- Email: [email protected]
Magnetic resonance imaging at different time periods following hamstring harvest for anterior cruciate ligament reconstruction
Magnetic Resonance Imaging at Different Time Periods Following Hamstring Harvest for Anterior Cruciate Ligament Reconstruction Damian M. Rispoli, M.D., Maj USAF, MC, Timothy G. Sanders, M.D., Lt Col USAF, MC, Mark D. Miller, M.D., Lt Col USAF, MC, and William B. Morrison, M.D.
Purpose: The purpose of this study was to evaluate the magnetic resonance imaging (MRI) appearance of the hamstring graft harvest site after harvesting the hamstring tendons to reconstruct a torn anterior cruciate ligament (ACL). Type of Study: Case series. Methods: We performed MRI on 21 patients who had previously undergone hamstring harvest and ACL reconstruction. Twenty of the patients (7 female and 13 male; mean age, 37 years; range, 16 to 84 years), all volunteers, were selected from a series of 45 ACL reconstructions performed by the senior author during a 20-month period. Another patient, a 32-year-old man, underwent ACL reconstruction elsewhere 32 months before. Both the semitendinosus and gracilis tendons were harvested in all cases. All MRIs were obtained on a 1.5-T magnet and were prospectively evaluated by 2 experienced musculoskeletal radiologists who were blinded to the time interval between graft harvest and MRI. Results: Two weeks after graft harvest, MRI showed ill-defined intermediate signal on T1-weighted images and increased signal on T2-weighted images, consistent with fluid in the harvest site, with no discernable tendon. At 6 weeks, structures were seen at the level of the superior pole of the patella that had morphology and signal characteristics similar to native tendon. By 3 months, structures with normal morphology and signal characteristics were seen to the level of the joint line, and by 12 months, to the level of 1 to 3 cm above that of the tibial attachment. At 32 months, the tendons appeared on MRI to normalize to a level of 1 to 2 cm above their tibial attachment. Conclusion: Following hamstring tendon harvest, MRI demonstrates an apparent regeneration of tendons beginning proximally and extending distally over time. Key Words: Anterior cruciate ligament—Surgery—Magnetic resonance imaging—Hamstring tendons.
H
amstring tendon grafts have been used increasingly to reconstruct torn anterior cruciate ligament (ACL).1 The hamstring graft is especially useful
From the Department of Orthopaedics, Wilford Hall Medical Center, San Antonio, Texas (D.M.R., M.D.M.); the Departments of Radiology, The University of Texas, Health Science Center at San Antonio and Wilford Hall Medical Center (T.G.S.); and the Department of Radiology, Thomas Jefferson University Hospital, Philadelphia, Pennsylvania (W.B.M.), U.S.A. Address correspondence and reprint requests to Timothy G. Sanders, M.D., Lt Col USAF, MC, Department of Radiology, Wilford Hall Medical Center, 759th MDTS/MTRD, 2200 Bergquist Dr, STE 1, Lackland AFB, TX 78236-5300, U.S.A. E-mail: [email protected] This is a US government work. There are no restrictions on its use. 0749-8063/01/1701-2454$0.00/0 doi:10.1053/jars.2001.19460
2
in older patients with patellofemoral arthrosis, patients with patellofemoral arthralgia or a history of anterior knee pain, and in revision surgery following patellar tendon harvest.2,3 Regrowth of patellar tendon graft defects has been well described4,5 and, although reharvesting of this tendon has been proposed,6,7 patellar tendon grafts taken from a previous harvest site have been shown to be both biomechanically4 and clinically inferior.6 Hamstring grafts have been shown to regenerate to the level of the medial head of the gastrocnemius both clinically and, in a few patients, on magnetic resonance imaging (MRI).8,9 Hamstring strength has been reported to return to almost normal values within 3 months of ACL reconstruction using semitendinosus and gracilis graft in 1 recent study.10 MRI evaluation
Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol 17, No 1 (January), 2001: pp 2– 8
MRI OF HAMSTRING TENDON HARVEST SITES of the cross-sectional area of the nonharvested hamstring muscles has been performed revealing no significant difference versus the operative side at 12 months.11 One recent study using ultrasound showed that the semitendinosus tendon may actually regenerate more distally than previously believed.12 The authors reported that, sonographically, the previously harvested tendons begin to show characteristics identical to native tendon throughout their normal anatomic course within 1 year. Reharvesting of the semitendinosus tendon for revision ACL reconstruction has been reported (P. Scranton, personal communication, October 1999). We chose to assess the previously harvested semitendinosus and gracilis tendons over a time continuum to more accurately define the time course and extent of their regeneration. METHODS Patients We performed a single MRI study on each of 21 patients who had previously undergone hamstring harvest and ACL reconstruction for either primary or revision surgery. Twenty of the patients (7 female and 13 male; mean age, 37 years; range, 16 to 84 years), all volunteers, were selected from a series of 45 ACL reconstructions performed by the senior author (M.D.M.) over a 20-month period. One patient, a 32-year-old man, underwent primary ACL reconstruction elsewhere using a hamstring graft 32 months before. The time interval between surgery and MRI ranged from 2 weeks to 32 months (2 examinations at 2 weeks, 3 at 6 weeks, 5 at 3 months, 6 between 4 and 10 months, 3 at 12 months, 1 at 15 months, and 1 at 32 months). Surgical Technique Both the semitendinosus and gracilis tendons were harvested. A 3-cm incision was centered midway between the tibial tubercle and the posteromedial aspect of the tibia, approximately 5 cm below the joint line. The tendons were carefully dissected and all overlying fascia and bands were carefully incised. A Krakow stitch (No. 5 nonabsorbable suture and tapered needle) was placed in the tendons beginning at their insertions, carried up the tendon 4 or 5 throws, and then brought down the opposite side. The tendons were then sharply incised at their insertions, preserving as much as possible. A closed-end tendon stripper (Smith & Nephew Dyonics, Memphis TN) was then “pushed” proximally while the distal ends were “pulled” and the
3
