Secondary Logo

Journal Logo


Saphenous Nerve Block From Within the Knee Is Feasible for TKA: MRI and Cadaveric Study

Kavolus, Joseph J. MD; Sia, David MD; Potter, Hollis G. MD; Attarian, David E. MD; Lachiewicz, Paul F. MD

Author Information
Clinical Orthopaedics and Related Research: January 2018 - Volume 476 - Issue 1 - p 30-36
doi: 10.1007/s11999.0000000000000006
  • Free



Multimodal analgesia, including peripheral block of sensory nerve fibers to the surgical site, is widely used for control of perioperative pain after TKA [2, 18]. However, the best method of providing that peripheral block as well as who should administer it remains unclear [4, 6]. Femoral nerve blocks provide excellent analgesia but at the expense of motor blockade and a possible increase in the risk of in-hospital falls [10, 12, 15, 19, 23]. A block of the saphenous nerve only at the adductor canal was first described in 1989 using both palpation and anatomic landmarks [26]. An adductor canal block in patients undergoing TKA described in 2009 [17] should provide comparable analgesia and reductions in opioid consumption to a femoral nerve block but without the attendant loss of motor function [13]. However, having these administered by an anesthesiologist will result in increased costs and time, particularly if they are performed with ultrasound guidance [28].

Several studies have evaluated the anatomy of the saphenous nerve and the adductor canal and shown consistent relationships to nearby structures [14, 22]. However, to our knowledge, there is no study of the relationship between the anatomy of the distal femur and the distal location of the saphenous nerve.

We therefore performed a study of distal thigh and knee MRIs as well as a proof-of-concept study in cadaver limbs and asked the following questions: (1) Can the saphenous nerve consistently be identified distally, and is there a consistent relationship between the width of the femoral transepicondylar axis (TEA) and the proximal (cephalad) distance at which the saphenous nerve emerges from the adductor canal? (2) Can we utilize the former anatomic relationship to simulate a surgeon-performed intraoperative block of the distal saphenous nerve from inside the knee with injections of dyes after implantation of trial TKA components in cadaveric lower extremity specimens?

Materials and Methods

After obtaining institutional review board approval from the Hospital for Special Surgery, a retrospective review of one orthopaedic hospital’s MRI library of distal thigh and knee images was performed. Using a PACS-based search system, relevant MRI examinations were collated by inputting the words “distal thigh” into the search filters, limiting the search to between January 1, 2008, and January 1, 2016. Only examinations that contained axial proton density MR sequences that depicted the saphenous nerve from the distal adductor canal to the femoral epicondyles were included. Examinations that showed pathology or treatment effects substantial enough to distort regional anatomy were excluded, examples of which include osteochondroma, neoplasms, and hematomas. Images containing metallic implants were also excluded as a result of the presence of susceptibility artifact.

There were 94 MRI studies of 42 female and 52 male patients. The patient ages ranged from 18 to 83 years for females (mean, 46 years) and 18 to 94 years for males (mean, 48 years). The studies were performed to evaluate thigh muscle or tendon tears in 55 patients, a mass in 16, nonspecific pain in 16, infection in four, suspected fracture in two, and sciatic neuropathy in one patient. Using a modified Noyes score, 11 knees had severe osteoarthritis, 21 moderate, 23 mild, and 39 had no osteoarthritis [21]. All MR studies were performed on clinical GE Healthcare (Chicago, IL, USA) scanners using a standardized transmit-receive eight-channel extremity coil (In Vivo, Gainesville, FL, USA). Axial fast spin echo acquisition was performed using a repetition time of 3500 to 5000 msec, echo time of 28 to 34 msec, matrix of 512 x 320, field of view ranging between 16 and 18 cm2 (women) and 18 and 22 cm2 (men) yielding a maximum in-plane resolution of 273 µm in the frequency direction, slice thickness of 3 to 4 mm with no interslice gap, and at two excitations. Of the 94 MRIs, 71 were performed using a 1.5-T scanner and 23 were performed using a 3-T scanner.

Transverse width of the distal femur at the TEA was chosen and measured as the reference point as a result of its reliability as a landmark, which is easily palpable intraoperatively. The saphenous nerve in the ipsilateral thigh was then followed proximally into the distal adductor canal, where it runs alongside the superficial femoral vessels (Fig. 1). The most inferior (caudally acquired) axial image where the saphenous nerve was clearly seen within the adductor canal was chosen as the proximal endpoint. By crossreferencing to a sagittal or coronal sequence, the cephalad distance from the TEA to the proximal endpoint was measured. This measurement was then divided by the TEA to obtain a ratio for the location of the saphenous nerve.

