Investigation Performed at Imperial College London, London, United Kingdom

1 Length change patterns of the lateral extra-articular structures of the knee and related 2 reconstructions.

3 C Kittl1,2, C Halewood1, JM Stephen1, CM Gupte3, A Weiler4, A Williams5, AA Amis1,3

4 Investigation performed at Imperial College London, London, United Kingdom

5 1 The Biomechanics Group, Department of Mechanical Engineering, Imperial College London, UK.

6 2Department of Trauma Surgery, Landeskrankenhaus Steyr, 4400 Steyr, Austria

7 3The MSk Lab, Department of Surgery and Cancer, Imperial College London, UK.

8 4Sporthopaedicum Berlin, 10627 Berlin, Germany

9 5Fortius Clinic, 17 Fitzhardinge Street, London W1H 6EQ, UK.

11 Correspondence:

12 Prof Andrew Amis

13 The Biomechanics Group

14 Department of Mechanical Engineering

15 Imperial College London

16 London SW7 2AZ

17 United Kingdom

18 Tel +44 (0)20 7594 7062

19 [email protected] 2

20 Abstract

21 Background: Lateral extra-articular soft-tissue reconstructions in the knee may be used as a

22 combined procedure in revision anterior cruciate ligament surgery, and in primary treatment of

23 patients who demonstrate excessive anterolateral rotatory instability. Only a few studies examining

24 length change patterns and isometry in lateral extra-articular reconstructions have been published.

25 Purpose: To determine a recommended femoral insertion area and graft path for lateral extra-

26 articular reconstructions by measuring length change patterns through a range of knee flexion

27 angles of several combinations of tibial and femoral insertion points on the lateral side of the knee.

28 Study design: Controlled laboratory study.

29 Methods: Eight fresh-frozen cadaver knees were freed of skin and subcutaneous fat. The knee was

30 then mounted in a kinematics rig, which loaded the quadriceps muscles and simulated openchain

31 knee flexion. The length changes of several combinations of tibio-femoral points were measured at

32 knee flexion angles between 0° and 90° using linear variable displacement transducers .The changes

33 in length relative to the 0° measurement were recorded.

34 Results: The anterior fiber region of the iliotibial tract displayed a significantly different (P<.001)

35 length change pattern compared to the posterior fiber region. The reconstructions that had a

36 femoral insertion site located proximal to the lateral epicondyle and with the grafts passed deep to

37 the lateral collateral ligament displayed similar length change patterns to each other, with small

38 length increases during knee extension. They also showed a significantly lower total strain range

39 compared to the reconstruction located anterior to the epicondyle (P<.001).

40 Conclusion: These findings show that the selection of graft attachment points and graft course affect

41 length change pattern during knee flexion. A graft attached proximal to the lateral femoral

42 epicondyle and running deep to the lateral collateral ligament will provide desirable graft behavior,

43 such that it will not suffer excessive tightening or slackening during knee motion. 3

44 Clinical relevance: These results provide a surgical rationale for lateral extra-articular soft-tissue

45 reconstruction in terms of femoral graft fixation site and graft route.

46 Key terms: Lateral extra-articular soft-tissue reconstruction, anterolateral rotatory instability,

47 isometry, length change pattern, knee, anterior cruciate ligament, anterolateral ligament.

48 What is known about the subject: The existing literature on lateral extra-articular soft-tissue

49 reconstruction isometry is inconsistent and authors have addressed various femoral insertion points,

50 which are to some extent not clinically applicable.

51 What this study adds to existing knowledge: The present study reports on length change patterns

52 and total strain range values of various tibio-femoral point combinations on the lateral side of the

53 knee. This is used to recommend a graft course and femoral insertion area in lateral soft-tissue

54 extra-articular reconstructions. This is the first study to address the actual graft path deep to the LCL

55 rather than calculating a theoretical straight line distance. It also questions the biomechanical

56 rationale of some suggested techniques. 4

57 Introduction

58 Anterolateral rotatory instability (ALRI) is a combined anterior translational and internal rotational

59 movement of the tibia, following injury to the anterior cruciate ligament (ACL) and the anterolateral

60 structures of the knee.13, 21, 29 Injuries to the mid-third lateral capsular ligament, the lateral meniscus,

61 the capsulo-osseus and deep layers of the ilio-tibial tract (ITT) and the biceps femoris muscle

62 complex have been suggested to cause ALRI in combination with an ACL tear.44, 45 The concept of the

63 ligamentous capsulo-osseus layer, together with the ACL, forming a ‘horseshoe’, or sling, around the

64 lateral femoral condyle thereby preventing anterior subluxation of the lateral tibial plateau has been

65 described by several authors.46, 49 This structure was considered to ‘act as an anterolateral ligament’.

66 Recently, four independent research groups have identified different distinct capsular and extra-

67 capsular structures on the anterolateral side of the knee, and three of them suggested a possible link

68 between damage to these structures, ALRI and the Segond fracture.6, 8, 20, 50 While each group named

69 their identified structure the ‘anterolateral ligament’ (ALL), three different femoral attachment sites

70 were reported: Claes et al.6 and Vincent et al.50 described the femoral attachment site anterior and

71 distal to the LCL insertion, Dodds et al.8 found it proximal and posterior. Thus, there are

72 inconsistencies in the anatomical literature describing the lateral aspect of the knee.