tendons were delivered. Any resistance that was felt during harvest called for additional dissection before repeating the technique. The tendons were brought to the back table and the muscle was removed from them using a curette. The proximal ends were tubularized and a “shoelace” type stitch was placed in the tendon using No. 5 nonabsorbable suture and an atraumatic needle. The grafts were fixed proximally using a 20-mm CL Endobutton (Smith & Nephew Dyonics) and distally with a screw and soft-tissue washer and a staple. MRI All MRIs were acquired on a Signa 1.5-T scanner (General Electric Medical Systems, Milwaukee, WI). The MRI protocols were as follows. By using a 6.5inch extremity coil for the knee (Quadrature coil; GE Medical Systems), fast spin-echo (FSE) T2-weighted images with frequency-selective fat saturation were acquired of the operated knee in the sagittal and axial planes (repetition time msec [TR]/echo time msec [TE] ⫽ 3,100-5,100/53-80 effective, echo train length 8, 4-mm section thickness, 1-mm slice intersection gap, 3 signals averaged, 256 ⫻ 192 matrix, and 15-cm field of view). T1-weighted spin echo (SE) sequences were performed of the operated knee in the sagittal and axial planes (TR/TE ⫽ 600-700/14-16, 4-mm section thickness, 1-mm slice intersection gap, 2 signals averaged, 256 ⫻ 192 matrix, and 15-cm field of view). T1-weighted SE images were then acquired through the thighs bilaterally from the level of the knee joint to the subtrochanteric region using a torso coil (TR/TE ⫽ 600/14, 1-cm section thickness, 3-mm slice intersection gap, 2 signal averages, 256 ⫻ 192 matrix, and 36-cm field of view). All MRIs were prospectively evaluated by 2 experienced musculoskeletal radiologists and a consensus reading was obtained. The radiologists knew that the hamstring tendon had been harvested, but they were blinded to the time interval between graft harvest and MRI. In each case, the normal anatomic location of the sartorius, gracilis, and semitendinosus muscles and tendons were evaluated separately on the MRIs. The muscles and/or tendons were evaluated at 2 locations: (1) between the superior pole of the patella and the level of the joint line, and (2) distal to the joint line. The presence or absence of a structure in the normal anatomic location of each of the hamstring tendons was noted. If a structure was present in the harvest site, the morphology and T1- and T2-weighted imaging characteristics were recorded. The presence or
4
D. M. RISPOLI ET AL.
FIGURE 1. MRIs of the knee of a 35-year-old man 2 weeks after hamstring tendon harvest. (A) T1-weighted sagittal image shows the normal course and appearance of the sartorius tendon (white arrows). There is no tendon located in the harvest site (black arrowheads). (B) T2-weighted sagittal image with fat-saturation again reveals a normal appearing sartorius tendon (long arrow). The tendon harvest site, marked by short arrows, shows edema within the tissue planes and no identifiable tendon. (C) T2-weighted axial image with fat-saturation at the level of the upper pole of the patella shows a normal cross-sectional view of the sartorius muscle at this level (arrow 1). Arrows 2 and 3 mark the normal location of the gracilis and semitendinosus tendons, respectively. There is no identifiable tendon and edema remains in the surgical bed.
absence of fluid or edema in the surgical harvest site was also noted. Using axial images through the thighs bilaterally, anterior-posterior and transverse diameter measurements were obtained using electronic calipers through each of the 3 hamstring muscles separately at the mid-thigh level and at the subtrochanteric level. Both radiologists and 1 orthopaedic surgeon (D.M.R.) evaluated the images collectively at a single setting and agreed on the measurements. Measurements were compared between the oper-
ative and nonoperative sides to evaluate for muscular atrophy. In addition, the presence of increased signal within the hamstring muscles on T1-weighted images was noted to document the presence of fatty atrophy on the operated versus the nonoperated side. Statistical Methods Differences in the measured cross-sectional areas of the proximal and distal sartorius, gracilis, and
MRI OF HAMSTRING TENDON HARVEST SITES semitendinosus muscles, between the harvested side and the nonoperated side, were evaluated using the Friedman analysis of variance test (a statistical method used to evaluate sets of related samples). They were calculated using Simstat for Windows, version 1.1, with probability of significance set a P ⱕ .05. RESULTS In general, the MRI findings can be considered chronologically. MRI was performed on 2 patients 2 weeks after hamstring tendon harvest. The studies showed ill-defined intermediate signal on T1weighted images and increased signal on T2-weighted images consistent with fluid or edema in the semitendinosus and gracilis tracts. There was no discernible tendon seen from the level of the superior pole of the patella inferiorly (Fig 1). Images acquired of 2 patients approximately 6 weeks after surgery showed tissue at the level of the superior pole of the patella with the normal morphologic appearance of tendon and decreased signal intensity on T1- and T2-weighted images. The more distal tissue, below the level of the mid-patella, was ill defined and blended into the adjacent fascia. There was no definite tendon-like structure seen below the level of the joint line and edema persisted within the
5
harvest site in the distal 3 to 4 cm. Images acquired on 5 patients 3 months after tendon harvest showed a mixed picture; some patients displayed only residual fluid along the distal 3 to 4 cm of the harvest tract, whereas in others, a distinct structure could be clearly identified. In this second group, the hamstring tendons had MR signal characteristics and morphology similar to the native tendon to the level of the joint line, but below the joint line, the tendons remained morphologically abnormal with indistinct poorly visualized fibers gradually blending into the adjacent fascia. By 7 to 12 months after surgery, 7 patients had normal tendon morphology and signal characteristics to the level of the distal 1 to 3 cm, just above the pes anserinus. The tendons remained ill defined in their most distal 3 to 4 cm (Fig 2). At 32 months postoperatively, MRI of a single patient showed complete normalization of the proximal tendon to the level approximately 1 to 2 cm above their tibial attachment sites (Fig 3). Cross-sectional areas as measured on axial MRIs showed no statistically significant atrophy of the sartorius or semitendinosus muscles (Table 1). The gracilis muscle at the mid-thigh level had a degree of atrophy that was statistically significant (P ⫽ .01). Asymmetric fatty atrophy of the ipsilateral hamstring muscles was inconsistently seen on MRI as increased signal within the muscle belly on T-1 imaging, but this
FIGURE 2. MRIs of the knee of a 16-year-old boy 8 months after hamstring tendon harvest. (A) T1-weighted axial image just below the level of the inferior pole of the patella shows a normal appearing sartorius muscle (arrow 1). Arrows 2 and 3 show the gracilis and semitendinosus tendons, respectively. They are in normal anatomic location and show normal morphology and signal characteristics. (B) T1-weighted axial image distal to image A, just above the level of the tibial insertion of the pes anserine. The sartorius is normal in appearance (arrow 1). The gracilis and semitendinosus tendons (arrows 1 and 2, respectively), although present, are less distinct than normal and are slightly attenuated.
