Biomechanical and histologic properties of the canine patellar tendon after removal of its medial third

Biomechanical and histologic properties of the canine patellar tendon after removal of its medial third

Lucinder H. Linder

Autogenous reconstruction of a torn or otherwise damaged ACL is a common procedure today. Although this approach has been used for the past 75 years, [15] questions persist about which tissue to use [20] and whether removal of the graft influences knee and limb function. In 1963, Jones [17] outlined a surgical procedure using the central third of the patellar tendon as an ACL autograft. Since then, several variations of the Jones procedure have been described using either the central or medial third of the patellar tendon.[1,10,14,18,19,21] The most popular reconstruction is with the central one-third portion of the patellar tendon since the strength and stiffness of a 14-mm wide portion of tendon exceed that of the ACL. [20] The patellar tendon is also favorable because of the bony ends for bone-to-bone healing. [22]

Most experimental studies of patellar tendon reconstruction for ACL-deficient knees have focused solely on the fate of the intraarticular graft. Cabaud et al. [8] and Burks et al. [7] have examined the remaining patellar tendon after harvesting one third for ACL reconstruction. In the Cabaud study, the ACL was reconstructed with the medial third of the patellar tendon in 11 dogs. Biomechanical tests were conducted on the remaining patellar tendon. The dogs were immobilized for 6 weeks postoperatively and then allowed unrestricted activity in a run. Patellar tendons from 6 dogs were examined at 4 months and showed only a slight decrease in maximum load compared with contralateral controls. In 5 knees tested at 8 months, analysis indicated that the operated patella-patellar tendon-tibia preparation was stronger than contralateral controls. Histologic evaluation of the operated tendons showed microscopically normal collagen. The tendons showed no abnormalities on gross inspection except for a thickening of the undersurface where the fat pad had been removed.

Burks et al. [7] studied the host patellar tendon after removal of its central third in 25 adult mongrel dogs. The dogs were allowed free use of their limbs postoperatively. Unlike Cabaud et al., [8] Burks et al. [7] documented a significant increase in cross-sectional area of the operated tendon at 6 months when compared with contralateral controls. The failure load (strength) was 70% of controls at 3 months and 60% of controls at 6 months. Also at 6 months, the structural stiffness and tensile modulus of the operated tendon, within the physiologic range, were dramatically reduced to 70% and 33% of controls, respectively. The defect was completely filled in with scar tissue on gross inspection.

The conclusions from the study by Cabaud et al. suggest better surgical results after removal of the medial versus central portion of the patellar tendon. However, variations in the test and data analyses make it difficult to directly compare these two studies. We undertook a study using the same surgical procedures, postsurgical housing, and mechanical test and data analysis procedures as Burks et al. The exception was that the medial third of the patellar tendon was removed rather than the central third as in the earlier study.


This study comprised 19 adult mongrel dogs with an average weight of 27.7 [+ or -] 5.6 kg. The dogs were divided into 3 groups. The first group (time zero dogs) consisted of 5 dogs for immediate sacrifice and testing of the patellar tendon after removal of its medial third. The contralateral limb served as a control. The other 2 groups were sacrificed at 3 and 6 months and consisted of 7 dogs each. One dog from each group was used for histologic testing, and the remaining 6 from each group were used in mechanical tests.

The surgery, tissue preparations, and testing procedures were the same as those used by Burks et al. [7] with two exceptions. The first exception was that the medial third of the tendon was removed rather than the central third as in the Burks et al. study. The second was that the specimens were wrapped in phosphate-buffered saline-soaked gauze and shipped overnight in refrigerated packaging to Michigan State University for mechanical tests. In the Burks et al. study, specimens were hand-delivered several hours after extraction. The experiments in the current study were performed within 30 hours of sacrifice.

Failure mechanisms were determined from direct posttest examination of the preparation. The elongation and force data were used to determine the following properties: tangent stiffness of the preparation before failure and maximum force and energy absorbed to failure. Structural stiffness was determined by the slope of the load-elongation response just before failure in the midrange, where a linear response was observed between approximately 2 and 8 mm of deformation. [7] These data were normalized by initial length and average cross-sectional area of the tendon to yield stress-strain responses and tangent modulus.