73 Lateral extra-articular procedures are sometimes used in revision ACL surgery47 and in primary cases

74 displaying excessive ALRI,33, 51 following a combined injury of the ACL and the peripheral structures

75 and it is judged that an isolated intra-articular ACL reconstruction may prove to be inadequate to

76 control rotational instability. Such reconstructions typically involve routing a strip of the ITT left

77 attached to Gerdy’s tubercle, which is passed deep to the lateral (fibular) collateral ligament (LCL),

78 before being attached to the lateral femur. 5, 25 This technique provides a lateral ‘check-rein’ against

79 anterior tibial subluxation, by positioning the graft posterior to the transverse axis of rotation

80 throughout the entire range of motion.5 However, these extra-articular reconstructions have lost

81 popularity, due to a number of perceived and actual drawbacks: excessive constraint of internal 5

82 tibial rotation,11, 12, 15, 32 failure to restore normal AP stability,11, 12 alteration of kinematics,14, 15 and

83 unsatisfactory clinical results7, 17, 24, 38 even in combination with an intra-articular procedure.3, 35, 39, 43

84 On the contrary, other authors have found reduced internal rotational laxity and good clinical

85 outcomes after extra-articular reconstructions,4, 33 both at follow up26, 31, 34, 40, 48 and at the time of

86 revision surgery.47 Long plaster cast immobilizations, non-isometric graft positioning and graft

87 tensioning may have contributed to the poor results in the past.9 Also, of course, historically these

88 lateral extra-articular soft-tissue reconstructions were often employed without simultaneous intra-

89 articular ACL reconstruction, thereby leaving a major ligamentous restraint unaddressed.

90 The principle of isometry indicates a constant distance between two moving points, where the

91 points are on either side of a joint. Exactly isometric behavior rarely exists and has not been found

92 for ACL36 or lateral extra-articular soft-tissue reconstructions.41 However, it is widely accepted that a

93 degree of isometry of a ligament reconstruction reduces the likelihood of unwanted graft behavior. 1

94 Inappropriate graft positioning and tensioning of a lateral extra-articular soft-tissue reconstruction

95 may excessively stretch the graft at certain knee flexion angles. This may over-constrain the lateral

96 compartment of the tibiofemoral joint and ultimately lead to graft failure, excessive compressive

97 load on the articular cartilage in the lateral compartment, and compromise graft healing to the

98 surrounding bone.2 It has been found that an increase in separation distance between insertion

99 points of just 6% could lead to permanent graft stretching. 37 Conversely, if a graft becomes slack at a

100 particular knee flexion angle, it may not be able to adequately replicate the function of the

101 reconstructed ligament. Only a few studies of the isometry of lateral extra-articular reconstructions

102 have been published, with none of these considering a graft passing deep to the LCL.10, 22, 23, 41 Given

103 that a lateral reconstruction is intended to stabilize the weight-bearing knee, and that the knee has

104 greater rotational laxity when flexed, it may be desirable for a graft to tend to be longer/tighter in

105 extension and shorter/slacker in flexion. 6

106 The aims of the present study were to: 1. Determine the effect of changing tibial and femoral

107 attachment points on ligament and graft length change pattern through knee range of motion. 2.

108 Investigate the effect of altering the path of the graft in relation to the LCL (superficial or deep). 3.

109 Examine the length change pattern of native tissue structures, extra-articular soft-tissue

110 reconstructions and previously-described isometric combinations10, 22, 41 on the anterolateral side of

111 the knee. 4. Compare the length change patterns of the different tibio-femoral point

112 combinations.10, 22, 41

113 Materials and Methods

114 Specimen preparation

115 Eight fresh-frozen left knees from donors with a mean age of 76 years (range: 69-86, 5 male and 3

116 female) were obtained from a tissue bank after ethical approval was given by the local research

117 ethics committee. Prior to testing, the specimens were thawed for 24 hours; the femur was cut

118 approximately 180 mm from the joint line, and the tibia 160 mm from the joint line. An

119 intramedullary rod was then cemented into the femur and another into the tibia using

120 polymethylmethacrylate (PMMA) bone cement. The skin and subcutaneous fat were then removed,

121 leaving the muscles and the fascia intact. The ITT was then dissected away from the vastus lateralis

122 longus and vastus lateralis obliquus muscles. The lateral retinaculum (iliopatellar band) was

123 horizontally incised to the point where the lateral femoral condyle became clearly visible. The lateral

124 retinaculum was sutured back after the preparation was finished. The ITT was then cut from the

125 intermuscular septum and its deep layer (Kaplan fibers). The capsulo-osseus layer of the ITT was

126 resected from its proximal attachment at the supraepicondylar region and its distal attachment at

127 the lateral tibial joint margin. Thus, only the superficial ITT layer was left attached at Gerdy’s

128 tubercle. Consistent with the technique used in previous studies, the quadriceps muscle was

129 separated into its six anatomical parts: Rectus femoris, Vastus intermedius, Vastus medialis longus,

130 Vastus lateralis longus, Vastus medialis obliquus and Vastus lateralis obliquus.18, 42 Cloth strips were 7

131 sutured to the proximal parts of the quadriceps and to the ITT, to augment the soft tissues and to

132 prevent slippage of the loading cables.

133 The femoral intramedullary rod was secured into a knee extension test rig (Figure 1). The posterior

134 condylar axis of the femur was aligned parallel to the base of the rig.42 The anatomical parts of the

135 quadriceps muscle and the ITT were then loaded according to their fiber orientation, using hanging

136 weights and a pulley system. Based on previous studies, a total of 175 N was applied to the

137 quadriceps muscle parts16 and 30 N to the ITT.18, 42 This tension extended the knee fully, which could

138 then be flexed and held at up to 90° of flexion (in 10° increments) using a horizontal bar anterior to

139 the tibial rod. Prior to the length change measurements, the loaded knee was cycled ten times

140 between 0° and 90° flexion in order to minimize the effects of soft tissue hysteresis.