Fig. 1
Fig. 1:
A-B (A) Sagittal and (B) axial MRI demonstrate the location of the saphenous nerve as it exits the adductor canal. The red arrow (B) points to the location of the saphenous nerve.

Adhering to all local, state, and federal laws regarding work with cadaveric specimens and exemption from the Duke institutional review board, 11 fresh frozen limbs were studied at the Duke Human Fresh Tissue Laboratory. None of the cadavers had a history of surgery about the knee or femur, and none was noted to have severe deformity or arthritis on gross visual examination. As a result of time constraints and the fatigue associated with this work, at most two cadaveric limbs were studied in each laboratory day. In random sequence, limbs were permitted to thaw for at least 24 hours. There were six male and five female cadavers studied. The average age at death was 70 years (range, 57-80 years) and the average body mass index was 20 kg/m2 (range, 15-26 kg/m2). There were six right and five left lower extremity specimens. Six of 11 limbs were matched such that the right and left limbs were from the same cadaver (Table 1).

Table 1.
Table 1.:
Demographic and anatomic characteristics of cadaver specimens

A standard medial parapatellar approach to the knee was performed by one of two authors (JJK, PFL). The knee arthrotomies were made from 5 cm (three fingerbreadths) proximal to the superior patellar pole to the tibial tubercle. After adequate exposure, a surgical caliper was used to measure the width of the TEA of the specimen and this number was recorded (Table 2). To simulate the intraoperative manipulation of the lower limb, the TKA technique previously described by one of the authors (PFL) was performed on each specimen, but without patella resurfacing [16]. We performed this to show that the components implanted would not preclude injections from inside the knee. Component size for each specimen was recorded after insertion of trial components of one specific manufacturer (NexGen® PS; Zimmer, Warsaw, IN, USA) with adequate knee stability throughout ROM.

Table 2.
Table 2.:
Specimen measurements and implant sizing

The surgeon-performed saphenous nerve block from within the knee with local anesthetic was simulated using latex and dye using the MRI data of anatomic relationships for guidance. The injections were performed following the techniques of previous investigations correlating advanced imaging with cadaveric procedures and injection assays [3, 5, 7-9]. With the knee trial components in full extension, the proposed location of the saphenous nerve, as it exited the adductor canal, was marked on the skin. This location was defined as 1.3 or 1.5 x TEA in males and females, respectively, proximal to the medial epicondyle. One author (JJK) performed all injections. The first injection was performed with 10 mL pink latex using a 10-cc syringe and a 1.5-inch blunt 18-gauge needle. This injection was aimed between the point of the saphenous nerve egress from the adductor canal and 1 x the TEA proximally. The first injection was always through the substance of the vastus medialis obliquus as it originates off the distal femur and angled 30° to 45° medial. Insufflation of the pink latex was done after the needle was buried relative to the anterior femoral cortex and continued as the needle was withdrawn at least 1 cm. After injection, this needle was retained in the tissue at the site of the injection. A second injection was repeated with 10 mL green dye using a 10-cc syringe and a 1.5-inch, 19-gauge needle. This injection was aimed 1 x the TEA proximal to the medial epicondyle and 20° to 30° medial; the dye was similarly insufflated during partial withdrawal of the needle. This needle was also left in situ at the site of injection (Fig. 2).

Fig. 2
Fig. 2:
This is a photograph of the two different intraarticular injections with the knee in extension. The more proximal injection is the pink latex with slightly more medial angulation of the syringe. The injection is performed after trial knee arthroplasty components have been inserted.

After these injections, the capsule and skin were closed with clamps and the latex was allowed to consolidate for 30 minutes. Thereafter, the femoral nerve of each specimen was located and dissected starting proximally about the femoral head. To help identify the femoral artery at the time of dissection, a blunt metal instrument was placed into the femoral artery if possible. The course of the femoral nerve as it becomes the saphenous nerve was followed (Fig. 3). The sartorius muscle was elevated and the branches of the femoral nerve were dissected to locate the saphenous nerve and the adductor canal. At the level of the adductor canal, the location of pink latex or green dye was recorded and the presence of either dye about the saphenous nerve or superficial femoral artery was noted (Fig. 4). The saphenous nerve was also dissected distal to the vastoadductor membrane, recording the presence and staining of the nerve by either injection. The superficial femoral artery was also evaluated for perforation by either of the needles. We did not routinely measure the distance of our needle from the femoral artery or vein, because these were both posterior to our injection sites. We did not dissect the sciatic nerve or its branches.