The Student’s t-test for paired observations was used for comparisons between operated and contralateral control preparations (Systat, Systat, Inc., Evanston, IL). One-way analysis of variance (ANOVA) with subsequent contrast post hoc tests was used to determine statistical differences between the control and test data for each variable. Statistical significance was set at a level of P < 0.05. Experimental data were documented as the mean [+ or -] standard deviation.


Tissue dimensions

There was no difference in the length of the control tendons between groups. Patellar tendons from the control limb of all groups averaged 33.7 [+ or -] 2.3 mm (N = 17) in length (Table 1). At time zero, no difference between the lengths of the control and test patellar tendons was noted. The operated tendon was, on average, 11% shorter than contralateral controls 3 months after surgical removal of the medial third. Operated tendons were 6% shorter than controls 6 months after the surgery. The operated tendon lengths were statistically shorter than controls at 3 and 6 months.

The average cross-sectional area of the control patellar tendons from all groups was an average of 33.6 [+ or -] 8.7 [mm.sup.2]. The average thickness and width was 2.7 [+ or -] 0.4 mm and 12.5 [+ or -] 1.4 mm, respectively. The cross-sectional area of the patellar tendon was decreased 27%, reflecting a decrease in the width dimension after removal of the medial third. Three months after surgery, the cross-sectional area averaged 275% of controls. In two animals with extensive scarring the cross-sectional area averaged 424% [+ or -] 20% of controls, while in the remaining four the cross-sectional area was 227% [+ or -] 100% of controls. The large increase in cross-sectional area for these two particular animals was not specifically caused by an increase in the thickness or width, per se. Six months after surgery the cross-sectional area of the patellar tendon was, on average, 288% of controls. In one tendon exhibiting extensive scar, the cross-sectional area was 487% of the contralateral control, compared with 249% [+ or -] 73% of controls for the remaining five in this group. No differences in postsurgical behavior could explain the extensive degree of scarring in these few animals.

Histologic observations

One animal from each of the 3- and 6-month groups was designated for histologic study. These appeared to be representative of the majority of tendons. Control tendons were difficult to section (Fig. 1). At the time of harvest, the 3-month control and experimental tendons appeared grossly normal; the experimental ones were slightly larger. On microscopic examination of the experimental tissue, normal collagen was noted in nearly half of the tendon (Fig. 2). Remodeled collagen was evident at one side and extended along the upper surface, overriding a portion of normal-appearing collagen. There was an area of “transitional” collagen between the original and the neotendon that was hypercellular and disorganized (Fig. 3). Numerous small blood vessels were seen in areas of newly synthesized collagen.

At 6 months, there was a moderate increase in thickness of the experimental tendon compared with its contralateral control. Histologically, approximately half of the structure was original tendon and distinct from newly synthesized areas of neotendon. The collagen in the “transition” area was organized into bundles somewhat larger than the original tendon bundles. This formation was also noted in other areas of the scar. While still hypercellular, the nuclei had a more elongated shape approaching the general appearance of normal tendon (Fig. 4). Little adipose tissue was noted. Control tendons at 3 and 6 months were normal in appearance.

Tensile response data

The tensile response curves for all preparations qualitatively appeared similar, exhibiting an initial nonlinear region, a range of linear response, and a short range of nonlinearity immediately before an abrupt unloading and gross structural failure of the preparation. The tangent stiffness of control specimens was 323 [+ or -] 42 kN/m (N = 17). Interestingly, the stiffness of control tendons was statistically greater after surgery (Table 2). After removal of the medial third the tensile response at time zero was significantly different than controls for deformations above approximately 4 mm (Fig. 5). This was reflected in the values of tangent stiffness, which were 29% less than contralateral controls at time zero. At 3 and 6 months after surgery, the tangent stiffness of the operated patellar tendon preparations was on the average 85% and 91%, respectively, of controls. These parameters were not statistically less than the control preparations.

When these data were normalized for initial length and cross-sectional area to produce stress-strain data, the tensile responses of operated tendons were shown to be significantly lower than control and time zero responses (Fig. 6). The tensile modulus of control specimens was 340 [+ or -] 100 MPa (N = 17), this was not statistically altered by surgical removal of the medial third. At 3 and 6 months after the operation, the modulus of operated tendons was statistically less: 31% and 28% of controls, respectively (Table 2).