141 Length changes between tibial and femoral attachments were measured by attaching small pins to

142 the tibia and eyelets on the femur. Sutures connected the pins to a displacement transducer via the

143 eyelets. One tibial pin was positioned at the tip of Gerdy’s tubercle (pinG), and another at the rim of

144 the lateral tibial condyle halfway between the fibular head and Gerdy’s tubercle (pinA), which has

145 been reported as the tibial attachment site of the capsule-osseous layer of the ITT44, the mid-third

146 lateral capsular ligament21 and the ALL described by Claes et al.6 and Dodds et al.8. A monofilament

147 suture was attached to each of these two pins. Six femoral eyelets, termed E1 to E6, were positioned

148 according to the anatomical structures, lateral extra-articular soft-tissue reconstruction methods,

149 and previously-defined isometric points on the anterolateral side of the knee. (Figure 2, Table 1).

150 These lateral extra-articular soft-tissue reconstructions typically route a strip of the ITT beneath the

151 LCL and loop it back to Gerdy’s tubercle via a bone tunnel in the lateral femoral condyle (Lemaire,

152 Losee), via a suture fixation on the intermuscular septum (MacIntosh), or via the over-the-top

153 position after an intra-articular ACL reconstruction (Rowe-Zarins).

154 The monofilament suture was collinearly attached to a linear variable displacement transducer

155 (LVDT) (Solartron Metrology, Bognor Regis, UK), thereby enabling measurement of length changes 8

156 between a pin and an eyelet at knee flexion angles between 0 and 90° (in 10° increments). The

157 monofilament suture was constantly under a small tension due to the weight of the sliding core of

158 the LVDT (0.5 N). Depending on the structure, lateral reconstruction method or isometric point being

159 assessed, the suture was guided either superficial or deep to the LCL. (Table 1) Length change

160 measurement data were collected for each of the 16 ligaments or reconstructions in each of the 8

161 knees, then processed using Solatron “Orbit” Excel software (Solatron Metrology). Each

162 measurement was repeated three times and the average results were used for analysis.

163 Data analysis

164 Absolute lengths between the different tibio-femoral point combinations were measured using a

165 ruler, to +/- 0.5 mm at 0°. The length change data were then normalized to percentage (strain= )

166 with reference to the length at 0° knee flexion.

167 In order to compare the isometry of the different tibio-femoral combinations, the total strain range

168 (TSR= maxStrain – minStrain) was calculated for each knee and then averaged. Low values of TSR

169 reflect near-isometry, and high values non-isometry.

170 Statistical Analysis:

171 1. Overall effects of changing the graft attachment position on the tibia (pinA and pinG) and the

172 femur (E1 to E6) and knee flexion (0°-90°) were calculated using a repeated measures ANOVA. The

173 course past the LCL was constant (superficial).

174 2. A repeated measures ANOVA was performed to investigate the effects of: graft path relative to

175 the LCL (deep and superficial), femoral position (E3 to E6) and knee flexion (0°-90°). The tibial

176 position was constant (pinG)

177 3. Three 2-way repeated measures ANOVAs were conducted to compare length changes on: 9

178 a) Native tissue structures of the anterolateral side (pinA/E1, pinA/E3, pinA/E6, and pinG/E6)

179 vs. flexion angle (0°-90°).

180 b) Lateral extra-articular soft-tissue reconstructions (pinG/E1, pinG/E3*, pinG/E5*) (* deep

181 to the LCL) vs. flexion angle (0°-90°).

182 c) Femoral isometric combinations (pinG/E2, pinG/E4 and pinG/E5) vs. flexion angle (0°-90°)

183 4. Three one-way ANOVAs were performed comparing TSR for native tissue structures, lateral extra-

184 articular soft-tissue reconstructions and femoral isometric points.

185 Pairwise comparisons with Bonferroni corrections were performed where appropriate. Statistical

186 analysis was performed in SPSS (Statistical Package for the Social Sciences, IBM Corp., Armonk, New,

187 York, U.S.) version 21, with significance level set at P<0.05.

188 Results

189 1. Attachment sites.

190 Altering the femoral attachment site had a large effect on the length changes (P<.001): for example,

191 changing from A-E1 to A-E3 changed the pattern from slackening to tightening with knee flexion

192 (Figure 3). The tibial attachment location had a smaller effect on length change pattern, but was still

193 significant (P<.001), for example changing from A-E6 to G-E6, Figure 3.

194 2. Graft course:

195 Graft length change patterns were significantly different depending on whether the graft ran

196 superficial or deep to the LCL (P<.001, Figure 4). There was a tendency for the superficial grafts to

197 lengthen during early knee flexion, whereas those running deep to the LCL tended to decrease in

198 length.

199 3.+ 4. Length change pattern and total strain range

200 a) Native tissue structures of the anterolateral side of the knee 10

201 The attachment points of the ALL described by Dodds et al.8 (pinA/E3 combination) was most

202 isometric among the ligaments on the anterolateral side (TSR = 8.7 ± 5.7%; mean ±SD), decreasing in

203 length between 30° and 80° of knee flexion (where a decrease in length between the points means a

204 tendency for a graft to slacken, and vice-versa). However, no significant difference in TSR value was

205 found compared to the other ligament combinations tested (Figures 3 and 5). The length between

206 the attachment points of the ALL described by Claes et al.6 (the pinA/E1 combination) increased

207 between 10° and 90° flexion, with the greatest overall TSR of 20.2 ± 8.4% among all ligament

208 combinations. This length change pattern was significantly different compared to both the ALL of

209 Dodds et al8 (pinA/E3: P < .001) and the posterior fibers of the ITT (pinA/E6: P < .001).

210 The posterior fibers of the ITT (pinA/E6 combination) were almost isometric at flexion angles

211 between 0° and 50°, then displayed a decrease in length from 50° to 90° (TSR = 9.4 ± 3.3%).

212 Conversely, the anterior fibers of the ITT (pinG/E6 combination) increased in length between 0° and

213 40° of flexion (TSR = 10.4 ± 3.4%), and was then almost isometric from 40° to 90° (Figure 3). There

214 was a significant difference in overall length change pattern between these two combinations (P < .

215 001).