Fig. 3
Fig. 3:
This is a photograph of a typical specimen with dissection of the femoral nerve from the hip to the knee. The sartorius muscle is mobilized anteriorly to show the distal saphenous nerve and other branches of the femoral nerve. The pink latex dye is present about the distal aspect of the adductor canal.
Fig. 4
Fig. 4:
This is a photograph of another specimen after the sartorius muscle has been mobilized anteriorly to demonstrate the saphenous nerve. In this specimen, both the proximal pink latex and the distal green dye injections bathe the saphenous nerve.

Statistical analysis of the differences between the MRI measurements between male and female patients was analyzed using the Wilcoxon rank-sum test.


The saphenous nerve was clearly visible in the distal thigh in all 94 MRI studies. The mean ± SD clinical TEA was smaller in women than in men (75 ± 4 mm versus 87 ± 4 mm, mean difference 12 mm; 95% confidence interval [CI], 10-13 mm; p < 0.001). With the numbers available, there was no difference between female and male knees in terms of the mean ± SD cephalad distance from the medial epicondyle to the nerve location (111 ± 12 versus 114 ± 11, mean difference 3 mm; 95% CI, -6 to 2 mm; p > 0.24). The ratio of TEA to cephalad nerve location was higher in female knees than in male knees (1.5 ± 0.16 versus 1.3 ± 0.13, mean difference 0.2; 95% CI, 0.12-0.23; p < 0.001).

The pink latex injection, performed at an angle of 30° to 45° medial, was noted to bathe the saphenous nerve in eight of 11 specimens (72%). The green dye injection, performed at an angle of 20° to 30° medial, was noted to bathe the saphenous nerve distal to the latex in six of 11 specimens (54%). The proximal needle from the latex injection was noted to be directly adjacent to the superficial femoral artery in one specimen, but it did not puncture the artery. This was one of the three specimens in which the nerve was not covered with latex. We did not measure the distance of the latex or dye to the femoral artery or vein routinely, because both structures had passed posteriorly at the injection sites. All three specimens were cachectic small cadaver limbs, and the dye was located in the body of the hamstring muscles. In all other cadaveric specimens, both the proximal and distal injection needles were remote from the femoral artery and well medial to the artery as it coursed posterior to the distal femur.


Multimodal analgesia, including a nerve block or periarticular injection, is widely used for control of pain after TKA [2, 18]. However, the best method of peripheral block of sensory nerve fibers as well as who should administer it remains unclear. Among many available strategies, periarticular injection by the surgeon and an anesthesiologist-performed femoral nerve or adductor canal block seem to be the two most common techniques. Although one study reported uniform success with the latter technique, another review reported 55% to 59% success of ultrasound-guided saphenous nerve block [1]. As a result of cost, time, and possible complications, a surgeon-performed periarticular injection with a variety of agents has been advocated [11]. With a retrospective study of thigh-knee MRI scans, we found a consistent relationship between the width of the TEA and the cephalad distance at which the saphenous nerve emerges from the adductor canal. With these data, we successfully simulated a surgeon-performed block of the distal saphenous nerve, from within the knee, with injections of latex, in eight of 11 cadaver specimens.

There are several limitations of this study. First, the MRI studies were not specifically performed in patients undergoing elective TKA, although there was some degree of osteoarthritis in 55 of 94 knees. However, there is no reason to believe that the width of the femur at the TEA or the location of the saphenous nerve and the adductor canal would be different in patients with a variety of injuries and disorders than patients with arthritis undergoing TKA. Second, as a result of limitations and cost of cadaver limbs, we were only able to perform the injection study and dissection in 11 lower extremities. However, this number is similar to other anatomic studies of the adductor canal. Third, none of the cadaver limbs had osteoarthritis of the knee or any notable axial deformity. However, there is no reason to believe that the anatomic relationship and location of the saphenous nerve would be notably different in these two populations. Fourth, because the cadaveric specimens came from subjects who had a body mass index of 26 kg/m2 or less, we do not know if this technique would be successful in a patient with obesity or morbid obesity. The number of overweight or obese patients undergoing TKA in the United States is well documented [27]. Although these cadaver specimen body mass indices are smaller, many patients have central obesity with relatively normal or smaller distal thigh girth. Patients with large body mass indices and large thighs may also be difficult or challenging for an anesthesiologist-performed adductor canal block. Fifth, we cannot comment on a possible “learning curve” necessary to be proficient with this injection technique. However, intraoperative measurements, based on MRI findings, should obviate a lengthy one. Finally, the finding that the saphenous vein was bathed in latex dye in the cadaver laboratory does not assure that a block of the nerve can be accomplished in patients.