Tensile failure mechanisms

All specimens were stretched sufficiently to cause a readily observable failure of the patella-patellar tendon-tibia preparation. In the time zero specimens, all control preparations failed by avulsion of the patellar tendon at the patella. In the time zero preparations that had removal of the medial third patellar tendon, avulsion fractures were evident in three of five cases. In the remaining two specimens from this group, failure was in the substance, and it appeared to originate from the insertion points without bone avulsion.

In six of the six cases at 3 months and five of six cases at 6 months, control preparations failed by avulsion of bone from the distal patella. On the operated sides, avulsion fractures occurred in five of six cases at both 3 and 6 months after surgery (Table 2).

Structural failure properties

Since the preparation did not consistently fail in the substance of the patellar tendon, it is only appropriate to discuss structural rather than material failure properties of the tendon (Table 3). The average load required to cause failure in the control preparations was 2.7 [+ or -] 0.6 kN (N = 17), and this was not significantly different between groups. After removal of the medial third of the tendon, the load to failure for the preparation decreased 29%, on average. Three months after surgery the operated tendons failed at a load that was 83% of controls, and at 6 months the failure load was 79% of contralateral controls. The failure loads were statistically lower than controls on the operated preparation at 6 months. The energy to failure for control preparations was 13.7 [+ or -] 4.5 J (N = 17). Immediately after surgical removal of the medial third, the energy required to fail the tendon was decreased 26%, but the difference between test and control tendon preparations was not statistically significant. The energy required to cause failure for the preparation was decreased 15% and 17% at 3 and 6 months postoperatively, respectively, but these changes were not statistically significant.


The primary objective of the current study was to repeat the experiment of Cabaud et al. [8] using the same procedures as Burks et al. [7] A review of these two studies indicated several significant differences. Firstly, Cabaud et al. showed that the network of collagen in the host tendon appeared histologically normal after surgery, whereas Burks et al. indicated a significant amount of hypertrophic scar that encompassed the entire cross-section. Secondly, Cabaud et al. documented that the strength and stiffness of the host tendon equalled or exceeded that of the contralateral control tendon 8 months after surgery, while Burks et al. showed 21% and 9% reductions in these parameters, respectively, at 6 months after surgery.

The current study showed that removal of a medial third of the patellar tendon resulted in significant scar tissue. Interestingly, the maximum force was decreased only 20% and the structural stiffness was depressed only 10% 6 months after surgery compared with controls. The normalized mechanical properties of the operated tendon, however, were significantly decreased compared with controls at 6 months. This probably reflects the poor quality of alignment and individual physical properties of newly synthesized collagen in the bulk of the host tendon after surgery. These results are in contrast to those of Cabaud et al. One major difference in the experimental protocols of the two studies was the period of cast immobilization used by Cabaud et al. It is uncertain whether activity of the dogs immediately after surgery could have influenced the differences between our study and that or Cabaud et al. Many reports have been published on the effects of immobilization and early exercise on the strength of ligaments and tendons. [2,25] Perhaps early immobilization is an advantage, protecting the initially weakened patellar tendon from developing excessive scar tissue. Yet, there are many known detrimental effects of such immobilization. [2,16,25] Frank et al. [13] recently showed in a healing model of the medial collateral ligament that scar tissue recovered to normal alignment after 14 weeks in the nonimmobilized limb. Surprisingly, they found that the immobilized limbs were even better, with the mean alignment values being statistically within normal limits at 3, 6, and 14 weeks after surgery. Future research must be directed toward more knowledge and better understanding of the consequences of postsurgical immobilization and exercise on the healing patellar tendon. It is clear that immobilization has negative effects, but interestingly, immediate postsurgical exercise causes excessive hypertrophic scar tissue. During the early stages of healing, collagen fibers are weak. What the tendon seems to lack in tensile quality it attempts to make up in quantity of tissue produced during early healing.