6
D. M. RISPOLI ET AL.
FIGURE 3. MRIs of the knee of a 32-year-old man 32 months after hamstring tendon harvest. (A) T1-weighted axial image at the level of the superior pole of the patella shows a normal sartorius muscle (arrow 1). The semitendinosus tendon is also normal at this level (arrow 2). The gracilis is not seen at this level. (B) T1-weighted axial image below the level of the joint line again shows normal appearing sartorius and semitendinosus tendons (arrows 1 and 2, respectively). No distinct gracilis tendon is seen. (C) T1-weighted image at the level of the tibial attachment of the pes anserine shows a normal appearing attachment (arrow).
finding appears to be unrelated to the time interval between surgery and imaging. DISCUSSION Hamstring tendons have been shown to have adequate strength13 and incorporation capability14 to function as an ACL graft and have done well in intermediate-term studies.15 Patellar tendon grafts have been associated with relatively high morbidity,
including patellofemoral pain and chondrosis,16-18 patellar fractures,19,20 patellar tendon rupture,21,22 and even combined fracture and tendon rupture.23,24 Hamstring graft harvest, however, is reported as having a very low associated morbidity.25,26 The patellar tendon has been shown to reconstitute its central one third following bone–patellar tendon– bone harvest.4,5 Recent studies on the healing of central third patellar tendon donor sites in a goat model have raised questions as to the advisability of a second
MRI OF HAMSTRING TENDON HARVEST SITES TABLE 1. Cross-Sectional Measurements of Hamstring Muscles
Sartorius Proximal Nonoperated Harvest site Distal Nonoperated Harvest site Gracilis Proximal Nonoperated Harvest site Distal Nonoperated Harvest site Semitendinosus Proximal Nonoperated Harvest site Distal Nonoperated Harvest site
side
side
side
side
side
side
Mean Area (mm)
SD
P Value
0.95 0.94
⫾0.13 ⫾0.14
.64
0.91 1.00
⫾0.26 ⫾0.27
.49
1.00 0.86
⫾0.26 ⫾0.16
.11
0.84 0.70
⫾0.35 ⫾0.30
.01 (significant)
0.87 0.88
⫾0.23 ⫾0.23
.64
0.98 0.96
⫾0.12 ⫾0.13
.81
NOTE. Measured cross-sectional areas on MRIs of sartorius, gracilis, and semitendinosus, proximally (between the superior pole of the patella and the level of the joint line) and distally (below the level of the joint line). Significance of differences between the harvest site and the nonoperated side of patients with previous ACL reconstruction (P ⱕ .05 defined as significant difference).
harvest of the same tendon because it has been shown to be biomechanically inferior.4 The normal gracilis and semitendinosus tendons lie invested in fascia of the medial aspect of the knee, between layers I and II as described by Warren and Marshall.27 The tendons become conjoined as they insert into the pes anserinus. The tendons exist as distinct structures at an average of 18 mm proximal to this insertion.28 The tendons lie adherent to the deep layer of the thigh fascia. This fascia, 8 to 10 cm proximal to the pes anserinus, forms a 3- to 4-cm band around the tendons.15 This area has been shown to connect with the medial intermuscular septum and the sheath invests the semitendinosus. The tissue plane for the semitendinosus (and to a lesser extent the gracilis) may help to explain the regeneration of tissue in the normal anatomic location of the harvested tendon. We theorize that the tendons first regenerate proximally in a more vascularized area, which then proceeds distally along the fascial planes. Although these MRI findings seem to support this theory, a true prospective study with repeated MRI measurements on the same individuals would be required for confirmation. This phe-
7
nomenon may be analogous to nerve regrowth through intact neural sheaths. Hamstring tendon regrowth undoubtedly resembles normal tendon because it has been reharvested in select cases. With the exception of the gracilis muscle at the mid-thigh level, the hamstring muscles show no significant decrease in the cross-sectional area in the operative versus the nonoperative extremity. This may help explain why, as shown in previous studies,10 the hamstring strength returns to nearly normal values during the first 3 months following tendon harvest. Potential weaknesses of this study are as follows. Although the MRIs were obtained and interpreted in a prospective fashion, sets of images at each time interval were performed on a different group of patients. It may be beneficial to image the same group of patients sequentially at specific intervals following hamstring tendon harvest to determine the true time course of tendon regeneration. Another potential weakness is that, although the tendons on MRI regain normal morphology and signal characteristics over time, the exact histologic composition of these regenerated tendons remains unknown. The structures seen on MRI in the location of the native tendons may represent a structurally inferior tendon or scar tissue. Furthermore, the functional ability of these tendons and their potential for reharvest remains uncertain. We plan further studies using an in vivo animal model to help answer these questions. In conclusion, we found in this study that, after hamstring tendon harvest, MRI shows an apparent regeneration of the tendons beginning proximally and extending distally over time. By 15 months, the tendons have normal MR signal characteristics on T1and T2-weighting, normal morphology, and normal location to the level of the distal 1 to 2 cm of the native tendon. It remains uncertain what the functional ability of these regenerated tendons is, and further studies are needed to help resolve this issue. REFERENCES 1. Brown CH Jr., Steiner ME, Carson EW. The use of hamstring tendons for anterior cruciate ligament reconstruction. Technique and results. Clin Sports Med 1993;12:723-756. 2. Brown CH Jr, Carson EW. Revision anterior cruciate ligament surgery. Clin Sports Med 1999;18:109-171. 3. Day B. Revision anterior cruciate ligament reconstruction after failed bone patellar tendon bone reconstruction. Arthroscopy 1996;12:381-382 (abstr). 4. LaPrade RF, Hamilton CD, Montgomery RD, Wentoff F, Hawkins HD. The reharvested central third of the patellar tendon: A histological and biomechanical analysis. Am J Sports Med 1997;25:779-785. 5. Proctor CS, Jackson DW, Simon TM. Characterization of the repair tissue after removal of the central one-third of the
8
6.