216 No significant differences in TSR values were found among all ligament combinations tested.

217 b) Lateral Extra-articular soft-tissue reconstructions

218 The sutures following the course of the MacIntosh procedure30 ( pinG/E6 combination), routed deep

219 to the LCL, was closest to isometry among all tibio-femoral point combinations tested (Figures 5c

220 and 6), with an overall TSR of 5.5 ± 2.4%. The length change patterns of the MacIntosh, Rowe-

221 Zarins52/posterior part of the Losee29 and Lemaire25 lateral extra-articular soft-tissue reconstruction

222 (pinG/E6, pinG/E5 and pinG/E3) were all similar when guided deep to the LCL (Figure 6), particularly

223 between 0° and 30°. The three corresponding femoral attachments (eyelets E3, E5 and E6) were

224 located on a straight oblique line on the lateral aspect. 11

225 The anterior part of the Losee reconstruction29 (pinG/E1 combination) displayed a uniform increase

226 in length between 0° and 90° of knee flexion (TSR = 25.9 ± 9.8%), and its length change pattern was

227 significantly different (P < .001) compared to those of all other tested reconstructions. Also the TSR

228 value was significantly higher than that of the Lemaire (pinG/E3; P = .024), Rowe-Zarins/posterior

229 part of the Losee reconstruction (pinG/E5; P = .017), and the MacIntosh reconstruction (pinG/E6; P =

230 .010).

231 c) Femoral isometric points

232 The isometric pair of points of Krackow and Brooks22 (pinG/E5, passing superficial to the LCL)

233 displayed a slight length increase between 0° and 30° and a slight decrease between 40° and 80°

234 (TSR = 8.0 ± 3.2%; Figure 7). The length changes of this combination, the isometric points of Sidles et

235 al.41 (pinG/E4; TSR = 8.0 ± 3.9%), and of Draganich et al.10 (pinG/E2; TSR = 9.5 ± 3.4) all followed a

236 broadly similar pattern.

237 Tables with detailed data and results of statistical testing are available with the online version of this

238 paper.

240 Discussion

241 The purpose of the present study was to assess length change patterns and isometry of several

242 combinations of tibial and femoral points on the lateral side of the knee. This is the first study to our

243 knowledge to investigate the course of a graft running deep to the LCL, which makes it relevant to

244 previously described surgical techniques of lateral extra-articular soft-tissue reconstructions. All

245 tibio-femoral reconstruction combinations inserting proximal to the lateral epicondyle and with a

246 course deep to the LCL (pinG/E3, pinG/E5, and pinG/E6) were close to being isometric between 0°

247 and 90° knee flexion, with only a slight increase in length as the knee was extended. These are ideal

248 properties for a lateral extra-articular soft-tissue reconstruction. These reconstruction combinations 12

249 had very similar length change patterns. This similarity was because their course was deep to the

250 LCL, and therefore the lateral epicondyle, with the proximal LCL attachment, acted as a pulley,

251 retaining the graft posterior to the knee flexion axis of rotation within the investigated range of

252 motion. Conversely, there was much greater variability in length change patterns when the suture

253 was guided superficial to the LCL. In this case, the epicondyle acted as a ‘hump’ and the suture

254 remained anterior for low flexion angles and moved posteriorly as flexion angle increased.

255 Krackow and Brooks22 examined various tibio-femoral point combinations with a flexible ruler. They

256 applied a central load to the whole quadriceps muscle group and did not load the ITT, in contrast to

257 this study. The length change pattern of their ‘T3 to F9’ combination was close to isometric, and a

258 similar result was found in this study when reproducing it using the pinG/E5 combination.

259 Furthermore, we also observed Gerdy’s tubercle moving slightly laterally/posteriorly in terminal

260 knee extension due to the ‘screw home mechanism’. Thus, length slightly decreased at low flexion

261 angles (Figure 3). With regard to the effect of alteration of the femoral eyelet observed by Krackow

262 and Brooks, the length change plot displayed a uniform lengthening during knee flexion when the

263 femoral insertion site was distal and anterior to the lateral epicondyle (pinG/E1). Conversely, when

264 moving the femoral insertion site proximal and posterior to the lateral femoral epicondyle (pinG/E3

265 and pinG/E4), the length change plot showed an increase in length in low flexion angles, and then

266 decreasing length in high flexion angles.

267 The most isometric combination of the present study was the pinG/E6 combination, corresponding

268 to the MacIntosh reconstruction.30 This suture path deep to the LCL does not represent a native

269 structure of the knee. However, the MacIntosh reconstruction anchors the ITT to its natural insertion

270 on the femur at the distal termination of the intermuscular septum (Kaplan fibers), which may

271 explain the high degree of isometry. The most isometric femoral point combination with Gerdy’s

272 tubercle reported by Sidles et al.41 (corresponding to the pinG/E4 combination) in a quadriceps-

273 loaded knee was approximately 10 mm proximal and 6 mm posterior to the lateral femoral 13

274 epicondyle. That reported by Draganich et al.10 (corresponding to the pinG/E2 combination) was 4

275 mm distal and 10 mm posterior to the lateral femoral epicondyle. However, analogous to Sidles et al.

276 and Draganich et al., no perfectly isometric combination was found. Further comparisons of the data

277 in this study to those two studies are difficult, because both of them involved calculations of a

278 theoretical 3-D straight-line distance, rather than the actual path accounting for anatomical

279 irregularities and the course deep to the LCL.

280 The data for the length change measurements of the capsular ALL of Claes et al.,6 who found an

281 average length increase from full extension to 90° flexion of 3 mm, were consistent with this study.