The current study indicates relatively better results after removal of the medial third versus the central third using surgical and experimental protocols similar to Burks et al. [7] Failure load, stiffness, and modulus of the tendon were all statistically closer to their controls after removal of the medial third than after removal of the central third. The explanation for this interesting result is yet unknown to us. We have, however, formulated a number of explanations that we believe warrant further study. One potential factor, suggested earlier by Burks et al., [7] is the process of making two incisions in the central procedure versus making one incision for the medial procedure. Two incisions have the potential for damaging two surfaces of collagen versus one surface in the medial procedure. The damaged collagen is broken down enzymatically, as are some of the newly formed elements. Phagocytes digest damaged and new collagen up to 5 mm from the incision line. [24] The degree of damage generated in the fat pad and surrounding blood supplies in the central versus medial procedure is also unknown. It is possible that removal of the central portion of the patellar tendon may cause greater damage to the fat pad than removing the medial third. Studies have shown that the tendon does not receive its nutrients from its osseous attachment; instead, revascularization originates from the soft tissues of the infrapatellar fat pad and posterior soft tissues of the joint. [3,4]

Another potential factor not investigated to date may deal with the transfer of mechanical load across the patella and patellar tendon after removal of the central versus medial portion of the tendon. The loads existing in the defect can play a significant role in the rate of healing and alignment of collagen in the scar. The distribution of load across the patellar tendon is not currently known for either the human or canine model. Chun et al., [9] however, have recently documented spatial variations across the width of the human patellar tendon. In their study, they showed that the tensile modulus of the central portion of the tendon exceeds that of the medial portion by a factor of 2. Noyes et al. [20] examined 10 different ligament graft tissues and also determined that the central third of the patellar tendon was stiffer and stronger than the medial third. From an elementary strength of materials analysis, removal of the central (stiffer) portion would transfer more load to the original tendon substance than removal of the medial portion. We also found that relatively more tissue was removed during the earlier central procedure than during the current medial procedure. While these data were based on time zero results from the 2 studies, we have speculated that the surgeons were consistent for specimens in the 3- and 6-month groups for both studies. The cross-sectional area of host tendon with the central procedure was thereby more reduced, resulting in larger stresses in the host two thirds compared with stress levels in the remaining tendon after the medial procedure. These larger stresses may influence the healing process to produce excessive scar, as has been observed in past studies. [11, 12, 23] The extent of postsurgical scarring may depend on how much the host is compromised during surgery. The resultant effects of these increased levels of stress on the host tendon are currently unknown but warrant future study.

As a final point of interest, we found that the structural stillnesses of time zero control tendons were statistically less than those at 3 and 6 months after both surgical procedures. If the dogs favored the operated limb, the loads on the control limb may have been enhanced, thereby resulting in an increase in the stiffness of control tendons after surgery. [25] While altered gait patterns were not observed, it is possible that a more refined gait analysis may pick up alterations in limb and joint function after autogenous reconstruction of the torn ACL with patellar tendon. [5,6] The results may help determine negative consequences on the biomechanics of limb function after surgical removal of a portion of the patellar tendon to repair a torn or otherwise damaged ACL.


1. In a dog model, removal of the medial third patellar tendon caused decreased length, failure load, stiffness, modulus, and failure energy of the tendon when compared with controls.

2. There was a significantly greater cross-sectional area of the operated tendon compared with controls.

3. The difference between control and test tendons was statistically larger after removal of the central third (Burks) in terms of the failure load, stiffness, and modulus of the tendons.

4. The differences between removing the central and medial thirds might, in part, be explained by the fact that more tissue was removed during the earlier study after central third extraction than during the current medial third extraction. One procedure may not necessarily be better than another from the host tendon’s perspective. The amount of tissue removed during surgery may be a primary factor along with the postsurgical rehabilitation program.


The authors acknowledge the help of Ms. Jane Walsh, Department of Biomechanics, Michigan State University, in the preparation and analysis of the histologic sections. We also thank Dr. James Render, DVM, PhD, Department of Pathology, for his help in interpretation of histology. We thank Ms. Tammy Haut for help on the statistical analysis and presentation of experimental data for this study. The research was supported by grants from the Orthopaedic Research and Education Foundation and the Foundation for Sports Medicine Education and Research and the Centers for Disease Control (R49/CCR-503607).