7. 8.
9.
10. 11.
12.
13.
14.
15. 16.
D. M. RISPOLI ET AL. patellar ligament. An experimental study in a goat model. J Bone Joint Surg Am 1997;79:997-1006. Kartus J, Stener S, Lindahl S, Eriksson BI, Karlsson J. Ipsi- or contralateral patellar tendon graft in anterior cruciate ligament revision surgery. A comparison of two methods. Am J Sports Med 1998;26:499-504. Karns DJ, Heidt RS Jr, Holladay BR, Colosimo AJ. Revision anterior cruciate ligament reconstruction. Arthroscopy 1994; 10:148-151. Cross MJ, Roger G, Kujawa P, Anderson IF. Regeneration of the semitendinosus and gracilis tendons following their transection for repair of the anterior cruciate ligament. Am J Sports Med 1992;20:221-223. Erickson K, Larson H, Wredmark T, Hamberg P. Semitendinosus tendon regeneration after harvesting for ACL reconstruction. a prospective MRI study. Knee Surg Sports Traumatol Arthrosc 1999;7:220-225. Iacono F, Loreti I, Maracci M, et at: Evaluation of hamstring strength following gracilis and semitendinosus reconstruction at 3 months follow-up. Arthroscopy 14:526,1998 (abstr). Simonian PT, Harrison SD, Cooley VJ, Escabedo EM, Deneka DA, Larson RV. Assessment of morbidity of semitendinosus and gracilis tendon harvest for ACL reconstructions. Am J Knee Surg 1997;10:54-59. Papandrea P, Vulpiana MC, Conteduca F, et al. Regeneration of the semitendinosus following its use for anterior cruciate ligament reconstruction: Ultrasound evaluation. Presented at the Annual Meeting of the American Academy of Orthopaedic Surgery, Anaheim, CA, March 1999. Recker AT, Brown CR, Deffner KT, et at: Tensile properties of young multiple stranded hamstring tendon grafts. Presented at the American Orthopaedic Society for Sports Medicine Specialty Day, Book of Abstracts p 8. San Francisco, February 1997. Pinczewski LA, Clingeleffer AJ, Otto DD, Bonar SF, Cory IS. Integration of hamstring tendon graft with bone in reconstruction of the anterior cruciate ligament. Arthroscopy 1997;13: 641-643. Siegel MG, Barber-Westin SD. Arthroscopic-assisted outpatient anterior cruciate ligament reconstruction using the semitendinosus and gracilis tendons. Arthroscopy 1998;14:268-277. Aglietti P, Buzzi R, D’Andria S, Zaccherotti G. Patellofemoral
17. 18.
19. 20. 21.
22.
23. 24. 25. 26. 27. 28.
problems after intraarticular anterior cruciate ligament reconstruction. Clin Orthop 1993;288:195-204. Sachs RA, Daniels DM, Stone ML, Garfein RF. Patellofemoral problems after anterior cruciate ligament reconstruction. Am J Sports Med 1989;17:760-765. Shino K, Nakagawa S, Inoue M, Horibe S, Yoneda M. Deterioration of patellofemoral articular surfaces after anterior cruciate ligament reconstruction. Am J Sports Med 1993;21:206211. Christen B, Jakob RP. Fractures associated with patellar ligament grafts in cruciate ligament surgery. J Bone Joint Surg Br 1992;74:617-619. McCarroll JR. Fracture of the patella during a golf swing following reconstruction of the anterior cruciate ligament: A case report. Am J Sports Med 1983;11:26-27. Bonamo JJ, Krinick RM, Sporn AA. Rupture of the patellar ligament after use of its central third for anterior cruciate reconstruction. A report of two cases. J Bone Joint Surg Am 1984;66:1294-1297. Marumoto JM, Mitsunaga MM, Richardson AB, Medoff RJ, Mayfield GW. Late patellar tendon rupture after removal of the central third for anterior cruciate ligament reconstruction. A report of two cases. Am J Sports Med 1996;24:698-701. Bonatus TJ, Alexander AH. Patellar fracture and avulsion of the patellar ligament complicating arthroscopic anterior cruciate ligament reconstruction. Orthop Rev 1991;20:770-774. Miller MD, Nichols T, Butler CA. Patella fracture and proximal patellar tendon rupture following arthroscopic anterior cruciate ligament ronstruction. Arthroscopy 1999;15:640-643. Larson RV, Erickson D: Complications in the use of hamstring tendons for anterior cruciate ligament reconstruction. Sports Med Arthrosc Rev 23:83-90, 1997 Yasuda K, Tsujino J, Ohkoshi Y, Tanabe Y, Kaneda K. Graft site morbidity with autogenous semitendinosus and gracilis tendons. Am J Sports Med 1995;23:706-714. Warren LF, Marshall JL. The supporting structures and layers on the medial side of the knee: an anatomical analysis. J Bone Joint Surg Am 1979;61:56-62. Pagnani MJ, Warner JJ, O’Brien SJ, Warren RF. Anatomic considerations in harvesting the semitendinosus and gracilis tendons and a technique of harvest. Am J Sports Med 1993; 21:565-571.