282 However, this study measured a length increase in the pinA/E1 combination of more than double

283 that amount (7.4 ± 3.0 mm), resulting in a strain of 19.0 ± 8.8% at 90° flexion. These results imply

284 that the ALL described by Claes et al. and the mid-third lateral capsular ligament are slack in low

285 flexion angles, which is where the pivot-shift occurs, because soft tissues cannot sustain large strain

286 cycles. Dodds et al.8 found a mild decrease in length of their ALL of 5.8 ± 4.1mm from 0°-90°, and the

287 matching pinA/E3 combination in this study had a similar length decrease of 4.0 ± 3.5mm.

288 In good agreement with length change measurements in previous studies of the ITT,19, 28 the anterior

289 fibers (pinG/E6) displayed a plateau of increased length in high flexion angles, whereas the posterior

290 fibers (pinA/E6) had a plateau of increased length in low flexion angles. This implies that different ITT

291 fiber areas are taut in different flexion ranges. These findings suggest that it is not the ALL alone that

292 controls anterolateral rotation of the tibia, and that other structures such as parts of the ITT may

293 have a role. This is emphasized by the findings of Terry and Laprade, that the anterior arm of the

294 short head of the biceps femoris muscle, the capsulo-osseus layer of the ITT and the mid-third

295 lateral capsular ligament were attached at the site of the Segond fragment.45

296 It is known from previous clinical studies that extra-articular soft-tissue reconstructions in

297 combination with an intra-articular ACL reconstruction are capable of controlling the anterior

298 subluxation of the lateral tibial plateau.26, 33, 40, 48 This is supported by this study, as all tested 14

299 reconstructions except for the anterior part of the Losee reconstruction showed a lengthening as the

300 knee approached full extension. The femoral insertion sites of these reconstructions, passing deep to

301 the LCL, were all located on an oblique line on the distal femur from just proximal to the lateral

302 epicondyle to the posterior edge of the lateral aspect of the femur at the metaphysis. It is possible

303 that all points on this line may have similar length change patterns and low TSR values, thereby

304 presenting a safe area for positioning the extra-articular femoral insertion point. However, care must

305 be taken when combining length change data with other factors such as femoral graft fixation and

306 tensioning. The length change plots only represent the mean length change of eight specimens,

307 which included minor inter-specimen variability. This may be due to different knee kinematics and

308 anatomic variations. For example, in one knee there was a thicker femoral insertion of the lateral

309 gastrocnemius tendon, which changed the path of the suture.

310 In addition to the age and number of knees, there are some limitations of this study to note. First, an

311 active loading state was created by loading the quadriceps muscle parts and the ITT according to

312 their fiber directions. However, only one loading state was tested, and others were not considered.

313 A second limitation was the use of a suture to measure tibio-femoral point length changes. This

314 reduced the complex fiber bundle structure of a ligament or a graft to effectively a single fiber,

315 which may have had an effect, particularly when passing the suture deep to the LCL. This study has

316 provided data on the changes of length between attachment points in the intact knee; use of tendon

317 grafts in actual reconstructions would add further variables: the type of graft, the tension, the

318 fixation method, the angle of knee flexion and of tibial internal-external rotation would all affect the

319 results and may be studied separately. Thirdly, the maximum unloaded length27 of each ligament

320 was not measured (That is: the point of transition from the ligament being slack, to being taut.);

321 hence, we can only speculate on the actual tensile strain. However, the strain of reconstruction

322 grafts can also be influenced by varying the flexion angle of graft fixation, and pre-tensioning.

323 Additionally, we felt that it was preferable to ‘normalize’ all femoral eyelet locations. This proved to

324 be less straightforward than imagined, because of the variable ligament attachment sites, 15

325 ambiguous anatomical descriptions and different knee sizes. For example, the main femoral

326 insertion site of the mid-third lateral capsular ligament has been described at the tip of the lateral

327 femoral epicondyle.6, 21 However we observed an attachment site slightly anterior and distal,

328 resulting in a 2mm distance between actual and described attachment sites (Table 1). Finally: lateral

329 extra-articular reconstruction is usually used to control tibial internal rotational laxity, and so data

330 showing the effects of tibial rotation on the structures examined would be of interest. However, we

331 faced the practical limitation of the time required to make those extra measurements on the set of

332 knees, which would have been excessive, and also the consideration that it would be more clinically

333 relevant to perform measurements of restraint to tibial rotation, rather than length changes.

335 Conclusion

The results of this study provide a rationale regarding the course and the behavior of the graft for an extra-articular lateral-based reconstruction. The sutures representing grafts that ran deep to the LCL, with insertion sites proximal to the lateral epicondyle, showed desirable length-change patterns, having relatively low length changes during knee flexion-extension, and being longer (tighter) near knee extension, implying an ability to prevent anterior subluxation of the lateral tibial plateau. This path and femoral attachment site did not correspond to any anatomical structure. The ALL of Dodds et al 8 passed over the LCL and gave a similar length-change pattern, but some other ligaments or lateral extra-articular soft-tissue reconstructions did not provide this. Further studies should determine which of the lateral structures resist the loads that tend to cause tibial internal rotational subluxation, and address the biomechanical behavior of lateral extra-articular reconstructions. Table 1. Femoral eyelet positioning and corresponding tibio-femoral point combinations (four native tissue structures, four reconstructions and three femoral isometric points). Femoral Position (from lateral Tibial pin Eyelet femoral epicondyle) pinG pinA

mid-third lateral capsular 21 2mm anterior, anterior part of the ligament E1 2mm distal Losee reconstruction29 ALL defined by Claes et al.6

10mm posterior, E2 Isometric point Draganich et al.10 4mm distal 16

4mm posterior, E3 Lemaire reconstruction25 * ALL defined by Dodds et al.8 8mm proximal

6mm posterior, E4 Isometric point Sidles et al.41 10mm proximal

Rowe-Zarins reconstruction52 * Isometric point F9 over-the -top E5 Krackow and Brooks22 position posterior part of the Losee reconstruction29 *

posterior femoral cortex anterior fibers of the ITT posterior fibers of the ITT E6 at the distal termination of the intramuscular septum MacIntosh reconstruction30 * * indicates course deep to the LCL 336