1. Alm A, Gillquist J: Reconstruction of the anterior cruciate ligament by using the medial third of the patellar ligament. Acta Chir Scand 140: 289-296, 1974

2. Amiel D, Woo SL-Y, Harwood FL, et al: The effect of immobilization on collagen turnover in connective tissue: A biochemical-biomechanical correlation. Acta Orthop Scand 53: 325-332, 1982

3. Arnoczky SP, Rubin RM, Marshall JL: Microvasculature of the cruciate ligaments and its response to injury. J Bone Joint Surg 61A: 1221-1229, 1979

4. Arnoczky SP, Tarvin GB, Marshall JL: Anterior cruciate ligament replacement using patellar tendon. An evaluation of graft revascularization in the dog. J Bone Joint Surg 64A: 217-224, 1982

5. Budsberg SC, Verstraete MC, Soutas-Little RW, et al: Force plate analyses before and after stabilization of canine stifles for cruciate injury. Am J Vet Res 49: 1522-1524. 1988

6. Budsberg SC, Verstraete MC, Soutas-Little RW: Force plate analysis of the walking gait in healthy dogs. Am J Vet Res 48: 915-918, 1987

7. Burks RT, Haut RC, Lancaster RL: Biomechanical and histological observations of the dog patellar tendon after removal of its central one-third. Am J Sports Med 18: 146-153, 1990

8. Cabaud HE, Feagin JA, Rodkey WG: Acute anterior cruciate ligament injury and augmented repair. Am J Sports Med 8: 395-401, 1980

9. Chun KJ, Butler DL, Bukovec DB, et al: Spatial variations in material properties in fascicle-bone units from human patellar tendon. Trans Orthop Res Soc 14: 214, 1989

10. Clancy WG Jr: Intra-articular reconstruction of the anterior cruciate ligament. Orthop Clin North Am 16: 181-189, 1985

11. Dahners LE, Torke MD, Gilbert JA, et al: The effect of motion on collagen synthesis, DNA synthesis and fiber orientation during ligament healing. Trans Orthop Res Soc 14: 299, 1989

12. Frank CB, Amiel D, Woo SLY, et al: Normal ligament properties and ligament healing. Clin Orthop 196: 15-25, 1985

13. Frank C, MacFarlane B, Edwards P, et al: A quantitative analysis of matrix alignment in ligament scars: A comparison of movement versus immobilization in an immature rabbit model. J Orthop Res 9: 219-227, 1991

14. Fried JA, Bergfeld JA, Weiker G, et al: Anterior cruciate reconstruction using the Jones-Ellison procedure. J Bone Joint Surg 67A: 1029-1033, 1985

15. Hey Groves EW: Operation for the repair of the cruciate ligaments. Lancet 2: 674-675, 1917

16. Inoue M, Woo SLY, Gomez MA, et al: Effects of surgical treatment and immobilization on the healing on the medial collateral ligament: A long-term multidisciplinary study. Conn Tiss Res 25: 13-26, 1990

17. Jones KG: Reconstruction of the anterior cruciate ligament. A technique using the central one-third of the patellar tendon. J Bone Joint Surg 45A: 925-932, 1963

18. Lambert KL: Vascularized patellar tendon graft with rigid internal fixation for anterior cruciate ligament insufficiency. Clin Orthop 172: 85-89, 1983

19. Marshall JL, Warren RF, Wickiewicz TL, et al: The anterior cruciate ligament: A technique of repair and reconstruction. Clin Orthop 143: 97-106, 1979

20. Noyes FR, Butler DL, Grood ES, et al: Biornechanical analysis of human ligament grafts used in knee ligament repairs and reconstructions. J Bone Joint Surg 66A: 344-352, 1984

21. Paterson FWN, Trickey EL: Anterior cruciate ligament reconstruction using part of the patellar tendon as a free graft. J Bone Joint Surg 68B: 453-457, 1986

22. Paulos LE, Butler DL, Noyes FR, et al: Intra-articular cruciate reconstruction II: Replacement with vascularized patellar tendon. Clin Orthop 172: 78-84, 1983

23. Viidik A: Biomechanical behavior of soft connective tissue, in Akkas N (ed): Progress in Biomechanics. Rockville, MD, Siijthoff & Noordhoff, 1979, 75-108

24. Viidik AV: Gottrup F: Mechanics of healing soft tissue wounds, in ScmidSchonbein G, Woo S, Zweifach B (eds): Frontiers in Biomechanics, New York, Springer-Verlag, 263-278

25. Woo SLY, Gomez MA, Woo YK, et al: Mechanical properties of tendons and ligaments. II. The relationships of immobilization and exercise on tissue remodeling. Biorheology 19: 397-408, 1982


COPYRIGHT 1994 American Orthopaedic Society for Sports Medicine

COPYRIGHT 2004 Gale Group