Purpose: The purpose of this study was to evaluate the magnetic resonance imaging (MRI) appearance of the hamstring graft harvest site after harvesting the hamstring tendons to reconstruct a torn anterior cruciate ligament (ACL). Type of Study: Case series. Methods: We performed MRI on 21 patients who had previously undergone hamstring harvest and ACL reconstruction. Twenty of the patients (7 female and 13 male; mean age, 37 years; range, 16 to 84 years), all volunteers, were selected from a series of 45 ACL reconstructions performed by the senior author during a 20-month period. Another patient, a 32-year-old man, underwent ACL reconstruction elsewhere 32 months before. Both the semitendinosus and gracilis tendons were harvested in all cases. All MRIs were obtained on a 1.5-T magnet and were prospectively evaluated by 2 experienced musculoskeletal radiologists who were blinded to the time interval between graft harvest and MRI. Results: Two weeks after graft harvest, MRI showed ill-defined intermediate signal on T1-weighted images and increased signal on T2-weighted images, consistent with fluid in the harvest site, with no discernable tendon. At 6 weeks, structures were seen at the level of the superior pole of the patella that had morphology and signal characteristics similar to native tendon. By 3 months, structures with normal morphology and signal characteristics were seen to the level of the joint line, and by 12 months, to the level of 1 to 3 cm above that of the tibial attachment. At 32 months, the tendons appeared on MRI to normalize to a level of 1 to 2 cm above their tibial attachment. Conclusion: Following hamstring tendon harvest, MRI demonstrates an apparent regeneration of tendons beginning proximally and extending distally over time. Key Words: Anterior cruciate ligament—Surgery—Magnetic resonance imaging—Hamstring tendons.
H
amstring tendon grafts have been used increasingly to reconstruct torn anterior cruciate ligament (ACL).1 The hamstring graft is especially useful
From the Department of Orthopaedics, Wilford Hall Medical Center, San Antonio, Texas (D.M.R., M.D.M.); the Departments of Radiology, The University of Texas, Health Science Center at San Antonio and Wilford Hall Medical Center (T.G.S.); and the Department of Radiology, Thomas Jefferson University Hospital, Philadelphia, Pennsylvania (W.B.M.), U.S.A. Address correspondence and reprint requests to Timothy G. Sanders, M.D., Lt Col USAF, MC, Department of Radiology, Wilford Hall Medical Center, 759th MDTS/MTRD, 2200 Bergquist Dr, STE 1, Lackland AFB, TX 78236-5300, U.S.A. E-mail: [email protected] This is a US government work. There are no restrictions on its use. 0749-8063/01/1701-2454$0.00/0 doi:10.1053/jars.2001.19460
2
in older patients with patellofemoral arthrosis, patients with patellofemoral arthralgia or a history of anterior knee pain, and in revision surgery following patellar tendon harvest.2,3 Regrowth of patellar tendon graft defects has been well described4,5 and, although reharvesting of this tendon has been proposed,6,7 patellar tendon grafts taken from a previous harvest site have been shown to be both biomechanically4 and clinically inferior.6 Hamstring grafts have been shown to regenerate to the level of the medial head of the gastrocnemius both clinically and, in a few patients, on magnetic resonance imaging (MRI).8,9 Hamstring strength has been reported to return to almost normal values within 3 months of ACL reconstruction using semitendinosus and gracilis graft in 1 recent study.10 MRI evaluation
Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol 17, No 1 (January), 2001: pp 2– 8
MRI OF HAMSTRING TENDON HARVEST SITES of the cross-sectional area of the nonharvested hamstring muscles has been performed revealing no significant difference versus the operative side at 12 months.11 One recent study using ultrasound showed that the semitendinosus tendon may actually regenerate more distally than previously believed.12 The authors reported that, sonographically, the previously harvested tendons begin to show characteristics identical to native tendon throughout their normal anatomic course within 1 year. Reharvesting of the semitendinosus tendon for revision ACL reconstruction has been reported (P. Scranton, personal communication, October 1999). We chose to assess the previously harvested semitendinosus and gracilis tendons over a time continuum to more accurately define the time course and extent of their regeneration. METHODS Patients We performed a single MRI study on each of 21 patients who had previously undergone hamstring harvest and ACL reconstruction for either primary or revision surgery. Twenty of the patients (7 female and 13 male; mean age, 37 years; range, 16 to 84 years), all volunteers, were selected from a series of 45 ACL reconstructions performed by the senior author (M.D.M.) over a 20-month period. One patient, a 32-year-old man, underwent primary ACL reconstruction elsewhere using a hamstring graft 32 months before. The time interval between surgery and MRI ranged from 2 weeks to 32 months (2 examinations at 2 weeks, 3 at 6 weeks, 5 at 3 months, 6 between 4 and 10 months, 3 at 12 months, 1 at 15 months, and 1 at 32 months). Surgical Technique Both the semitendinosus and gracilis tendons were harvested. A 3-cm incision was centered midway between the tibial tubercle and the posteromedial aspect of the tibia, approximately 5 cm below the joint line. The tendons were carefully dissected and all overlying fascia and bands were carefully incised. A Krakow stitch (No. 5 nonabsorbable suture and tapered needle) was placed in the tendons beginning at their insertions, carried up the tendon 4 or 5 throws, and then brought down the opposite side. The tendons were then sharply incised at their insertions, preserving as much as possible. A closed-end tendon stripper (Smith & Nephew Dyonics, Memphis TN) was then “pushed” proximally while the distal ends were “pulled” and the
3