Table 2. Length change (%) and significant differences at each flexion angle for ligament tibio-femoral point combinations. Native tissue structures pinA/E1 #,† pinA/E3 *,§ pinG/E6 #,† pinA/E6 *,§ Knee flexion angle (°) Mean SD Mean SD Mean SD Mean SD 0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 10 -0.8 § 1.7 -1.0 § 1.5 2.4 *,#,† 1.1 -0.2 § 0.5 20 1.6 #,§ 1.8 -0.9 *,§,† 1.4 5.3 *,#,† 1.2 0.6 #,§ 0.8 30 4.7 #,† 2.6 -1.4 *,§,† 1.9 7.6 #,† 1.3 0.9 *,#,§ 1.1 40 7.6 #,† 3.4 -2.5 *,§,† 2.6 8.9 #,† 2.0 0.2 *,#,§ 1.8 50 10.4 #,† 4.4 -3.7 *,§,† 3.2 9.5 #,† 2.9 -0.9 *,#,§ 2.3 60 12.8 #,† 5.4 -5.0 *,§ 3.8 9.4 #,† 3.3 -2.6 *,§ 2.7 70 15.3 #,† 6.5 -6.3 *,§ 4.6 8.8 #,† 4.0 -4.8 *,§ 2.8 80 17.4 #,§,† 8.0 -7.5 *,§ 5.4 7.8 *,#,† 4.7 -6.9 *,§ 3.3 90 19.0 #,§,† 8.8 -7.8 *,§ 6.5 7.4 *,#,† 5.5 -8.3 *,§ 4.1 * indicates statistical significance from pinA/E1 # indicates statistical significance from pinA/E3 § indicates statistical significance from pinG/E6 † indicates statistical significance from pinA/E6 SD: Standard Deviation 337

Table 3. Total strain range values of each tested tibio-femoral point combination. Superficial to LCL Deep to LCL pinG pinA pinG 17

Femoral eyelet Mean SD Mean SD Mean SD E1 25.9 9.8 20.2 8.4 E2 9.5 3.4 14.2 5.9 E3 7.2 1.9 8.7 5.7 7.1 3.5 E4 8.0 3.9 15.3 5.3 8.7 4.3 E5 8.0 3.2 14.1 4.2 7.3 2.7 E6 10.4 3.4 9.4 3.3 5.5 2.4 338

Table 4. Length change (%) and significant differences at each flexion angle for reconstruction tibio- femoral point combinations. The course of the suture was deep to the lateral collateral ligament. Reconstructions pinG/E1 #,§,† pinG/E3 * pinG/E5 *,† pinG/E6 *,§ Knee flexion angle (°) Mean SD Mean SD Mean SD Mean SD 0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 10 3.3 #,§,† 1.5 -0.5 * 1.8 -0.1 * 1.0 -0.1 * 0.9 20 6.9 #,§,† 2.1 -0.7 * 2.7 -0.2 * 1.7 -0.1 * 1.4 30 9.4 #,§,† 3.3 -2.0 * 3.1 -1.2 * 2.3 -1.0 * 1.6 40 11.6 #,§,† 4.0 -3.1 * 2.8 -2.3 *,† 1.9 -1.6 *,§ 1.5 50 14.3 #,§,† 4.9 -3.9 *,† 2.5 -3.4 *,† 1.8 -2.2 *,§ 1.5 60 17.4 #,§,† 6.3 -4.5 * 2.8 -4.4 *,† 1.9 -2.9 *,§ 1.5 70 20.3 #,§,† 7.7 -5.0 * 3.3 -5.4 *,† 2.1 -3.7 *,§ 1.7 80 23.4 #,§,† 9.2 -5.3 * 3.8 -6.3 *,† 2.6 -4.5 *,§ 1.9 90 25.9 #,§,† 9.8 -5.1 * 4.1 -6.4 *,† 2.8 -4.8 *,§ 2.4 * indicates statistical significance from pinG/E1 # indicates statistical significance from pinG/E3 § indicates statistical significance from pinG/E5 † indicates statistical significance from pinG/E6 SD: Standard Deviation 339

Table 5. Length change (%) and significances at each flexion angle for tested femoral isometric points. Isometric Points pinG/E2 #,§ pinG/E4 *,§ pinG/E5 *,# Knee flexion angle (°) Mean SD Mean SD Mean SD 0 0.0 0.0 0.0 0.0 0.0 0.0 10 0.9 § 1.6 1.7 1.8 2.1 * 1.6 20 0.5 § 3.0 2.4 § 2.4 4.5 *,# 2.6 30 -1.8 #,§ 4.1 1.6 *,§ 3.3 5.5 *,# 3.7 40 -4.4 #,§ 4.4 0.4 *,§ 3.4 5.5 *,# 4.5 50 -5.8 #,§ 4.0 -0.8 *,§ 3.5 4.7 *,# 4.8 60 -6.6 #,§ 3.9 -1.9 *,§ 3.8 4.0 *,# 5.0 70 -6.5 #,§ 4.6 -2.9 *,§ 4.4 2.8 *,# 5.2 18

80 -5.5 § 5.1 -3.8 § 5.1 2.0 *,# 5.6 90 -3.6 § 5.5 -3.8 5.5 2.0 * 6.5 * indicates statistical significance from pinG/E2 # indicates statistical significance from pinG/E4 § indicates statistical significance from pinG/E5 340