Using MRI studies of the measurement of the TEA has been shown to be reliable and accurate before and after TKA [20, 25]. The location of the saphenous nerve and its relationship to the width measurement of the TEA has not been previously studied to our knowledge. Olson and Holt reported the statistical relationship of distances among the anterosuperior iliac spine, the adductor tubercle, and the adductor hiatus and the influence of cadaveric gender on this relationship [22].

This study suggests that with the measurement of the TEA, meticulous technique, and knowledge of local anatomy, it is feasible for a surgeon to perform an injection of local anesthetic solution or “cocktail” from inside the knee, near the saphenous nerve distal to the adductor canal. One technique study of periarticular analgesia mentioned injection of “several millimeters…in the region of Hunter canal,” but not a specific location [11]. A study by Pepper et al. [24] also studied 11 cadaveric knees and injected 10-mL aliquots of dye toward the proximal and distal adductor canal (AC) using a variety of needles. This study reported an accuracy of 86% using a blunt needle toward the “distal AC,” 57% accuracy with a blunt needle in the “proximal AC,” and only 14% accuracy using a spinal needle in the “proximal AC” [24]. This study is limited in that the AC locations were descriptive only, and TKA bone resections and trial component implantations were not performed. One cadaveric dissection study reported a consistent relationship of the saphenous nerve to the femoral artery and sartorius muscle [14]. In that study, after the saphenous nerve crosses the superficial femoral artery to lie medial to the vasculature, 73% of specimens had given off the most distal branch to the vastus medialis muscle [14]. A block distal to the adductor canal has the potential benefit of avoiding puncture and injection of the femoral artery.

In conclusion, the anatomy of the distal femur and the saphenous nerve in the distal thigh seems reasonably consistent, and the injection technique seems feasible. Based on the findings of this study, we propose that a prospective randomized clinical trial comparing surgeon-performed nerve block (with a variety of solutions) from within the knee with an anesthesiologist-performed, ultrasound-guided nerve block after TKA should be the next step. This trial should evaluate the postoperative level of pain, cost and time involved, and complications after each technique.


We thank Stephen Perlman for help with the literature search, the staff of the Duke HFTL for their assistance with obtaining cadaveric specimens, and Nick Kimmons and Fred Kelly, Zimmer North Carolina, for providing the TKA instruments and trial components.


