than 10 days) of neurological symptoms was as high as 14.4%, the incidence of long-term nerve damage attributable to a PNB (and lasting longer than 10 days) was 3 cases out of 19,219 patients [11]. This study did have direct patient outreach, either by phone or in person depending on the patient’s inpatient or ambulatory status after surgery. A total of 20 patients had PONS lasting greater than 10 days, though it was determined that only 3 cases were due PNB, and 2 of these completely resolved in 15 months or fewer. This resulted in an incidence of prolonged nerve injury attributable to PNB of approximately 2 in 10,000 [11], which is similar to published estimates [17], but lower than our estimate. This is likely because the former study reported the incidence of PONS after all types of orthopedic/plastic surgery with any PNB, whereas our data is selective for ACB and knee surgery.

Knee surgery has previously been associated with saphenous nerve injury, including after ACL reconstruction or TKA, with some series reporting 22% to 59% likelihood of injury [5,16,19]. The infrapatellar branch of the saphenous nerve (IBSN) is at high risk of injury from surgical incision, placement of port sites, and/or during tendon harvest [4,5,20,25]. Several systematic reviews/meta-analyses suggest the risk to the IBSN from the arthrotomy can be modified via use of an oblique versus vertical incision [3,4,20]. However, the risk of saphenous nerve injury is probably not simply due to the arthrotomy, per se, since we found the lowest incidence of saphenous injury after TKA. Indeed, the IBSN is considered at particularly high risk of iatrogenic injury during hamstring harvesting for ACL reconstruction [18]. Consistent with this, we found the highest incidence of saphenous nerve injury was following ACL reconstruction with hamstring autograft, and none of these were deemed likely to be secondary to the ACB.

Of note, the 2 cases that developed CRPS in the surgical limb (#35 and #70) had ACB performed as rescue analgesia in the PACU after surgery. This presents the possibility that there was saphenous nerve damage from surgery causing poorly controlled pain in the immediate postoperative period. Disproportionate pain to surgical insult may potentially be a surrogate for saphenous nerve injury incurred during surgery, thereby increasing a patient’s risk for CRPS [23]. Notwithstanding, of the 2 CRPS cases identified here, 1 was determined to be secondary to PNB. Provocatively, providing an ACB in the PACU might have affected the recovery trajectory in this patient. A prior retrospective analysis of nerve injury after 12,329 elective TKAs found that a nerve block was not correlated with increased risk of nerve injury, but of patients who did develop an injury, those who received a block had significantly lower odds of complete neurologic recovery [8].

Given the extremely low risk of saphenous nerve injury secondary to ACB, it is prudent to examine efficacy of the block to evaluate the role and value of providing the block for patients undergoing knee arthroscopy and ACL reconstruction. A 2019 systematic review and meta-analysis by Sehmbi et al [21] examined 10 randomized controlled trials that compared ACB with placebo or femoral nerve block. Compared with placebo, ACBs led to minor reductions in opioid consumption in the first 24 hours after knee arthroscopy, but not after ACL reconstruction. ACBs also lowered numeric rating scale (NRS) pain scores at 0 and 6 hours after knee arthroscopy, but not after ACL reconstruction. Although these results impugn the value of ACB for these procedures, calls for further high quality studies on the topic have been made [15], and are emerging [15]: Indeed, in a recent prospective randomized trial, combining IPACK with femoral triangle block reduced intravenous morphine consumption during the first 24 hours after ACL reconstruction, compared with local infiltration analgesia alone, and helps to clarify the optimal regimen of regional techniques for these procedures [14].

Future studies on procedure-specific risk would ideally be prospective in nature with direct patient outreach to ensure all cases of potential nerve injury can be accounted for, regardless of severity. In this way, a more complete risk: benefit analysis may be applied to individual patients undergoing specific surgeries in order to improve outcomes and better inform patients. In addition, increasing the number of surgical, anesthetic, and patient variables analyzed could be used to further risk stratify patients.

In conclusion, this retrospective review found that saphenous nerve injury after knee surgery with ACB was overall an infrequent event and that the incidence of saphenous nerve injury attributed to the block was even less common. However, the incidence of injury was higher among patients undergoing surgery for ACL reconstruction, and clinical outcomes were poor among patients who received ACB postoperatively for rescue analgesia. The latter finding suggests that performing ACB in the postoperative patient (particularly where pain is out of proportion to the magnitude of the surgical insult) should be approached with caution. These data help to clarify the relative procedure-specific risks and benefits of ACB and likelihood of complications after knee surgery.