343 19

344 20

356 21

376 22

403 23

405 References:

406 1. Amis A, Zavras T. Isometricity and graft placement during anterior cruciate ligament 407 reconstruction. Knee. 1995;2(1):5-17. 408 2. Amis AA, Jakob RP. Anterior cruciate ligament graft positioning, tensioning and twisting. 409 Knee Surg Sports Traumatol Arthrosc. 1998;6:S2-S12. 410 3. Barrett G, Richardson K. The effect of added extra-articular procedure on results of ACL 411 reconstruction. Am J Knee Surg. 1994;8:1–6. 412 4. Bignozzi S, Zaffagnini S, Lopomo N, Martelli S, Iacono F, Marcacci M. Does a lateral plasty 413 control coupled translation during antero-posterior stress in single-bundle ACL 414 reconstruction? An in vivo study. Knee Surg Sports Traumatol Arthrosc. 2009;17:65-70. 415 5. Carson WG, Jr. Extra-articular reconstruction of the anterior cruciate ligament: lateral 416 procedures. Orthop Clin North Am. 1985;16:191-211. 417 6. Claes S, Vereecke E, Maes M, Victor J, Verdonk P, Bellemans J. Anatomy of the anterolateral 418 ligament of the knee. J Anat. 2013;223:321–328. 419 7. Dahlstedt LJ, Dalén N, Jonsson U. Extraarticular repair of the unstable knee: Disappointing 6- 420 year results of the Slocum and Ellison operations. Acta Orthop. 1988;59:687–691. 421 8. Dodds A, Halewood C, Gupte C, Williams A, Amis A. The anterolateral ligament Anatomy, 422 length changes and association with the Segond fracture. Bone Joint J. 2014;96(3):325-331. 423 9. Dodds AL, Gupte CM, Neyret P, Williams AM, Amis AA. Extra-articular techniques in anterior 424 cruciate ligament reconstruction a literature review. J Bone Joint Surg Br. 2011;93-B:1440- 425 1448. 426 10. Draganich LF, Hsieh Y-F, Reider B. Iliotibial Band Tenodesis: A New Strategy for Attachment. 427 Am J Sports Med. 1995;23:186-195. 428 11. Draganich LF, Reider B, Ling M, Samuelson M. An in vitro study of an intraarticular and 429 extraarticular reconstruction in the anterior cruciate ligament deficient knee. Am J Sports 430 Med. 1990;18:262-266. 431 12. Draganich LF, Reider B, Miller PR. An in vitro study of the Müller anterolateral femorotibial 432 ligament tenodesis in the anterior cruciate ligament deficient knee. Am J Sports Med. 433 1989;17:357-362. 434 13. Ellison A. Distal iliotibial-band transfer for anterolateral rotatory instability of the knee. J 435 Bone Joint Surg Am. 1979;61(3):330-337. 436 14. Engebretsen L, Lew WD, Lewis JL, Hunter RE. The effect of an iliotibial tenodesis on 437 intraarticular graft forces and knee joint motion. Am J Sports Med. 1990;18:169-176. 438 15. Engebretsen L, Lew WD, Lewis JL, Hunter RE, Benum Pa. Anterolateral rotatory instability of 439 the knee: cadaver study of extraarticular patellar-tendon transposition. Acta Orthop. 440 1990;61:225–230. 441 16. Farahmand F, Sejiavongse W, Amis AA. Quantitative study of the quadriceps muscles and 442 trochlear groove geometry related to instability of the patellofemoral joint. J Orthop Res. 443 1998;16:136–143. 444 17. Garcia R, Jr, Brunet ME, Timon S, Barrack RL. Lateral extra-articular knee reconstruction: 445 long-term patient outcome and satisfaction. J South Orthop Assoc. 2000;9:19-23. 446 18. Ghosh KM, Merican AM, Iranpour-Boroujeni F, Deehan DJ, Amis AA. Length change patterns 447 of the extensor retinaculum and the effect of total knee replacement. J Orthop Res. 448 2009;27:865–870. 449 19. Hassler H, Jakob RP. Ein beitrag zur ursache der anterolateralen instabilität des kniegelenkes. 450 Arch Orthop Trauma Surg. 1981;98:45-50. 451 20. Helito CP, Demange MK, Bonadio MB, et al. Anatomy and Histology of the Knee 452 Anterolateral Ligament. Orthop J Sports Med. 2013;1:2325967113513546. 24