tendons were delivered. Any resistance that was felt during harvest called for additional dissection before repeating the technique. The tendons were brought to the back table and the muscle was removed from them using a curette. The proximal ends were tubularized and a “shoelace” type stitch was placed in the tendon using No. 5 nonabsorbable suture and an atraumatic needle. The grafts were fixed proximally using a 20-mm CL Endobutton (Smith & Nephew Dyonics) and distally with a screw and soft-tissue washer and a staple. MRI All MRIs were acquired on a Signa 1.5-T scanner (General Electric Medical Systems, Milwaukee, WI). The MRI protocols were as follows. By using a 6.5inch extremity coil for the knee (Quadrature coil; GE Medical Systems), fast spin-echo (FSE) T2-weighted images with frequency-selective fat saturation were acquired of the operated knee in the sagittal and axial planes (repetition time msec [TR]/echo time msec [TE] ⫽ 3,100-5,100/53-80 effective, echo train length 8, 4-mm section thickness, 1-mm slice intersection gap, 3 signals averaged, 256 ⫻ 192 matrix, and 15-cm field of view). T1-weighted spin echo (SE) sequences were performed of the operated knee in the sagittal and axial planes (TR/TE ⫽ 600-700/14-16, 4-mm section thickness, 1-mm slice intersection gap, 2 signals averaged, 256 ⫻ 192 matrix, and 15-cm field of view). T1-weighted SE images were then acquired through the thighs bilaterally from the level of the knee joint to the subtrochanteric region using a torso coil (TR/TE ⫽ 600/14, 1-cm section thickness, 3-mm slice intersection gap, 2 signal averages, 256 ⫻ 192 matrix, and 36-cm field of view). All MRIs were prospectively evaluated by 2 experienced musculoskeletal radiologists and a consensus reading was obtained. The radiologists knew that the hamstring tendon had been harvested, but they were blinded to the time interval between graft harvest and MRI. In each case, the normal anatomic location of the sartorius, gracilis, and semitendinosus muscles and tendons were evaluated separately on the MRIs. The muscles and/or tendons were evaluated at 2 locations: (1) between the superior pole of the patella and the level of the joint line, and (2) distal to the joint line. The presence or absence of a structure in the normal anatomic location of each of the hamstring tendons was noted. If a structure was present in the harvest site, the morphology and T1- and T2-weighted imaging characteristics were recorded. The presence or
4
D. M. RISPOLI ET AL.
FIGURE 1. MRIs of the knee of a 35-year-old man 2 weeks after hamstring tendon harvest. (A) T1-weighted sagittal image shows the normal course and appearance of the sartorius tendon (white arrows). There is no tendon located in the harvest site (black arrowheads). (B) T2-weighted sagittal image with fat-saturation again reveals a normal appearing sartorius tendon (long arrow). The tendon harvest site, marked by short arrows, shows edema within the tissue planes and no identifiable tendon. (C) T2-weighted axial image with fat-saturation at the level of the upper pole of the patella shows a normal cross-sectional view of the sartorius muscle at this level (arrow 1). Arrows 2 and 3 mark the normal location of the gracilis and semitendinosus tendons, respectively. There is no identifiable tendon and edema remains in the surgical bed.
absence of fluid or edema in the surgical harvest site was also noted. Using axial images through the thighs bilaterally, anterior-posterior and transverse diameter measurements were obtained using electronic calipers through each of the 3 hamstring muscles separately at the mid-thigh level and at the subtrochanteric level. Both radiologists and 1 orthopaedic surgeon (D.M.R.) evaluated the images collectively at a single setting and agreed on the measurements. Measurements were compared between the oper-
ative and nonoperative sides to evaluate for muscular atrophy. In addition, the presence of increased signal within the hamstring muscles on T1-weighted images was noted to document the presence of fatty atrophy on the operated versus the nonoperated side. Statistical Methods Differences in the measured cross-sectional areas of the proximal and distal sartorius, gracilis, and
MRI OF HAMSTRING TENDON HARVEST SITES semitendinosus muscles, between the harvested side and the nonoperated side, were evaluated using the Friedman analysis of variance test (a statistical method used to evaluate sets of related samples). They were calculated using Simstat for Windows, version 1.1, with probability of significance set a P ⱕ .05. RESULTS In general, the MRI findings can be considered chronologically. MRI was performed on 2 patients 2 weeks after hamstring tendon harvest. The studies showed ill-defined intermediate signal on T1weighted images and increased signal on T2-weighted images consistent with fluid or edema in the semitendinosus and gracilis tracts. There was no discernible tendon seen from the level of the superior pole of the patella inferiorly (Fig 1). Images acquired of 2 patients approximately 6 weeks after surgery showed tissue at the level of the superior pole of the patella with the normal morphologic appearance of tendon and decreased signal intensity on T1- and T2-weighted images. The more distal tissue, below the level of the mid-patella, was ill defined and blended into the adjacent fascia. There was no definite tendon-like structure seen below the level of the joint line and edema persisted within the
5
harvest site in the distal 3 to 4 cm. Images acquired on 5 patients 3 months after tendon harvest showed a mixed picture; some patients displayed only residual fluid along the distal 3 to 4 cm of the harvest tract, whereas in others, a distinct structure could be clearly identified. In this second group, the hamstring tendons had MR signal characteristics and morphology similar to the native tendon to the level of the joint line, but below the joint line, the tendons remained morphologically abnormal with indistinct poorly visualized fibers gradually blending into the adjacent fascia. By 7 to 12 months after surgery, 7 patients had normal tendon morphology and signal characteristics to the level of the distal 1 to 3 cm, just above the pes anserinus. The tendons remained ill defined in their most distal 3 to 4 cm (Fig 2). At 32 months postoperatively, MRI of a single patient showed complete normalization of the proximal tendon to the level approximately 1 to 2 cm above their tibial attachment sites (Fig 3). Cross-sectional areas as measured on axial MRIs showed no statistically significant atrophy of the sartorius or semitendinosus muscles (Table 1). The gracilis muscle at the mid-thigh level had a degree of atrophy that was statistically significant (P ⫽ .01). Asymmetric fatty atrophy of the ipsilateral hamstring muscles was inconsistently seen on MRI as increased signal within the muscle belly on T-1 imaging, but this
FIGURE 2. MRIs of the knee of a 16-year-old boy 8 months after hamstring tendon harvest. (A) T1-weighted axial image just below the level of the inferior pole of the patella shows a normal appearing sartorius muscle (arrow 1). Arrows 2 and 3 show the gracilis and semitendinosus tendons, respectively. They are in normal anatomic location and show normal morphology and signal characteristics. (B) T1-weighted axial image distal to image A, just above the level of the tibial insertion of the pes anserine. The sartorius is normal in appearance (arrow 1). The gracilis and semitendinosus tendons (arrows 1 and 2, respectively), although present, are less distinct than normal and are slightly attenuated.