1. Adoni A, Paraskeuopoulos T, Saranteas T, Sidiropoulou T, Mastrokalos D, Kostopanagiotou G. Prospective randomized comparison between ultrasound-guided saphenous nerve block within and distal to the adductor canal with low volume of local anesthetic. J Anaesthesiol Clin Pharmacol. 2014;30:378–382.
2. American Society of Anesthesiologists. Practice guidelines for acute pain management in the perioperative setting: an updated report by the American Society of Anesthesiologists Task Force on Acute Pain Management. Anesthesiology. 2012;116:248–273.
3. Andersen HL, Andersen SL, Tranum-Jensen J. The spread of injectate during saphenous nerve block at the adductor canal: a cadaver study. Acta Anaesthesiol Scand. 2015;59:238–245.
4. Benzon HT, Sharma S, Calimaran A. Comparison of the different approaches to saphenous nerve block. Anesthesiology. 2005;102:633–638.
5. Bhatia A, Gofeld M, Ganapathy S, Hanlon J, Johnson M. Comparison of anatomic landmarks and ultrasound guidance for intercostal nerve injections in cadavers. Reg Anesth Pain Med. 2013;38:503–507.
6. Busch CA, Shore BJ, Bhandari R, Ganapathy S, MacDonald SJ, Bourne RB, Rorabeck CH, McCalden RW. Efficacy of periarticular multimodal drug injection in total knee arthroplasty. A randomized trial. J Bone Joint Surg Am. 2006;88:959–963.
7. Dunaway DJ, Steensen RN, Wiand W, Dopirak RM. The sartorial branch of the saphenous nerve: its anatomy at the joint line of the knee. Arthroscopy. 2005;21:547–551.
8. Finnoff JT, Hurdle MF, Smith J. Accuracy of ultrasound-guided versus fluoroscopically guided contrast-controlled piriformis injections: a cadaveric study. J Ultrasound Med. 2008;27:1157–1163.
9. Gofeld M, Bhatia A, Abbas S, Ganapathy S, Johnson M. Development and validation of a new technique for ultrasound-guided stellate ganglion block. Reg Anesth Pain Med. 2009;34:475–479.
10. Grevstad U, Mathiesen O, Valentiner LS, Jaeger P, Hilsted KL, Dahl JB. Effect of adductor canal block versus femoral nerve block on quadriceps strength, mobilization, and pain after total knee arthroplasty: a randomized, blinded study. Reg Anesth Pain Med. 2015;40:3–10.
11. Guild GN 3rd, Galindo RP, Marino J, Cushner FD, Scuderi GR. Periarticular regional analgesia in total knee arthroplasty: a review of the neuroanatomy and injection technique. Orthop Clin North Am. 2015;46:1–8.
12. Jaeger P, Nielsen ZJ, Henningsen MH, Hilsted KL, Mathiesen O, Dahl JB. Adductor canal block versus femoral nerve block and quadriceps strength: a randomized, double-blind, placebo-controlled, crossover study in healthy volunteers. Anesthesiology. 2013;118:409–415.
13. Jin SQ, Ding XB, Tong Y, Ren H, Chen ZX, Wang X, Li Q. Effect of saphenous nerve block for postoperative pain on knee surgery: a meta-analysis. Int J Clin Exp Med. 2015;8:368–376.
14. Kapoor R, Adhikary SD, Siefring C, McQuillan PM. The saphenous nerve and its relationship to the nerve to the vastus medialis in and around the adductor canal: an anatomical study. Acta Anaesthesiol Scand. 2012;56:365–367.
15. Kim DH, Lin Y, Goytizolo EA, Kahn RL, Maalouf DB, Manohar A, Patt ML, Goon AK, Lee YY, Ma Y, Yadeau JT. Adductor canal block versus femoral nerve block for total knee arthroplasty: a prospective, randomized, controlled trial. Anesthesiology. 2014;120:540–550.
16. Lachiewicz PF, Soileau ES. Patella maltracking in posterior-stabilized total knee arthroplasty. Clin Orthop Relat Res. 2006;452:155–158.
17. Lund J, Jenstrup MT, Jaeger P, Sorensen AM, Dahl JB. Continuous adductor-canal-blockade for adjuvant post-operative analgesia after major knee surgery: preliminary results. Acta Anaesthesiol Scand. 2011;55:14–19.
18. Moucha CS, Weiser MC, Levin EJ. Current strategies in anesthesia and analgesia for total knee arthroplasty. J Am Acad Orthop Surg. 2016;24:60–73.
19. Mudumbai SC, Kim TE, Howard SK, Workman JJ, Giori N, Woolson S, Ganaway T, King R, Mariano ER. Continuous adductor canal blocks are superior to continuous femoral nerve blocks in promoting early ambulation after TKA. Clin Orthop Relat Res. 2014;472:1377–1383.
20. Murakami AM, Hash TW, Hepinstall MS, Lyman S, Nestor BJ, Potter HG. MRI evaluation of rotational alignment and synovitis in patients with pain after total knee replacement. J Bone Joint Surg Br. 2012;94:1209–1215.
21. Noyes FR, Stabler CL. A system for grading articular cartilage lesions at arthroscopy. Am J Sports Med. 1989;17:505–513.
22. Olson SA, Holt BT. Anatomy of the medial distal femur: a study of the adductor hiatus. J Orthop Trauma. 1995;9:63–65.
23. Paul JE, Arya A, Hurlburt L, Cheng J, Thabane L, Tidy A, Murthy Y. Femoral nerve block improves analgesia outcomes after total knee arthroplasty: a meta-analysis of randomized controlled trials. Anesthesiology. 2010;113:1144–1162.
24. Pepper AM, North TW, Sunderland AM, Davis JJ. Intraoperative adductor canal block for augmentation of periarticular injection in total knee arthroplasty: a cadaveric study. J Arthroplasty. 2016;31:2072–2076.
25. Potter HG, Foo LF. Magnetic resonance imaging of joint arthroplasty. Orthop Clin North Am. 2006;37:361–373, vi-vii.
26. Romanoff ME, Cory PC Jr, Kalenak A, Keyser GC, Marshall WK. Saphenous nerve entrapment at the adductor canal. Am J Sports Med. 1989;17:478–481.
27. Vasarhelyi EM, MacDonald SJ. The influence of obesity on total joint arthroplasty. J Bone Joint Surg Br. 2012;94:100–102.
28. Wang Y, Klein MS, Mathis S, Fahim G. Adductor canal block with bupivacaine liposome versus ropivacaine pain ball for pain control in total knee arthroplasty: a retrospective cohort study. Ann Pharmacother. 2016;50:194–202.
Copyright © 2017 by the Association of Bone and Joint Surgeons