453 21. Hughston J, Andrews J, Cross M, Moschi A. Classification of knee ligament instabilities. Part 454 II. The lateral compartment. J Bone Joint Surg Am. 1976;58(2):173. 455 22. Krackow KA, Brooks RL. Optimization of knee ligament position for lateral extraarticular 456 reconstruction. Am J Sports Med. 1983;11:293-302. 457 23. Kurosawa H, Yasuda K, Yamakoshi K-I, Kamiya A, Kaneda K. An experimental evaluation of 458 isometric placement for extraarticular reconstructions of the anterior cruciate ligament. Am 459 J Sports Med. 1991;19:384-388. 460 24. Larsen E, Blyme P, Hede A. Pes anserinus and iliotibial band transfer for anterior cruciate 461 insufficiency. Am J Sports Med. 1991;19:601-604. 462 25. Lemaire M, Combelles F. Technique actuelle de plastie ligamentaire pour rupture ancienne 463 du ligament croisé antérieur. Rev Chir Orthop. 1980;66:523–525. 464 26. Lerat JL, Chotel F, Besse JL, et al. [The results after 10-16 years of the treatment of chronic 465 anterior laxity of the knee using reconstruction of the anterior cruciate ligament with a 466 patellar tendon graft combined with an external extra-articular reconstruction]. Rev Chir 467 Orthop Reparatrice Appar Mot. 1998;84:712-727. 468 27. Lewis JL. Maximum unloaded length (MUL) and graft force as criteria for anterior cruciate 469 ligament graft fixation. Knee Surg Sports Traumatol Arthrosc. 1998;6:S25-S29. 470 28. Lobenhoffer P, Posel P, Witt S, Piehler J, Wirth CJ. Distal femoral fixation of the iliotibial 471 tract. Arch Orthop Trauma Surg. 1987;106:285-290. 472 29. Losee R, Johnson T, Southwick W. Anterior subluxation of the lateral tibial plateau. A 473 diagnostic test and operative repair. J Bone Joint Surg Am. 1978;60:1015-1030. 474 30. MacIntosh D, Darby T. Lateral substitution reconstruction. Proceedings of the Canadian 475 Orthopaedic Association. J Bone Joint Surg Br. 1976;58:142. 476 31. Marcacci M, Zaffagnini S, Iacono F, Neri MP, Loreti I, Petitto A. Arthroscopic intra- and extra- 477 articular anterior cruciate ligament reconstruction with gracilis and semitendinosus tendons. 478 Knee Surg Sports Traumatol Arthrosc. 1998;6:68-75. 479 32. Matsumoto H, Seedhom B. Treatment of the Pivot-Shift Intraarticular Versus Extraarticular 480 or Combined Reconstruction Procedures A Biomechanical Study. Clin Orthop Relat Res. 481 1994;299:298–304. 482 33. Monaco E, Labianca L, Conteduca F, Carli AD, Ferretti A. Double bundle or single bundle plus 483 extraarticular tenodesis in ACL reconstruction? Knee Surg Sports Traumatol Arthrosc. 484 2007;15:1168-1174. 485 34. Noyes F, Barber S. The effect of an extra-articular procedure on allograft reconstructions for 486 chronic ruptures of the anterior cruciate ligament. J Bone Joint Surg Am. 1991;73:882-892. 487 35. O'Brien SJ, Warren RF, Wickiewicz TL, et al. The iliotibial band lateral sling procedure and its 488 effect on the results of anterior cruciate ligament reconstruction. Orthop J Sports Med. 489 1991;19:21-25. 490 36. O'Meara PM, O'Brien WR, Henning CE. Anterior cruciate ligament reconstruction stability 491 with continuous passive motion. The role of isometric graft placement. Clin Orthop Relat 492 Res. 1992(277):201-209. 493 37. Penner DA, Daniel DM, Wood P, Mishra D. An in vitro study of anterior cruciate ligament 494 graft placement and isometry. Am J Sports Med. 1988;16(3):238-243. 495 38. Reid J, Hanks G, Kalenak A, Kottmeier S, Aronoff V. The Ellison iliotibial-band transfer for a 496 torn anterior cruciate ligament of the knee. Long-term follow-up. J Bone Joint Surg Am. 497 1992;74:1392-1402. 498 39. Roth J, Kennedy J, Lockstadt H, McCallum C, Cunning L. Intra-articular reconstruction of the 499 anterior cruciate ligament with and without extra-articular supplementation by transfer of 500 the biceps femoris tendon. J Bone Joint Surg Am. 1987;69:275-278. 501 40. Saragaglia D, Pison A, Refaie R. Lateral tenodesis combined with anterior cruciate ligament 502 reconstruction using a unique semitendinosus and gracilis transplant. Int Orthop. 503 2013;37:1575-1581. 25

504 41. Sidles JA, Larson RV, Garbini JL, Downey DJ, Matsen FA. Ligament length relationships in the 505 moving knee. J Orthop Res. 1988;6:593–610. 506 42. Stephen JM, Lumpaopong P, Deehan DJ, Kader D, Amis AA. The Medial Patellofemoral 507 Ligament Location of Femoral Attachment and Length Change Patterns Resulting From 508 Anatomic and Nonanatomic Attachments. Am J Sports Med. 2012;40:1871-1879. 509 43. Strum GM, Fox JM, Ferkel RD, et al. Intraarticular versus intraarticular and extraarticular 510 reconstruction for chronic anterior cruciate ligament instability. Clin Orthop Relat Res. 511 1989;245:188–198. 512 44. Terry GC, Hughston JC, Norwood LA. The anatomy of the iliopatellar band and iliotibial tract. 513 Am J Sports Med. 1986;14:39-45. 514 45. Terry GC, LaPrade RF. The biceps femoris muscle complex at the knee Its anatomy and injury 515 patterns associated with acute anterolateral-anteromedial rotatory instability. Am J Sports 516 Med. 1996;24(1):2-8. 517 46. Terry GC, Norwood LA, Hughston JC, Caldwell KM. How iliotibial tract injuries of the knee 518 combine with acute anterior cruciate ligament tears to influence abnormal anterior tibial 519 displacement. Am J Sports Med. 1993;21:55-60. 520 47. Trojani C, Beaufils P, Burdin G, et al. Revision ACL reconstruction: influence of a lateral 521 tenodesis. Knee Surg Sports Traumatol Arthrosc. 2012;20:1565-1570. 522 48. Vadalà AP, Iorio R, Carli AD, et al. An extra-articular procedure improves the clinical outcome 523 in anterior cruciate ligament reconstruction with hamstrings in female athletes. Int Orthop. 524 2013;37:187-192. 525 49. Vieira ELC, Vieira EÁ, Teixeira da Silva R, dos Santos Berlfein PA, Abdalla RJ, Cohen M. An 526 Anatomic Study of the Iliotibial Tract. Arthroscopy. 2007;23:269-274. 527 50. Vincent J-P, Magnussen RA, Gezmez F, et al. The anterolateral ligament of the human knee: 528 an anatomic and histologic study. Knee Surg Sports Traumatol Arthrosc. 2012;20:147-152. 529 51. Zaffagnini S, Marcacci M, Presti ML, Giordano G, Iacono F, Neri MP. Prospective and 530 randomized evaluation of ACL reconstruction with three techniques: a clinical and 531 radiographic evaluation at 5 years follow-up. Knee Surg Sports Traumatol Arthrosc. 532 2006;14:1060-1069. 533 52. Zarins B, Rowe C. Combined anterior cruciate-ligament reconstruction using semitendinosus 534 tendon and iliotibial tract. J Bone Joint Surg Am. 1986;68:160-177.