6
D. M. RISPOLI ET AL.
FIGURE 3. MRIs of the knee of a 32-year-old man 32 months after hamstring tendon harvest. (A) T1-weighted axial image at the level of the superior pole of the patella shows a normal sartorius muscle (arrow 1). The semitendinosus tendon is also normal at this level (arrow 2). The gracilis is not seen at this level. (B) T1-weighted axial image below the level of the joint line again shows normal appearing sartorius and semitendinosus tendons (arrows 1 and 2, respectively). No distinct gracilis tendon is seen. (C) T1-weighted image at the level of the tibial attachment of the pes anserine shows a normal appearing attachment (arrow).
finding appears to be unrelated to the time interval between surgery and imaging. DISCUSSION Hamstring tendons have been shown to have adequate strength13 and incorporation capability14 to function as an ACL graft and have done well in intermediate-term studies.15 Patellar tendon grafts have been associated with relatively high morbidity,
including patellofemoral pain and chondrosis,16-18 patellar fractures,19,20 patellar tendon rupture,21,22 and even combined fracture and tendon rupture.23,24 Hamstring graft harvest, however, is reported as having a very low associated morbidity.25,26 The patellar tendon has been shown to reconstitute its central one third following bone–patellar tendon– bone harvest.4,5 Recent studies on the healing of central third patellar tendon donor sites in a goat model have raised questions as to the advisability of a second
MRI OF HAMSTRING TENDON HARVEST SITES TABLE 1. Cross-Sectional Measurements of Hamstring Muscles
Sartorius Proximal Nonoperated Harvest site Distal Nonoperated Harvest site Gracilis Proximal Nonoperated Harvest site Distal Nonoperated Harvest site Semitendinosus Proximal Nonoperated Harvest site Distal Nonoperated Harvest site
side
side
side
side
side
side
Mean Area (mm)
SD
P Value
0.95 0.94
⫾0.13 ⫾0.14
.64
0.91 1.00
⫾0.26 ⫾0.27
.49
1.00 0.86
⫾0.26 ⫾0.16
.11
0.84 0.70
⫾0.35 ⫾0.30
.01 (significant)
0.87 0.88
⫾0.23 ⫾0.23
.64
0.98 0.96
⫾0.12 ⫾0.13
.81
NOTE. Measured cross-sectional areas on MRIs of sartorius, gracilis, and semitendinosus, proximally (between the superior pole of the patella and the level of the joint line) and distally (below the level of the joint line). Significance of differences between the harvest site and the nonoperated side of patients with previous ACL reconstruction (P ⱕ .05 defined as significant difference).
harvest of the same tendon because it has been shown to be biomechanically inferior.4 The normal gracilis and semitendinosus tendons lie invested in fascia of the medial aspect of the knee, between layers I and II as described by Warren and Marshall.27 The tendons become conjoined as they insert into the pes anserinus. The tendons exist as distinct structures at an average of 18 mm proximal to this insertion.28 The tendons lie adherent to the deep layer of the thigh fascia. This fascia, 8 to 10 cm proximal to the pes anserinus, forms a 3- to 4-cm band around the tendons.15 This area has been shown to connect with the medial intermuscular septum and the sheath invests the semitendinosus. The tissue plane for the semitendinosus (and to a lesser extent the gracilis) may help to explain the regeneration of tissue in the normal anatomic location of the harvested tendon. We theorize that the tendons first regenerate proximally in a more vascularized area, which then proceeds distally along the fascial planes. Although these MRI findings seem to support this theory, a true prospective study with repeated MRI measurements on the same individuals would be required for confirmation. This phe-
7
nomenon may be analogous to nerve regrowth through intact neural sheaths. Hamstring tendon regrowth undoubtedly resembles normal tendon because it has been reharvested in select cases. With the exception of the gracilis muscle at the mid-thigh level, the hamstring muscles show no significant decrease in the cross-sectional area in the operative versus the nonoperative extremity. This may help explain why, as shown in previous studies,10 the hamstring strength returns to nearly normal values during the first 3 months following tendon harvest. Potential weaknesses of this study are as follows. Although the MRIs were obtained and interpreted in a prospective fashion, sets of images at each time interval were performed on a different group of patients. It may be beneficial to image the same group of patients sequentially at specific intervals following hamstring tendon harvest to determine the true time course of tendon regeneration. Another potential weakness is that, although the tendons on MRI regain normal morphology and signal characteristics over time, the exact histologic composition of these regenerated tendons remains unknown. The structures seen on MRI in the location of the native tendons may represent a structurally inferior tendon or scar tissue. Furthermore, the functional ability of these tendons and their potential for reharvest remains uncertain. We plan further studies using an in vivo animal model to help answer these questions. In conclusion, we found in this study that, after hamstring tendon harvest, MRI shows an apparent regeneration of the tendons beginning proximally and extending distally over time. By 15 months, the tendons have normal MR signal characteristics on T1and T2-weighting, normal morphology, and normal location to the level of the distal 1 to 2 cm of the native tendon. It remains uncertain what the functional ability of these regenerated tendons is, and further studies are needed to help resolve this issue. REFERENCES 1. Brown CH Jr., Steiner ME, Carson EW. The use of hamstring tendons for anterior cruciate ligament reconstruction. Technique and results. Clin Sports Med 1993;12:723-756. 2. Brown CH Jr, Carson EW. Revision anterior cruciate ligament surgery. Clin Sports Med 1999;18:109-171. 3. Day B. Revision anterior cruciate ligament reconstruction after failed bone patellar tendon bone reconstruction. Arthroscopy 1996;12:381-382 (abstr). 4. LaPrade RF, Hamilton CD, Montgomery RD, Wentoff F, Hawkins HD. The reharvested central third of the patellar tendon: A histological and biomechanical analysis. Am J Sports Med 1997;25:779-785. 5. Proctor CS, Jackson DW, Simon TM. Characterization of the repair tissue after removal of the central one-third of the
8
6.
7. 8.
9.
10. 11.
12.
13.
14.
15. 16.
D. M. RISPOLI ET AL. patellar ligament. An experimental study in a goat model. J Bone Joint Surg Am 1997;79:997-1006. Kartus J, Stener S, Lindahl S, Eriksson BI, Karlsson J. Ipsi- or contralateral patellar tendon graft in anterior cruciate ligament revision surgery. A comparison of two methods. Am J Sports Med 1998;26:499-504. Karns DJ, Heidt RS Jr, Holladay BR, Colosimo AJ. Revision anterior cruciate ligament reconstruction. Arthroscopy 1994; 10:148-151. Cross MJ, Roger G, Kujawa P, Anderson IF. Regeneration of the semitendinosus and gracilis tendons following their transection for repair of the anterior cruciate ligament. Am J Sports Med 1992;20:221-223. Erickson K, Larson H, Wredmark T, Hamberg P. Semitendinosus tendon regeneration after harvesting for ACL reconstruction. a prospective MRI study. Knee Surg Sports Traumatol Arthrosc 1999;7:220-225. Iacono F, Loreti I, Maracci M, et at: Evaluation of hamstring strength following gracilis and semitendinosus reconstruction at 3 months follow-up. Arthroscopy 14:526,1998 (abstr). Simonian PT, Harrison SD, Cooley VJ, Escabedo EM, Deneka DA, Larson RV. Assessment of morbidity of semitendinosus and gracilis tendon harvest for ACL reconstructions. Am J Knee Surg 1997;10:54-59. Papandrea P, Vulpiana MC, Conteduca F, et al. Regeneration of the semitendinosus following its use for anterior cruciate ligament reconstruction: Ultrasound evaluation. Presented at the Annual Meeting of the American Academy of Orthopaedic Surgery, Anaheim, CA, March 1999. Recker AT, Brown CR, Deffner KT, et at: Tensile properties of young multiple stranded hamstring tendon grafts. Presented at the American Orthopaedic Society for Sports Medicine Specialty Day, Book of Abstracts p 8. San Francisco, February 1997. Pinczewski LA, Clingeleffer AJ, Otto DD, Bonar SF, Cory IS. Integration of hamstring tendon graft with bone in reconstruction of the anterior cruciate ligament. Arthroscopy 1997;13: 641-643. Siegel MG, Barber-Westin SD. Arthroscopic-assisted outpatient anterior cruciate ligament reconstruction using the semitendinosus and gracilis tendons. Arthroscopy 1998;14:268-277. Aglietti P, Buzzi R, D’Andria S, Zaccherotti G. Patellofemoral
17. 18.
19. 20. 21.
22.
23. 24. 25. 26. 27. 28.
problems after intraarticular anterior cruciate ligament reconstruction. Clin Orthop 1993;288:195-204. Sachs RA, Daniels DM, Stone ML, Garfein RF. Patellofemoral problems after anterior cruciate ligament reconstruction. Am J Sports Med 1989;17:760-765. Shino K, Nakagawa S, Inoue M, Horibe S, Yoneda M. Deterioration of patellofemoral articular surfaces after anterior cruciate ligament reconstruction. Am J Sports Med 1993;21:206211. Christen B, Jakob RP. Fractures associated with patellar ligament grafts in cruciate ligament surgery. J Bone Joint Surg Br 1992;74:617-619. McCarroll JR. Fracture of the patella during a golf swing following reconstruction of the anterior cruciate ligament: A case report. Am J Sports Med 1983;11:26-27. Bonamo JJ, Krinick RM, Sporn AA. Rupture of the patellar ligament after use of its central third for anterior cruciate reconstruction. A report of two cases. J Bone Joint Surg Am 1984;66:1294-1297. Marumoto JM, Mitsunaga MM, Richardson AB, Medoff RJ, Mayfield GW. Late patellar tendon rupture after removal of the central third for anterior cruciate ligament reconstruction. A report of two cases. Am J Sports Med 1996;24:698-701. Bonatus TJ, Alexander AH. Patellar fracture and avulsion of the patellar ligament complicating arthroscopic anterior cruciate ligament reconstruction. Orthop Rev 1991;20:770-774. Miller MD, Nichols T, Butler CA. Patella fracture and proximal patellar tendon rupture following arthroscopic anterior cruciate ligament ronstruction. Arthroscopy 1999;15:640-643. Larson RV, Erickson D: Complications in the use of hamstring tendons for anterior cruciate ligament reconstruction. Sports Med Arthrosc Rev 23:83-90, 1997 Yasuda K, Tsujino J, Ohkoshi Y, Tanabe Y, Kaneda K. Graft site morbidity with autogenous semitendinosus and gracilis tendons. Am J Sports Med 1995;23:706-714. Warren LF, Marshall JL. The supporting structures and layers on the medial side of the knee: an anatomical analysis. J Bone Joint Surg Am 1979;61:56-62. Pagnani MJ, Warner JJ, O’Brien SJ, Warren RF. Anatomic considerations in harvesting the semitendinosus and gracilis tendons and a technique of harvest. Am J Sports Med 1993; 21:565-571.